BUILDING SYSTEM WITH AIR QUALITY CONTROL BASED ON ENERGY, MOLD RISK, AND PRODUCTIVITY MODELING

BACKGROUND

The present disclosure relates generally to a building system in a building. The present disclosure relates more particularly to maintaining occupant comfort in a building through environmental control.

Maintaining occupant comfort, productivity, and mold risk minimization in a building requires building equipment (e.g., HVAC equipment) to be operated to change environmental conditions in the building. In some systems, occupants are required to make any desired changes to the environmental conditions themselves if they are not comfortable. When operating building equipment to change specific environmental conditions, other environmental conditions may be affected as a result. Maintaining occupant comfort, productivity, and mold risk minimization can be expensive if not performed correctly. Thus, systems and methods are needed to maintain occupant comfort and provide sufficient productivity and mold risk minimization for multiple environmental conditions while reducing expenses related to maintaining occupant comfort and mold risk minimization.

SUMMARY

At least one aspect relates to a controller for heating, ventilation, or air conditioning (HVAC) equipment operable to affect an environmental condition of a building. The controller includes: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform operations including: obtaining one or more predictive models configured to predict values of an air quality control objective and another control objective as a function of control decision variables for the HVAC equipment, executing an optimization process using the one or more predictive models to produce multiple sets of optimization results corresponding to different values of the control decision variables, the air quality control objective, and the other control objective, selecting one or more of the sets of optimization results based on the values of the air quality control objective and the other control objective, and operating the HVAC equipment to affect the environmental condition of the building in accordance with the values of the control decision variables corresponding to a selected set of the optimization results.

In some embodiments, the air quality control objective includes a mold risk control objective and the other control objective includes at least one of an operating cost predicted to result from operating the HVAC equipment in accordance with the control decision variables or a capital cost of purchasing or installing the HVAC equipment. In some embodiments, the air quality control objective includes a mold risk control objective and the other control objective includes an infection risk predicted to result from operating the HVAC equipment in accordance with the control decision variables.

In some embodiments, the air quality control objective includes a productivity control objective and the other control objective includes an energy consumption or energy cost predicted to result from operating the HVAC equipment in accordance with the control decision variables. In some embodiments, the air quality control objective includes a productivity control objective and the other control objective includes at least one of an operating cost predicted to result from operating the HVAC equipment in accordance with the control decision variables or a capital cost of purchasing or installing the HVAC equipment.

In some embodiments, the air quality control objective comprises a productivity control objective and the one or more predictive models are configured to predict values of a productivity score predicted to result from operating the HVAC equipment in accordance with the control decision variables. In some embodiments, executing the optimization process includes executing multiple optimization processes using different sets of constraints for the control decision variables or different search spaces for the control decision variables, the multiple optimization processes producing corresponding sets of the multiple sets of optimization results. In some embodiments, selecting one or more of the sets of optimization results includes selecting one or more of the sets of optimization results for which the values of the air quality control objective and the other control objective are not both improved by another of the sets of optimization results.

In some embodiments, selecting one or more of the sets of optimization results includes: classifying the multiple sets of optimization results as either Pareto-optimal optimization results or non-Pareto-optimal optimization results with respect to the air quality control objective and the other control objective, and selecting the Pareto-optimal optimization results. In some embodiments, selecting one or more of the sets of optimization results includes selecting: a first set of optimization results that prioritizes the air quality control objective over the other control objective, a second set of optimization results that prioritizes the other control objective over the air quality control objective, and a third set of optimization results that balances the air quality control objective and the other control objective.

In some embodiments, the operations further include: presenting the values of the air quality control objective and the other control objective associated with the first set of optimization results, the second set of optimization results, and the third set of optimization results as selectable options via a user interface, and determining the selected set of the optimization results responsive to a user selecting one of the selectable options via the user interface.

At least one aspect relates to a method for controlling heating, ventilation, or air conditioning (HVAC) equipment operable to affect an environmental condition of a building. The method includes obtaining, by one or more processors, one or more predictive models configured to predict values of an air quality control objective and another control objective as a function of control decision variables for the HVAC equipment, executing, by one or more processors, an optimization process using the one or more predictive models to produce multiple sets of optimization results corresponding to different values of the control decision variables, the air quality control objective, and the other control objective, selecting, by one or more processors, one or more of the sets of optimization results based on the values of the air quality control objective and the other control objective, and operating, by one or more processors, the HVAC equipment to affect the environmental condition of the building in accordance with the values of the control decision variables corresponding to a selected set of the optimization results.

In some embodiments, the air quality control objective includes a mold risk control objective and the other control objective includes at least one of: an operating cost predicted to result from operating the HVAC equipment in accordance with the control decision variables, a capital cost of purchasing or installing the HVAC equipment, or an infection risk predicted to result from operating the HVAC equipment in accordance with the control decision variables. In some embodiments, the air quality control objective includes a productivity control objective and the other control objective includes at least one of: an energy consumption or energy cost predicted to result from operating the HVAC equipment in accordance with the control decision variables, an operating cost predicted to result from operating the HVAC equipment in accordance with the control decision variables, or a capital cost of purchasing or installing the HVAC equipment.

In some embodiments, executing the optimization process includes executing, by the one or more processors, multiple optimization processes using different sets of constraints for the control decision variables or different search spaces for the control decision variables, the multiple optimization processes producing corresponding sets of the multiple sets of optimization results. In some embodiments, selecting one or more of the sets of optimization results includes selecting, by the one or more processors, one or more of the sets of optimization results for which the values of the air quality control objective and the other control objective are not both improved by another of the sets of optimization results.

In some embodiments, selecting one or more of the sets of optimization results includes: classifying, by the one or more processors, the multiple sets of optimization results as either Pareto-optimal optimization results or non-Pareto-optimal optimization results with respect to the air quality control objective and the other control objective, and selecting, by the one or more processors, the Pareto-optimal optimization results. In some embodiments, selecting one or more of the sets of optimization results includes selecting, by the one or more processors: a first set of optimization results that prioritizes the air quality control objective over the other control objective, a second set of optimization results that prioritizes the other control objective over the air quality control objective, and a third set of optimization results that balances the air quality control objective and the other control objective.

In some embodiments, the method further includes presenting, by the one or more processors, the values of the air quality control objective and the other control objective associated with the first set of optimization results, the second set of optimization results, and the third set of optimization results as selectable options via a user interface, and determining, by the one or more processors, the selected set of the optimization results responsive to a user selecting one of the selectable options via the user interface.

At least one aspect relates to one or more non-transitory computer-readable media for controlling heating, ventilation, or air conditioning (HVAC) equipment operable to affect an environmental condition of a building storing instructions thereon that, when executed by one or more processors, cause the one or more processors to: obtain one or more predictive models configured to predict values of an air quality control objective and another control objective as a function of control decision variables for the HVAC equipment, execute an optimization process using the one or more predictive models to produce multiple sets of optimization results corresponding to different values of the control decision variables, the air quality control objective, and the other control objective, select one or more of the sets of optimization results based on the values of the air quality control objective and the other control objective, and operate the HVAC equipment to affect the environmental condition of the building in accordance with the values of the control decision variables corresponding to a selected set of the optimization results.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a drawing of a building equipped with a HVAC system, according to some embodiments.

FIG. 2 is a block diagram of an airside system which can be implemented in the building of FIG. 1, according to some embodiments.

FIG. 3 is a block diagram of a building automation system (BAS) that may be used to monitor and/or control the building of FIG. 1, according to an exemplary embodiment.

FIG. 4 is a block diagram of a building analysis system that may generate one or more air quality occupant impact assessments and/or recommendations of spaces of a building based on air quality measurements of sensors, according to an exemplary embodiment.

FIG. 5 is a flow diagram of a process where air quality sensors are installed in spaces of a building to collect air quality measurements of the spaces and generate one or more air quality occupant impact assessments and/or recommendations, according to an exemplary embodiment.

FIG. 6 is a flow diagram of a process where one or more productivity scores may be analyzed, according to an exemplary embodiment.

FIG. 7 is a block diagram of an HVAC system including a controller configured to operate an air-handling unit (AHU) of the HVAC system of FIG. 1, according to some embodiments.

FIG. 8 is a block diagram illustrating the controller of FIG. 7 in greater detail, showing operations performed when the controller is used in an on-line mode or real-time implementation for making control decisions to minimize energy consumption of the HVAC system and provide sufficient productivity scores, according to some embodiments.

FIG. 9 is a block diagram illustrating the controller of FIG. 7 in greater detail, showing operations performed when the controller is used in an off-line or planning mode for making design suggestions to minimize energy consumption of the HVAC system and provide sufficient productivity scores, according to some embodiments.

FIG. 10 is a flow diagram of a process which can be performed by the controller of FIG. 7 for determining control decisions for an HVAC system to minimize energy consumption and provide sufficient disinfection, according to some embodiments.

FIG. 11 is a flow diagram of a process which can be performed by the controller of FIG. 7 for determining design suggestions for an HVAC system to minimize energy consumption and provide sufficient disinfection, according to some embodiments.

FIG. 12 is a graph of various design suggestions or information that can be provided by the controller of FIG. 7, according to some embodiments.

FIG. 13 is a drawing of a user interface that can be used to specify building options and disinfection options and provide simulation results, according to some embodiments.

FIG. 14 is a graph illustrating a technique which can be used by the controller of FIG. 7 to estimate a Pareto front of a tradeoff curve for relative energy cost vs. infection probability, according to some embodiments.

FIG. 15 is a block diagram illustrating the controller of FIG. 7 including a Pareto optimizer, according to some embodiments.

FIG. 16 is a block diagram illustrating the functionality of the controller of FIG. 15, according to some embodiments.

FIG. 17 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including energy cost and productivity for each of the different combinations of decision variables, according to some embodiments.

FIG. 18 is a diagram including the second graph of FIG. 17 and a third graph illustrating which of the simulation results are infeasible, feasible but not Pareto optimal, and feasible and Pareto optimal, according to some embodiments.

FIG. 19 is a diagram including the third graph of FIG. 18 and a fourth graph that illustrates a minimum productivity solution, a minimum energy cost solution, and an equal priority productivity/energy cost solution of the Pareto optimal simulation results, according to some embodiments.

FIG. 20 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including energy cost and productivity for each of the different combinations of decision variables, according to some embodiments.

FIG. 21 is a flow diagram of a process for performing a Pareto optimization to determine different Pareto optimal solutions in terms of energy cost and productivity scores for a BMS system, according to some embodiments.

FIG. 22 is a user interface showing results of a productivity score analysis for display on a user device, according to some embodiments.

FIG. 23 is a user interface showing results of the Pareto optimization, according to some embodiments.

FIG. 24 is a user interface showing operating adjustments for a building administrator to perform as a result of the Pareto optimization, according to some embodiments.

FIG. 25 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including energy cost and infection risk for each of the different combinations of decision variables, according to some embodiments.

FIG. 26 is a diagram including the second graph of FIG. 25 and a third graph illustrating which of the simulation results are infeasible, feasible but not Pareto optimal, and feasible and Pareto optimal, according to some embodiments.

FIG. 27 is a diagram including the third graph of FIG. 26 and a fourth graph that illustrates a minimum infection risk solution, a minimum energy cost solution, and an equal priority infection risk/energy cost solution of the Pareto optimal simulation results, according to some embodiments.

FIG. 28 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including energy cost and infection risk for each of the different combinations of decision variables, according to some embodiments.

FIG. 29 is a flow diagram of a process for performing a Pareto optimization to determine different Pareto optimal solutions in terms of energy cost and infection risk for an HVAC system, according to some embodiments.

FIG. 30 is a flow diagram of a process for performing an infection metric analysis of an HVAC system over a previous time period, according to some embodiments.

FIG. 31 is a flow diagram of a process for performing a Pareto optimization in terms of energy cost and infection risk over a future time period, according to some embodiments.

FIG. 32 is a flow diagram of a process for performing an infection metrics analysis of an HVAC system over a previous time period and a Pareto optimization for the HVAC system over a future time period, according to some embodiments.

FIG. 33 is a user interface showing results of an infection metric analysis for display on a user device, according to some embodiments.

FIG. 34 is a user interface showing results of the Pareto optimization of FIG. 29 or 31, according to some embodiments.

FIG. 35 is a user interface showing operating adjustments for a building administrator to perform as a result of the Pareto optimization of FIG. 29 or 31, according to some embodiments.

FIG. 36 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including a sustainability metric and infection risk for each of the different combinations of decision variables, according to some embodiments.

FIG. 37 is a graph showing a relationship between energy cost and a carbon equivalent, according to some embodiments.

FIG. 38 is another graph showing a relationship between total cost and a carbon equivalent, according to some embodiments.

FIG. 39 is a block diagram illustrating the functionality of the controller of FIG. 15 including functionality for converting between energy cost and a sustainability metric, according to some embodiments.

FIG. 40 is a flow diagram of a process for performing a Pareto optimization to determine different Pareto optimal solutions in terms of energy cost converted to a sustainability metric and infection risk for an HVAC system, according to some embodiments.

FIG. 41 is a flow diagram of a process for performing a Pareto optimization to determine different Pareto optimal solutions in terms of a sustainability metric and infection risk for an HVAC system, according to some embodiments.

FIG. 42 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including a sustainability metric and an energy cost for each of the different combinations of decision variables, according to some embodiments.

FIG. 43 is a diagram including a plurality of graphs showing air temperature, relative humidity, and dew point versus time, according to some embodiments.

FIG. 44 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including energy cost and mold risk for each of the different combinations of decision variables, according to some embodiments.

FIG. 45 is a diagram including the second graph of FIG. 44 and a third graph illustrating which of the simulation results are infeasible, feasible but not Pareto optimal, and feasible and Pareto optimal, according to some embodiments.

FIG. 46 is a diagram including the third graph of FIG. 45 and a fourth graph that illustrates a minimum mold risk solution, a minimum energy cost solution, and an equal priority mold risk/energy cost solution of the Pareto optimal simulation results, according to some embodiments.

FIG. 47 is a diagram including a first graph that shows different combinations of decision variables, and a second graph that shows simulation results including energy cost and mold risk for each of the different combinations of decision variables, according to some embodiments.

FIG. 48 is a flow diagram of a process for performing a Pareto optimization to determine different Pareto optimal solutions in terms of energy cost and mold risk for an HVAC system, according to some embodiments.

FIG. 49 is a flow diagram of a process for performing a mold metric analysis of an HVAC system over a previous time period, according to some embodiments.

FIG. 50 is a flow diagram of a process for performing a Pareto optimization in terms of energy cost and mold risk over a future time period, according to some embodiments.

FIG. 51 is a flow diagram of a process for performing a mold metrics analysis of an HVAC system over a previous time period and a Pareto optimization for the HVAC system over a future time period, according to some embodiments.

FIG. 52 is a user interface showing results of a mold metric analysis for display on a user device, according to some embodiments.

FIG. 53 is a user interface showing results of the Pareto optimization of FIG. 48 or 50, according to some embodiments.

FIG. 54 is a flow diagram of a process performing either a two control objective Pareto optimization or a three control objective Pareto optimization, according to some embodiments.

DETAILED DESCRIPTION
Overview

Referring generally to the FIGURES, systems and methods for minimizing energy consumption of an HVAC system while maintaining a desired level of disinfection and/or productivity are shown. The system may include an AHU that serves multiple zones, a controller, one or more UV lights that disinfect air before it is provided from the AHU to the zones, and/or a filter that is configured to filter air to provide additional disinfection for the air before it is provided to the zones. In some embodiments, the system also includes one or more zone sensors (e.g., temperature and/or humidity sensors, etc.) and one or more ambient or outdoor sensors (e.g., outdoor temperature and/or outdoor humidity sensors, etc.).

The controller uses a model-based design and optimization framework to integrate building disinfection and/or air quality control with existing temperature regulation in building HVAC systems. The controller uses the Wells-Riley equation, or any infection risk management standard, such as ASHRAE 241, to transform a required upper limit of infection probability into constraints on indoor concentration of infectious particles, according to some embodiments. ASHRAE standard 241, Control of Infectious Aerosols, utilizes a model similar to that described with respect to the Wells-Riley equation or model to provide a guideline for an amount of “Equivalent Clean Air” needed to maintain an acceptable risk of infection. In some embodiments, the controller uses a dynamic model for infectious agent concentration to impose these constraints on an optimization problem similar to temperature and humidity constraints. By modeling effects of various types of optional infection control equipment (e.g., UV lights and/or filters), the controller may utilize a combination of fresh-air ventilation and direct filtration/disinfection to achieve desired infection constraints. In some embodiments, the controller can use this composite model for optimal design (e.g., in an off-line implementation of the controller) to determine which additional disinfection strategies are desirable, cost effective, or necessary. The controller can also be used for on-line control to determine control decisions for various controllable equipment (e.g., dampers of the AHU) in real-time to minimize energy consumption or energy costs of the HVAC system while meeting temperature, humidity, and infectious quanta concentration constraints.

The systems and methods described herein treat infection control as an integral part of building HVAC operation rather than a short term or independent control objective, according to some embodiments. While it may be possible to achieve disinfection by the addition of UV lights and filters running at full capacity, such a strategy may be costly and consume excessive amounts of energy. However, the systems and methods described herein couple both objectives (disinfection control and minimal energy consumption) to assess optimal design and operational decisions on a case-by-case basis also taking into account climate, energy and disinfection goals of particular buildings.

The controller can be implemented in an off-line mode as a design tool. With the emergence of various strategies for building disinfection, building designers and operators now have a wide array of options for retrofitting a building to reduce the spread of infectious diseases to building occupants. This is typically accomplished by lowering the concentration of infectious particles in the air space, which can be accomplished by killing the microbes via UV radiation, trapping them via filtration, or simply forcing them out of the building via fresh-air ventilation. While any one of these strategies individually can provide desired levels of disinfection, it may do so at unnecessarily high cost or with negative consequences for thermal comfort of building occupants. Thus, to help evaluate the tradeoff and potential synergies between the various disinfection options, the model-based design tool can estimate annualized capital and energy costs for a given set of disinfection equipment. For a given AHU, this includes dynamic models for temperature, humidity, and infectious particle concentration, and it employs the Wells-Riley infection equation, or any infection risk management standard, such as ASHRAE 241, to enforce constraints on maximum occupant infection probability. By being able to quickly simulate a variety of simulation instances, the controller (when operating as the design tool in the off-line mode) can present building designers with the tradeoff between cost and disinfection, allowing them to make informed decisions about retrofit.

A key feature of the design tool is that it shows to what extent the inherent flexibility of the existing HVAC system can be used to provide disinfection. In particular, in months when infectivity is of biggest concern, a presence of free cooling from fresh outdoor air means that the energy landscape is relatively flat regardless of how the controller determines to operate the HVAC system. Thus, the controller could potentially increase fresh-air intake significantly to provide sufficient disinfection without UV or advanced filtration while incurring only a small energy penalty. The design tool can provide estimates to customers to allow them to make informed decisions about what additional disinfection equipment (if any) to install and then provide the modified control systems needed to implement the desired infection control.

The controller can also be implemented in an on-line mode as a real-time controller. Although equipment like UV lamps and advanced filtration can be installed in buildings to mitigate the spread of infectious diseases, it is often unclear how to best operate that equipment to achieve desired disinfection goals in a cost-effective manner. A common strategy is to take the robust approach of opting for the highest-efficiency filters and running UV lamps constantly. While this strategy will indeed reduce infection probability to its lowest possible value, it is likely to do so at exorbitant cost due to the constant energy penalties of both strategies. Building managers may potentially choose to completely disable filters and UV lamps to conserve energy consumption. Thus, the building may end up in a worst-of-both-words situation where the building manager has paid for disinfection equipment but the zones are no longer receiving any disinfection. To remove this burden from building operators, the controller can automate infection control by integrating disinfection control (e.g., based on the Wells-Riley equation or any infection risk management standard, such as ASHRAE 241) in a model based control scheme. In this way, the controller can simultaneously achieve thermal comfort and provide adequate disinfection at the lowest possible cost given currently available equipment.

Advantageously, the control strategy can optimize in real time the energy and disinfection tradeoffs of all possible control variables. Specifically, the controller may choose to raise fresh-air intake fraction even though it incurs a slight energy penalty because it allows a significant reduction of infectious particle concentrations while still maintaining comfortable temperatures. Thus, in some climates it may be possible to provide disinfection without additional equipment, but this strategy is only possible if the existing control infrastructure can be guided or constrained so as to provide desired disinfection. Alternatively, in buildings that have chosen to add UV lamps and/or filtration, the controller can find the optimal combination of techniques to achieve desired control objectives at the lowest possible cost. In addition, because the constraint on infection probability is configurable, the controller can empower building operators to make their own choices regarding disinfection and energy use (e.g. opting for a loose constraint in the summer when disease is rare and energy use is intensive, while transitioning to a tight constraint in winter when disease is prevalent and energy less of a concern). Advantageously, the controller can provide integrated comfort, disinfection, and energy management to customers to achieve better outcomes in all three areas compared to other narrow and individual solutions.

In some embodiments, the models used to predict temperature, humidity, and/or infectious quanta are dynamic models. The term “dynamic model” and variants thereof (e.g., dynamic temperature model, dynamic humidity model, dynamic infectious quanta model, etc.) are used throughout the present disclosure to refer to any type of model that predicts the value of a quantity (e.g., temperature, humidity, infectious quanta) at various points in time as a function of zero or more input variables. A dynamic model may be “dynamic” as a result of the input variables changing over time even if the model itself does not change. For example, a steady-state model that uses ambient temperature or any other variable that changes over time as an input may be considered a dynamic model. Dynamic models may also include models that vary over time. For example, models that are retrained periodically, configured to adapt to changing conditions over time, and/or configured to use different relationships between input variables and predicted outputs (e.g., a first set of relationships for winter months and a second set of relationships for summer months) may also be considered dynamic models. Dynamic models may also include ordinary differential equation (ODE) models or other types of models having input variables that change over time and/or input variables that represent the rate of change of a variable.

Building Management System and HVAC System

Referring now to FIG. 1, a perspective view of a building 10 is shown. Building 10 can be served by a building management system (BMS). A BMS is, in general, a system of devices configured to control, monitor, and manage equipment in or around a building or building area. A BMS can include, for example, a HVAC system, a security system, a lighting system, a fire alerting system, any other system that is capable of managing building functions or devices, or any combination thereof. An example of a BMS which can be used to monitor and control building 10 is described in U.S. patent application Ser. No. 14/717,593 filed May 20, 2015, the entire disclosure of which is incorporated by reference herein.

The BMS that serves building 10 may include a HVAC system 100. HVAC system 100 can include a plurality of HVAC devices (e.g., heaters, chillers, air handling units, pumps, fans, thermal energy storage, etc.) configured to provide heating, cooling, ventilation, or other services for building 10. For example, HVAC system 100 is shown to include a waterside system 120 and an airside system 130. Waterside system 120 may provide a heated or chilled fluid to an air handling unit of airside system 130. Airside system 130 may use the heated or chilled fluid to heat or cool an airflow provided to building 10. In some embodiments, waterside system 120 can be replaced with or supplemented by a central plant or central energy facility (described in greater detail with reference to FIG. 2). An example of an airside system which can be used in HVAC system 100 is described in greater detail with reference to FIG. 2.

HVAC system 100 is shown to include a chiller 102, a boiler 104, and a rooftop air handling unit (AHU) 106. Waterside system 120 may use boiler 104 and chiller 102 to heat or cool a working fluid (e.g., water, glycol, etc.) and may circulate the working fluid to AHU 106. In various embodiments, the HVAC devices of waterside system 120 can be located in or around building 10 (as shown in FIG. 1) or at an offsite location such as a central plant (e.g., a chiller plant, a steam plant, a heat plant, etc.). The working fluid can be heated in boiler 104 or cooled in chiller 102, depending on whether heating or cooling is required in building 10. Boiler 104 may add heat to the circulated fluid, for example, by burning a combustible material (e.g., natural gas) or using an electric heating element. Chiller 102 may place the circulated fluid in a heat exchange relationship with another fluid (e.g., a refrigerant) in a heat exchanger (e.g., an evaporator) to absorb heat from the circulated fluid. The working fluid from chiller 102 and/or boiler 104 can be transported to AHU 106 via piping 108.

AHU 106 may place the working fluid in a heat exchange relationship with an airflow passing through AHU 106 (e.g., via one or more stages of cooling coils and/or heating coils). The airflow can be, for example, outside air, return air from within building 10, or a combination of both. AHU 106 may transfer heat between the airflow and the working fluid to provide heating or cooling for the airflow. For example, AHU 106 can include one or more fans or blowers configured to pass the airflow over or through a heat exchanger containing the working fluid. The working fluid may then return to chiller 102 or boiler 104 via piping 110.

Airside system 130 may deliver the airflow supplied by AHU 106 (i.e., the supply airflow) to building 10 via air supply ducts 112 and may provide return air from building 10 to AHU 106 via air return ducts 114. In some embodiments, airside system 130 includes multiple variable air volume (VAV) units 116. For example, airside system 130 is shown to include a separate VAV unit 116 on each floor or zone of building 10. VAV units 116 can include dampers or other flow control elements that can be operated to control an amount of the supply airflow provided to individual zones of building 10. In other embodiments, airside system 130 delivers the supply airflow into one or more zones of building 10 (e.g., via supply ducts 112) without using intermediate VAV units 116 or other flow control elements. AHU 106 can include various sensors (e.g., temperature sensors, pressure sensors, etc.) configured to measure attributes of the supply airflow. AHU 106 may receive input from sensors located within AHU 106 and/or within the building zone and may adjust the flow rate, temperature, or other attributes of the supply airflow through AHU 106 to achieve setpoint conditions for the building zone.

Airside System

Referring now to FIG. 2, a block diagram of an airside system 200 is shown, according to some embodiments. In various embodiments, airside system 200 may supplement or replace airside system 130 in HVAC system 100 or can be implemented separate from HVAC system 100. When implemented in HVAC system 100, airside system 200 can include a subset of the HVAC devices in HVAC system 100 (e.g., AHU 106, VAV units 116, ducts 112-114, fans, dampers, etc.) and can be located in or around building 10. Airside system 200 may operate to heat, cool, humidify, dehumidify, filter, and/or disinfect an airflow provided to building 10 in some embodiments.

Airside system 200 is shown to include an economizer-type air handling unit (AHU) 202. Economizer-type AHUs vary the amount of outside air and return air used by the air handling unit for heating or cooling. For example, AHU 202 may receive return air 204 from building zone 206 via return air duct 208 and may deliver supply air 210 to building zone 206 via supply air duct 212. In some embodiments, AHU 202 is a rooftop unit located on the roof of building 10 (e.g., AHU 106 as shown in FIG. 1) or otherwise positioned to receive both return air 204 and outside air 214. AHU 202 can be configured to operate exhaust air damper 216, mixing damper 218, and outside air damper 220 to control an amount of outside air 214 and return air 204 that combine to form supply air 210. Any return air 204 that does not pass through mixing damper 218 can be exhausted from AHU 202 through exhaust damper 216 as exhaust air 222.

Each of dampers 216-220 can be operated by an actuator. For example, exhaust air damper 216 can be operated by actuator 224, mixing damper 218 can be operated by actuator 226, and outside air damper 220 can be operated by actuator 228. Actuators 224-228 may communicate with an AHU controller 230 via a communications link 232. Actuators 224-228 may receive control signals from AHU controller 230 and may provide feedback signals to AHU controller 230. Feedback signals can include, for example, an indication of a current actuator or damper position, an amount of torque or force exerted by the actuator, diagnostic information (e.g., results of diagnostic tests performed by actuators 224-228), status information, commissioning information, configuration settings, calibration data, and/or other types of information or data that can be collected, stored, or used by actuators 224-228. AHU controller 230 can be an economizer controller configured to use one or more control algorithms (e.g., state-based algorithms, extremum seeking control (ESC) algorithms, proportional-integral (PI) control algorithms, proportional-integral-derivative (PID) control algorithms, model predictive control (MPC) algorithms, feedback control algorithms, etc.) to control actuators 224-228.

Still referring to FIG. 2, AHU 202 is shown to include a cooling coil 234, a heating coil 236, and a fan 238 positioned within supply air duct 212. Fan 238 can be configured to force supply air 210 through cooling coil 234 and/or heating coil 236 and provide supply air 210 to building zone 206. AHU controller 230 may communicate with fan 238 via communications link 240 to control a flow rate of supply air 210. In some embodiments, AHU controller 230 controls an amount of heating or cooling applied to supply air 210 by modulating a speed of fan 238. In some embodiments, AHU 202 includes one or more air filters (e.g., filter 708) and/or one or more ultraviolet (UV) lights (e.g., UV lights 706) as described in greater detail with reference to FIG. 7. AHU controller 230 can be configured to control the UV lights and route the airflow through the air filters to disinfect the airflow as described in greater detail below.

Cooling coil 234 may receive a chilled fluid from central plant 200 (e.g., from cold water loop 216) via piping 242 and may return the chilled fluid to central plant 200 via piping 244. Valve 246 can be positioned along piping 242 or piping 244 to control a flow rate of the chilled fluid through cooling coil 234. In some embodiments, cooling coil 234 includes multiple stages of cooling coils that can be independently activated and deactivated (e.g., by AHU controller 230, by BMS controller 266, etc.) to modulate an amount of cooling applied to supply air 210.

Heating coil 236 may receive a heated fluid from central plant 200 (e.g., from hot water loop 214) via piping 248 and may return the heated fluid to central plant 200 via piping 250. Valve 252 can be positioned along piping 248 or piping 250 to control a flow rate of the heated fluid through heating coil 236. In some embodiments, heating coil 236 includes multiple stages of heating coils that can be independently activated and deactivated (e.g., by AHU controller 230, by BMS controller 266, etc.) to modulate an amount of heating applied to supply air 210.

Each of valves 246 and 252 can be controlled by an actuator. For example, valve 246 can be controlled by actuator 254 and valve 252 can be controlled by actuator 256. Actuators 254-256 may communicate with AHU controller 230 via communications links 258-260. Actuators 254-256 may receive control signals from AHU controller 230 and may provide feedback signals to controller 230. In some embodiments, AHU controller 230 receives a measurement of the supply air temperature from a temperature sensor 262 positioned in supply air duct 212 (e.g., downstream of cooling coil 234 and/or heating coil 236). AHU controller 230 may also receive a measurement of the temperature of building zone 206 from a temperature sensor 264 located in building zone 206.

In some embodiments, AHU controller 230 operates valves 246 and 252 via actuators 254-256 to modulate an amount of heating or cooling provided to supply air 210 (e.g., to achieve a setpoint temperature for supply air 210 or to maintain the temperature of supply air 210 within a setpoint temperature range). The positions of valves 246 and 252 affect the amount of heating or cooling provided to supply air 210 by cooling coil 234 or heating coil 236 and may correlate with the amount of energy consumed to achieve a desired supply air temperature. AHU controller 230 may control the temperature of supply air 210 and/or building zone 206 by activating or deactivating coils 234-236, adjusting a speed of fan 238, or a combination of both.

Still referring to FIG. 2, airside system 200 is shown to include a building management system (BMS) controller 266 and a client device 268. BMS controller 266 can include one or more computer systems (e.g., servers, supervisory controllers, subsystem controllers, etc.) that serve as system level controllers, application or data servers, head nodes, or master controllers for airside system 200, central plant 200, HVAC system 100, and/or other controllable systems that serve building 10. BMS controller 266 may communicate with multiple downstream building systems or subsystems (e.g., HVAC system 100, a security system, a lighting system, central plant 200, etc.) via a communications link 270 according to like or disparate protocols (e.g., LON, BACnet, etc.). In various embodiments, AHU controller 230 and BMS controller 266 can be separate (as shown in FIG. 2) or integrated. In an integrated implementation, AHU controller 230 can be a software module configured for execution by a processor of BMS controller 266.

In some embodiments, AHU controller 230 receives information from BMS controller 266 (e.g., commands, setpoints, operating boundaries, etc.) and provides information to BMS controller 266 (e.g., temperature measurements, valve or actuator positions, operating statuses, diagnostics, etc.). For example, AHU controller 230 may provide BMS controller 266 with temperature measurements from temperature sensors 262-264, equipment on/off states, equipment operating capacities, and/or any other information that can be used by BMS controller 266 to monitor or control a variable state or condition within building zone 206.

Client device 268 can include one or more human-machine interfaces or client interfaces (e.g., graphical user interfaces, reporting interfaces, text-based computer interfaces, client-facing web services, web servers that provide pages to web clients, etc.) for controlling, viewing, or otherwise interacting with HVAC system 100, its subsystems, and/or devices. Client device 268 can be a computer workstation, a client terminal, a remote or local interface, or any other type of user interface device. Client device 268 can be a stationary terminal or a mobile device. For example, client device 268 can be a desktop computer, a computer server with a user interface, a laptop computer, a tablet, a smartphone, a PDA, or any other type of mobile or non-mobile device. Client device 268 may communicate with BMS controller 266 and/or AHU controller 230 via communications link 272.

Referring now to FIG. 3, a block diagram of a building automation system (BAS) 300 is shown, according to an exemplary embodiment. BAS 300 can be implemented in building 10 to automatically monitor and control various building functions. BAS 300 is shown to include BAS controller 302 and building subsystems 328. Building subsystems 328 are shown to include a building electrical subsystem 334, an information communication technology (ICT) subsystem 336, a security subsystem 338, a HVAC subsystem 340, a lighting subsystem 342, a lift/escalators subsystem 332, and a fire safety subsystem 330. In various embodiments, building subsystems 328 can include fewer, additional, or alternative subsystems. For example, building subsystems 328 can also or alternatively include a refrigeration subsystem, an advertising or signage subsystem, a cooking subsystem, a vending subsystem, a printer or copy service subsystem, or any other type of building subsystem that uses controllable equipment and/or sensors to monitor or control building 10. In some embodiments, building subsystems 328 include a waterside system and/or an airside system. A waterside system and an airside system are described with further reference to U.S. patent application Ser. No. 15/631,830 filed Jun. 23, 2017, the entirety of which is incorporated by reference herein.

Each of building subsystems 328 can include any number of devices, controllers, and connections for completing its individual functions and control activities. HVAC subsystem 340 can include many of the same components as HVAC system 100, as described with reference to FIG. 1. For example, HVAC subsystem 340 can include a chiller, a boiler, any number of air handling units, economizers, field controllers, supervisory controllers, actuators, temperature sensors, and other devices for controlling the temperature, humidity, airflow, or other variable conditions within building 10. Lighting subsystem 342 can include any number of light fixtures, ballasts, lighting sensors, dimmers, or other devices configured to controllably adjust the amount of light provided to a building space. Security subsystem 338 can include occupancy sensors, video surveillance cameras, digital video recorders, video processing servers, intrusion detection devices, access control devices and servers, or other security-related devices.

Still referring to FIG. 3, BAS controller 302 is shown to include a communications interface 307 and a BAS interface 309. Interface 307 can facilitate communications between BAS controller 302 and external applications (e.g., monitoring and reporting applications 322, enterprise control applications 326, remote systems and applications 344, applications residing on client devices 348, etc.) for allowing user control, monitoring, and adjustment to BAS controller 302 and/or subsystems 328. Interface 307 can also facilitate communications between BAS controller 302 and client devices 348. BAS interface 309 can facilitate communications between BAS controller 302 and building subsystems 328 (e.g., HVAC, lighting security, lifts, power distribution, business, etc.).

Interfaces 307, 309 can be or include wired or wireless communications interfaces (e.g., jacks, antennas, transmitters, receivers, transceivers, wire terminals, etc.) for conducting data communications with building subsystems 328 or other external systems or devices. In various embodiments, communications via interfaces 307, 309 can be direct (e.g., local wired or wireless communications) or via a communications network 346 (e.g., a WAN, the Internet, a cellular network, etc.). For example, interfaces 307, 309 can include an Ethernet card and port for sending and receiving data via an Ethernet-based communications link or network. In another example, interfaces 307, 309 can include a Wi-Fi transceiver for communicating via a wireless communications network. In another example, one or both of interfaces 307, 309 can include cellular or mobile phone communications transceivers. In one embodiment, communications interface 307 is a power line communications interface and BAS interface 309 is an Ethernet interface. In other embodiments, both communications interface 307 and BAS interface 309 are Ethernet interfaces or are the same Ethernet interface.

Still referring to FIG. 3, BAS controller 302 is shown to include a processing circuit 304 including a processor 306 and memory 308. Processing circuit 304 can be communicably connected to BAS interface 309 and/or communications interface 307 such that processing circuit 304 and the various components thereof can send and receive data via interfaces 307, 309. Processor 306 can be implemented as a general purpose processor, an application specific integrated circuit (ASIC), one or more field programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components.

Memory 308 (e.g., memory, memory unit, storage device, etc.) can include one or more devices (e.g., RAM, ROM, Flash memory, hard disk storage, etc.) for storing data and/or computer code for completing or facilitating the various processes, layers and modules described in the present application. Memory 308 can be or include volatile memory or non-volatile memory. Memory 308 can include database components, object code components, script components, or any other type of information structure for supporting the various activities and information structures described in the present application. According to an exemplary embodiment, memory 308 is communicably connected to processor 306 via processing circuit 304 and includes computer code for executing (e.g., by processing circuit 304 and/or processor 306) one or more processes described herein.

In some embodiments, BAS controller 302 is implemented within a single computer (e.g., one server, one housing, etc.). In various other embodiments BAS controller 302 can be distributed across multiple servers or computers (e.g., that can exist in distributed locations). Further, while FIG. 3 shows applications 322 and 326 as existing outside of BAS controller 302, in some embodiments, applications 322 and 326 can be hosted within BAS controller 302 (e.g., within memory 308).

Still referring to FIG. 3, memory 308 is shown to include an enterprise integration layer 310, an automated measurement and validation (AM&V) layer 312, a demand response (DR) layer 314, a fault detection and diagnostics (FDD) layer 316, an integrated control layer 318, and a building subsystem integration later 320. Layers 310-320 is configured to receive inputs from building subsystems 328 and other data sources, determine optimal control actions for building subsystems 328 based on the inputs, generate control signals based on the optimal control actions, and provide the generated control signals to building subsystems 328 in some embodiments. The following paragraphs describe some of the general functions performed by each of layers 310-320 in BAS 300.

Enterprise integration layer 310 can be configured to serve clients or local applications with information and services to support a variety of enterprise-level applications. For example, enterprise control applications 326 can be configured to provide subsystem-spanning control to a graphical user interface (GUI) or to any number of enterprise-level business applications (e.g., accounting systems, user identification systems, etc.). Enterprise control applications 326 can also or alternatively be configured to provide configuration GUIs for configuring BAS controller 302. In yet other embodiments, enterprise control applications 326 can work with layers 310-320 to optimize building performance (e.g., efficiency, energy use, comfort, or safety) based on inputs received at interface 307 and/or BAS interface 309.

Building subsystem integration layer 320 can be configured to manage communications between BAS controller 302 and building subsystems 328. For example, building subsystem integration layer 320 can receive sensor data and input signals from building subsystems 328 and provide output data and control signals to building subsystems 328. Building subsystem integration layer 320 can also be configured to manage communications between building subsystems 328. Building subsystem integration layer 320 translate communications (e.g., sensor data, input signals, output signals, etc.) across multi-vendor/multi-protocol systems.

Demand response layer 314 can be configured to optimize resource usage (e.g., electricity use, natural gas use, water use, etc.) and/or the monetary cost of such resource usage in response to satisfy the demand of building 10. The optimization can be based on time-of-use prices, curtailment signals, energy availability, or other data received from utility providers, distributed energy generation systems 324, from energy storage 327, or from other sources. Demand response layer 314 can receive inputs from other layers of BAS controller 302 (e.g., building subsystem integration layer 320, integrated control layer 318, etc.). The inputs received from other layers can include environmental or sensor inputs such as temperature, carbon dioxide levels, relative humidity levels, air quality sensor outputs, occupancy sensor outputs, room schedules, and the like. The inputs can also include inputs such as electrical use (e.g., expressed in kWh), thermal load measurements, pricing information, projected pricing, smoothed pricing, curtailment signals from utilities, and the like.

According to an exemplary embodiment, demand response layer 314 includes control logic for responding to the data and signals it receives. These responses can include communicating with the control algorithms in integrated control layer 318, changing control strategies, changing setpoints, or activating/deactivating building equipment or subsystems in a controlled manner. Demand response layer 314 can also include control logic configured to determine when to utilize stored energy. For example, demand response layer 314 can determine to begin using energy from energy storage 327 just prior to the beginning of a peak use hour.

In some embodiments, demand response layer 314 includes a control module configured to actively initiate control actions (e.g., automatically changing setpoints) which minimize energy costs or energy consumptions based on one or more inputs representative of or based on demand (e.g., price, a curtailment signal, a demand level, etc.). In some embodiments, demand response layer 314 uses equipment models to determine an optimal set of control actions. The equipment models can include, for example, thermodynamic models describing the inputs, outputs, and/or functions performed by various sets of building equipment. Equipment models can represent collections of building equipment (e.g., subplants, chiller arrays, etc.) or individual devices (e.g., individual chillers, heaters, pumps, etc.).

Demand response layer 314 can further include or draw upon one or more demand response policy definitions (e.g., databases, XML files, etc.). The policy definitions can be edited or adjusted by a user (e.g., via a graphical user interface) so that the control actions initiated in response to demand inputs can be tailored for the user's application, desired comfort level, particular building equipment, or based on other concerns. For example, the demand response policy definitions can specify which equipment can be turned on or off in response to particular demand inputs, how long a system or piece of equipment should be turned off, what setpoints can be changed, what the allowable setpoint adjustment range is, how long to hold a high demand setpoint before returning to a normally scheduled setpoint, how close to approach capacity limits, which equipment modes to utilize, the energy transfer rates (e.g., the maximum rate, an alarm rate, other rate boundary information, etc.) into and out of energy storage devices (e.g., thermal storage tanks, battery banks, etc.), and when to dispatch on-site generation of energy (e.g., via fuel cells, a motor generator set, etc.).

Integrated control layer 318 can be configured to use the data input or output of building subsystem integration layer 320 and/or demand response later 314 to make control decisions. Due to the subsystem integration provided by building subsystem integration layer 320, integrated control layer 318 can integrate control activities of the subsystems 328 such that the subsystems 328 behave as a single integrated supersystem. In an exemplary embodiment, integrated control layer 318 includes control logic that uses inputs and outputs from building subsystems to provide greater comfort and energy savings relative to the comfort and energy savings that separate subsystems could provide alone. For example, integrated control layer 318 can be configured to use an input from a first subsystem to make an energy-saving control decision for a second subsystem. Results of these decisions can be communicated back to building subsystem integration layer 320.

Integrated control layer 318 is shown to be logically below demand response layer 314. Integrated control layer 318 can be configured to enhance the effectiveness of demand response layer 314 by enabling building subsystems 328 and their respective control loops to be controlled in coordination with demand response layer 314. This configuration can reduce disruptive demand response behavior relative to conventional systems. For example, integrated control layer 318 can be configured to assure that a demand response-driven upward adjustment to the setpoint for chilled water temperature (or another component that directly or indirectly affects temperature) does not result in an increase in fan energy (or other energy used to cool a space) that would result in greater total building energy use than was saved at the chiller.

Integrated control layer 318 can be configured to provide feedback to demand response layer 314 so that demand response layer 314 checks that constraints (e.g., temperature, lighting levels, etc.) are properly maintained even while demanded load shedding is in progress. The constraints can also include setpoint or sensed boundaries relating to safety, equipment operating limits and performance, comfort, fire codes, electrical codes, energy codes, and the like. Integrated control layer 318 is also logically below fault detection and diagnostics layer 316 and automated measurement and validation layer 312. Integrated control layer 318 can be configured to provide calculated inputs (e.g., aggregations) to these higher levels based on outputs from more than one building subsystem.

Automated measurement and validation (AM&V) layer 312 can be configured to verify that control strategies commanded by integrated control layer 318 or demand response layer 314 are working properly (e.g., using data aggregated by AM&V layer 312, integrated control layer 318, building subsystem integration layer 320, FDD layer 316, or otherwise). The calculations made by AM&V layer 312 can be based on building system energy models and/or equipment models for individual BAS devices or subsystems. For example, AM&V layer 312 can compare a model-predicted output with an actual output from building subsystems 328 to determine an accuracy of the model.

Fault detection and diagnostics (FDD) layer 316 can be configured to provide on-going fault detection for building subsystems 328, building subsystem devices (i.e., building equipment), and control algorithms used by demand response layer 314 and integrated control layer 318. FDD layer 316 can receive data inputs from integrated control layer 318, directly from one or more building subsystems or devices, or from another data source. FDD layer 316 can automatically diagnose and respond to detected faults. The responses to detected or diagnosed faults can include providing an alarm message to a user, a maintenance scheduling system, or a control algorithm configured to attempt to repair the fault or to work-around the fault.

FDD layer 316 can be configured to output a specific identification of the faulty component or cause of the fault (e.g., loose damper linkage) using detailed subsystem inputs available at building subsystem integration layer 320. In other exemplary embodiments, FDD layer 316 is configured to provide “fault” events to integrated control layer 318 which executes control strategies and policies in response to the received fault events. According to an exemplary embodiment, FDD layer 316 (or a policy executed by an integrated control engine or business rules engine) can shut-down systems or direct control activities around faulty devices or systems to reduce energy waste, extend equipment life, or assure proper control response.

FDD layer 316 can be configured to store or access a variety of different system data stores (or data points for live data). FDD layer 316 can use some content of the data stores to identify faults at the equipment level (e.g., specific chiller, specific AHU, specific terminal unit, etc.) and other content to identify faults at component or subsystem levels. For example, building subsystems 328 can generate temporal (i.e., time-series) data indicating the performance of BAS 300 and the various components thereof. The data generated by building subsystems 328 can include measured or calculated values that exhibit statistical characteristics and provide information about how the corresponding system or process (e.g., a temperature control process, a flow control process, etc.) is performing in terms of error from its setpoint. These processes can be examined by FDD layer 316 to expose when the system begins to degrade in performance and alarm a user to repair the fault before it becomes more severe.

Referring now to FIG. 4, a system 400 including a building analysis system 404 that generates recommendations and air quality occupant impact assessments for spaces of an office building 401 based on air quality measurements of sensors 402 is shown, according to an exemplary embodiment. In some embodiments, a technician can install the sensors 402 in the office building 401 on a temporary basis, e.g., for one week, two weeks, a month, etc. In some embodiments, the sensors 402 can be permanently installed so that the environment can be permanently monitored. The sensors 402 can be spread out through various spaces of the office building 401 in order to record air quality measurements of each space of the office building 401. In some embodiments, the office building 401 can be a school building, a commercial building, an apartment building, a hospital, etc. The systems and methods discussed herein can be applied to various types of buildings and are not limited to office buildings.

Each sensor of the temporary air quality sensors 402 can measure one or multiple air quality metrics, e.g., can include one sensors or a set of sensors. For example, the sensors 402 can measure ventilation for a space, occupancy for a space, CO2 for a space, particulate matter PM1, particulate matter PM10 for a space, particulate matter PM2.5 for a space, volatile organic compounds (VOC) for the space, total volatile organic compound (TVOC) for the space, thermal measurements for the space, temperature for the space, relative humidity for the space, dew point for the space, ozone for the space, carbon monoxide (CO) for the space, formaldehyde for the space, acetone for the space, mold/mildew for the space, pollen for the space, bacteria for the space, microbial flora (including allergens), viruses for the space and/or any other types of metrics/characteristics for the space. In some embodiments, the sensors 402 are permanent sensors that are installed in a permanent manner. In this regard, if the sensors 402 are permanent, the air quality occupant impact assessments and/or recommendations can be generated over a requested period of time, e.g., a particular day, week, year, etc.

The measurements of the sensors 402 can be communicated to a cloud platform that can perform an analysis on the air quality measurements of the various spaces of the office building 401. For example, the sensors 402 can be wireless sensors (or wired sensors) that communicate across a network 414 which may include local networks within the office building 401 and/or external networks. For example, various routers, switches, servers, cellular towers, LAN networks, WAN networks, Wi-Fi networks, etc. can be included within the network 414 and can communicate the measurements of the sensors 402 to the building analysis system 404.

In some embodiments one or more canisters may be used to collect the air quality metrics. For example, the canisters may collect data regarding ventilation for a space, occupancy for a space, CO2 for a space, particulate matter PM1, particulate matter PM10 for a space, particulate matter PM2.5 for a space, volatile organic compounds (VOC) for the space, total volatile organic compound (TVOC) for the space, thermal measurements for the space, temperature for the space, relative humidity for the space, dew point for the space, ozone for the space, carbon monoxide (CO) for the space, formaldehyde for the space, acetone for the space, mold/mildew for the space, pollen for the space, bacteria for the space, microbial flora (including allergens), viruses for the space and/or any other types of metrics/characteristics for the space. Additionally, the contents of the canisters may be provided to the temporary air quality sensors 402. In some embodiments, the temporary air quality sensors 402 may take measurements of the contents of the canisters. The measurements relating to the contents of the canisters may be analyzed by communicating the measurements to a remote (e.g., on-premises or off-premises) system, such as the cloud platform described herein. In some embodiments, the measurements relating to the contents of the canisters may be communicated to the building analysis system 404.

Furthermore, information describing physical characteristics of the office building 401 and various spaces of the office building 401 can be provided to the building analysis system 404 via a mobile application of a user device 412, a web browser of the user device 412, and/or any another application of the user device 412. The information can be manually collected site data, photos of the office building 401, equipment information of the office building 401, schematic diagrams of the office building 401, user information, desired metrics, desired building performance, floor plans of the spaces assessed via the sensors 402, AHU zone maps indicating each AHU and the spaces the AHUs serve, an AHU list/schedule indicating lists of AHUs with sizes and service information, etc. The user device 412 can be a smartphone, a tablet, a laptop computer, a desktop computer, etc. The user device 412 can communicate with the building analysis system 404 via the network 414.

The building analysis system 404 can be a cloud based system, a remote system, a local on-premises system with the office building 401, etc. The building analysis system 404 can include processors 406 and/or memory devices 408. Processors 406 can be implemented as a general purpose processor, an application specific integrated circuit (ASIC), one or more field programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components.

Memory devices 408 (e.g., memory, memory unit, storage device, etc.) can include one or more devices (e.g., RAM, ROM, Flash memory, hard disk storage, etc.) for storing data and/or computer code for completing or facilitating the various processes, layers and modules described in the present application. Memory devices 408 can be or include volatile memory or non-volatile memory. Memory devices 408 can include database components, object code components, script components, or any other type of information structure for supporting the various activities and information structures described in the present application. According to an exemplary embodiment, memory devices 408 is communicably connected to processors 406 via the processors 406 and includes computer code for executing (e.g., by the processors 406) one or more processes described herein.

The building analysis system 404 includes a space air quality analyzer 411, a recommendation generator 407, and an air quality occupant impact assessment generator 410. The space air quality analyzer 411 can record measurements of the various sensors 402 and create air quality profiles of the various spaces of the office building 401. For example, the space air quality analyzer 411 can record air quality for the spaces and generate a trend over time (e.g., timeseries data or other time correlated data) in which the temporary sensors 402 are installed, e.g., over the two weeks that the sensors 402 are installed. The trends created by the space air quality analyzer 411 are described with further reference to U.S. Provisional Application No. 63/230,608 filed Aug. 6, 2021, the entirety of which is incorporated by reference herein. The trends created by the space air quality analyzer 411 can be provided to the recommendation generator 407 and the air quality occupant impact assessment generator 410.

In some embodiments, the space air quality analyzer 411 can generate space hierarchy air quality information. For example, rooms, hallways, and closets may be basic units of space in the office building 401. However, a hierarchy of spaces can be built from the basic space unit. For example, a group of rooms could form a zone of a floor. A group of zones could form a floor of a building. A group of floors could form a building of a campus. The space air quality analyzer 411 can generate higher level space air quality metrics for a particular space based on the basic space units that make up the particular space. For example, a CO2 metric for a floor could be generated by averaging the CO2 metrics for all rooms that make up the floor. Similarly, the CO2 metrics for a building could be made based on averaging CO2 metrics for all the floors of the building. In some embodiments, the metrics may be used to generate productivity scores for the spaces (e.g., the rooms themselves, floors/buildings/campuses that include the rooms, etc.). In some embodiments, the space productivity scores may be specific to air quality. In some embodiments, the metrics may be used in combination with other metrics to generate an overall space productivity score. In some embodiments, the air quality metrics may be used in combination to generate a combined air quality productivity score, and that score may in turn be used as a component score to generate an overall space/building productivity score that includes air quality as a component. Examples of such features that may be used in conjunction with the features of the present disclosure can be found in U.S. patent application Ser. No. 17/354,583, filed Jun. 22, 2021, and Ser. No. 17/354,565, filed Jun. 22, 2021, both of which are incorporated herein by reference in their entireties. For example, in some embodiments, the productivity score may be presented within similar interfaces and/or could be a component score of a healthy person score and/or overall building health score as described in accordance with various example embodiments therein. Similarly, in some embodiments the recommendations may be provided using a framework similar to that shown therein.

In some embodiments, the productivity scores may be generated by taking the air quality measurements and comparing the air quality measurements to stored data correlating the air quality measurements to predictive occupant productivity. In some embodiments, correlating the air quality measurements to predictive occupant productivity may be done by way of a lookup table. In some embodiments, a mechanism, that directly correlates the air quality measurements to predictive occupant productivity, may involve using weighted combinations of the air quality measurements and/or the predictive occupant productivity. In some embodiments, a model and/or other type of algorithm may be used that takes the air quality measurements as inputs. In some embodiments, the model may be generated using historical data for the building and/or other buildings (e.g., sharing similar characteristics).

In some embodiments, the productivity scores may be occupant productivity scores. Additionally, an occupant productivity score may provide an expected occupant productivity of one or more occupants (e.g., a relative expected occupant productivity, such as on a relative scale, e.g., via a normalized score such as from 1-10). In some embodiments, occupant productivity may be determined by an occupant's work output in a given timeframe. For example, occupant productivity may be the total number of tasks an occupant completed within one work day, wherein the number of completed tasks may be the work output and one work day may be the given timeframe. Additionally, if the occupant productivity were to increase, the number of tasks completed within one work day may increase. Similarly, if occupant productivity were to decrease, the number of tasks completed within one work day may decrease.

The air quality occupant impact assessment generator 410 may generate reports that summarize the air quality trends of the spaces of the building and/or include recommendations. The report generated by the air quality occupant impact assessment generator 410 may provide the report to the user device 412 for review by a user. The report may further indicate areas of the office building 401, recommendations for improving indoor air quality (IAQ), recommendations for saving energy in the office building 401, etc. In some embodiments, the report is a user interface including various charts, graphs, trends, recommendations, or other information. The interface may be displayed on a display device of the user device 412.

The air quality occupant impact assessment generator 410 can generate a report including recommendations generated by the recommendation generator 407 indicating actionable data that can be implemented by the building analysis system 404 and/or a BMS system of the office building 401 (e.g., the BMS system described in FIGS. 1 and 3). The recommendations can indicate a recommendation to improve ventilation in a room of the office building 401 (e.g., operate ventilation equipment to increase a ventilation rate), can recommend opportunities for energy savings where adjusting ventilation when a space is unoccupied would save energy (e.g., operate ventilation equipment to decrease a ventilation rate), recommendations which identify equipment which could provide better ventilation and/or filtering for spaces (e.g., install VAVs or unit ventilators (UVs) based on which type of equipment is performing better), assessments of adequacy of outdoor air filtration, recommendation to filter mixed air, etc. The report generated by the air quality occupant impact assessment generator 410 can include a summary indicating key findings, testing details, testing results, photographs, conclusions, recommendations, etc.

The report generated by the air quality occupant impact assessment generator 410 can also include a detailed building data summary report that indicates building size and use, recent renovation, special use areas, number of AHU's, filtration type and schedule, air supply system type, and specific areas of concern. The report can indicate a technicians visual inspection of representative AHU's, fan coil units, induction units, filter type/installation/condition, air supply diffusers, exhaust systems, and/or return air grilles. The report can indicate whether air systems of the office building 401 are under proper control, sequence of operations is being followed, and all controls are operating per the desired setpoint and schedule.

The report generated by the air quality occupant impact assessment generator 410 may include an expected amount of occupant absences. The expected amount of occupant absences may be based on the air quality measurements. The expected amount of occupant absences may be given an occupant absence value. The occupant absence value may be compared to a predetermined occupant absence value range. In some embodiments, the productivity scores may be impacted by the occupant absence value. The occupant absence value may be above, below, and/or within the predetermined occupant absence value range. In some embodiments, the productivity scores may decrease as a result of the occupant absence value being above the predetermined occupant absence value range. In some embodiments, the productivity scores may increase as a result of the occupant absence value being above the predetermined absence value range. In some embodiments, the productivity scores may decrease as a result of the occupant absence value being below the predetermined occupant absence value range. In some embodiments the productivity scores may increase as a results of the occupant absence value being below the predetermined occupant absence value range.

The report can include air quality tests of the sensors 402, e.g., CO2, TVOC, CO, PM 1, PM2.5, PM 10, viruses, bacteria, acetone, mold/mildew, pollen, microbial flora, temperature, relative humidity, NO2, SO2, O3, VOC's, airflow vectors, air pressure differentials, etc. The report can indicate a ventilation assessment indicating the results of testing that ensures outside air intake, supply air fan, and/or ventilation system is supplying minimum outdoor air ventilation rate detailed by ASHRAE 62.1-2016. Ventilation needs based on space type, square footage, and occupancy. The report can indicate an infection risk assessment indicating DNA-tagged bioaerosolstracers safely simulate respiratory emissions to identify potential infection hotspots, verify ventilation and filtration system performance for mitigating airborne exposures, and optimize enhancements.

The recommendations generated by the recommendation generator 407 and included within the report generated by the air quality occupant impact assessment generator 410 can further include recommendations to investigate ventilation rates of rooms of the office building 401 with CO2 levels above a particular level (e.g., 1100 ppm). The recommendations can indicate a current ventilation rate of a space along with comparisons to other ventilation rates of other spaces, inconsistencies can indicate that a user should consider adjusting the ventilation rates of the spaces. If all of the ventilation rates are similar, the recommendation can recommend changing a ventilation policy for the entire office building 401. The recommendations could further be to analyze a source of TVOC for a space where TVOC is above a particular amount, investigate a source of VOCs in a space with TVOCs above a particular amount, etc.

The recommendations in some embodiments, can include recommendations to improve ventilation, e.g., diluting dirty air with clean air as available from outside the office building 401. This recommendation can ensure the delivery of ASHRAE required ventilation rates. The recommendations can be recommendations to improve filtration for spaces. Filtration may mechanically remove particles from the air of the space. The recommendation can be a recommendation to increase particle collection with options with filters such as Koch filters, MAC-10 fan filter units, enviro portable HEPA filtration units, etc.

The recommendations can include recommendations for improving disinfection for a space, e.g., deactivating bacteria and/or viruses in the space. The recommendations can be recommendations to install and/or operate disinfectant systems such as disinfectant light systems (e.g., ultraviolet (UV), ultraviolet-C (UVC), etc.). The recommendations can be recommendations to implement isolation of certain spaces of the office building 401. For example, cause one space to be an isolated space that contains particles and prevents the particles going elsewhere in the office building 401. This can be implemented through creating a negative-pressure isolation environments. The recommendations can be recommendations for performing monitoring and maintenance of equipment, e.g., to inspect equipment at a particular frequently and/or track results for maintenance and monitoring to maintain clean air.

In some embodiments, the CO2 measurements of the sensors 402 can be used by the recommendation generator 407 to determine how well a space is being ventilated. If the CO2 levels are higher than particular amounts, a recommendation to increase ventilation can be generated and/or implemented. The TVOC measurements can indicate how safe a space is for human beings and/or animals. If TVOC is above a particular level, an alert can be generated to evacuate the space and/or address the high TVOC level. The PM2.5 levels can indicate how well filtering equipment is operating. If PM2.5 is greater than a particular amount, this may indicate that the space is not being properly filtered and that a filter of equipment serving the space needs to be replaced and/or changed to a higher quality filter.

In some embodiments, the recommendation generator 407 can perform an analysis on equipment type for the spaces. For example, the recommendation generator 407 could analyze that spaces with low PM2.5 use unit ventilators while spaces with high PM2.5 use VAVs. This improvement in performance of the unit ventilators vs. the VAVs can be used in a recommendation for the recommendation generator 407 to recommend that unit ventilators replace the VAVS in the office building 401.

In some embodiments, the recommendation generator 407 could recommend that persons with allergies be assigned to areas of a building with low VOC, TVOC, PM2.5, PM10 levels, pollen, and/or allergens. This may allow the allergenic persons to avoid having an asthma attack or other breathing problems. In some embodiments, scheduling can be set up and/or recommended by the building analysis system 404 such that occupants are not assigned spaces with high VOC, TVOC, PM2.5 levels for a long duration.

In some embodiments, the air quality occupant impact assessment generator 410 may generate one or more productivity scores for the entire office building 401 or one or more spaces or zones of the office building 401. The productivity scores may be generated using any possible combination of one or more of the air quality metrics described herein. For example, the productivity scores may be generated by using the CO2 metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the TVOC metric. Similarly, the productivity scores may be generated by using the CO2 metric in combination with the measured ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the PM1 metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the PM2.5 metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the PM10 metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the VOC metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the thermal metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the temperature metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the relative humidity metric. Similarly, the productivity scores may be generated by using the CO2 metric in combination with the dew point metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the Ozone metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the CO metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the formaldehyde metric.

In some embodiments, the productivity scores may be generated by adding or removing metrics used to generate the productivity scores. For example, the productivity score may be generated by using the CO2 metric in combination with the TVOC metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the PM1 metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the PM2.5 metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the PM10 metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the VOC metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the thermal metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the temperature metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the relative humidity metric and the ventilation metric. Similarly, the productivity scores may be generated by using the CO2 metric in combination with the dew point metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the Ozone metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the CO metric and the ventilation metric. Similarly, the productivity score may be generated by using the CO2 metric in combination with the formaldehyde metric and the ventilation metric. Similarly, the productivity scores may be generated in combination of one or more of the air quality metrics described herein. While the present disclosure discusses various combinations of air quality metrics that may be used to generate the productivity scores, it should be noted that the features of the present disclosure are equally applicable to any one or more air quality metrics and/or any one or more combination of air quality metrics.

In some embodiments, the productivity scores may be generated by applying one or more weights to the air quality metrics described herein. For example, a weight value of X may be applied to the CO2 metric. Additionally, a weight value of Y may be applied to the PM1 metric. The value of X may be equal to the value of Y. Additionally, the value of X may be less than or equal to the value of Y. Additionally, the value of X may be greater than or equal to the value of Y. Similarly, the value of Y may be equal to the value of X. Additionally, the value of Y may be less than or equal to the value of X. Additionally, the value of Y may be greater than or equal to the value of X. Similarly, the weight value of X and/or the weight value of Y may be applied to the CO2 metric. While the present disclosure discusses various combinations of applying weight values to air quality metrics that may be used to generate the productivity scores, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combination of applying weight values to air quality metrics.

In some embodiments, the applied weight values may be used to increase the priority of the air quality metrics when generating the productivity scores. For example, the ventilation metric may have a weight value of 2 applied. Additionally, the CO2 metric may have a weight value of 1 applied. The ventilation metric may contribute to the productivity scores at twice the proportion to that of the CO2 metric.

In some embodiments, the air quality metrics may have the same weight value applied which may result in the air quality metrics having the same priority when generating the productivity scores. For example, the CO2 metric, the ventilation metric, and the VOC metric may each have a weight value of 1 applied. The CO2 metric, the ventilation metric and the VOC metric may now contribute equally to the productivity scores.

In some embodiments, the applied weight values may be used to decrease the priority of the air quality metrics when generating the productivity scores. For example, the ventilation metric may have a weight value of 2 applied. Additionally, the CO2 metric may have a weight value of 1 applied. The ventilation metric may now contribute to the productivity scores at half the proportion to that of the CO2 metric.

In some embodiments, the productivity scores may be generated by normalizing the air quality metrics described herein. For example, the CO2 metric and the acetone metric may be normalized by determining a ratio of CO2 to acetone. Similarly in some embodiments, the CO2 metric and the acetone metric may be normalized by determining a ratio of acetone to CO2. While the present disclosure discusses various combinations of normalizing the air quality metrics that may be used to generate the productivity scores, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combination of normalizing air quality metrics.

In some embodiments, the recommendation generator 407 may perform an analysis of the productivity scores. The recommendation generator 407 may determine that the productivity scores are within one or more acceptable productivity scores ranges. For example, the office building 401 may have one or more predetermined productivity scores. Additionally, the office building 401 may have one or more productivity scores above and/or below the predetermined productivity scores that may create the acceptable productivity scores ranges. Additionally, the recommendation generator 407 may determine if any action is recommended to maintain or improve the productivity scores. Similarly, the recommendation generator 407 may determine that the productivity scores are not within the acceptable productivity scores ranges. Additionally, the recommendation generator 407 may determine if any action is recommended to adjust the productivity scores.

In some embodiments, the recommendation generator 407 may determine that the one or more air quality metrics described herein may be controlled in order to maintain the productivity scores. The building analysis system 404 and/or a BMS system of the office building 401 (e.g., the BMS system described in FIGS. 1 and 3) may control the one or more air quality metrics described herein. For example, the recommendation generator 407 may determine that the CO2 metric may be controlled in order to maintain the productivity scores. The building analysis system 404 and/or the BMS system may then control the CO2 metric. Similarly, the recommendation generator 407 may determine that the VOC metric may be controlled in order to maintain the productivity scores. The building analysis system 404 and/or the BMS system may then control the VOC metric in order to maintain the productivity scores. Similarly, the recommendation generator 407 may determine that the CO2 metric and/or the VOC metric may be controlled in order to maintain the productivity scores. The building analysis system 404 and/or the BMS system may then control the CO2 metric and/or the VOC metric in order to maintain the productivity scores. While the present disclosure discusses various combinations of controlling one or more air quality metrics described herein to maintain the productivity scores, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combinations of controlling the air quality metrics described herein in order to maintain the productivity scores.

In some embodiments, the recommendation generator 407 may determine that the one or more air quality metrics described herein may be controlled in order to adjust the productivity scores. The building analysis system 404 and/or the BMS system may then control the one or more air quality metrics described herein in order to adjust the productivity scores. For example, the recommendation generator 407 may determine that the CO2 metric may be controlled in order to adjust the productivity scores. The building analysis system 404 and/or the BMS system may then control the CO2 metric in order to adjust the productivity score. Similarly, the recommendation generator 407 may determine that the VOC metric may be controlled in order to adjust the productivity scores. The building analysis system 404 and/or the BMS system may then control the VOC metric. Similarly, the recommendation generator 407 may determine that the CO2 metric and/or the VOC metric may be controlled in order to adjust the productivity scores. The building analysis system 404 and/or the BMS system may then control the CO2 metric and/or the VOC metric in order to adjust the productivity scores. While the present disclosure discusses various combinations of using one or more air quality metrics described herein to adjust the productivity scores, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combinations of controlling the air quality metrics described herein in order to adjust the productivity scores.

In some embodiments, the air quality metrics described herein may be controlled by performing one or more control operations. For example, the air quality metrics described herein may be controlled by using one or more ventilation rates. Similarly, the air quality metrics described herein may be controlled by using one or more filtration rates. Additionally, the filtration rates may involve Air-handler filtration, in zone filtration, and/or any combination of thereof. Similarly, the air quality metrics described herein may be controlled by using one or more disinfection rates. Additionally, the disinfection rates may involve Air-handler disinfection, in zone disinfection, and/or any combination thereof. Similarly, the air quality metrics described herein may be controlled by using the ventilation rates in combination with the filtration rates.

In some embodiments, the recommendation generator 407 may determine that the air quality metrics described herein may be controlled by performing one or more control operations described herein. For example, the recommendation generator 407 may determine that the CO2 metric may be controlled by adjusting the ventilation rates of the spaces within the office building 401. The building analysis system 404 and/or the BMS system may then control the CO2 metric by adjusting the ventilation rates of the office spaces. Similarly, the recommendation generator 407 may determine that the VOC metric may be controlled by adjusting the ventilation rates of the spaces within the office building 401. The building analysis system 404 and/or the BMS system may then control the VOC metric by adjusting the ventilation rates of the office spaces. Similarly, the recommendation generator 407 may determine that the CO2 metric and/or the VOC metric may be controlled by adjusting the ventilation rates of the spaces within the office building 401. The building analysis system 404 and/or the BMS system may then control the CO2 metric and/or the VOC metric by adjusting the ventilation rates of the office spaces. While the present disclosure discusses various combinations of the air metrics described herein that may be controlled by performing the actions described herein, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combinations of the air quality metrics that may be controlled by performing the actions described herein.

In some embodiments, the recommendation generator 407 may determine that one or more control operations may be performed in order to control the air quality metrics described herein. For example, the recommendation generator 407 may determine that adjusting the filtration rate of the office spaces may control the CO2 metric. The building analysis system 404 and/or the BMS system may then adjust the filtration rate of the office spaces in order to control the CO2 metric. Similarly, the recommendation generator 407 may determine that adjusting the air disinfection rate may control the CO2 metric. The building analysis system 404 and/or the BMS system may then adjust the air disinfection rate in order to control the CO2 metric. Similarly, the recommendation generator 407 may determine that adjusting the filtration rate and/or adjusting the air disinfection rate of the office spaces may control the CO2 metric. The building analysis system 404 and/or the BMS system may then adjust the filtration rate and/or adjust the air disinfection rate of the office spaces in order to control the CO2 metric. Similarly, the recommendation generator 407 may determine that adjusting the filtration rate and/or adjusting the air disinfection rate of the office spaces may control the TVOC metric and/or the PM1 metric. The building analysis system 404 and/or the BMS system may then adjust the filtration rate and/or adjust the air disinfection rate of the office spaces in order to control the TVOC metric and/or the PM1 metric. While the present disclosure discusses various combinations of one or more actions that may be performed in order to control the air quality metrics described herein, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combinations of the actions that may be performed in order to control the air quality metrics described herein.

In some embodiments the air quality metrics described herein may be controlled if occupants occupy a different location within the office building 401. In some embodiments, the air quality metrics described herein may be controlled if one or more occupants occupy a location that is not the office building 401.

In some embodiments, the recommendation generator 407 may determine that the air quality metrics described herein may be controlled by recommending that one or more occupants occupy a different location within the office building 401. For example, the recommendation generator 407 may determine that the CO2 metric may be controlled by adjusting occupant location within the office building 401. The recommendation generator 407 may then control the CO2 metric by recommending one or more locations within the office building 401 that the occupants may occupy. Similarly, the recommendation generator 407 may determine that the VOC metric may be controlled by adjusting occupant location within the office building 401. The recommendation generator 407 may then control the VOC metric by recommending one or more locations within the office building 401 that the occupants may occupy. Similarly, the recommendation generator 407 may determine that the CO2 metric and/or the VOC metric may be controlled by adjusting occupant location within the office building 401. The recommendation generator 407 may then control the CO2 metric and/or the VOC metric by recommending one or more locations within the office building 401 that the occupants may occupy. While the present disclosure discusses various combinations of the air quality metrics described herein that may be controlled by adjusting occupant location within an office building, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combinations of the air quality metrics that may be controlled by adjusting occupant location within an office building.

In some embodiments, the recommendation generator 407 may determine that the air quality metrics described herein may be controlled by recommending that one or more occupants work from the occupants' home. For example, the recommendation generator 407 may determine that the CO2 metric may be controlled by having one or more occupants work from home. The recommendation generator 407 may then control the CO2 metric by recommending the occupants work from home. Similarly, the recommendation generator 407 may determine that the VOC metric may be controlled by having the occupants work from home. The recommendation generator 407 may then control the VOC metric by recommending the occupants work from home. Similarly, the recommendation generator 407 may determine that the CO2 metric and/or the VOC metric may be controlled by having the occupants work from home. The recommendation generator 407 may then control the CO2 metric and/or the VOC metric by recommending the occupants work from home. While the present disclosure discusses various combinations of the air quality metrics described herein that may be controlled by having the occupants work from home, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combinations of the air quality metrics that may be controlled by adjusting occupant location.

In some embodiments, the recommendation generator 407 may compare the air quality metrics of the office building 401 with the air quality metrics of one or more occupants' homes. The recommendation generator 407 may determine that the productivity scores may be higher if the occupant works from home. The recommendation generator 407 may recommend that the occupants work from the occupants' homes. In some embodiments, the recommendation generator 407 may compare expected air quality metrics of the office building 401 with the expected air quality metrics of the occupants' homes. The recommendation generator 407 may determine using the expected air quality metrics which location may have the higher productivity score. The recommendation generator may then recommend that the occupants occupy the location with the higher productivity score.

In some embodiments, the recommendation generator 407 may use factors other than the air quality metrics of the office building 401 and the air quality metrics of the occupants' homes. In some embodiments, the recommendation generator 407 may compare expected distractions of the office building 401 with the expected distractions of the occupants' home. The recommendation generator 407 may recommend that the occupants work from the location with the least amount of distractions.

In some embodiments, the recommendation generator 407 may determine that the productivity scores of the office building 401 are lower than the productivity scores of one or more locations not within the office building 401. Additionally, the recommendation generator 407 may recommend that the occupants work from the locations with the higher productivity score.

In some embodiments, the recommendation generator 407 may determine that the productivity scores of the office building 401 are the same as the productivity scores of one or more locations not within the office building 401. Additionally, the recommendation generator 407 may recommend that the occupants work at a location not within the office building 401. Similarly, in some embodiments, the recommendation generator 407 may recommend that the occupants work from a location within the office building 401.

In some embodiments, the recommendation generator 407 may determine that the productivity scores of the office building 401 may decrease if the occupants work in the office building 401. Additionally, the recommendation generator 407 may recommend that the occupants work from a location not within the office building 401.

In some embodiments, the recommendation generator 407 may determine that the productivity scores of the office building may decrease if the occupants work in the office building 401. Additionally, the recommendation generator 407 may recommend that the occupant work from a location not within the office building 401.

In some embodiments, various data discussed herein may be stored in, retrieved from, or processed in the context of digital twins. In some such embodiments, the digital twins may be provided within an infrastructure such as those described in U.S. patent application Ser. No. 17/134,661 filed Dec. 28, 2020, 63/289,499 filed Dec. 14, 2021, and Ser. No. 17/537,046 filed Nov. 29, 2021, the entireties of each of which are incorporated herein by reference.

In some embodiments, various data discussed herein may be processed at (e.g., processed using models executed at) a cloud or other off-premises computing system/device or group of systems/devices, an edge or other on-premises system/device or group of systems/devices, or a hybrid thereof in which some processing occurs off-premises and some occurs on-premises. In some example implementations, the data may be processed using systems and/or methods such as those described in U.S. patent application Ser. No. 17/710,458 filed Mar. 31, 2022, which is incorporated herein by reference in its entirety.

Referring now to FIG. 5, a process 500 where temporary air quality sensors may be installed in spaces of an office building to collect air quality measurements of the spaces and may generate one or more air quality occupant impact assessments and/or recommendations is shown, according to an exemplary embodiment. The process 500 may be performed by the building analysis system 404. Furthermore, the process 500 may be performed by any computing device described herein.

In step 502, the building analysis system 404 may connect to the temporary air quality sensors 402 via the network 414, the temporary air quality sensors 402 being installed by a technician in the office building 401 on a temporary basis (e.g., for two weeks, three weeks, etc.). Connecting to the sensors 402 may include sending a message to the sensors 402 requesting a response, receiving an indication from the sensors 402 indicating that the sensors 402 are online, receiving measurements from the sensors 402 for a first time, creating a data point to store measurements of the sensor in, etc. In step 504, the air quality measurements may be received by the building analysis system 404.

In step 506, the building analysis system 404 may generate one or more air quality occupant impact assessments and/or recommendations for the air quality occupant impact assessments of the office building 401. The air quality occupant impact assessments may summarize air quality for various spaces of the office building 401. The recommendations may be included within the air quality occupant impact assessments and may indicate control operations for implementation for various spaces (e.g., new ventilation rates, air flow rates, air change rates, etc.). The recommendations may recommend investigation into various sources of TVOCs, VOCs, etc. in various places of the office building 401, etc. The one or more air quality occupant impact assessments and/or recommendations may be used to generate one or more productivity scores for the office building 401 and/or space/spaces within the office building 401 in step 508.

In some embodiments, the air quality occupant impact assessments may include infection risk for the office building 401, spaces of the office building 401, and/or occupants of the office building 401. The infection risk may be a risk level of contracting an infectious disease present in a population (e.g., COVID19, influenza, the bird influenza, etc.). The infection risk may be based on current ventilation rates, filter performance, etc. which may be derived from the air quality measurements of the sensors 402. The air quality occupant impact assessment may indicate a control profile, e.g., guidelines for implementing physical control of AHUs, VAVs, unit ventilators, temporary space filters, etc. The guidelines may be ranges for operating settings, recommended operating settings, specific control algorithms to be used, etc. This control profile may operate equipment to reduce the infection risk. In some embodiments, the control profile may operate to provide energy savings. The control profile may be used to determine operating settings that are implemented at a time after the sensors 402 are disconnected from and/or uninstalled. In some embodiments, features described in U.S. patent application Ser. Nos. 16/927,759 and 16,927,318, both filed Jul. 13, 2020 and both incorporated by reference herein in their entireties, may be utilized in conjunction with the features of the present disclosure. For example, in some embodiments, the infection risk may be estimated using the readings collected by the sensors 402 and processed using the Wells-Riley equation, or any infection risk management standard, such as ASHRAE 241, as described in detail in the aforementioned applications and described herein.

In step 510, the productivity scores may be analyzed in order to determine if the productivity scores are within one or more predetermined productivity scores ranges. Analysis that may be done during this step may be found herein.

In step 512, the building analysis system 404 may disconnect from the sensors 402 as the sensors 402 are to be removed and uninstalled by a technician. Disconnecting from the sensors 402 can include sending a shutdown message to the sensor 402, sending a disconnect message from the sensors 402, not receiving new data from the sensors 402, etc. The sensors 402 can be uninstalled by the technician and disconnected from at the end of the temporary installation period. After the sensors are disconnected from, the building analysis system 404 and/or the BMS system may begin operating with operating settings and/or control algorithms based on the control profile generated by the building analysis system 404. While FIG. 5 provides various steps that may be performed, it should be noted that the features of FIG. 5 are equally applicable to any one or more possible combinations of actions that may be performed and all such combinations and permutations are contemplated in the scope of the disclosure. For example, in some embodiments, one or more steps may be added. Similarly, in some embodiments, one or more steps may be omitted.

Referring now to FIG. 6, a process 600 where the productivity scores may be analyzed. The process 600 may be performed by the building analysis system 404. Furthermore, the process may be performed by any computing devices described herein.

In step 602 the productivity scores are provided for analysis. In step 604, the productivity scores are compared to the predetermined productivity scores ranges. If the productivity scores are within the predetermined productivity scores ranges then the process moves to step 612. If the productivity scores are not within the predetermined productivity scores ranges the process moves to step 606.

In step 606, the building analysis system 404 determines if action will be taken to adjust the productivity scores. If no action will be taken then the process moves to step 608. If action will be taken then the process moves to step 610. In step 608, the building analysis system 404 does not to take any action. The process will return to step 604.

In step 610, one or more control operations may be performed in order to adjust the productivity scores. Once the control operations are performed the process will return to step 604. In step 612, the building analysis system 404 determines if action will be taken to maintain the productivity scores. If no action will be taken then the process moves to step 614. If action will be taken then the process will move to step 616.

In step 614, the building analysis system 404 does not take any action. The process will return to step 604. In step 616, the building analysis system 404 will perform one or more control operations. Once the control operation are performed the process will return to step 604. The process 600 may be performed one or more times. While the present disclosure discusses various combinations of one or more steps that may be performed while analyzing the productivity scores, it should be noted that the features of the present disclosure are equally applicable to any one or more possible combinations of steps that may be performed in order to analyze the productivity scores.

Pareto Optimization Techniques

In some embodiments, the building analysis system 404 and/or the BMS may use Pareto optimization techniques to evaluate various potential control strategies that can be used to affect the air quality metrics described herein as well as other control objectives such as energy consumption, cost, productivity, infection risk, etc. A Pareto optimization controller and Pareto optimization techniques are described herein and with further reference to U.S. patent application Ser. No. 17/483,078 filed Sep. 23, 2021, the entirety of which is incorporated by reference herein. In some embodiments, the BAS controller 302 may be similar to the Pareto optimization controller described therein.

In some embodiments, control operations for building equipment may be used as decision variables when performing Pareto optimization. The control operations may include any setpoint, setting, operating mode, or other control strategy for various types of equipment in order to affect air quality attributes as well as other building conditions. For example, control operations may include temperature setpoints, humidity setpoints, filtration setpoints or modes, UV light sanitation setpoints, ventilation rate setpoints, air flow rate setpoints, air change rate setpoints, or any other setting that can be adjusted to affect conditions within the building. In some embodiments, one or more possible combinations of the control operations may be used as decision variables to define various potential control strategies which can be used to control building conditions. Each potential control strategy may include different sets of control operations (e.g., different values of the decision variables) that define how the building equipment should be operated according to that control strategy. Pareto optimization may seek to identify a set of potential control strategies which are Pareto optimal (e.g., along the efficient frontier) with respect to a given set of control objectives.

In some embodiments, the productivity scores and energy cost or energy consumption may be used as control objectives. In some embodiments, the productivity scores and sustainability may be used as control objectives. In some embodiments, the productivity scores and carbon emissions may be used as control objectives. In some embodiments, the productivity scores and infection risk may be used as control objectives. In some embodiments, the productivity scores and occupant comfort may be used as control objectives. In some embodiments, air quality and energy consumption or energy cost may be used as control objectives. In some embodiments, control objectives may be a PM2.5 value and a CO2 value. While the present disclosure discusses various combinations of control objectives, it should be noted that the features of the present disclosure are equally applicable to any possible combination of control objectives. In some embodiments, the control objectives are competing in the sense that a given control strategy may benefit one control objective at the expense of another control objective. For example, a control strategy which provides maximum ventilation may increase air quality and benefit the occupant productivity control objectives, but may consume a high amount of energy and thus may be poor with respect to the energy consumption control objective.

The values of the control objectives can be affected by operating building equipment which consume energy or other resources to affect building conditions (e.g., temperature, humidity, air quality, etc.). The value of a control objective can be calculated or quantified using an objective function. For example, the value of the energy consumption control objective can be calculated using an objective function which sums energy consumption by the building equipment over a given time period. Similarly, the productivity score control objective can be calculated using an objective function which sums productivity scores over a given time period. As discussed above, the productivity score at a given time can be calculated as a function of various air quality metrics at that time. For control objectives that depend on building conditions over a future time period (e.g., occupant comfort, occupant productivity, infection risk, etc.), the values of the decision variables over the future time period can be used as inputs to one or more predictive models to predict the values of the control objectives. Each control objective may have its own predictive model which quantifies the value of that control objective as a function of the decision variables. For example, a dynamic temperature model can be used to predict building temperature as a function of control decisions for HVAC equipment over a future time period. Similarly, a dynamic air quality model can be used to predict air quality metrics (e.g., CO2 levels, particulate matter concentrations, etc.) as a function of control decisions for HVAC equipment over the future time period.

HVAC System with Building Infection Control

Overview

Referring particularly to FIG. 7, a HVAC system 700 that is configured to provide disinfection for various zones of a building (e.g., building 10) is shown, according to some embodiments. The HVAC system may be or may be similar to the HVAC system 100. HVAC system 700 can include an air handling unit (AHU) 704 (e.g., AHU 106, etc.) that can provide conditioned air (e.g., cooled air, supply air, etc.) to various building zones 716. The AHU 704 may draw air from the zones 716 in combination with drawing air from outside (e.g., outside air) to provide conditioned or clean air to the zones 716. The HVAC system 700 includes a controller 710 (e.g., BAS controller 302) that is configured to determine a fraction x of outdoor air to recirculated air that the AHU 704 should use to provide a desired amount of disinfection to one or more building zones 716. In some embodiments, controller 710 can generate control signals for various dampers of AHU 704 so that AHU 704 operates to provide the conditioned air to the building zones 716 using the fraction x.

The HVAC system 700 can also include ultraviolet (UV) lights 706 that are configured to provide UV light to the conditioned air before it is provided to building zones 716. The UV lights 706 can provide disinfection as determined by controller 710 and/or based on user operating preferences. For example, the controller 710 can determine control signals for UV lights 706 in combination with the fraction x of outdoor air to provide a desired amount of disinfection and satisfy an infection probability constraint. Although UV lights 706 are referred to throughout the present disclosure, the systems and methods described herein can use any type of disinfection lighting using any frequency, wavelength, or luminosity of light effective for disinfection. It should be understood that UV lights 706 (and any references to UV lights 706 throughout the present disclosure) can be replaced with disinfection lighting of any type without departing from the teachings of the present disclosure.

The HVAC system 700 can also include one or more filters 708 or filtration devices (e.g., air purifiers). In some embodiments, the filters 708 are configured to filter the conditioned air or recirculated air before it is provided to building zones 716 to provide a certain amount of disinfection. In this way, controller 710 can perform an optimization in real-time or as a planning tool to determine control signals for AHU 704 (e.g., the fraction x) and control signals for UV lights 706 (e.g., on/off commands) to provide disinfection for building zones 716 and reduce a probability of infection of individuals that are occupying building zones 716. Controller 710 can also function as a design tool that is configured to determine suggestions for building managers regarding benefits of installing or using filters 708, and/or specific benefits that may arise from using or installing a particular type or size of filter. Controller 710 can thereby facilitate informed design decisions to maintain sterilization of air that is provided to building zones 716 and reduce a likelihood of infection or spreading of infectious matter, thus increasing productivity and productivity scores of occupants of the building 10.

Wells-Riley Airborne Transmission

The systems and methods described herein may use an infection probability constraint in various optimizations (e.g., in on-line or real-time optimizations or in off-line optimizations) to facilitate reducing infection probability among residents or occupants of spaces that the HVAC system serves. In some embodiments, the infection probability may be utilized to determine a productivity score. For example, reducing infection probabilities may lead to an increase in productivity scores. The infection probability constraint can be based on a steady-state Wells-Riley equation (or any other infection risk management standard) for a probability of airborne transmission of an infectious agent given by:

$P := \frac{D}{S} = 1 - \exp (- \frac{Ipqt}{Q})$

where P is a probability that an individual becomes infected (e.g., in a zone, space, room, environment, etc.), D is a number of infected individuals (e.g., in the zone, space, room, environment, etc.), S is a total number of susceptible individuals (e.g., in the zone, space, room, environment, etc.), I is a number of infectious individuals (e.g., in the zone, space, room, environment, etc.), q is a disease quanta generation rate (e.g., with units of 1/sec), p is a volumetric breath rate of one individual (e.g., in m³/sec), t is a total exposure time (e.g., in seconds), and Q is an outdoor ventilation rate (e.g., in m³/see). For example, Q may be a volumetric flow rate of fresh outdoor air that is provided to the building zones 716 by AHU 704. In various embodiments, the variables of the Wells-Riley equation may correspond to different values. For example, a Wells-Riley equation or any infection risk management standard (e.g., ASHRAE 241) may be used for a constraint other than infection probability, and different variables may therefore be defined and used.

In various embodiments, a Wells-Riley equation and/or one or more models (e.g., any infection risk management standard, such as ASHRAE 241) may be used to model productivity scores for a Pareto optimization. For example, the Wells-Riley equation or other infection risk management standard can be used to calculate or predict infection probability and the infection probability can be used to calculate a productivity score. As discussed above, the productivity scores can be calculated as a function of various air quality metrics and infection probability can be used as an air quality metric in this regard. That is, the productivity score can be calculated as a function of infection probability. The equations and models described herein can be used to model or predict infection risk or infection probability, which can then be used to calculate a corresponding productivity score. As discussed herein, any infection risk management standard, such as ASHRAE 241, may be used in addition or alternative to a Wells-Riley equation or model.

When the Wells-Riley equation, or any infection risk management standard, such as ASHRAE 241, is implemented by controller 710 as described herein, controller 710 may use the Wells-Riley equation or any infection risk management standard, such as ASHRAE 241, (or a dynamic version of the Wells-Riley equation or other infection risk management standard) to determine an actual or current probability of infection P and operate the HVAC system 100 to maintain the actual probability of infection P below (or drive the actual probability of infection below) a constraint or maximum allowable value. The constraint value (e.g., P_max) may be a constant value, or may be adjustable by a user (e.g., a user-set value). For example, the user may set the constraint value of the probability of infection to a maximum desired probability of infection (e.g., either for on-line implementation of controller 710 to maintain the probability of infection below the maximum desired probability, or for an off-line implementation/simulation performed by controller 710 to determine various design parameters for HVAC system 100 such as filter size), or may select from various predetermined values (e.g., 3-5 different choices of the maximum desired probability of infection).

In some embodiments, the number of infectious individuals I can be determined by controller 710 based on data from the Centers for Disease and Control Prevention or a similar data source. The value of I may be typically set equal to 1 but may vary as a function of occupancy of building zones 716.

The disease quanta generation rate q may be a function of the infectious agent. For example, more infectious diseases may have a higher value of q, while less infectious diseases may have a lower value of q. For example, the value of q for COVID-19 may be 30-300 (e.g., 100).

The value of the volumetric breath rate p may be based on a type of building space 716. For example, the volumetric breath rate p may be higher if the building zone 716 is a gym as opposed to an office setting. In general, an expected level of occupant activity may determine the value of the volumetric breath rate p.

A difference between D (the number of infected individuals) and I (the number of infectious individuals) is that D is a number of individuals who are infected (e.g., infected with a disease), while I is a number of people that are infected and are actively contagious (e.g., individuals that may spread the disease to other individuals or spread infectious particles when they exhale). The disease quanta generation rate indicates a number of infectious droplets that give a 63.2% chance of infecting an individual (e.g., 1−exp(−1)). For example, if an individual inhales k infectious particles, the probability that the individual becomes infected (P) is given by

$1 - \exp (- \frac{k}{k_{0}})$

where k is the number of infectious particles that the individual has inhaled, and k₀is a quantum of particles for a particular disease (e.g., a predefined value for different diseases). The quanta generation rate q is the rate at which quanta are generated (e.g., K/k₀) where K is the rate of infectious particles exhaled by an infectious individual. It should be noted that values of the disease quanta generation rate q may be back-calculated from epidemiological data or may be tabulated for well-known diseases.

The Wells-Riley equation (shown above) is derived by assuming steady-state concentrations for infectious particles in the air. Assuming a well-mixed space:

$V \frac{dN}{dt} = Iq - NQ$

where V is a total air volume (e.g., in m³), N is a quantum concentration in the air, I is the number of infectious individuals, q is the disease quanta generation rate, and Q is the outdoor ventilation rate. The term Iq is quanta production by infectious individuals (e.g., as the individuals breathe out or exhale), and the term NQ is the quanta removal rate due to ventilation (e.g., due to operation of AHU 704).

Assuming steady-state conditions, the steady state quantum concentration in the air is expressed as:

$N_{ss} = \frac{Iq}{Q}$

according to some embodiments.

Therefore, if an individual inhales at an average rate of p (e.g., in m³/sec), over a period of length t the individual inhales a total volume pt or N_ssptk₀infectious particles. Therefore, based on a probability model used to define the quanta, the infectious probability is given by:

$P = 1 - \exp (- \frac{k}{k_{0}}) = 1 - \exp (- N_{SS} pt) = 1 - \exp (- \frac{Iqpt}{Q})$

where P is the probability that an individual becomes infected, k is the number of infectious particles that the individual has inhaled, and k₀is the quantum of particles for the particular disease.

Carbon Dioxide for Infectious Particles Proxy

While the above equations may rely on in-air infectious quanta concentration, measuring in-air infectious quanta concentration may be difficult. However, carbon dioxide (CO2) is a readily-measurable parameter that can be a proxy species, measured by zone sensors 712. In some embodiments, a concentration of CO2 in the zones 716 may be directly related to a concentration of the infectious quanta. In various embodiments, a concentration of CO2 in the zones 716 may be utilized to determine an air quality and, subsequently, a productivity score for an occupant and/or building.

A quantity q that defines a ratio of an infected particle concentration in the building air to the infected particle concentration in the exhaled breath of an infectious individual is defined:

$ϕ := \frac{pN}{q}$

where p is the volumetric breath rate for an individual, N is the quantum concentration in the air, and q is the disease quanta generation rate. Deriving the above equation with respect to time yields:

$\frac{d ϕ}{dt} = \frac{p}{q} (\frac{dN}{dt}) = \frac{Ip}{V} - ϕ (\frac{Q}{V})$

where p is the volumetric breath rate for the individual, q is disease quanta generation rate, N is the quantum concentration in the air, t is time, I is the number of infectious individuals, V is the total air volume, q ϕ is the ratio, and Q is the outdoor ventilation rate. Since it can be difficult to measure the ratio q ϕ of the air, CO2 can be used as a proxy species.

Humans release CO2 when exhaling, which is ultimately transferred to the ambient via ventilation of an HVAC system. Therefore, the difference between CO2 particles and infectious particles is that all individuals (and not only the infectious population) release CO2 and that the outdoor air CO2 concentration is non-zero. However, it may be assumed that the ambient CO2 concentration is constant with respect to time, which implies that a new quantity C can be defined as the net indoor CO2 concentration (e.g., the indoor concentration minus the outdoor concentration). With this assumption, the following differential equation can be derived:

$V \frac{dC}{dt} = Spc - QC$

where V is the total air volume (e.g. in m³), C is the net indoor CO2 concentration, t is time, S is the total number of susceptible individuals (e.g., in building zone 716, or a modeled space, or all of building zones 716, or building 10), p is the volumetric breath rate for one individual, c is the net concentration of exhaled CO2, and Q is the outdoor ventilation rate. This equation assumes that the only way to remove infectious particles is with fresh air ventilation (e.g., by operating AHU 704 to draw outdoor air and use the outdoor air with recirculated air). A new quantity ψ can be defined that gives the ratio of net CO2 concentration in the building air to net CO2 concentration in the exhaled air:

$ψ = \frac{C}{c}$

where ψ is the ratio, C is the net indoor CO2 concentration, and c is the net concentration of exhaled CO2.

Deriving the ratio 4 with respect to time yields:

$\frac{d ψ}{dt} = \frac{1}{c} (\frac{dC}{dt}) = \frac{Sp}{V} - ψ (\frac{Q}{V})$

according to some embodiments.

Combining the above equation with the quantity, it can be derived that:

$\frac{d}{dt} \log (\frac{ϕ}{ψ}) = \frac{1}{ϕ} (\frac{d ϕ}{dt}) - \frac{1}{ψ} (\frac{d ψ}{dt}) = \frac{p}{V} (\frac{I}{ϕ} - \frac{S}{ψ})$

according to some embodiments. Assuming that the initial condition satisfies:

$ϕ (0) = \frac{1}{S} ψ (0)$

it can be determined that the right-hand side of the

$\frac{d}{dt} \log (\frac{ϕ}{ψ})$

equation becomes zero. This implies that the term

$\log (\frac{ϕ}{ψ})$

and therefore

$\frac{ϕ}{ψ}$

is a constant. Therefore, ϕ/ψ is constant for all times t and not merely initial conditions when t=0.

The

$\frac{d}{dt} \log (\frac{ϕ}{ψ})$

relationship only holds true when fresh outdoor air is used as the only disinfection mechanism. However, in many cases HVAC system 100 may include one or more filters 708, and UV lights 706 that can be operated to provide disinfection for building zones 716. If additional infection mitigation strategies are used, the ventilation rate may instead by an effective ventilation rate for infectious quanta that is different than that of the CO2. Additionally, the only way for the initial conditions ϕ(0) and ψ(0) to be in proportion is for both to be zero. This assumption can be reasonable if HVAC system 100 operates over a prolonged time period (such as overnight, when the concentrations have sufficient time to reach equilibrium zero values). However, ventilation is often partially or completely disabled overnight and therefore the two quantities ϕ and ψ are not related. However, CO2 concentration can be measured to determine common model parameters (e.g., for the overall system volume V) without being used to estimate current infectious particle concentrations. If fresh outdoor air ventilation is the only mechanism for disinfection of zones 716, and the HVAC system 100 is run so that the concentrations reach equilibrium, CO2 concentration can be measured and used to estimate current infectious particle concentrations.

In various embodiments, one or more models may be used to determine or predict the values of various air quality metrics over a future time period as a function of control decisions for building equipment. For example, control decisions such as ventilation rate, air filtration, air temperature setpoints, humidity setpoints, or other control decisions that affect building conditions can be used as inputs to predictive models (e.g., dynamic models) that predict the values of air quality metrics as a function of the control decisions over a future time period. For example, predictive models may be used to determine air quality metrics such as concentrations of TVOC, PM1, PM2.5, VOC, ozone, etc., in the air as a function of a time series of control decisions over a future time period. The predicted values of the air quality metrics can then be used to calculate a productivity score.

Dynamic Extension and Infection Probability Constraints

Referring still to FIG. 7, it may be desirable to model the infectious quanta concentration N of building zones 716 as a dynamic parameter rather than assuming N is equal to the steady state N_SSvalue. For example, if infectious individuals enter building zones 716, leave building zones 716, etc., the infectious quanta concentration N may change over time. This can also be due to the fact that the effective fresh air ventilation rate (which includes outdoor air intake as well as filtration or UV disinfection that affects the infectious agent concentration in the supply air that is provided by AHU 704 to zones 716) can vary as HVAC system 100 operates.

Therefore, assuming that the infectious quanta concentration N(t) is a time-varying quantity, for a given time period t∈[0,T], an individual breathes in:

$k_{[0, T]} = \int_{0}^{T} {pk}_{0} N (t) dt$

where k_[0,T] is the number of infectious particles that an individual inhales over the given time period [0,T], p is the volumetric breath rate of one individual, k₀is the quantum of particles for a particular disease, and N(t) is the time-varying quantum concentration of the infectious particle in the air.

Since

$P = 1 - \exp (- \frac{k}{k_{0}}),$

the above equation can be rearranged and substitution yields:

$- \log (1 - P_{[0, T]}) = \int_{0}^{T} pN (t) dt \approx Δ \sum_{t} {pN}_{t}$

according to some embodiments.

Assuming an upper boundary P_[0,T]^maxon acceptable or desirable infection probability, a constraint is defined as:

$\frac{Δ}{T} \sum_{t} N_{t} \leq - \frac{1}{pT} \log (1 - P_{[0, T]})$

according to some embodiments. The constraint can define a fixed upper boundary on an average value of N_tover the given time interval.

Control Formulation

Referring particularly to FIG. 8, controller 710 is shown in greater detail, according to some embodiments. Controller 710 is configured to generate control signals for any of UV lights 706, filter 708, and/or AHU 704. AHU 704 operates to draw outdoor air and/or recirculated air (e.g., from zones 716) to output conditioned (e.g., cooled) air. The conditioned air may be filtered by passing through filter 708 (e.g., which may include fans to draw the air through the filter 708) to output filtered air. The filtered air and/or the conditioned air can be disinfected through operation of UV lights 706. The AHU 704, filter 708, and UV lights 706 can operate in unison to provide supply air to zones 716.

Controller 710 includes a processing circuit 802 including a processor 804 and memory 806. Processing circuit 802 can be communicably connected with a communications interface of controller 710 such that processing circuit 802 and the various components thereof can send and receive data via the communications interface. Processor 804 can be implemented as a general purpose processor, an application specific integrated circuit (ASIC), one or more field programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components.

Memory 806 (e.g., memory, memory unit, storage device, etc.) can include one or more devices (e.g., RAM, ROM, Flash memory, hard disk storage, etc.) for storing data and/or computer code for completing or facilitating the various processes, layers and modules described in the present application. Memory 806 can be or include volatile memory or non-volatile memory. Memory 806 can include database components, object code components, script components, or any other type of information structure for supporting the various activities and information structures described in the present application. According to some embodiments, memory 806 is communicably connected to processor 804 via processing circuit 802 and includes computer code for executing (e.g., by processing circuit 802 and/or processor 804) one or more processes described herein.

In some embodiments, controller 710 is implemented within a single computer (e.g., one server, one housing, etc.). In various other embodiments, controller 710 can be distributed across multiple servers or computers (e.g., that can exist in distributed locations).

Memory 806 can include a constraint generator 810, a model manager 816, a sensor manager 814, an optimization manager 812, and a control signal generator 808. Sensor manager 814 can be configured to obtain zone sensor data from zone sensors 712 and/or ambient sensor data from ambient sensors 714 (e.g., environmental conditions, outdoor temperature, outdoor humidity, etc.) and distribute required sensor data to the various components of memory 806 thereof. Constraint generator 810 is configured to generate one or more constraints for an optimization problem (e.g., an infection probability constraint, a productivity score constraint) and provide the constraints to optimization manager 812. In some embodiments, the constraints define minimum or maximum acceptable values of the corresponding control objective (e.g., minimum acceptable productivity score, maximum acceptable infection probability, etc.). Model manager 816 can be configured to generate dynamic models (e.g., individual or zone-by-zone dynamic models, aggregate models, etc.) and provide the dynamic models to optimization manager 812. Optimization manager 812 can be configured to use the constraints provided by constraint generator 810 and the dynamic models provided by model manager 816 to formulate an optimization problem. Optimization manager 812 can also define an objective function for the optimization problem, and minimize or optimize the objective function subject to the one or more constraints and the dynamic models. The objective function may be a function that indicates an amount of energy consumption, energy consumption costs, carbon footprint, or any other optimization goals over a time interval or time horizon (e.g., a future time horizon) as a function of various control decisions. Optimization manager 812 can output optimizations results to control signal generator 808. Control signal generator 808 can generate control signals based on the optimization results and provide the control signals to any of AHU 704, filter 708, and/or UV lights 706.

Referring particularly to FIGS. 7 and 8, AHU 704 can be configured to serve multiple building zones 716. For example, AHU 704 can be configured to serve a collection of zones 716 that are numbered k=1, . . . , K. Each zone 716 can have a temperature, referred to as temperature T_k(the temperature of the kth zone 716), a humidity ω_k(the humidity of the kth zone 716), and an infectious quanta concentration N_k(the infectious quanta concentration of the kth zone 716). Using this notation, the following dynamic models of individual zones 716 can be derived:

$ρ {cV}_{k} (\frac{{dT}_{k}}{dt}) = ρ {cf}_{k} (T_{0} - T_{k}) + Q_{k} (T_{k})$

$ρ V_{k} (\frac{d ω_{k}}{dt}) = ρ f (ω_{0} - T_{0}) + w_{k}$

$V_{k} (\frac{{dN}_{k}}{dt}) = f_{k} (N_{0} - N_{k}) + I_{k} q$

where f_kis a volumetric flow of air into the kth zone, ρ is a mass density of air (e.g., in kg per cubic meters), c is the heat capacity of air (e.g., in kJ/kg·K), Q_k(⋅) is heat load on the kth zone 716 (which may depend on the temperature T_k), w_kis the moisture gain of the kth zone 716, and I_kis the number of infectious individuals in the kth zone 716. T₀is the temperature of the air provided to the kth zone (e.g., as discharged by a VAV box of AHU 704), ω₀is the humidity of the air provided to the kth zone 716, and N₀is the infectious quanta concentration of the air provided to the kth zone 716. In various embodiments, one or more additional or alternative dynamic models may be generated for one or more air quality metrics (e.g., CO2, TVOC, etc.) and/or one or more productivity scores.

The temperature T₀of air output by the AHU 704, the humidity ω₀of air output by the AHU 704, and the infectious quanta concentration N₀of air output by the AHU 704 is calculated using the equations:

$T_{0} = {xT}_{a} + (1 - x) \frac{\sum_{k} f_{k} T_{k}}{\sum_{k} f_{k}} - Δ T_{c}$

$ω_{0} = x ω_{a} + (1 - x) \frac{\sum_{k} f_{k} T_{k}}{\sum_{k} f_{k}} - Δ ω_{c}$

$N_{0} = (1 - λ) (1 - x) \frac{\sum_{k} f_{k} N_{k}}{\sum_{k} f_{k}}$

where x is the fresh-air intake fraction of AHU 704, T_ais current ambient temperature, ω_ais current ambient humidity, ΔT_cis temperature reductions from the cooling coil of AHU 704, Δω_cis humidity reduction from the cooling coil of AHU 704, and λ is a fractional reduction of infectious quanta due to filtration (e.g., operation of filter 708) and/or UV treatment (e.g., operation of UV lights 706) at AHU 704 (but not due to ventilation which is accounted for in the factor 1−x, according to some embodiments.

In some embodiments, the dynamic models of the individual zones are stored by and used by model manager 816. Model manager 816 can store the individual dynamic models shown above and/or aggregated models (described in greater detail below) and populate the models. The populated models can then be provided by model manager 816 to optimization manager 812 for use in performing an optimization.

In some embodiments, model manager 816 is configured to receive sensor data from sensor manager 814. Sensor manager 814 may receive sensor data from zone sensors 712 and/or ambient sensors 714 and provide appropriate or required sensor data to the various managers, modules, generators, components, etc., of memory 806. For example, sensor manager 814 can obtain values of the current ambient temperature T_a, the current ambient humidity ω_a, the temperature reduction ΔT_cresulting from the cooling coil of AHU 704, and/or the humidity reduction Δω_cresulting from the cooling coil of AHU 704, and provide these values to model manager 816 for use in determining T₀, ω₀, and N_σor for populating the dynamic models of the individual zones 716.

In some embodiments, various parameters or values of the variables of the dynamic models of the individual zones 716 are predefined, predetermined, or stored values, or may be determined (e.g., using a function, an equation, a table, a look-up table, a graph, a chart, etc.) based on sensor data (e.g., current environmental conditions of the ambient or outdoor area, environmental conditions of the zones 716, etc.). For example, the mass density of air p may be a predetermined value or may be determined based on sensor data. In some embodiments, model manager 816 can use stored values, sensor data, etc., to fully populate the dynamic models of the individual zones 716 (except for control or adjustable variables of the dynamic models of the individual zones 716 that are determined by performing the optimization). Once the models are populated so that only the control objectives remain undefined or undetermined, model manager 816 can provide the populated models to optimization manager 812.

The number of infectious individuals I_kcan be populated based on sensor data (e.g., based on biometric data of occupants or individuals of the building zones 716), or can be estimated based on sensor data. In some embodiments, model manager 816 can use an expected number of occupants and various database regarding a number of infected individuals in an area. For example, model manager 816 can query an online database regarding potential infection spread in the area (e.g., number of positive tests of a particular virus or contagious illness) and estimate if it is likely that an infectious individual is in the building zone 716.

In some embodiments, it can be difficult to obtain zone-by-zone values of the number of infectious individuals I_kin the modeled space (e.g., the zones 716). In some embodiments, model manager 816 is configured to use an overall approximation of the model for N_k. Model manager 816 can store and use volume-averaged variables:

$\overline{N} = \frac{\sum_{k} V_{k} N_{k}}{\overline{V}}$

$\overline{f} = \sum_{k} f_{k}$

$\overline{V} = \sum_{k} V_{k}$

$\overline{I} = \sum_{k} I_{k}$

according to some embodiments. Specifically, the equations shown above aggregate the variables N, f, V, and Ī across multiple zones 716 by calculating a weighted average based on the volume of zones 716.

The K separate ordinary differential equation models (i.e., the dynamic models of the individual zones 716) can be added for N_kto determine an aggregate quantum concentration model:

$\begin{matrix} \overline{V} \frac{d \overline{N}}{dt} = \sum_{k} V_{k} \frac{{dN}_{k}}{dt} \\ = \sum_{k} (f_{k} (N_{0} - N_{k}) + I_{k} q) \\ = \overline{I} q + \sum_{k} f_{k} ((1 - λ) (1 - x) \frac{\sum_{k'} f_{k'} N_{k'}}{\sum_{k'} f_{k'}} - N_{k}) \\ = \overline{I} q + (1 - λ) (1 - x) \sum_{k'} f_{k'} N_{k'} - \sum_{k} f_{k} N_{k} \\ = \overline{I} q - (λ + x - λ x) \sum_{k} f_{k} N_{k} \\ \approx \overline{I} q - (λ + x - λ x) \overline{f} \overline{N} \end{matrix}$

according to some embodiments, assuming that N_k≈N for each zone 716. The aggregate quantum concentration model is shown below:

$\overline{V} \frac{d \overline{N}}{dt} = \overline{I} q - (λ + x - λ x) \sum_{k} f_{k} N_{k} \approx \overline{I} q - (λ + x - λ x) \overline{f} \overline{N}$

according to some embodiments.

Defining aggregate temperature, humidity, heat load, and moisture gain parameters:

$\overline{T} = \frac{\sum_{k} V_{k} T_{k}}{\overline{V}}$

$\overline{ω} = \frac{\sum_{k} V_{k} ω_{k}}{\overline{V}}$

$\overline{Q} (\cdot) = \sum_{k} Q_{k} (\cdot)$

$\overline{w} = \sum_{k} w_{k}$

allows the k thermal models

$ρ {cV}_{k} (\frac{{dT}_{k}}{dt})$

to be added:

$\begin{matrix} ρ c \overline{V} \frac{d \overline{T}}{dt} = \sum_{k} ρ {cV}_{k} \frac{{dT}_{k}}{dt} \\ = \sum_{k} (ρ {cf}_{k} (T_{0} - T_{k}) + Q_{k} (T_{k})) \\ = \sum_{k} Q_{k} (T_{k}) + \sum_{k} ρ {cf}_{k} ({xT}_{a} + (1 - x) \frac{\sum_{k'} f_{k'} T_{k'}}{\sum_{k'} f_{k'}} - T_{k} - Δ T_{c}) \\ = \sum_{k} Q_{k} (T_{k}) + (1 - x) \sum_{k'} ρ {cf}_{k'} T_{k'} + \sum_{k} ρ f_{k} ({xT}_{a} - T_{k} - Δ T_{c}) \\ = \sum_{k} Q_{k} (T_{k}) + ρ c \sum_{k} f_{k} (x (T_{a} - T_{k}) - Δ T_{c}) \\ \approx \overline{Q} (\overline{T}) + ρ c \overline{f} (x (T_{a} - \overline{T}) - Δ T_{c}) \end{matrix}$

according to some embodiments (assuming that T_k≈T for each zone 716). This yields the aggregate thermal model:

$ρ c \overline{V} \frac{d \overline{T}}{dt} = \sum_{k} Q_{k} (T_{k}) + ρ c \sum_{k} f_{k} (x (T_{a} - T_{k}) - Δ T_{c}) \approx \overline{Q} (\overline{T}) + ρ c \overline{f} (x (T_{a} - \overline{T}) - Δ T_{c})$

according to some embodiments.

The moisture model

$ρ V_{k} (\frac{d ω_{k}}{dt})$

can similarly be aggregated to yield an aggregate moisture model:

$\begin{matrix} ρ \overline{V} \frac{d \overline{ω}}{dt} = \overline{w} + ρ \sum_{k} f_{k} (x (ω_{a} - ω_{k}) - {Δω}_{c}) \\ \approx \overline{w} + ρ \overline{f} (x (ω_{a} - \overline{ω}) - {Δω}_{c}) \end{matrix}$

to predict an evolution of volume-averaged humidity, according to some embodiments.

In some embodiments, model manager 816 stores and uses the aggregate quantum concentration model, the aggregate thermal model, and/or the aggregate moisture model described hereinabove. Model manager 816 can populate the various parameters of the aggregate models and provide the aggregate models to optimization manager 812 for use in the optimization.

Referring still to FIG. 8, memory 806 includes optimization manager 812. Optimization manager 812 can be configured to use the models provided by model manager 816 and various constraints provided by constraint generator 810 to construct an optimization problem for HVAC system 100 (e.g., to determine design outputs and/or to determine control parameters, setpoints, control decisions, etc., for UV lights 706 and/or AHU 704). Optimization manager 812 can construct an optimization problem that uses the individual or aggregated temperature, humidity, and/or quantum concentration models subject to constraints to minimize energy use. In some embodiments, decision variables of the optimization problem formulated and solved by optimization manager 812 are the flows f_k(or the aggregate f if the optimization problem uses the aggregate models), the outdoor air fraction x and the infectious quanta removal fraction λ.

The infectious quanta removal fraction λ is defined as:

$λ = λ_{filter} + λ_{UV}$

where λ_filteris an infectious quanta removal fraction that results from using filter 708 (e.g., an amount or fraction of infectious quanta that is removed by filter 708), and L, is an infectious quanta removal fraction that results from using UV lights 706 (e.g., an amount or fraction of infectious quanta that is removed by operation of UV lights 706). In some embodiments, λ_filteris a design-time constant (e.g., determined based on the chosen filter 708), whereas λ_UVis an adjustable or controllable variable that can be determined by optimization manager 812 by performing the optimization of the optimization problem. In some embodiments, λ_UVis a discrete variable. In some embodiments, λ_UVis a continuous variable.

Instantaneous electricity or energy consumption of HVAC system 100 is modeled using the equation (e.g., an objective function that is minimized):

$E = η_{coil} ρ \overline{f} (c Δ T_{c} + L {Δω}_{c}) + η_{fan} \overline{f} Δ P + η_{UV} λ_{UV}$

where L is a latent heat of water, ΔP is a duct pressure drop, η_coilis an efficiency of the cooling coil of AHU 704, η_fanis an efficiency of a fan of AHU 704, and η_UVis an efficiency of the UV lights 706, according to some embodiments. In some embodiments, optimization manager 812 is configured to store and use the energy consumption model shown above for formulating and performing the optimization. In some embodiments, the term η_coilρf(cΔT_c+LΔω_c) is an amount of energy consumed by the cooling coil or heating coil of the AHU 704 (e.g., over an optimization time period or time horizon), the term η_fanfΔP is an amount of energy consumed by the fan of the AHU 704, and η_UVλ_UVis the amount of energy consumed by the UV lights 706. In some embodiments, the duct pressure drop ΔP is affected by or related to a choice of a type of filter 708, where higher efficiency filters 708 (e.g., filters 708 that have a higher value of η_filter) generally resulting in a higher value of the duct pressure drop ΔP and therefore greater energy consumption. In some embodiments, a more complex model of the energy consumption is used by optimization manager 812 to formulate the optimization problem (e.g., a non-linear fan model and a time-varying or temperature-dependent coil efficiency model).

In some embodiments, the variables ΔT_cand Δω_cfor the cooling coil of the AHU 704 are implicit dependent decision variables. In some embodiments, a value of a supply temperature T_AHUis selected for the AHU 704 and is used to determine the variables ΔT_cand Δψ_cbased on inlet conditions to the AHU 704 (e.g., based on sensor data obtained by sensor manager 814). In such an implementation, model manager 816 or optimization manager 812 may determine that T₀=T_AHUand an equation for ω₀.

Optimization manager 812 can use the models (e.g., the individual models, or the aggregated models) provided by model manager 816, and constraints provided by constraint generator 810 to construct the optimization problem. Optimization manager 812 may formulate an optimization problem to minimize energy consumption subject to constraints on the modeled parameters, ω, and N and additional constraints:

$\min_{f_{t}, x_{t}, λ_{t}} \sum_{t} E_{t} (Energy Cost)$

$s . t . \dots (Dynamic Models for T_{t}, ω_{t}, and N_{t})$

$\dots (Infection Probability Constraint)$

$T_{t}^{\min} \leq T_{t} \leq T_{t}^{\max} (Temperature Bounds)$

$ω_{t}^{\min} \leq ω_{t} \leq ω_{t}^{\max} (Humidity Bounds)$

$x_{t} f_{t} \geq F_{t}^{\min} (Fresh - Air Ventilation Bound)$

$f_{t}^{\min} \leq f_{t} \leq f_{t}^{\max} (VAV Flow Bounds)$

$0 \leq x_{t} \leq 1 (Outdoor - Air Damper Bounds)$

where Σ_tE_tis the summation of instantaneous electricity or energy consumption of the HVAC system 100 over an optimization time period, subject to the dynamic models for T_t, ω_t, and N_t(either zone-by-zone individual models, or aggregated models as described above), an infection probability constraint (described in greater detail below), temperature boundary constraints (T_t^min≤T_t≤T_t^max, maintaining T_tbetween a minimum temperature boundary T_t^minand a maximum temperature boundary T_t^max), humidity boundary constraints (ω_t^min≤w_t≤ω_t^max, maintaining the humidity ω_tbetween a minimum humidity boundary ω_t^minand a maximum humidity boundary ω_t^max), a fresh air ventilation boundary (x_tf_t≥F_t^min, maintaining the fresh air ventilation x_tf_tabove or equal to a minimum required amount F_t^min), a VAV flow boundary (f_t^min≤f_t≤f_t^maxmaintaining the volumetric flow rate f_tbetween a minimum boundary f_t^minand a maximum boundary f_t^max), and an outdoor air damper bound/constraint (0≤x_t≤1 maintaining the outdoor air fraction x_tbetween 0 and 1). In some embodiments, optimization manager 812 is configured to discretize the dynamic models (e.g., the individual dynamic models or the aggregate dynamic models) using matrix exponentials or approximately using collocation methods. In various embodiments, additional constraints may include air quality metrics, such as CO2, ozone, TVOC, PM2.5, etc.

The boundaries on temperature (T_t^min≤T_t≤T_t^max) and humidity (ω_t^min≤ω_t≤ω_t^max) can be determined by optimization manager 812 based on user inputs or derived from comfort requirements. The temperature and humidity bounds may be enforced by optimization manager 812 as soft constraints. The remaining bounds (e.g., the fresh-air ventilation bound, the VAV flow bounds, and the outdoor-air damper bounds) can be applied to input quantities (e.g., decision variables) by optimization manager 812 as hard constraints for the optimization. In some embodiments, the fresh-air ventilation bound is enforced by optimization manager 812 to meet the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE) standards. In some embodiments, the fresh-air ventilation bound is replaced with a model and corresponding bounds for CO2 concentration.

In some embodiments, the various constraints generated by constraint generator 810 or other constraints imposed on the optimization problem can be implemented as soft constraints, hard constraints, or a combination thereof. Hard constraints may impose rigid boundaries (e.g., maximum value, minimum value) on one or more variables in the optimization problem such that any feasible solution to the optimization problem must maintain the hard constrained variables within the limits defined by the hard constraints. Conversely, soft constraints may be implemented as penalties that contribute to the value of the objective function (e.g., adding to the objective function if the optimization problem seeks to minimize the objective function or subtracting from the objective function if the optimization problem seeks to maximize the objective function). Soft constraints may be violated when solving the optimization problem, but doing so will incur a penalty that affects the value of the objective function. Accordingly, soft constraints may encourage optimization manager 812 to maintain the values of the soft constrained variables within the limits defined by the soft constraints whenever possible to avoid the penalties, but may allow optimization manager 812 to violate the soft constraints when necessary or when doing so would result in a more optimal solution.

In some embodiments, constraint generator 810 may impose soft constraints on the optimization problem by defining large penalty coefficients (relative to the scale of the other terms in the objective function) so that optimization manager 812 only violates the soft constraints when absolutely necessary. However, it is contemplated that the values of the penalty coefficients can be adjusted or tuned (e.g., by a person or automatically by constraint generator 810) to provide an optimal tradeoff between maintaining the soft constrained variables within limits and the resulting cost (e.g., energy cost, energy consumption, monetary cost) defined by the objective function. One approach which can be used by constraint generator 810 is to use penalties proportional to amount by which the soft constraint is violated (i.e., static penalty coefficients). For example, a penalty coefficient of 0.1 $/° C.·hr for a soft constrained temperature variable would add a cost of $0.10 to the objective function for every 1° C. that the temperature variable is outside the soft constraint limit for every hour of the optimization period. Another approach which can be used by constraint generator 810 is to use variable or progressive penalty coefficients that depend on the amount by which the soft constraint is violated. For example, a variable or progressive penalty coefficient could define a penalty cost of 0.1 $/° C.·hr for the first 1° C. by which a soft constrained temperature variable is outside the defined limit, but a relatively higher penalty cost of 1.0 $/° C.·hr for any violations of the soft constrained temperature limit outside the first 1° C.

Another approach which can be used by constraint generator 810 is to provide a “constraint violation budget” for one or more of the constrained variables. The constraint violation budget may define a total (e.g., cumulative) amount by which a constrained variable is allowed to violate a defined constraint limit within a given time period. For example a constraint violation budget for a constrained temperature variable may define 30° C.·hr (or any other value) as the cumulative amount by which the constrained temperature variable is allowed to violate the temperature limit within a given time period (e.g., a day, a week, a month, etc.). This would allow the temperature to violate the temperature constraint by 30° C. for a single hour, 1° C. for each of 30 separate hours, 0.1° C. for each of 300 separate hours, 10° C. for one hour and 1° C. for each of 20 separate hours, or any other distribution of the 30° C.·hr amount across the hours of the given time period, provided that the cumulative temperature constraint violation sums to 30° C.·hr or less. As long as the cumulative constraint violation amount is within (e.g., less than or equal to) the constraint violation budget, constraint generator 810 may not add a penalty to the objective function or subtract a penalty from the objective function. However, any further violations of the constraint that exceed the constraint violation budget may trigger a penalty. The penalty can be defined using static penalty coefficients or variable/progressive penalty coefficients as discussed above.

The infection probability constraint (described in greater detail below) is linear, according to some embodiments. In various embodiments, one or more constraints (e.g., productivity score constraints, air quality metric constraints, etc.) may be linear. In some embodiments, two sources of nonlinearity in the optimization problem are the dynamic models and a calculation of the coil humidity reduction Δω_c. In some embodiments, the optimization problem can be solved using nonlinear programming techniques provided sufficient bounds are applied to the input variables.

Infection Probability Constraint

Referring still to FIG. 8, memory 806 is shown to include a constraint generator 810. Constraint generator 810 can be configured to generate the infection probability constraint, and provide the infection probability constraint to optimization manager 812. In some embodiments, constraint generator 810 is configured to also generate the temperature bounds, the humidity bounds, the fresh-air ventilation bound, the VAV flow bounds, and the outdoor-air damper bounds and provide these bounds to optimization manager 812 for performing the optimization. In various embodiments, the constraint generator 810 may also be configured to generate constraints for productivity scores and/or air quality metrics and provide the constraints to optimization manager 812.

For the infection probability constraint, the dynamic extension of the Wells-Riley equation implies that there should be an average constraint over a time interval during which an individual is in the building. An individual i's probability of infection P_i,[0,T] over a time interval [0, T] is given by:

$P_{i, [0, T]} = 1 - \exp (- p Δ \sum_{t} δ_{it} N_{t}), δ_{it} = {\begin{matrix} 1 & if i present at time t \\ 0 & else \end{matrix}$

according to some embodiments. Assuming that the individual's probability of infection P_i[0,T] is a known value, an upper bound P^maxcan be chosen for P_i,[0,T] and can be implemented as a linear constraint:

$\sum_{t} δ_{it} N_{t} \leq - \frac{1}{p Δ} \log (1 - P^{\max})$

according to some embodiments. In some embodiments, the variable δ_itmay be different for each individual in the building 10 but can be approximated as described herein.

The above linear constraint is an average constraint that gives optimization manager 812 (e.g., an optimizer) a maximum amount of flexibility since the average constraint may allow a higher concentration of infectious quanta during certain times of the day (e.g., when extra fresh-air ventilation is expensive due to outdoor ambient conditions) as long as the higher concentrations are balanced by lower concentrations of the infectious quanta during other times of the day. However, the δ_itsequence may be different for each individual in the building 10. For purposes of the example described herein it is assumed that generally each individual is present a total of 8 hours (e.g., if the building 10 is an office building). However, the estimated amount of time the individual is within the building can be adjusted or set to other values for other types of buildings. For example, when the systems and methods described herein are implemented in a restaurant or store, the amount of time the individual is assumed to be present in the building can be set to an average or estimated amount of time required to complete the corresponding activities (e.g., eating a meal, shopping, etc.). While an occupancy time of the building by each individual may be reasonably known, the times that the individual is present in the building may vary (e.g., the individual may be present from 7 AM to 3 PM, 9 AM to 5 PM, etc.). Therefore, to ensure that the constraint is satisfied for all possible Sit sequences, the constraint may be required to be satisfied when summing over 8 hours of the day that have a highest concentration.

This constraint is represented using linear constraints as:

$M η + \sum_{t} μ_{t} \leq - \frac{1}{p Δ} \log (1 - P^{\max})$

$μ_{t} + η \geq N_{t} \forall t$

where η and μ_tare new auxiliary variables in the optimization problem, and M is a number of discrete timesteps corresponding to 8 hours (or any other amount of time that an individual is expected to occupy building 10 or one of building zones 716). This formulation may work since η is set to an Mth highest value of N_tand each of the μ_tsatisfy μ_t=max(N_t−η,0). Advantageously, this implementation of the infection probability constraint can be generated by constraint generator 810 and provided to optimization manager 812 for use in the optimization problem when controller 710 is implemented to perform control decisions for HVAC system 100 (e.g., when controller 710 operates in an on-line mode).

An alternative implementation of the infection probability constraint is shown below that uses a pointwise constraint:

$N_{t} \leq N_{t}^{\max} = - \frac{1}{Mp Δ} \log (1 - P^{\max})$

where N_tis constrained to be less than or equal to N_t^maxfor a maximum infection probability value P^max. In some embodiments, the pointwise constraint shown above is generated by constraint generator 810 for when optimization manager 812 is used in an off-line or design implementation. In some embodiments, the pointwise constraint shown above, if satisfied in all zones 716, ensures that any individual will meet the infection probability constraint. Such a constraint may sacrifice flexibility compared to the other implementation of the infection probability constraint described herein, but translates to a simple box constraint similar to the other bounds in the optimization problem, thereby facilitating a simpler optimization process.

In some embodiments, the maximum allowable or desirable infection probability P^maxis a predetermined value that is used by constraint generator 810 to generate the infection probability constraints described herein. In some embodiments, constraint generator 810 is configured to receive the maximum allowable or desirable infection probability P^maxfrom a user as a user input. In some embodiments, the maximum allowable or desirable infection probability P^maxis an adjustable parameter that can be set by a user or automatically generated based on the type of infection, time of year, type or use of the building, or any of a variety of other factors. For example, some buildings (e.g., hospitals) may be more sensitive to preventing disease spread than other types of buildings and may use lower values of P^max. Similarly, some types of diseases may be more serious or life-threatening than others and therefore the value of P^maxcan be set to relatively lower values as the severity of the disease increases. In some embodiments, the value of P^maxcan be adjusted by a user and the systems and methods described herein can run a plurality of simulations or optimizations for a variety of different values of P^maxto determine the impact on cost and disease spread. A user can select the desired value of P^maxin view of the estimated cost and impact on disease spread using the results of the simulations or optimizations. In various embodiments, the constraint may be a productivity score constraint or an air quality metric constraint.

Model Enhancements

Referring still to FIG. 8, optimization manager 812, constraint generator 810, and/or model manager 816 can implement various model enhancements in the optimization. In some embodiments, optimization manager 812 is configured to add a decision variable for auxiliary (e.g., controlled) heating (e.g., via baseboard heat or VAV reheat coils). In some embodiments, an effect of the auxiliary heating is included in the dynamic model of temperature similar to the disturbance heat load Q_k(⋅). Similar to the other decision variables, the auxiliary heating decision variable may be subject to bounds (e.g., with both set to zero during cooling season to disable auxiliary heating) that are generated by constraint generator 810 and used by optimization manager 812 in the optimization problem formulation and solving. In some embodiments, the auxiliary heating also results in optimization manager 812 including another term for associated energy consumption in the energy consumption equation (shown above) that is minimized.

In some embodiments, certain regions or areas may have variable electricity prices and/or peak demand charges. In some embodiments, the objective function (e.g., the energy consumption equation) can be augmented by optimization manager 812 to account for such cost structures. For example, the existing energy consumption E_tthat is minimized by optimization manager 812 may be multiplied by a corresponding electricity price P_t. A peak demand charge may require the use of an additional parameter et that represents a base electric load of building 10 (e.g., for non-HVAC purposes). Optimization manager 812 can include such cost structures and may minimize overall cost associated with electricity consumption rather than merely minimizing electrical consumption. In some embodiments, optimization manager 812 accounts for revenue which can be generated by participating in incentive based demand response (IBDR) programs, frequency regulation (FR) programs, economic load demand response (ELDR) programs, or other sources of revenue when generating the objective function. In some embodiments, optimization manager 812 accounts for the time value of money by discounting future costs or future gains to their net present value. These and other factors which can be considered by optimization manager 812 are described in detail in U.S. Pat. No. 10,359,748 granted Jul. 23, 2019, U.S. Patent Application Publication No. 2019/0347622 published Nov. 14, 2019, and U.S. Patent Application Publication No. 2018/0357577 published Dec. 13, 2018, each of which is incorporated by reference herein in its entirety.

In some embodiments, certain locations have time-varying electricity pricing, and therefore there exists a potential to significantly reduce cooling costs by using a solid mass of building 10 for thermal energy storage. In some embodiments, controller 710 can operate to pre-cool the solid mass of building 10 when electricity is cheap so that the solid mass can later provide passive cooling later in the day when electricity is less expensive. In some embodiments, optimization manager 812 and/or model manager 816 are configured to model this effect using a model augmentation by adding a new variable T_k^mto represent the solid mass of the zone 716 evolving as:

$ρ c_{m} V_{k}^{m} \frac{{dT}_{k}^{m}}{dt} = h_{k}^{m} (T_{k} - T_{k}^{m})$

with a corresponding term:

$ρ {cV}_{k} \frac{{dT}_{k}}{dt} = \dots + h_{k}^{m} (T_{k}^{m} - T_{k})$

added to the air temperature model (shown above). This quantity can also be aggregated by model manager 816 to an average value T^msimilar to T.

For some diseases, infectious particles may naturally become deactivated or otherwise removed from the air over time. To consider these effects, controller 710 can add a proportional decay term to the infectious quanta model (in addition to the other terms of the infectious quanta model discussed above). An example is shown in the following equation:

$V \frac{dN}{dt} = \dots - V β N$

where β represents the natural decay rate (in s⁻¹) of the infectious species and the ellipsis represents the other terms of the infectious quanta model as discussed above. Because the natural decay subtracts from the total amount of infectious particles, the natural decay term is subtracted from the other terms in the infectious quanta model. For example, if a given infectious agent has a half-life t_1/2of one hour (i.e., t_1/2=1 hr=3600 s), then the corresponding decay rate is given by:

$β = \frac{\ln (2)}{t_{1 / 2}} \approx 1.925 \times 10^{- 4} s^{- 1}$

This extra term can ensure that infectious particle concentrations do not accumulate indefinitely over extremely long periods of time.

In various embodiments, air quality metrics may also naturally be removed from the air. The decay of air quality metrics may be modeled using the same or similar model as described above.

Off-Line Optimization

Referring particularly to FIG. 9, controller 710 can be configured for use as a design or planning tool for determining various design parameters of HVAC system 700 (e.g., for determining a size of filter 708, UV lights 706, etc.). In some embodiments, controller 710 implemented as a design tool, a planning tool, a recommendation tool, etc., (e.g., in an off-line mode) functions similarly to controller 710 implemented as a real-time control device (e.g., in an on-line mode). However, model manager 816, constraint generator 810, and optimization manager 812 may receive required sensor input data (i.e., model population data) from a simulation database 824. Simulation database 824 can store values of the various parameters of the constraints or boundaries, the dynamic models, or typical energy consumption costs or operational parameters for energy-consuming devices of the HVAC system 100. In some embodiments, simulation database 824 also stores predicted or historical values as obtained from sensors of HVAC system 100. For example, simulation database 824 can store typical ambient temperature, humidity, etc., conditions for use in performing the off-line simulation.

When controller 710 is configured for use as the design tool (shown in FIG. 9), controller 710 may receive user inputs from user input device 820. The user inputs may be initial inputs for various constraints (e.g., a maximum value of the probability of infection for the simulation, a minimum productivity score, etc.) or various required input parameters. The user can also provide simulation data for simulation database 824 used to populate the models or constraints, etc. Controller 710 can output suggestions of whether to use a particular piece of equipment (e.g., whether or not to use or install UV lights 706), whether to use AHU 704 to draw outside air, etc., or other factors to minimize cost (e.g., to optimize the objective function, minimize energy consumption, minimize energy consumption monetary cost, etc.), to meet disinfection goals (e.g., to provide a desired level of infection probability), and to meet productivity goals (e.g., have a productivity score at or above a predefined threshold). In some embodiments, controller 710 may provide different recommendations or suggestions based on a location of building 10. In some embodiments, the recommendations notify the user regarding what equipment is needed to keep the infection probability of zones 716 within the threshold while not increasing energy cost, energy consumption, or carbon footprint.

Compared to the on-line optimization (described in greater detail below), the optimization problem formulated by optimization manager 812 for the off-line implementation includes an additional constraint on the infectious quanta concentration (as described in greater detail above). In some embodiments, the infectious quanta concentration can be controlled or adjusted by (a) changing the airflow into each zone 716 (e.g., adjusting f_i), (b) changing the fresh-air intake fraction (e.g., adjusting x), or (c) destroying infectious particles in the AHU 704 via filtration or UV light (e.g., adjusting λ).

It should be noted that the first and second control or adjustments (e.g., (a) and (b)) may also affect temperature and humidity of the zones 716 of building 10. However, the third control option (c) (e.g., adjusting the infectious quanta concentration through filtration and/or operation of UV lights) is independent of the temperature and humidity of the zones 716 of building 10 (e.g., does not affect the temperature or humidity of zones 716 of building 10). In some embodiments, optimization manager 812 may determine results that rely heavily or completely on maintaining the infectious quanta concentration below its corresponding threshold or bound through operation of filter 708 and/or UV lights 706. However, there may be sufficient flexibility in the temperature and humidity of building zone 716 so that optimization manager 812 can determine adjustments to (a), (b), and (c) simultaneously to achieve lowest or minimal operating costs (e.g., energy consumption). Additionally, since purchasing filters 708 and/or UV lights 706 may incur significant capital costs (e.g., to purchase such devices), controller 710 may perform the optimization as a simulation to determine if purchasing filters 708 and/or UV lights 706 is cost effective.

When controller 710 is configured as the design tool shown in FIG. 9, controller 710 may provide an estimate of a total cost (both capital costs and operating costs) to achieve a desired level of infection control (e.g., to maintain the infection probability below or at a desired amount) and/or a desired productivity level (e.g., to maintain the productivity score at or above a desired level). The purpose is to run a series of independent simulations, assuming different equipment configurations (e.g., as stored and provided by simulation database 824) and for different infection probability constraints and/or productivity score constraints given typical climate and occupancy data (e.g., as stored and provided by simulation database 824). In some embodiments, the different equipment configurations include scenarios when filters 708 and/or UV lights 706 are installed in the HVAC system 100, or when filters 708 and/or UV lights 706 are not installed in the HVAC system 100.

After performing the simulation for different equipment configuration scenarios and/or different infection probability constraints/productivity score constraints, controller 710 can perform a cost benefit analysis based on global design decisions (e.g., whether or not to install UV lights 706 and/or filters 708). The cost benefit analysis may be performed by results manager 818 and the cost benefit analysis results can be output as display data to a building manager via display device 822. These results may aid the building manager or a building designer in assessing potential options for infection control of building 10 (as shown in FIG. 12).

Referring particularly to FIGS. 9 and 12, graph 1200 illustrates a potential output of results manager 818 that can be displayed by display device 822. Graph 1200 illustrates relative cost (the Y-axis) with respect to infection probability (the X-axis) for a case when both filtration and UV lights are used for infection control (represented by series 1208), a case when filtration is used for infection control without using UV lights (represented by series 1202), a case when UV lights are used for infection control without using filtration (represented by series 1206), and a case when neither UV lights and filtration are used for infection control (represented by series 1204). In some embodiments, each of the cases illustrated by series 1202-1208 assume that fresh-air intake is used to control infection probability. Data associated with graph 1200 can be output by results manager 818 so that graph 1200 can be generated and displayed on display device 822.

In some embodiments, the off-line optimization performed by optimization manager 812 is faster or more computationally efficient than the on-line optimization performed by optimization manager 812. In some embodiments, the simulation is performed using conventional rule-based control rather than a model-predictive control scheme used for the on-line optimization. Additionally, the simulation may be performed over shorter time horizons than when the optimization is performed in the on-line mode to facilitate simulation of a wide variety of design conditions.

In some embodiments, optimization manager 812 is configured to use the aggregate dynamic models as generated, populated, and provided by model manager 816 for the off-line optimization (e.g., the design optimization). When optimization manager 812 uses the aggregate dynamic models, this implies that there are three decision variables of the optimization: f, x, and λ. The variable λ can include two positions at each timestep (e.g., corresponding to the UV lights 706 being on or the UV lights 706 being off). A reasonable grid size of f and x may be 100. Accordingly, this leads to 100×100×2=20,000 possible combinations of control decisions at each step, which is computationally manageable. Therefore, optimization manager 812 can select values of the variables f, x, and λ via a one-step restriction of the optimization problem by simply evaluating all possible sets of control inputs and selecting the set of control inputs that achieves a lowest cost.

If additional variables are used, such as an auxiliary heating variable, this may increase the dimensionality of the optimization problem. However, optimization manager 812 can select a coarser grid (e.g., 5 to 10 choices) for the additional variable.

In some embodiments, optimization manager 812 is configured to solve a number of one-step optimization problems (e.g., formulate different optimization problems for different sets of the control objectives and solve the optimization problem over a single timestep) in a training period, and then train a function approximator (e.g., a neural network) to recreate a mapping. This can improve an efficiency of the optimization. In some embodiments, optimization manager 812 is configured to apply a direct policy optimization to the dynamic models in order to directly learn a control law using multiple parallel optimization problems.

In some embodiments, when controller 710 functions as the design tool shown in FIG. 9, there are two design variables. The first design variable is whether it is cost effective or desirable to purchase and install UV lights 706, and the second design variable is whether it is cost effective or desirable to purchase and install filters 708 (e.g., advanced filtration devices).

In some embodiments, optimization manager 812 is configured to perform a variety of simulations subject to different simulation variables for each simulation month. These simulation variables can be separated into a design decision category and a random parameter category. The design decision category includes variables whose values are chosen by system designers, according to some embodiments. The random parameters category includes variables whose values are generated by external (e.g., random) processes.

The design decision category can include a first variable of whether to activate UV lights 706. The first variable may have two values (e.g., a first value for when UV lights 706 are activated and a second value for when UV lights 706 are deactivated). The design decision category can include a second decision variable of which of a variety of high-efficiency filters to use, if any. The second variable may have any number of values that the building manager wishes to simulate (e.g., 5) and can be provided via user input device 820. The design decisions category can also include a third variable of what value should be used for the infection probability constraint or the productivity score constraint (as provided by constraint generator 810 and used in the optimization problem by optimization manager 812). In some embodiments, various values of the third variable are also provided by the user input device 820. In some embodiments, various values of the third variable are predetermined or stored in simulation database 824 and provided to optimization manager 812 for use in the simulation. The third variable may have any number of values as desired by the user (e.g., 3 values).

The random parameters category can include an ambient weather and zone occupancy variable and a number of infected individuals that are present in building 10 variable. In some embodiments, the ambient weather and zone occupancy variable can have approximately 10 different values. In some embodiments, the number of infected individuals present can have approximately 5 different values.

In order to determine a lowest cost for a given month, optimization manager 812 can aggregate the variables in the random parameters category (e.g., average) and then perform an optimization to minimize the energy consumption or cost over feasible values of the variables of the design decisions category. In some embodiments, some of the design-decision scenarios are restricted by a choice of global design decisions. For example, for optimization manager 812 to calculate monthly operating costs assuming UV lights 706 are chosen to be installed but not filtration, optimization manager 812 may determine that a lowest cost scenario across all scenarios is with no filtration but with the UV lights 706 enabled or disabled. While this may be unusual (e.g., for the UV lights 706 to be disabled) even though the UV lights 706 are installed, various conditions (e.g., such as weather) may make this the most cost effective solution.

In some embodiments, simulation logic performed by optimization manager 812 may be performed in a Tensorflow (e.g., as operated by a laptop computer, or any other sufficiently computationally powerful processing device). In order to perform 1,500 simulation scenarios for each month, or 18,000 for an entire year, with a timestep of 15 minutes, this implies a total of approximately 52 million timesteps of scenarios for a given simulation year.

In some embodiments, optimization manager 812 requires various simulation data in order to perform the off-line simulation (e.g., to determine the design parameters). In some embodiments, the simulation data is stored in simulation database 824 and provided to any of constraint generator 810, model manager 816, and/or optimization manager 812 as required to perform their respective functions. The simulation data stored in simulation database 824 can include heat-transfer parameters for each zone 716, thermal and moisture loads for each zone 716, coil model parameters of the AHU 704, fan model parameters of the AHU 704, external temperature, humidity, and solar data, filtration efficiency, pressure drop, and cost versus types of the filter 708, disinfection fraction and energy consumption of the UV lights 706, installation costs for the UV lights 706 and the filter 708, typical breathing rate p, a number of infected individuals Ī in building zones 716, and disease quanta generation q values for various diseases. In some embodiments, the heat-transfer parameters for each zone 716 may be obtained by simulation database 824 from previous simulations or from user input device 820. In some embodiments, the thermal and moisture loads for each zone 716 are estimated based on an occupancy of the zones 716 and ASHRAE guidelines. After this simulation data is obtained in simulation database 824, controller 710 may perform the simulation (e.g., the off-line optimization) as described herein.

It should be understood that as used throughout this disclosure, the term “optimization” may signify a temporal optimization (e.g., across a time horizon) or a static optimization (e.g., at a particular moment in time, an instantaneous optimization). In some embodiments, optimization manager 812 is configured to either run multiple optimizations for different equipment selections, or is configured to treat equipment configurations as decision variables and perform a single optimization to determine optimal equipment configurations.

It should also be understood that the term “design” as used throughout this disclosure (e.g., “design data” and/or “design tool”) may include equipment recommendations (e.g., recommendations to purchase particular equipment or a particular type of equipment such as a particular filter) and/or operational recommendations for HVAC system 100. In other words, “design data” as used herein may refer to any information, metrics, operational data, guidance, suggestion, etc., for selecting equipment, an operating strategy, or any other options to improve financial metrics or other control objectives (e.g., comfort and/or infection probability).

For example, controller 710 as described in detail herein with reference to FIG. 9 may be configured to provide recommendations of specific models to purchase. In some embodiments, controller 710 is configured to communicate with an equipment performance database to provide product-specific selections. For example, controller 710 can search the database for equipment that has particular specifications as determined or selected by the optimization. In some embodiments, controller 710 may also provide recommended or suggested control algorithms (e.g., model predictive control) as the design data. In some embodiments, controller 710 may provide a recommendation or suggestion of a general type of equipment or a general equipment configuration without specifying a particular model. In some embodiments, controller 710 may also recommend a specific filter or a specific filter rating. For example, optimization manager 812 can perform multiple optimizations with different filter ratings and select the filter ratings associated with an optimal result.

On-Line Optimization

Referring again to FIG. 8, controller 710 can be implemented as an on-line controller that is configured to determine optimal control for the equipment of building 10. Specifically, controller 710 may determine optimal operation for UV lights 706 and AHU 704 to minimize energy consumption after UV lights 706 and/or filter 708 are installed and HVAC system 100 is operational. When controller 710 is configured as an on-line controller, controller 710 may function similarly to controller 710 as configured for off-line optimization and described in greater detail above with reference to FIG. 9. However, controller 710 can determine optimal control decisions for the particular equipment configuration of building 10.

In some embodiments, optimization manager 812 is configured to perform model predictive control similar to the techniques described in U.S. patent application Ser. No. 15/473,496, filed Mar. 29, 2017, the entire disclosure of which is incorporated by reference herein.

While optimization manager 812 can construct and optimize the optimization problem described in greater detail above, and shown below, using MPC techniques, a major difference is that optimization manager 812 performs the optimization with an infectious quanta concentration model as described in greater detail above.

Therefore, the resulting optimization problem has additional constraints on this new variable (the infectious quanta concentration) but also new flexibility by determined decisions for activating UV lights 706. In some embodiments, the optimization performed by optimization manager 812 can balance, in real time, a tradeoff between takin gin additional outdoor air (which generally incurs a cooling energy penalty) and activating the UV lights 706 (which requires electricity consumption). Additionally, the addition of infectious agent control can also provide additional room optimization of HVAC system 100 during a heating season (e.g., during winter). Without considering infectious quanta concentrations, solutions generally lead to a minimum outdoor airflow below a certain break-even temperature, below which heating is required throughout building 10. However, since the optimization problem formulated by optimization manager 812 can determine to increase outdoor air intake, this can provide an additional benefit of disinfection.

For purposes of real-time or on-line optimization, the HVAC system 100 can be modeled on a zone-by-zone basis due to zones 716 each having separate temperature controllers and VAV boxes. In some embodiments, zone-by-zone temperature measurements are obtained by controller 710 from zone sensors 712 (e.g., a collection of temperature, humidity, CO2, air quality, etc., sensors that are positioned at each of the multiple zones 716). In some embodiments, optimization manager 812 is configured to use zone-level temperature models but aggregate humidity and infectious quanta models for on-line optimization. Advantageously, this can reduce a necessary modeling effort and a number of decision variables in the optimization problem. In some embodiments, if the AHU 704 serves an excessive number of zones 716, the zone-level thermal modeling may be too computationally challenging so optimization manager 812 can use aggregate temperature models.

After optimization manager 812 has selected whether to use individual or aggregate models (or some combination thereof), optimization manager 812 can implement a constraint in the form:

$\frac{dx}{dt} = f (x (t), u (t), p (t)) for all t \in [0, T]$

given a horizon t, where u(t) is a decision, control, or adjustable variable, and p(t) are time-varying parameters (the values of which are forecasted ahead of time). In some embodiments, optimization manager 812 is configured to implement such a constraint by discretizing the u(t) and p(t) signals into piecewise-constant values u_nand p_nwhere the discrete index n represents the time interval t∈[nΔ,(n+1)Δ] for a fixed sample time Δ. Optimization manager 812 may then transform the constraint to:

$\frac{dx}{dt} = f (x (t), u_{j}, p_{j}) for all t \in [n Δ, (n + 1) Δ] and n \in {0, \dots, N - 1}$

where N=T/Δ the total number of timesteps. In some embodiments, optimization manager 812 is configured to evaluate this constraint using advanced quadrature techniques. For example, optimization manager 812 may transform the constraint to:

$x_{n + 1} = F (x_{n}, u_{n}, p_{n})$

where x(t) is discretized to x_nand F(⋅) represents a numerical quadrature routine. In some embodiments, if the models provided by model manager 816 are sufficiently simple, optimization manager 812 can derive an analytical expression for F(⋅) to perform this calculation directly.

In some embodiments, optimization manager 812 uses an approximate midpoint method to derive the analytical expression:

$x_{n + 1} = x_{k} + f (\frac{x_{n + 1} + x_{n}}{2}, u_{n}, p_{n}) Δ$

where the ordinary differential equation f(⋅) is evaluated at a midpoint of the time interval.

In some embodiments, optimization manager 812 is configured to repeatedly solve the optimization problem at regular intervals (e.g., every hour) to revise an optimized sequence of inputs for control signal generator 808. However, since the optimization is nonlinear and nonconvex, it may be advantageous to decrease a frequency at which the optimization problem is solved to provide additional time to retry failed solutions.

In some embodiments, optimization manager 812 uses a daily advisory capacity. For example, optimization manager 812 may construct and solve the optimization problem once per day (e.g., in the morning) to determine optimal damper positions (e.g., of AHU 704), UV utilizations (e.g., operation of UV lights 706), and zone-level airflows. Using the results of this optimization, optimization manager 812 may be configured to pre-schedule time-varying upper and lower bounds on the various variables of the optimized solution, but with a range above and below so that optimization manager 812 can have sufficient flexibility to reject local disturbances. In some embodiments, regulatory control systems of HVAC system 100 are maintained but may saturate at new tighter bounds obtained from the optimization problem. However, optimization manager 812 may be configured to re-optimize during a middle of the day if ambient sensor data from ambient sensors 714 (e.g., ambient temperature, outdoor temperature, outdoor humidity, etc.) and/or weather forecasts and/or occupancy forecasts indicate that the optimization should be re-performed (e.g., if the weather forecasts are incorrect or change).

In some embodiments, optimization manager 812 is configured to reduce an amount of optimization by training a neural network based on results of multiple offline optimal solutions (e.g., determined by controller 710 when performing off-line optimizations). In some embodiments, the neural network is trained to learn a mapping between initial states and disturbance forecasts to optimal control decisions. The neural network can be used in the online implementation of controller 710 as a substitute for solving the optimization problem. One advantage of using a neural network is that the neural network evaluation is faster than performing an optimization problem, and the neural network is unlikely to suggest poor-quality local optima (provided such solutions are excluded from the training data). The neural network may, however, return nonsensical values for disturbance sequences. However, this downside may be mitigated by configuring controller 710 to use a hybrid trust-region strategy in which optimization manager 812 solves the optimization problem via direct optimization at a beginning of the day, and then for the remainder of the day, controller 710 uses neural-network suggestions if they are within a predefined trust region of the optimal solution. If a neural-network suggestion is outside of the predefined trust region, optimization manager 812 may use a previous optimal solution that is within the predefined trust region.

In some embodiments, the optimization problem is formulated by optimization manager 812 assuming the zone-level VAV flows f_kare the decision variables. In some systems, however, a main interface between controller 710 and equipment of HVAC system 100 is temperature setpoints that are sent to zone-level thermostats. In some embodiments, optimization manager 812 and control signal generator 808 are configured to shift a predicted optimal temperature sequence backward by one time interval and then pass these values (e.g., results of the optimization) as the temperature setpoint. For example, if the forecasts over-estimate head loads in a particular zone 716, then a VAV damper for that zone will deliver less airflow to the zone 716, since less cooling is required to maintain a desired temperature.

When optimization manager 812 uses the constraint on infectious quanta concentration, controller 710 can now use the zone-level airflow to control two variables, while the local controllers are only aware of one. Therefore, in a hypothetical scenario, the reduced airflow may infection control result in a violation of the constraint on infection probability. In some embodiments, optimization manager 812 and/or control signal generator 808 are configured to maintain a higher flow rate at the VAV even though the resulting temperature may be lower than predicted. To address this situation, optimization manager 812 may use the minimum and maximum bounds on the zone-level VAV dampers, specifically setting them to a more narrow range so that the VAV dampers are forced to deliver (at least approximately) an optimized level of air circulation. In some embodiments, to meet the infectious quanta concentration, the relevant bound is the lower flow limit (as any higher flow will still satisfy the constraint, albeit at higher energy cost or energy consumption). In some embodiments, a suitable strategy is to set the VAV minimum position at the level that delivers 75% to 90% of the optimized flow. In some embodiments, a VAV controller is free to dip slightly below the optimized level when optimization manager 812 over-estimates heat loads, while also having the full flexibility to increase flow as necessary when optimization manager 812 under-estimates heat loads. In the former case, optimization manager 812 may slightly violate the infectious quanta constraint (which could potentially be mitigated via rule-based logic to activate UV lights 706 if flow drops below planned levels), while in the latter case, the optimal solution still satisfies the constraint on infectious quanta. Thus, optimization manager 812 can achieve both control goals without significant disruption to the low-level regulatory controls already in place.

HVAC System with Building Infection Control

On-Line Optimization Process

Referring particularly to FIG. 10, a process 1000 for performing an on-line optimization to minimize energy consumption and satisfy an infection probability constraint in a building is shown, according to some embodiments. Process 1000 can be performed by controller 710 when controller 710 is configured to perform an on-line optimization. In some embodiments, process 1000 is performed in real time for HVAC system 100 to determine optimal control of AHU 704 and/or UV lights 706. Process 1000 can be performed for an HVAC system that includes UV lights 706 configured to provide disinfection for supply air that is provided to one or more zones 206 of a building 10, filter 708 that filters an air output of an AHU, and/or an AHU (e.g., AHU 704). Process 1000 can also be performed for HVAC systems that do not include filter 708 and/or UV lights 706.

Process 1000 includes determining a temperature model for each of multiple zones to predict a temperature of a corresponding zone based on one or more conditions or parameters of the corresponding zone (step 1002), according to some embodiments. The temperature model can be generated or determined by model manager 816 for use in an optimization problem. In some embodiments, the temperature model is:

$ρ c V_{k} (\frac{{dT}_{k}}{dt}) = ρ c f_{k} (T_{0} - T_{k}) + Q_{k} (T_{k})$

where ρ is a mass density of air, c is a heat capacity of air, V_kis a volume of the kth zone, f_kis a volumetric flow of air into the kth zone, T₀is the temperature of air output by the AHU, T_kis the temperature of the kth zone, and Q_kis the heat load on the kth zone. Step 1002 can be performed by model manager 816 as described in greater detail above with reference to FIGS. 8-9.

Process 1000 includes determining a humidity model for each of the multiple zones to predict a humidity of the corresponding zone based on one or more conditions or parameters of the corresponding zone (step 1004), according to some embodiments. Step 1004 can be similar to step 1002 but for the humidity model instead of the temperature model. In some embodiments, the humidity model is:

$ρ V_{k} (\frac{d ω_{k}}{dt}) = ρ f (ω_{0} - T_{0}) + w_{k}$

for a kth zone 206. In some embodiments, step 1004 is performed by model manager 816 as described in greater detail above with reference to FIGS. 8-9.

Process 1000 incudes determining an infectious quanta concentration model for each of the multiple zones to predict an infectious quanta of the corresponding zone based on one or more conditions or parameters of the corresponding zone (step 1006), according to some embodiments. In some embodiments, the infectious quanta concentration model is similar to the humidity model of step 1004 or the temperature model of step 1002. The infectious quanta concentration model can be:

$V_{k} (\frac{{dN}_{k}}{dt}) = f_{k} (N_{0} - N_{k}) + I_{k} q$

according to some embodiments. In some embodiments, step 1006 is performed by model manager 816.

Process 1000 includes determining an aggregated temperature model, an aggregated humidity model, an aggregated infectious quanta model, an aggregated thermal model, and an aggregated moisture model (step 1008), according to some embodiments. In some embodiments, step 1008 is optional. Step 1008 can include generating or determining each of the aggregated models by determining a volume-average across zones 206. The aggregate infectious quanta model is:

$\bar{V} \frac{d \bar{N}}{dt} = \overline{I} q - (λ + x - λ x) \sum_{k} f_{k} N_{k} \approx \bar{I} q - (λ + x - λ x) \overline{f} \overline{N}$

according to some embodiments. The aggregated thermal model is:

$ρ c \bar{V} \frac{d \bar{T}}{dt} = \sum_{k} Q_{k} (T_{k}) + ρ c \sum_{k} f_{k} (x (T_{a} - T_{k}) - Δ T_{c}) \approx \bar{Q} (\bar{T}) + ρ c \bar{f} (x (T_{a} - \bar{T}) - Δ T_{c})$

according to some embodiments. The aggregated moisture model is:

$\begin{matrix} ρ \bar{V} \frac{d \bar{ω}}{dt} = \bar{w} + ρ \sum_{k} f_{k} (x (ω_{a} - ω_{k}) - Δ ω_{c}) \\ \approx \bar{w} + ρ \bar{f} (x (ω_{a} - \bar{ω}) - Δ ω_{c}) \end{matrix}$

according to some embodiments. In some embodiments, the aggregated thermal and moisture models are aggregate thermal models. Step 1008 can be optional. Step 1008 can be performed by model manager 816.

Process 1000 includes populating any of the temperature model, the humidity model, the infectious quanta model, or the aggregated models using sensor data or stored values (step 1010), according to some embodiments. In some embodiments, step 1010 is performed by model manager 816. In some embodiments, step 1010 is optional. Step 1010 can be performed based on sensor data obtained from zone sensors 712.

Process 1000 includes determining an objective function including a cost of operating an HVAC system that serves the zones (step 1012), according to some embodiments. In some embodiments, step 1012 is performed by optimization manager 812 using the dynamic models and/or the aggregated models provided by model manager 816. The objective function may be a summation of the energy consumption, energy cost, or other variable of interest over a given time period. The instantaneous energy consumption at a discrete time step is given by:

$E = η_{coil} ρ \bar{f} (c Δ T_{c} + L Δ ω_{c}) + η_{fan} \bar{f} Δ P + η_{UV} λ_{UV}$

which can be summed or integrated over all time steps of the given time period as follows:

$\int_{0}^{T} E (t) dt \approx Δ \sum_{t} E_{t}$

where Δ is the duration of a discrete time step, according to some embodiments.

Process 1000 includes determining one or more constraints for the objective function including an infection probability constraint (step 1014), according to some embodiments. In some embodiments, step 1014 is performed by constraint generator 810. The one or more constraints can include the infection probability constraint, temperature bounds or constraints, humidity bounds or constraints, fresh-air ventilation bounds or constraints, VAV flow bounds or constraints, and/or outdoor-air damper bounds or constraints. The infection probability constraint is:

$M η + \sum_{t} μ_{t} \leq - \frac{1}{p Δ} \log (1 - P^{m ax}) μ_{t} + η \geq N_{t} \forall t or: N_{t} \leq N_{t}^{m ax} = - \frac{1}{M p Δ} \log (1 - P^{m ax})$

according to some embodiments.

Process 1000 includes performing an optimization to determine control decisions for HVAC equipment of the HVAC system, and ultraviolet lights of the HVAC system such that the one or more constraints are met and the cost is minimized (step 1016), according to some embodiments. Step 1016 can be performed by optimization manager 812 by minimizing the objective function subject to the one or more constraints (e.g., the temperature, humidity, etc., bounds and the infection probability constraint). Step 1016 can also include constructing the optimization problem and constructing the optimization problem based on the objective function, the dynamic models (or the aggregated dynamic models), and the one or more constraints. The control decisions can include a fresh-air fraction x for an AHU of the HVAC system (e.g., AHU 704), whether to turn on or off the UV lights, etc.

Off-Line Optimization Process

Referring particularly to FIG. 11, a process for performing an off-line optimization to determine equipment configurations that minimize energy consumption or cost and satisfy an infection probability constraint is shown, according to some embodiments. Process 1100 may share similarities with process 1000 but can be performed in an off-line mode (e.g., without determining control decisions or based on real-time sensor data) to determine or assess various design decisions and provide design information to a building manager. Process 1100 can be performed by controller 710 when configured for the off-line mode (as shown in FIG. 9).

Process 1100 includes steps 1102-1108 that can be the same as steps 1002-1008 of process 1000. However, while step 1008 may be optional in process 1000 so that the optimization can be performed using a combination of individual dynamic models and aggregate dynamic models, step 1108 may be non-optional in process 1100. In some embodiments, using the aggregate dynamic models reduces a computational complexity of the optimization for process 1100. Process 1100 can be performed for a wide variety of design parameters (e.g., different equipment configurations) whereas process 1000 can be performed for a single equipment configuration (e.g., the equipment configuration that process 1000 is used to optimize). Therefore, it can be advantageous to use aggregate models in process 1100 to reduce a complexity of the optimization problem.

Process 1100 includes populating the aggregated models using simulation data (step 1110). In some embodiments, step 1110 is performed by model manager 816 using outputs from simulation database 824 (e.g., using values of various parameters of the aggregate models that are stored in simulation database 824). In some embodiments, step 1110 is performed using known, assumed, or predetermined values to populate the aggregated models.

Process 1100 includes determining an objective function including a cost of operating an HVAC system that serves the zones (step 1112), and determining one or more constraints for the objective function including an infection probability constraint (step 1114), according to some embodiments. In some embodiments, step 1112 and step 1114 are similar to or the same as steps 1012 and 1014 of process 1000.

Process 1100 includes performing a sequence of one-step optimizations for various equipment configurations to estimate an operating cost associated with that equipment configuration (step 1116), according to some embodiments. In some embodiments, step 1116 is performed by optimization manager 812. Optimization manager 812 can construct different optimization problems for different equipment configurations using the aggregate temperature model, the aggregated humidity model, the aggregated infectious quanta model, the one or more constraints, and the objective function. In some embodiments, optimization manager 812 is configured to solve the optimization problems for the different equipment configurations over a single time step. The results of the optimizations problems can be output to results manager 818 for displaying to a user.

Process 1100 includes outputting design suggestions or optimizations results to a user (step 1118), according to some embodiments. In some embodiments, step 1118 includes outputting costs associated with different equipment configurations (e.g., equipment configurations that include UV lights for disinfection and/or filters for disinfection) to a user (e.g., via a display device) so that the user (e.g., a building manager) can determine if they wish to purchase additional disinfection equipment (e.g., UV lights and/or filters). For example, step 1118 can include operating a display to provide graph 1200 (or a similar graph) to a user.

Although process 1100 is described primarily as an “off-line” process, it should be understood that process 1100 is not limited to off-line implementations only. In some embodiments, process 1100 can be used when controller 710 operates in an on-line mode (as described with reference to FIGS. 8 and 10). In some embodiments, the results generated by performing process 1100 and/or the results generated when operating controller 710 in the off-line mode (e.g., recommended equipment configurations, recommended operating parameters, etc.) can be used to perform on-line control of HVAC equipment or perform other automated actions. For example, controller 710 can use the recommended equipment configurations to automatically enable, disable, or alter the operation of HVAC equipment in accordance with the recommended equipment configurations (e.g., enabling the set of HVAC equipment associated with the lowest cost equipment configuration identified by the simulations/optimizations). Similarly, controller 710 can use the recommended operating parameters to generate and provide control signals to the HVAC equipment (e.g., operating the HVAC equipment in accordance with the recommended operating parameters).

In general, the controller 710 can use the optimization/simulation results generated when operating controller 710 in the off-line mode to generate design data including one or more recommended design parameters (e.g., whether to include or use UV lights 706 for disinfection, whether to include or use filter 708 for disinfection, whether to use fresh/outdoor air for disinfection, a recommended type or rating of UV lights 706 or filter 708, etc.) as well as operational data including one or more recommended operational parameters (e.g., the fraction of fresh/outdoor air that should exist in the supply air provided to the building zone, operating decisions for UV lights 706, an amount of airflow to send to each building zone, etc.). The design data may include a recommended equipment configuration that specifies which HVAC equipment to use in the HVAC system to optimize the energy consumption, energy cost, carbon footprint, or other variable of interest while ensuring that a desired level of disinfection is provided.

Controller 710 can perform or initiate one or more automated actions using the design data and/or the operational data. In some embodiments, the automated actions include automated control actions such as generating and providing control signals to UV lights 706, AHU 704, one or more VAV units, or other types of airside HVAC equipment that operate to provide airflow to one or more building zones. In some embodiments, the automated action include initiating a process to purchase or install the recommended set of HVAC equipment defined by the design data (e.g., providing information about the recommended set of HVAC equipment to a user, automatically scheduling equipment upgrades, etc.). In some embodiments, the automated actions include providing the design data and/or the operational data to a user interface device (e.g., display device 822) and/or obtaining user input provided via the user interface device. The user input may indicate a desired level of disinfection and/or a request to automatically update the results of the optimizations/simulations based on user-selected values that define the desired infection probability or level of disinfection. Controller 710 can be configured to provide any of a variety of user interfaces (examples of which are discussed below) to allow a user to interact with the results of the optimizations/simulations and adjust the operation or design of the HVAC system based on the results.

User Interfaces

Referring now to FIGS. 9 and 13, in some embodiments, user input device 820 is configured to provide a user interface 1300 to a user. An example of a user interface 1300 that can be generated and presented via user input device 820 is shown in FIG. 13. User interface 1300 may allow a user to provide one or more user inputs that define which equipment are available in the building or should be considered for design purposes (e.g., filtration, UV, etc.) as well as the desired infection probability (e.g., low, medium, high, percentages, etc.). The inputs provided via user interface 1300 can be used by controller 710 to set up the optimization problem or problems to be solved by optimization manager 812. For example, constraint generator 810 can use the inputs received via user interface 1300 to generate the various bounds, boundaries, constraints, infection probability constraint, etc., that are used by optimization manager 812 to perform the optimization. After completing all of the simulation scenarios, the results can be presented to the user via the “Results” portion of user interface 1300 that allows the user to explore various tradeoffs.

As an example, the “Building Options” portion of user interface 1300 allows the user to specify desired building and climate parameters such as the square footage of the building, the city in which the building is located, etc. The user may also specify whether UV disinfection and/or advanced filtration should be considered in the simulation scenarios (e.g., by selecting or deselecting the UV and filtration options). The “Disinfection Options” portion of user interface 1300 allows the user to specify the desired level of disinfection or infection probability. For example, the user can move the sliders within the Disinfection Options portion of user interface 1300 to define the desired level of disinfection for each month (e.g., low, high, an intermediate level, etc.). Alternatively, user interface 1300 may allow the user to define the desired level of disinfection by inputting infection probability percentages, via a drop-down menu, by selecting or deselecting checkboxes, or any other user interface element.

After specifying the desired parameters and clicking the “Run” button, optimization manager 812 may perform one or more simulations (e.g., by solving one or more optimization problems) using the specified parameters. Once the simulations have completed, results may be displayed in the “Results” portion of user interface 1300. The results may indicate the energy cost, energy consumption, carbon footprint, or any other metric which optimization manager 812 seeks to optimize for each of the design scenarios selected by the user (e.g., UV+Filtration, UV Only, Filtration Only, Neither). The results may also indicate the daily infection probability for each of the design scenarios (e.g., mean infection probability, minimum infection probability, maximum infection probability). In some embodiments, an initial simulation or simulations are run using default settings for the disinfection options. In some embodiments, the results include equipment recommendations (e.g., use UV+Filtration, use UV Only, use Filtration Only, use Neither). The results of each simulation can be sorted to present the most optimal results first and the least optimal results last. For example, user interface 1300 is shown presenting the simulation result with the least energy consumption first and the most energy consumption last. In other embodiments, the results can be sorted by other criteria such as infection probability or any other factor.

The user can adjust desired disinfection options on a monthly basis (e.g., by adjusting the sliders within the Disinfection Options portion of user interface 1300), at which point the results may be re-calculated by averaging over the appropriate subset of simulation instances, which can be performed in real time because the simulations need not repeated. Advantageously, this allows the user to adjust the disinfection options and easily see the impact on energy cost, energy consumption, carbon footprint, etc., as well as the impact on infection probability for each of the design scenarios. Additional display options beyond what is shown in FIG. 13 may be present in various embodiments, for example to selectively disable UV and/or filtration in certain months or to consider worst-case instances for each month rather than mean values. In addition, various other graphical displays could be added to provide more detailed results. User interface 1300 may initially present optimization results and/or equipment recommendations based on default settings, but then the user is free to refine those settings and immediately see updates to cost estimates and suggested equipment.

Although a specific embodiment of user interface 1300 is shown in FIG. 13, it should be understood that this example is merely one possible user interface that can be used in combination with the systems and methods described herein. In general, controller 710 can operate user input device 820 to provide a user interface that includes various sliders, input fields, etc., to receive a variety of user inputs from the user via user input device 820. In some embodiments, user input device 820 is configured to receive a desired level of disinfection, a desired level of infection probability, etc., from the user and provide the desired level of disinfection, or desired level of infection probability to constraint generator 810 as the user input(s). In some embodiments, the user interface includes a knob or a slider that allows the user to adjust between a level of energy savings and a level of infection control. For example, the user may adjust the knob or slider on the user input device 820 to adjust the infection probability constraint (e.g., to adjust thresholds or boundaries associated with the infection probability constraint). In some embodiments, the user

In some embodiments, an infection spread probability is treated by constraint generator 810 as a constraint, or as a value that is used by constraint generator 810 to determine the infection probability constraint. If a user desires to provide a higher level of disinfection (e.g., a lower level of infection spread probability) and therefore an increased energy consumption or energy consumption cost, the user may adjust the knob or slider on the user interface of user input device 820 to indicate a desired trade-off between energy consumption and infection probability. Likewise, if the user desired to provide a lower level of disinfection (e.g., a higher level of infection spread probability) and therefore a lower energy consumption or energy consumption cost, the user may adjust the knob or slider on the user interface of the user input device 820 to indicate such a desired tradeoff between energy consumption or energy consumption cost and disinfection control.

In some embodiments, user input device 820 is configured to provide analytics, data, display data, building data, operational data, diagnostics data, energy consumption data, simulation results, estimated energy consumption, or estimated energy consumption cost to the user via the user interface of user input device 820. For example, results manager 818 may operate the user input device 820 and/or the display device 822 to provide an estimated energy consumption or energy consumption cost to the user (e.g., results of the optimization of optimization manager 812 when operating in either the on-line or off-line mode/configuration). In some embodiments, user input device 820 and display device 822 are a same device (e.g., a touchscreen display device, etc.) that are configured to provide the user interface, while in other embodiments, user input device 820 and display device 822 are separate devices that are configured to each provide their own respective user interfaces.

For example, controller 710 can perform the off-line or planning or design tool functionality as described in greater detail above in real-time (e.g., as the user adjusts the knob or slider) to determine an estimated energy consumption or energy consumption cost given a particular position of the knob or slider (e.g., given a particular desired level of infection or disinfection control as indicated by the position of the knob or slider). In some embodiments, controller 710 is configured to operate the user input device 820 and/or the display device 822 to provide or display the estimated energy consumption or estimated energy consumption cost as the user adjusts the knob or slider. In this way, the user can be informed regarding an estimation of costs or energy consumption associated with a specific level of disinfection control (e.g., with a particular infection probability constraint). Advantageously, providing the estimation of costs or energy consumption associated with the specific level of disinfection control to the user in real-time or near real-time facilitates the user selecting a level of disinfection control that provides sufficient or desired disinfection control in addition to desired energy consumption or energy consumption costs.

Pareto Optimization

Referring now to FIG. 14, a graph 1400 illustrating a Pareto search technique which can be used by controller 710 is shown, according to an exemplary embodiment. In some cases, users may want a more detailed tradeoff analysis than merely comparing a set of optimization results for a set of selected infection probabilities. For such cases, controller 710 may use a more detailed Pareto search that iteratively determines points on a Pareto front 1402 for an energy cost vs. infection probability tradeoff curve. By running additional simulations, this tradeoff curve can be plotted as accurately as possible so that users can fully evaluate the entire continuum of infection probabilities, (e.g., to look for natural breakpoints where additional disinfection probability begins to get more expensive).

To determine the points on the Pareto front 1402, controller 710 may start with a small number of infection probabilities already simulated for a given month and plot them against monthly energy cost. Then, additional candidate infection probabilities can be selected (e.g., as the points furthest from already completed simulations). After simulating instances with the new infection probabilities, these points are added to the plot, and the process repeats to the desired accuracy. Many criteria for selecting new points are possible, but one possible strategy is to choose the midpoint of successive points with the largest area (i.e., of the rectangle whose opposite corners are given by the two existing points) between them. This strategy prioritizes regions where the curve is changing rapidly and leads to efficient convergence.

As an example, consider the case in graph 1400. The goal is to obtain an approximation of the true Pareto front 1402, which is illustrated in FIG. 14 for ease of explanation, but may not be truly known. The instances of the optimization run for the small number of infection probabilities result in the points marked with squares in graph 1400 for Iteration 0. This gives a very coarse approximation of the true front. Controller 710 may then select new points in each iteration, run those simulations, and add those points to graph 1400. For example, the points marked with diamond shapes in graph 1400 show the points selected for Iteration 1 the points marked with triangles in graph 1400 show the points selected for Iteration 2, the points marked with inverted triangles in graph 1400 show the points selected for Iteration 3, and the points marked with circles in graph 1400 show the points selected for Iteration 4. By the end of Iteration 4, the empirical Pareto front is a good approximation of the true front 1402, and of course additional iterations can be performed to further improve accuracy. The empirical Pareto front generated using this technique can be used by controller 710 to solve a Pareto optimization problem to determine an optimal tradeoff between the costs and benefits of selecting different infection probability values in the infection probability constraint.

In some embodiments, determining the infection probability constraint (e.g., to provide an optimal level of disinfection control, or an optimal level of infection probability spread) and the resulting energy consumption or energy consumption costs required for HVAC system 100 to operate to achieve the infection probability constraint is a Pareto optimization problem. For example, at a certain point, additional disinfection control may require undesirably high energy consumption or energy consumption costs. In some embodiments, controller 710 may solve a Pareto optimization problem given various inputs for the system to determine one or more inflection points along a curve between cost (e.g., energy consumption or energy consumption cost) and a benefit (e.g., disinfection control, infection probability, disinfection, etc.) or to determine an optimal tradeoff between the cost and the benefit.

In some embodiments, controller 710 is configured to operate display device 822 and/or user input device 820 to provide an infection probability constraint associated with the optimal tradeoff between the cost and the benefit. In some embodiments, controller 710 can operate according to various modes that can be selected by the user via the user interface of user input device 820. For example, the user may opt for a first mode where controller 710 solves the Pareto optimization problem to determine the infection probability constraint associated with the optimal tradeoff point between the cost (e.g., the energy consumption or energy consumption cost) and the benefit (e.g., the disinfection control, a provided level of disinfection, an infection probability, etc.). In the first mode, the controller 710 can automatically determine the infection probability constraint based on the results of the Pareto optimization problem. In some embodiments, controller 710 still operates display device 822 to provide estimated, actual, or current energy consumption or energy consumption costs and infection probability constraints.

In a second mode, controller 710 can provide the user the ability to manually adjust the tradeoff between the cost and the benefit (e.g., by adjusting the slider or knob as described in greater detail above). In some embodiments, the user may select the desired tradeoff between infection control and energy consumption or energy consumption costs based on the provided estimations of energy consumption or energy consumption costs.

In a third mode, controller 710 can provide the user additional manual abilities to adjust the infection probability constraint directly. In this way, the user may specifically select various boundaries (e.g., linear boundaries if the infection probability constraint is implemented as a linear constraint as described in greater detail above) for the infection probability constraint. In some embodiments, the user may select between the various modes (e.g., the first mode, the second mode, and/or the third mode).

It should be understood that while the Pareto optimization as described herein is described with reference to only two variables (e.g., energy consumption or energy consumption cost and disinfection control), the Pareto optimization may also account for various comfort parameters or variables (e.g., temperature and/or humidity of zones 206, either individually or aggregated). In some embodiments, controller 710 may also operate display device 822 to provide various comfort parameters that result from a particular position of the knob or slider that is provided on the user interface of user input device 820. In some embodiments, additional knobs, sliders, input fields, etc., are also provided on the user interface of user input device 820 to receive various inputs or adjustments for desired comfort parameters (e.g., temperature and/or humidity). In some embodiments, controller 710 (e.g., results manager 818) is configured to use the dynamic models for temperature or humidity as described above to determine estimations of the various comfort parameters as the user adjusts the knobs or sliders (e.g., the knobs or sliders associated with disinfection control and/or energy consumption or energy cost consumption). Similarly, controller 710 can solve the Pareto optimization problem as a multi-variable optimization problem to determine an inflection point or a Pareto efficiency on a surface (e.g., a 3d graph or a multi-variable optimization) which provides an optimal tradeoff between cost (e.g., the energy consumption, the energy consumption cost, etc.), comfort (e.g., temperature and/or humidity), and disinfection control (e.g., the infection probability constraint).

Pareto Optimization Controller

Referring particularly to FIG. 15, a controller 1510 is shown, according to some embodiments. The controller 1510 can be similar to the controller 710 and can be implemented in the HVAC system 700 as described in greater detail above. In some embodiments, the controller 1510 includes the constraint generator 810, the model manager 816, the simulation database 824, the optimization manager 812, and the results manager 818, similar to the controller 710. The controller 1510 additionally includes a Pareto optimizer 1512 that is configured to use optimization results from the optimization manager 812 and perform a Pareto optimization to determine feasible and infeasible operating points, and to determine, from the feasible operating point, which is the Pareto optimal point. In some embodiments, the optimization results provided to the Pareto optimizer 1512 are or include values of the objective function. Values of the objective function may be referred to as control objectives. For example, the values of the objective function can include values of two or more variables of interest. Generally, as described herein, control objectives may include an air quality control objective and a second control objective. In some embodiments, the control objectives can include energy cost and infection risk for an associated pair of decision variables, such as minimum ventilation setpoint and supply temperature setpoint. In some embodiments, the control objectives can include energy cost or energy consumption and productivity scores for an associated pair of decision variables, such as minimum ventilation setpoint and supply temperature setpoint. In some embodiments, both the values of the decision variables and the control objectives are provided to the Pareto optimizer 1512 for use in determining what values of the decision variables should be used to achieve the Pareto optimal values of the control objectives.

In some embodiments, the control objectives can include a PM2.5 value and a CO2 value. In some implementations, air quality may be affected by, for example, wildfires or other natural phenomena that affect air quality (within a building and/or external to a building). As such PM2.5 values may be a measure of negative health consequences of the natural phenomenon (e.g., a wildfire) or other event causing reduced air quality. CO2 values may be a measure of bio effluents generated by people or building occupants. Performing a Pareto optimization using such variables (e.g., PM2.5 and/or CO2 values) as the control objectives may reduce or lower air ventilation rates in a building when smoke or other particulate matter is in outside air, as health impacts of particulate matter caused by a natural phenomenon (e.g., a wildfire) may be more significant than bio effluents and other gas phase contaminants generated within the building 10.

In some embodiments, the controller 1510 is operable between an operational mode and a monitoring mode. For example, when the controller 1510 is in the operational mode, the controller 1510 may include the control signal generator 808 instead of the results manager 818 and may automatically determine control decisions and operate the AHU 704, the UV lights 706, etc., of the HVAC system 700 based on the determined control decisions. When the controller 1510 is in the monitoring mode, the controller 1510 may include the results manager 818 (as shown in FIG. 15) and can be configured to provide the display data to the display device 822. In some embodiments, the display device 822 may operate to display different control options for the HVAC system 700, and the user may select from the different control options. The selection can be provided to the controller 1510 or the control signal generator 808 and implemented by the controller 1510 to operate the HVAC system 700 according to the selected control option (e.g., over a future time horizon). In some embodiments, the controller 1510 is also operable

In some embodiments, the controller 1510 is configured to use a combination of domain knowledge and artificial intelligence for either the operational mode or the advisory mode. For example, the controller 1510 can use domain knowledge including physics-based models for HVAC heat and mass transfer, phenomenological models that match system behavior for regulatory control, and/or different default values of the various parameters described herein. In some embodiments, the controller 1510 uses the artificial intelligence to train key model parameters (e.g., of the physics-based models described herein) in an online mode (e.g., when the controller 1510 communicates with a remote device, processing circuitry, network, gateway, etc.) using one or more regression techniques. In some embodiments, the controller 1510 uses the artificial intelligence to predict future disturbances using recent data obtained from the HVAC system 700 and also using timeseries models.

Referring particularly to FIG. 16, a diagram 1600 illustrating the functionality of the controller 1510 is shown, according to some embodiments. The diagram 1600 includes a data model 1602, a model generator 1610, a timeseries resampler 1612, a model tuner 1614, an input generator 1616, modeling data 1618, a dynamic model simulator 1632, analysis mode outputs 1634, a Pareto optimizer 1636, and advisory mode outputs 1640, according to some embodiments. In some embodiments, the data model 1602 includes one or more zone configurations 1604 (e.g., of zones 206), operational data 1606 (e.g., of the HVAC system 700, or historical operational data thereof), and weather forecast data 1608. In some embodiments, the data model 1602 is stored in the memory 806 of the controller 1510. In some embodiments, the data model 1602 is populated using data received from various sensors or control decisions of the HVAC system 700 over a previous time period (e.g., the operational data 1606). In some embodiments, the data model 1602 is populated using system configuration information such as the zone configurations 1604 (e.g., proximity of the zones 206, which of the zones 206 are served by which AHUs, etc.). In some embodiments, the data model 1602 is populated using information obtained from a third party service such as a weather service.

In some embodiments, the modeling data 1618 includes psychometric data 1620 of the HVAC system 700, HVAC equipment data 1622, parameters 1624 (e.g., infection parameters and/or productivity score parameters), and/or disturbance schedules 1626. In some embodiments, the HVAC equipment data 1622 includes different performance curves, model identifiers, model numbers, models of HVAC equipment that predict one or more operational parameters (e.g., air delivery, temperature of air delivered to a zone, etc.) as a function of one or more input variables, etc. In some embodiments, the infection parameters 1624 are values of any of the variables of the Wells-Riley equations or derivations thereof or variables of any infection risk management standard, such as ASHRAE 241, quantum concentration models, infection probability models, CO2 concentration models, infection probability constraints, etc., as described in greater detail above with reference to FIGS. 7 and 8. For example, the parameters 1624 can include any of expected, actual, or hypothetical number of infected individuals D, total number of susceptivle individuals S, number of infectious individuals I, disease quanta generation rate q, total exposure time t, quantum concentration in the air N, net indoor CO2 concentration C, total air volume of one or more zones V, net concentration of exhaled CO2 c, number of infectious particles that an individual inhales over a given time k_[0,T], the upper boundary on acceptable or desirable infection probability P_[0,T]^max, infectious quanta removal fraction of a filter λ_filter, infectious quanta removal of UV lights λ_UV, etc. In embodiments where the parameters 1624 are productivity score parameters, the parameters may include air quality metrics (CO2, TVOC, PM2.5, PM10, ozone, ventilation, etc.). In some embodiments, the disturbance schedules 1626 include expected heat disturbances, CO2 disturbances, expected occupancy schedules, etc., or any other schedules of disturbances for various infections or productivity parameters, environmental parameters, HVAC parameters, etc. In some embodiments, the data model 1602 and/or the modeling data 1618 are the domain knowledge that is used for performing the optimization described herein. In some embodiments, the data model 1602 and the modeling data 1618 are stored in the simulation database 824.

In some embodiments, the modeling data 1618 and the zone configuration data 1604 is provided to the model generator 1610 for generation of a model (e.g., any of the models or constraints described in greater detail above with reference to FIGS. 7-8). In some embodiments, the model generator 1610 is configured to perform any of the functionality of the model manager 816. In some embodiments, the weather forecast data 1608 is provided to the timeseries resampler 1612 that is configured to resample the weather forecast data 1608 and output resampled data to the model tuner 1614 and the input generator 1616. The resampled data has a frequency or time interval that is different than the frequency or time interval of the weather forecast data 1608 provided to the timeseries resampler 1612, according to some embodiments. In some embodiments, the timeseries resampler 1612 operates to provide the resampled data to the model tuner 1614 and the input generator 1616 at an appropriate frequency or time interval (e.g., between data points of the weather forecast data) so that the model tuner 1614 and the input generator 1616 can use the weather forecast data 1608, provided as the resampled data. In some embodiments, the timeseries resampler 1612 is configured to perform interpolation and/or extrapolation techniques to generate the resampled data based on the weather forecast data 1608.

The model tuner 1614 is configured to use the resampled data (e.g., the resampled weather forecast data 1608) to determine data-derived parameters and provide the data-derived parameters to the model generator 1610, according to some embodiments. In some embodiments, the model tuner 1614 is configured to generate a disturbance model using the resampled data to predict disturbances that may be introduced to HVAC system (e.g., temperature fluctuations, humidity fluctuations, etc., due to weather) and provide parameters of the disturbance model or outputs of the disturbance model to the model generator 1610 as the data-derived parameters. In some embodiments, the model tuner 1614 is configured to output the disturbance model to the input generator 1616. In some embodiments, the data-derived parameters are adjustments, calibration factors, additional correction terms, etc., for the model generator 1610 so that the model generator 1610 outputs models that accurately predict temperature, humidity, energy consumption, infection probability, productivity probability, etc., while accounting for different weather conditions, disturbances, occupancy, etc. In some embodiments, the data-derived parameters are generated by the model tuner 1614 using a neural network, a machine learning technique, artificial intelligence, etc. For example, the model tuner 1614 can obtain the operational data 1606 or the weather forecast 1608 for a historical or previous time period, as well as predictions of the various models for the historical or previous time period (e.g., predictions of zone temperatures, infection risks, infection probability, productivity score, productivity, humidity, etc.), and actual values of the predictions of the various models for the historical or previous time period (e.g., actual zone temperatures, actual infection risks, actual infection probability, actual productivity score, actual productivity, etc., or any other environmental, infection, or productivity related parameter that can be sensed or determined based on sensor data), and determine adjustments for the models using the neural network, the machine learning technique, the artificial intelligence, etc., to improve accuracy of the models.

The model generator 1610 uses the data-derived parameters (e.g., disturbance parameters, adjustment parameters, correction factors, calibration factors, additional model terms, etc.), the zone configurations 1604, and the modeling data 1618 to generate one or more models and output model parameters (shown in FIG. 16 as “generic parameters”) to the dynamic model simulator 1632, according to some embodiments. In some embodiments, the model generator 1610 uses the parameters of the disturbance model to tune or adjust the model generated based on the zone configuration data 1604 and the modeling data 1618. In some embodiments, the models (e.g., the generic parameters) are the dynamic models (e.g., the dynamic temperature model, the dynamic humidity model, the dynamic infectious quanta model, etc.) as described in greater detail above with reference to FIGS. 7-9.

The input generator 1616 uses the resampled data provided by the timeseries resampler 1612 and the disturbance model provided by the model tuner 1614 to determine model timeseries inputs, according to some embodiments. In some embodiments, the input generator 1616 is configured to generate timeseries inputs for various extrinsic parameters such as ambient or outdoor temperature, ambient or outdoor humidity, price per unit of energy as provided by a utility provider, etc. The input generator 1616 may provide the model timeseries inputs to the dynamic model simulator 1632 for use in performing a simulation (e.g., either to specific model forms 1628 or to generic model simulation 1630), according to some embodiments. In some embodiments, the model timeseries inputs are predicted or estimated timeseries data for a future time period, or are historical data from a previous time period (e.g., for the advisory mode outputs 1640 and the analysis mode outputs 1634, respectively). The input generator 1616 can provide specific model parameters to the specific model forms 1628 of the dynamic model simulator 1632 so that different generic models can be simulated for a specific HVAC system, a specific building, a specific space or zone, etc. The specific model parameters can be various thermal characteristics (e.g., heat transfer or heat storage parameters), HVAC equipment model numbers, HVAC equipment operating curves, etc.

The dynamic model simulator 1632 is configured to use the specific model parameters and the model timeseries inputs to perform a simulation for a future time period, and to perform a simulation or analysis for a previous time period, according to some embodiments. In some embodiments, the dynamic model simulator 1632 includes specific model forms 1628 and generic model simulation 1630. In some embodiments, the specific model forms 1628 are determined based on predefined or generic models (e.g., generic versions of the dynamic models as described in greater detail above) with specific model parameters that are generated or adjusted based on outputs of the model tuner 1614 and real-world or actual data as provided by the data model 1602 (e.g., the zone configuration data 1604, the operational data 1606, the weather forecast data 1608, etc.).

In some embodiments, the generic model simulation 1630 is performed using the specific model forms 1628 for both a future time period (e.g., for the advisory mode outputs 1640) and for a previous time period (e.g., for the analysis mode outputs 1634). In some embodiments, the generic model simulation 1630 is performed to determine energy consumption or energy cost and associated infection risks or disinfection (e.g., a reduction in infection risks resulting from performing any of the fresh air intake operations of an AHU, UV light disinfection, or filtration). In some embodiments, the generic model simulation 1630 uses the dynamic infectious quanta model to simulate or assess infection risks for previously performed HVAC operations, or for predicted future HVAC operations. In some embodiments, the generic model simulation 1630 is performed to determine energy consumption or energy cost and associated productivity risks or productivity scores (e.g., an increase in productivity resulting from performing any of the fresh air intake operations of an AHU, UV light disinfection, filtration, or decrease in infection risk). In some embodiments, the generic model simulation 1630 uses the dynamic infectious quanta model to simulate or assess a productivity score for previously performed HVAC operations, or for predicted future HVAC operations.

The outputs of the generic model simulation 1630 (e.g., control objectives such as including but not limited to infection risk, productivity score, and energy cost or energy consumption) are provided to the Pareto optimizer 1636 and the analysis mode outputs 1634, according to some embodiments. In some embodiments, the outputs of the dynamic model simulator 1632 that use historical BMS data (e.g., infection risk, productivity score, and energy cost or energy consumption values associated with previous operation of the HVAC system over the previous time period) are provided as the analysis mode outputs 1634. In some embodiments, the output of the dynamic model simulator 1632 that are for the future time period (e.g., for different possible control decisions of the HVAC system or decision variables over the future time period) are provided to the Pareto optimizer 1636 for determination of the advisory mode outputs 1640 (e.g., using the Pareto optimization techniques described in greater detail below with reference to FIGS. 25-28).

Pareto Optimization Techniques for Productivity Scores

Referring particularly to FIGS. 17-20, the Pareto optimizer 1636 or the Pareto optimizer 1512 are configured to perform various Pareto optimization techniques as described herein to determine Pareto optimization results, according to some embodiments. It should be understood that while the techniques described herein with reference to FIGS. 17-20 are described as being performed by the Pareto optimizer 1636, the techniques can also be performed by the Pareto optimizer 1512 or processing circuitry 802 thereof. Additionally, it should be noted that the graphs of the energy cost and productivity scores shown in FIGS. 17-20 (i.e., graphs 1704, 1710, 1714, and 2004) are arranged such that energy cost is increasing along the vertical axis (i.e., higher values of energy cost are near the top) whereas the productivity scores are decreasing along the horizontal axis (i.e., higher values of productivity scores are near the left). Accordingly, the most desirable operating points in graphs 1704, 1710, 1714, and 2004 are the points near the bottom left as these points correspond to low energy cost and high productivity scores.

Referring particularly to FIG. 17, a diagram 1700 shows a graph 1702 of different decision variables, and a graph 1704 of corresponding control objectives for each of the different decision variables. The graph 1702 shows different combinations for a supply temperature setpoint and a minimum ventilation setpoint (shown on the Y and X axes, respectively) for the HVAC system 100, according to some embodiments. It should be understood that only two decision variables are shown for ease of explanation, and that any number of decision variables may be used. Different values and combinations of both the decision variables are represented in FIG. 17 as points 1706. For example, points 1706a-1706d have a same value for the minimum ventilation setpoint decision variables but different values of the supply temperature setpoint decision variable. Similarly, points 1706e-1706h have the same value for the minimum ventilation setpoint decision (different than the value of the minimum ventilation setpoint decision variable for points 1706a-1706d) but different values of the supply temperature setpoint decision variable. Points 1706i-17061 likewise have the same value of the minimum ventilation setpoint decision variable (different than the values of the minimum ventilation setpoint decision variable for points 1706a-1706d and 1706e-1706h) but different values of the supply temperature setpoint decision variable. In some embodiments, the values of the decision variables are a fixed set (e.g., generated as a grid using minimum and maximum allowed values for each of the multiple decision variables) or are generated iteratively based on simulation results (e.g., by adding additional points that are likely to be Pareto optimal based on simulation results of proximate points). In general, the control objectives considered by Pareto optimizer 1636 may include any control objectives that at least partially conflict or trade-off with each other such that optimizing one of the control objectives comes at the expense of the other control objective (i.e., control decisions that improve one control objective worsen the other control objective). In some embodiments, the control objectives considered by Pareto optimizer 1636 include two air quality control objectives. For example, in some implementations, the control objectives include minimizing a PM2.5 value and minimizing a CO2 value of air within the building. These control objectives may conflict with each other in scenarios where the control actions which decrease the PM2.5 value tend to increase the CO2 value and vice versa. This may occur, for example, when increasing the ventilation rate of air from outside the building reduces the level of CO2 of air within the building due to a lower concentration of CO2 in the outside air, but increases the PM2.5 value of the air within the building because the outside air has a higher level of pollutants than the indoor air (e.g., when the outdoor air is polluted with smoke, dust, or other contaminants). It should be understood that a Pareto optimization or other optimization technique performed using PM2.5 values and CO2 values as the control objectives may be performed using the same methods and processes as described with respect to different decision variables (e.g., minimum ventilation setpoint variables and supply temperature setpoint decision variables).

For example, PM2.5 may be a measure of negative health consequences of a natural phenomenon causing reduced air quality, and CO2 may be a measure of bio effluents generates by occupants of a building. A Pareto optimization using these metrics as the competing control objectives may cause building equipment to change operation, such as by, for example, reducing air ventilation rates when smoke or other particulates are in outside air. This may be due to the fact that health impacts of the outside particles (e.g., smoke particles caused by a wildfire) may be more significant (e.g., pose a greater risk) than bio effluents and/or other gas phase contaminants generated within the building. Conversely, other optimal operating points along the Pareto front may prioritize reducing CO2 levels within the building over reducing PM2.5 levels and can be achieved by selecting values of the control decision variables that result in higher values of the outdoor air ventilation rate.

In some embodiments, a simulation is performed to determine corresponding energy cost or energy consumption and productivity score for each of the different points 1706. The corresponding energy cost energy consumption and productivity score are shown as points 1708 in graph 1704. In some embodiments, points 1708a-1708l of graph 1704 correspond to points 1706a-17061 of graph 1702. For example, point 1708a illustrates the corresponding energy cost or energy consumption and productivity score for the values of the minimum ventilation setpoint and the supply temperature of the point 1706a. Likewise, points 1708b-1708l illustrate the various corresponding energy costs or energy consumptions and productivity score for each of the minimum ventilation setpoint and supply temperature setpoint values as represented by points 1706b-17061. In some embodiments, each of the points 1706a-17061 and the corresponding points 1708a-1708l correspond to a simulation performed by the optimization manager 812, or the dynamic model simulator 1632. For example, the dynamic model simulator 1632 may perform a simulation for each of the sets of values of the supply temperature setpoint and the minimum ventilation setpoint (e.g., the decision variables) and output values of the energy cost or energy consumption and productivity score for each simulation (shown as points 1708). It should be understood that while FIG. 17 shows only two objectives of the Pareto optimization (e.g., energy cost or energy consumption and productivity), the Pareto optimization may have any number of optimization objectives (e.g., more than two, etc.).

Referring particularly to FIG. 18, a diagram 1800 shows the graph 1704 and a graph 1710, according to some embodiments. The graph 1704 shows the points 1708 that illustrate the various combinations of energy cost or energy consumption and productivity score for the corresponding values of the decision variables (the supply temperature setpoint and the minimum ventilation setpoint shown in graph 1702 in FIG. 17). In some embodiments, the graph 1710 illustrates groups 1712 that are include the points 1708 grouped according to feasibility, and further group according to Pareto optimality. Specifically, groups 1712 includes a first group of points 1712a (e.g., points 1708i and 1708e), a second group of points 1712b (e.g., points 1708f, 1708a, 1708b, 1708c, and 1708d), and a third group of points 1712c (e.g., points 1708j, 1708k, 1708l, 1708g, and 1708h).

The first group of points 1712a are points that are infeasible, unfeasible, or non-feasible. In some embodiments, the Pareto optimizer 1636 is configured to determine or identify which of the points 1708 are infeasible and group such combinations of energy cost or energy consumption and productivity score as infeasible solutions. In some embodiments, the Pareto optimizer 1636 is configured to use threshold energy costs, energy consumptions, or productivity scores, and if some of the points 1708 are greater than a maximum allowable energy cost, energy consumption, or productivity score, or less than a minimum allowable productivity score, energy consumption, or energy cost, the Pareto optimizer 1636 can determine that such points are infeasible and group them accordingly as the first group of points 1712a. In some embodiments, the maximum or minimum allowable energy cost, energy consumption, or productivity score values used by the Pareto optimizer 1636 are user inputs, values set by legal regulations, or values determined based on abilities of the HVAC system 700. In some embodiments, the second group of points 1712b and the third group of points 1712c are feasible solutions.

In some embodiments, the Pareto optimizer 1636 is configured to perform a Pareto optimization based on the points 1708 to determine which of the points 1708 are Pareto optimal. A Pareto optimal point is a point where neither the energy cost or energy consumption, or the productivity score can be reduced without causing a corresponding increase in the productivity score or the energy cost or energy consumption. In the example shown in FIG. 18, the points 1708j, 1708k, 1708l, 1708g, and 1708h are Pareto optimal points, and the Pareto optimizer 1636 is configured to classify these points as such, thereby defining the third group of points 1712c, according to some embodiments. In some embodiments, the Pareto optimizer 1636 is configured to determine that points which are feasible but are not Pareto optimal (e.g., points 1708f, 1708a, 1708b, 1708c, and 1708d) should define the second group of points 1712b. In this way, the Pareto optimizer 1636 can define several groups of the points 1708 (e.g., groups of various solutions to be considered): (i) infeasible points, (ii) feasible points that are Pareto optimal, and (iii) feasible points that are not Pareto optimal.

Referring particularly to FIG. 19, the Pareto optimizer 1636 can further identify which of the Pareto optimal points, shown as the third group of points 1712c in graph 1710, result in minimum productivity score, minimum energy consumption, and an equal priority between productivity and energy consumption, according to some embodiments. In some embodiments, the Pareto optimizer 1636 is configured to determine which of the Pareto optimal points (i.e., the points 1708 of the third group of points 1712c) have a minimum productivity score, a minimum energy consumption, and an equal priority between energy consumption and productivity score. Specifically, in the example shown in FIGS. 17-19, the point 1708j is a Pareto optimal point that has a highest productivity score, and therefore the Pareto optimizer 1636 identifies point 1708j as a maximum productivity score point 1718. Similarly, in the example shown in FIGS. 17-19, the Pareto optimizer 1636 determines that the point 1708h is a Pareto optimal point associated with minimum energy consumption, and therefore the Pareto optimizer 1636 identifies the point 1708h as a minimum energy consumption point 1722. Finally, in the example shown in FIGS. 17-19, the Pareto optimizer 1636 determines that the point 1708l is a Pareto optimal point that results in an equal priority between the productivity score and the energy consumption, shown as equal priority point 1720.

In some embodiments, the maximum productivity score point 1718 (e.g., point 1708j), the minimum energy consumption (or energy costs) point 1722 (e.g., point 1708h), and the equal priority point 1720 (e.g., point 1708l) are the Pareto optimization results. In some embodiments, the Pareto optimization results also include the energy cost or energy consumption, productivity score, as well as the minimum ventilation setpoint, and the supply temperature setpoint for each of the maximum productivity score point 1718, the minimum energy consumption point 1722, and the equal priority point 1720. In some embodiments, the Pareto optimization results are provided to the user via the display device for selection. For example, the display device can provide different recommended operating possibilities such as a maximum productivity score operating possibility, a minimum energy consumption operating possibility, and an equal energy and productivity score. In some embodiments, the user may select one of the different operating recommendations, and provide the selection to the controller 1510 for use in operating the HVAC system 400 according to the selected operating possibility.

Referring particularly to FIG. 20, a diagram 2000 shows another example of a graph 2002 and a graph 2004 illustrating the functionality of the Pareto optimizer 1636, according to some embodiments. Graph 2002 includes points 2022 that illustrate different combinations of minimum ventilation and supply temperature setpoint (e.g., the decision variables) usable by the dynamic model simulator 1632 or the optimization manager 812 for performing a simulation to determine corresponding productivity score and energy costs or energy consumption (shown as points 2012 in graph 2004), according to some embodiments. In some embodiments, the points 2012 illustrated in the graph 2004 are mapped to the points 2022. For example, the simulation can be performed for each of the points 2022 to determine the corresponding productivity score and energy cost or energy consumption, as represented by points 2012 in graph 2004.

In the example shown in FIG. 20, the Pareto optimizer 1636 identifies a group 2020 of infeasible points, according to some embodiments. The infeasible points may be points that cannot be achieved due to constraints and operational ability of the HVAC system 400. In some embodiments, the Pareto optimizer 1636 further identifies which of the points 2012 are Pareto optimal points, and determines which of the Pareto optimal points are associated with highest productivity score, lowest energy costs or energy consumption, and a balanced priority point where energy costs or energy consumption and productivity score are equally prioritized. In the example shown in FIG. 20, the Pareto optimizer 1636 identifies that a point 2018 of graph 2004 is the Pareto optimal point that results in highest productivity score, which corresponds to point 2010 of graph 2002. Similarly, the Pareto optimizer 1636 may determine that the point 2014 is the Pareto optimal point that results in lowest energy cost or energy consumption, which corresponds to the point 2006 in graph 2002, according to some embodiments. Finally, the Pareto optimizer 1636 can determine that a point 2016 of graph 2004 is the Pareto optimal point that results in an equal priority between the productivity score and the energy cost or energy consumption, which corresponds to point 2008 of graph 2002, according to some embodiments.

In some embodiments, the Pareto optimizer 1636 is configured to provide all of the Pareto optimal points to the user via the display device (e.g., user device 412). In some embodiments, the Pareto optimizer 1636 or the Pareto optimizer 1512 is configured to provide the Pareto optimal points as the Pareto optimal results to the results manager 818 for use in display to the user. In some embodiments, the Pareto optimizer 1512 or the Pareto optimizer 1636 automatically selects one of the Pareto optimal points for use and provides the Pareto optimal points and its associated control decisions (e.g., the minimum ventilation and supply temperature setpoints) to the control signal generator 808.

Pareto Optimization Process

Referring now to FIG. 21, a process 2100 for performing Pareto optimization to determine operation of the BMS system is shown, according to some embodiments. Process 2100 includes steps 2102-2116 and can be performed by the building analysis system 404 and/or the BMS system, according to some embodiments. In some embodiments, process 2100 is performed to determine various Pareto optimal values, or to determine a Pareto optimal solution and associated operating parameters that results in optimal tradeoff between productivity scores and energy cost or energy consumption. The process 2100 may include processes described in greater detail with respect to FIGS. 15-20.

Process 2100 includes obtaining multiple sets of values of control decision variables, each set of values including a different combination of the control decision variables (step 2102), according to some embodiments. In some embodiments, the control decision variables include control operations describe herein. In some embodiments, the control decision variables include operating parameters or control decisions of the ventilation rates or the filtration rate. It should be understood that the control decision variables described herein are not limited to only two variables, and may include any number of variables. In some embodiments, each set of the values of the control decision variables is a unique combination of different values of the control decision variables. Similarly, regardless of a number of control decision variables, each set of the values of the control decision variables may be a unique combination, according to some embodiments.

Process 2100 includes performing a simulation for each set of the values of the control decision variables to determine sets of values of energy cost or energy consumption and productivity scores (step 2104), according to some embodiments. In some embodiments, the simulations are performed by the building analysis system 404 and/or the BMS system, or processing circuitry thereof. In some embodiments, the simulations are performed for a future time horizon to generate predicted or simulated values for the energy cost or energy consumption and productivity score. In some embodiments, the simulations are performed for a previous or historical time period to determined values of the energy cost or energy consumption and productivity score for analysis (e.g., for comparison with actual historical data of the energy cost or energy consumption and productivity score). In some embodiments, each of the sets of values of control decisions (e.g., as obtained in step 2102) is used for a separate simulation to determine a corresponding set of performance variables (e.g., the values of energy cost or energy consumption and productivity score). In some embodiments, the simulations are performed subject to one or more constraints.

Process 2100 includes determining which of the sets of values of energy cost or energy consumption and productivity scores are infeasible and which are feasible (step 2106), according to some embodiments. In some embodiments, step 2106 is performed by the building analysis system 404 and/or the BMS system. In some embodiments, step 2106 is performed using one or more constraints. The one or more constraints can be minimum or maximum allowable values of either of the energy cost or energy consumption and productivity scores, according to some embodiments. For example, if one of the sets of values of energy cost or energy consumption and productivity scores has an energy cost or energy consumption that exceeds a maximum allowable value of energy cost or energy consumption (e.g., exceeds a maximum threshold), then such a set of values of energy cost or energy consumption and productivity scores, and consequently the corresponding sets of values of the control decision variables, may be considered infeasible, according to some embodiments. In some embodiments, the constraints are set based on capabilities of an HVAC system that the process 2100 is performed to optimize user inputs, budgetary constraints, etc.

Process 2100 includes determining which of the feasible sets of values of energy cost or energy consumption and productivity scores are Pareto optimal solutions (step 2108), according to some embodiments. In some embodiments, the step 2108 is performed by the building analysis system 404 and/or the BMS system. In some embodiments, the step 2108 is performed to determine which of the sets of values of energy cost or energy consumption and productivity scores are Pareto optimal from the feasible sets of values of energy cost or energy consumption and productivity scores. In some embodiments, process 2100 includes performing steps 2102-2108 iteratively to determine sets of decision variables. For example, the decision variables can be iteratively generated based on simulation results (e.g., by generating additional points that are likely to be Pareto optimal based on the results of step 2108).

Process 2100 includes determining, based on the Pareto optimal solutions, a minimum energy cost or energy consumption solution, a maximum productivity score solution, and an equal priority energy cost or energy consumption/productivity score solution, according to some embodiments. In some embodiments, step 2110 is performed by the building analysis system 404 and/or the BMS system. In some embodiments, the minimum energy cost or energy consumption solution is the set of values of the energy cost or energy consumption and productivity scores that are Pareto optimal, feasible, and also have a lowest value of the energy cost or energy consumption. In some embodiments, the maximum productivity scores solution is selected from the set of values of the energy cost or energy consumption and productivity scores that are feasible and Pareto optimal, and that has a highest value of the productivity scores. In some embodiments, the equal priority energy cost or energy consumption/productivity scores solution is selected from the set of values of the energy cost or energy consumption and productivity scores that are feasible and Pareto optimal, and that equally prioritizes energy cost or energy consumption and productivity scores. For example, the energy cost or energy consumption/productivity scores solution can be a point that is proximate an inflection of a curve that is fit to the sets of values of energy cost or energy consumption and productivity scores (e.g., including the feasible and infeasible points, only the feasible points, only the Pareto optimal points, etc.).

Process 2100 includes providing one or more of the Pareto optimal solutions to a user via a display screen (step 2112), according to some embodiments. In some embodiments, step 2112 includes operating the display device 822 to display the Pareto optimal solutions to the user as different operational modes or available operating profiles. In some embodiments, step 2112 is performed by the display device 822 and the building analysis system 404 and/or the BMS system. In some embodiments, step 2112 includes providing the Pareto optimal solutions and historical data (e.g., historical data of actually used control decisions and the resulting energy cost or energy consumption and productivity scores). In some embodiments, step 2112 is optional. For example, if the user has already set a mode of operation (e.g., always use maximum productivity scores settings, always use minimum energy cost or energy consumption solution, always use equal priority energy cost or energy consumption/productivity scores solution, etc.), then step 2112 can be optional.

Process 2100 includes automatically selecting one of the Pareto optimal solutions or receiving a user input of a selected Pareto optimal solution (step 2114), according to some embodiments. In some embodiments, a user may select a setting for the building analysis system 404 and/or the BMS system to either automatically select one of the Pareto optimal solutions, or that the Pareto optimal solutions should be provided to the user for selection. In some embodiments, step 2114 is performed by the building analysis system 404 and/or the BMS system and the display device 822. For example, step 2114 can be performed by a user providing a selection of one of the Pareto optimal solutions (and therefore the corresponding control decisions) to be used by the BMS system for operation, according to some embodiments. In some embodiments, step 2114 is performed automatically (e.g., if a user or administrator has selected a predetermined mode of operation for the BMS system) by the building analysis system and/or the BMS system to select one of the Pareto optimal solutions and therefore the corresponding control decisions for operational use of the BMS system.

Process 2100 includes operating equipment of the BMS system according to the control decisions of the selected Pareto optimal solution (step 2116), according to some embodiments. In some embodiments, step 2116 includes operating the BAS system 300. More specifically, step 2116 can include operating the HVAC system 700 according to the control decisions of the selected Pareto optimal solution. Advantageously, using the control decisions of the selected Pareto optimal solution can facilitate optimal control of the HVAC system 700 in terms of risk reduction, energy consumption, or an equal priority between productivity scores maximization and energy consumption or energy cost reduction. It should be understood that while FIG. 21 only uses two objectives of the Pareto optimization (e.g., energy cost or energy consumption and productivity scores), the Pareto optimization may have any number of optimization objectives (e.g., more than two, etc.).

User Interfaces

Referring particularly to FIGS. 22-24, user interfaces 2200, 2300, and 2400 can display various outputs of the building analysis system 404 (e.g., the display data, the productivity scores, the reports, etc.), according to some embodiments. In some embodiments, the user interfaces 2200, 2300, and 2400 are displayed on the user device 412 and presented to a user or a building administrator.

Referring particularly to FIG. 22, the user interface 2200 includes a productivity score icon 2202, and an indoor air quality score icon 2204. In some embodiments, the productivity score 2202 is a scaled version of the productivity score for the previous time period. In some embodiments, the productivity score is a weighted average, a time-series average, etc., of the productivity scores of one/or more zones of the building 10 over the previous time period. In some embodiments, the indoor air quality score icon 2204 displays a similarly aggregated, average, etc., score of the indoor air quality of the zones of the building 10 over the previous time period. In some embodiments, the values of the productivity score and the indoor air quality score are normalized values from ranging from 0 or 1 to 100. In some embodiments, the indoor air quality score icon 2204 and the productivity score icon 2202 are graphical icons that display a bar or a circle chart and a textual or numeric value of the indoor air quality score and the productivity score for the spaces of the building 10 over the previous time period. In some embodiments, the indoor air quality score icon 2204 and the productivity score icon 2202 are color-coded based on their values. For example, if the indoor air quality score is between a first or normal range, then the color of the indoor air quality score icon 2204 may be yellow, according to some embodiments. In some embodiments, if the indoor air quality score is between a second range and less than a lower value of the first or normal range, this may indicate that the indoor air quality score is poor and the color of the indoor air quality score icon 2204 may be red. In some embodiments, if the indoor air quality score is between a third range or greater than a higher value of the first or normal range, this may indicate that the indoor air quality score is good and the color of the indoor air quality score 2204 may be green.

Referring still to FIG. 22, the user interface 2200 includes a list 2206 of one or more productivity score alerts, according to some embodiments. In some embodiments, the list 2206 includes different items 2208, each item corresponding to a different zone of the building 10 and a productivity score associated with the different space. In some embodiments, the items 2208 of the list 2206 are zone-specific and are determined based on the productivity score for each of the zones of the building 10. For example, if one of the zones has an associated productivity score that is below a threshold amount, then that zone may be added with the associated productivity score to the list 2206 as one of the items 2208.

Referring still to FIG. 22, the user interface 2200 also includes a list 2210 of one or more low indoor air quality alerts, according to some embodiments. In some embodiments, the list 2210 includes different items 2212, each item corresponding to a different zone of the building 10 and an individual indoor air quality associated with the different zones. In some embodiments, the items 2212 of the list 2210 are zone-specific and are determined based on the indoor air quality for each of the zones of the building 10. For example, if one of the zones has an associated indoor air quality that is below a threshold amount, then that zone may be added with the associated indoor air quality to the list 2210 as one of the items 2212.

Referring still to FIG. 22, the user interface 2200 includes a list 2214 of each of the zones of the building 10 (e.g., organized by zone type, floor of the building 10, etc.). Each of the items of the list 2214 includes an indication of the zone or floor, an associated productivity score for the zone or floor, a number of productivity score alerts for the zone or floor, an indoor air quality score for the zone or floor, an indoor air quality trend (e.g., a 30 day trend), a number of indoor air quality alerts, and/or an energy spend versus budget (e.g., for 30 days).

Referring to FIG. 23, the user interface 2300 includes different widgets 2302-2308 indicating the results of the Pareto optimization. Specifically, the user interface 2300 includes a current operational state widget 2302 illustrating current energy costs or energy consumption and associated productivity score with additional air flow, comfort, UV disinfection, and filtration specifics, according to some embodiments. The user interface 2300 includes a widget 2304 illustrating a first option, namely, the Pareto optimal result for optimizing the productivity score (e.g., maximizing productivity score), a widget 2306 illustrating a second option, namely, the Pareto optimal result for equal priority between productivity score and energy consumption, and a widget 2308 illustrating a third option, namely, the Pareto optimal result for operating with minimum energy cost or energy consumption. Each of the widgets 2304-2308 include graphical and/or textual information regarding a corresponding productivity score, an energy cost or energy consumption per a time period (e.g., a monthly time period), air flow parameters, required operational adjustments, optional design adjustments, etc., for each of the options. In some embodiments, the user or building administrator may select one of the options by selecting one of the widgets 2304-2308.

Referring particularly to FIG. 24, the user interface 2400 illustrates different operational adjustments for the HVAC system 700 that the building administrator should implement in order to configure the HVAC system 700 to perform the selected option, according to some embodiments. The user interface 2400 includes widgets 2402a-2402d, each of which illustrate a next step that should be performed to implement the selected option, according to some embodiments. In some embodiments, each widget 2402 includes a button 2404 which, when selected, navigates the user to a command and control panel where the user or building administrator can perform the specific operational adjustment (e.g., adjusting the supply temperature setpoint). Each widget 2402 also includes a button 2406 which, when selected, marks the task associated with the widget 2402 as completed, according to some embodiments.

Pareto Optimization Techniques for Infection

Referring particularly to FIGS. 25-28, the Pareto optimizer 1636 or the Pareto optimizer 1512 are configured to perform various Pareto optimization techniques as described herein to determine Pareto optimization results, according to some embodiments. It should be understood that while the techniques described herein with reference to FIGS. 25-28 are described as being performed by the Pareto optimizer 1636, the techniques can also be performed by the Pareto optimizer 1512 or processing circuitry 802 thereof.

Referring particularly to FIG. 25, a diagram 2500 shows a graph 2502 of different decision variables, and a graph 2504 of corresponding control objectives for each of the different decision variables. The graph 2502 shows different combinations for a supply temperature setpoint and a minimum ventilation setpoint (shown on the Y and X axes, respectively) for the HVAC system 700, according to some embodiments. It should be understood that only two decision variables are shown for ease of explanation, and that any number of decision variables may be used. Different values and combinations of both the decision variables are represented in FIG. 25 as points 2506. For example, points 2506a-2506d have a same value for the minimum ventilation setpoint decision variables but different values of the supply temperature setpoint decision variable. Similarly, points 2506e-2506h have the same value for the minimum ventilation setpoint decision (different than the value of the minimum ventilation setpoint decision variable for points 2506a-2506d) but different values of the supply temperature setpoint decision variable. Points 2506i-25061 likewise have the same value of the minimum ventilation setpoint decision variable (different than the values of the minimum ventilation setpoint decision variable for points 2506a-2506d and 2506e-2506h) but different values of the supply temperature setpoint decision variable. In some embodiments, the values of the decision variables are a fixed set (e.g., generated as a grid using minimum and maximum allowed values for each of the multiple decision variables) or are generated iteratively based on simulation results (e.g., by adding additional points that are likely to be Pareto optimal based on simulation results of proximate points).

In some embodiments, a simulation is performed to determine corresponding energy cost and infection risk for each of the different points 2506. The corresponding energy cost and infection risk are shown as points 2508 in graph 2504. In some embodiments, points 2508a-2508l of graph 2504 correspond to points 2506a-25061 of graph 2502. For example, point 2508a illustrates the corresponding energy cost and infection risk for the values of the minimum ventilation setpoint and the supply temperature of the point 2506a. Likewise, points 2508b-2508l illustrate the various corresponding energy costs and infection risks for each of the minimum ventilation setpoint and supply temperature setpoint values as represented by points 2506b-25061. In some embodiments, each of the points 2506a-25061 and the corresponding points 2508a-2508l correspond to a simulation performed by the optimization manager 812, or the dynamic model simulator 1632. For example, the dynamic model simulator 1632 may perform a simulation for each of the sets of values of the supply temperature setpoint and the minimum ventilation setpoint (e.g., the decision variables) and output values of the energy cost and infection risk for each simulation (shown as points 2508). It should be understood that while FIG. 25 shows only two objectives of the Pareto optimization (e.g., energy cost and infection risk), the Pareto optimization may have any number of optimization objectives (e.g., more than two, etc.).

Further, as described herein, in some implementations, the control objectives may be a PM2.5 value and a CO2 value as described above with reference to FIG. 17. It should be understood that a Pareto optimization or other optimization technique performed using PM2.5 values and CO2 values as the control objectives may be performed using the same methods and processes as described previously and may use the same or different decision variables (e.g., minimum ventilation setpoint variables and supply temperature setpoint decision variables) to affect the values of the control objectives.

Referring particularly to FIG. 26, a diagram 2600 shows the graph 2504 and a graph 2510, according to some embodiments. The graph 2504 shows the points 2508 that illustrate the various combinations of energy cost and infection risk for the corresponding values of the decision variables (the supply temperature setpoint and the minimum ventilation setpoint shown in graph 2502 in FIG. 25). In some embodiments, the graph 2510 illustrates groups 2512 that are include the points 2508 grouped according to feasibility, and further group according to Pareto optimality. Specifically, groups 2512 includes a first group of points 2512a (e.g., points 2508i and 2508e), a second group of points 2512b (e.g., points 2508f, 2508a, 2508b, 2508c, and 2508d), and a third group of points 2512c (e.g., points 2508j, 2508k, 2508l, 2508g, and 2508h).

The first group of points 2512a are points that are infeasible, unfeasible, or non-feasible. In some embodiments, the Pareto optimizer 1636 is configured to determine or identify which of the points 2508 are infeasible and group such combinations of energy cost and infection risk as infeasible solutions. In some embodiments, the Pareto optimizer 1636 is configured to use threshold energy costs or infection risks, and if some of the points 2508 are greater than a maximum allowable energy cost or infection risk, or less than a minimum allowable infection risk or energy cost, the Pareto optimizer 1636 can determine that such points are infeasible and group them accordingly as the first group of points 2512a. In some embodiments, the maximum or minimum allowable energy cost or infection risk values used by the Pareto optimizer 1636 are user inputs, values set by legal regulations, or values determined based on abilities of the HVAC system 700. In some embodiments, the second group of points 2512b and the third group of points 2512c are feasible solutions.

In some embodiments, the Pareto optimizer 1636 is configured to perform a Pareto optimization based on the points 2508 to determine which of the points 2508 are Pareto optimal. A Pareto optimal point is a point where neither the energy cost or the infection risk can be reduced without causing a corresponding increase in the infection risk or the energy cost. In the example shown in FIG. 26, the points 2508j, 2508k, 2508l, 2508g, and 2508h are Pareto optimal points, and the Pareto optimizer 1636 is configured to classify these points as such, thereby defining the third group of points 2512c, according to some embodiments. In some embodiments, the Pareto optimizer 1636 is configured to determine that points which are feasible but are not Pareto optimal (e.g., points 2508f, 2508a, 2508b, 2508c, and 2508d) should define the second group of points 2512b. In this way, the Pareto optimizer 1636 can define several groups of the points 2508 (e.g., groups of various solutions to be considered): (i) infeasible points, (ii) feasible points that are Pareto optimal, and (iii) feasible points that are not Pareto optimal.

Referring particularly to FIG. 27, the Pareto optimizer 1636 can further identify which of the Pareto optimal points, shown as the third group of points 2512c in graph 2510, result in maximum disinfection (e.g., minimal infection risk), minimum energy consumption, and an equal priority between disinfection and energy consumption, according to some embodiments. In some embodiments, the Pareto optimizer 1636 is configured to determine which of the Pareto optimal points (i.e., the points 2508 of the third group of points 2512c) have a minimum infection risk, a minimum energy consumption, and an equal priority between energy consumption and infection risk. Specifically, in the example shown in FIGS. 25-27, the point 2508j is a Pareto optimal point that has a lowest infection risk, and therefore the Pareto optimizer 1636 identifies point 2508j as a maximum disinfection point 2518. Similarly, in the example shown in FIGS. 25-27, the Pareto optimizer 1636 determines that the point 2508h is a Pareto optimal point associated with minimum energy consumption, and therefore the Pareto optimizer 1636 identifies the point 2508h as a minimum energy consumption point 2522. Finally, in the example shown in FIGS. 25-27, the Pareto optimizer 1636 determines that the point 2508l is a Pareto optimal point that results in an equal priority between the infection risk and the energy consumption, shown as equal priority point 2520.

In some embodiments, the maximum disinfection point 2518 (e.g., point 2508j), the minimum energy consumption (or energy costs) point 2522 (e.g., point 2508h), and the equal priority point 2520 (e.g., point 2508l) are the Pareto optimization results. In some embodiments, the Pareto optimization results also include the energy cost, infection risk, as well as the minimum ventilation setpoint, and the supply temperature setpoint for each of the maximum disinfection point 2518, the minimum energy consumption point 2522, and the equal priority point 2520. In some embodiments, the Pareto optimization results are provided to the user via the display device 822 for selection. For example, the display device 822 can provide different recommended operating possibilities such as a maximum disinfection operating possibility, a minimum energy consumption operating possibility, and an equal energy and infection risk. In some embodiments, the user may select one of the different operating recommendations, and provide the selection to the controller 1510 for use in operating the HVAC system 700 according to the selected operating possibility.

Referring particularly to FIG. 28, a diagram 2800 shows another example of a graph 2802 and a graph 2804 illustrating the functionality of the Pareto optimizer 1636, according to some embodiments. Graph 2802 includes points 2822 that illustrate different combinations of minimum ventilation and supply temperature setpoint (e.g., the decision variables) usable by the dynamic model simulator 1632 or the optimization manager 812 for performing a simulation to determine corresponding infection risk and energy costs (shown as points 2812 in graph 2804), according to some embodiments. In some embodiments, the points 2812 illustrated in the graph 2804 are mapped to the points 2822. For example, the simulation can be performed for each of the points 2822 to determine the corresponding infection risk and energy cost, as represented by points 2812 in graph 2804.

In the example shown in FIG. 28, the Pareto optimizer 1636 identifies a group 2820 of infeasible points, according to some embodiments. The infeasible points may be points that cannot be achieved due to constraints and operational ability of the HVAC system 700. In some embodiments, the Pareto optimizer 1636 further identifies which of the points 2812 are Pareto optimal points, and determines which of the Pareto optimal points are associated with lowest infection risk (e.g., maximum disinfection), lowest energy costs or energy consumption, and a balanced priority point where energy costs and infection risk are equally prioritized. In the example shown in FIG. 28, the Pareto optimizer 1636 identifies that a point 2818 of graph 2804 is the Pareto optimal point that results in lowest infection risk, which corresponds to point 2810 of graph 2802. Similarly, the Pareto optimizer 1636 may determine that the point 2814 is the Pareto optimal point that results in lowest energy cost, which corresponds to the point 2806 in graph 2802, according to some embodiments. Finally, the Pareto optimizer 1636 can determine that a point 2816 of graph 2804 is a Pareto optimal point that results in a solution with equal priority between the infection risk and the energy cost, which corresponds to point 2808 of graph 2802, according to some embodiments.

In some embodiments, the Pareto optimizer 1636 is configured to provide all of the Pareto optimal points to the user via the display device 822. In some embodiments, the Pareto optimizer 1636 or the Pareto optimizer 1512 is configured to provide the Pareto optimal points as the Pareto optimal results to the results manager 818 for use in display to the user. In some embodiments, the Pareto optimizer 1512 or the Pareto optimizer 1636 automatically selects one of the Pareto optimal points for use and provides the Pareto optimal points and its associated control decisions (e.g., the minimum ventilation and supply temperature setpoints) to the control signal generator 808.

Pareto Optimization Process

Referring particularly to FIG. 29, a process 2900 for performing a Pareto optimization to determine operation of a building HVAC system is shown, according to some embodiments. Process 2900 includes steps 2902-2916 and can be performed by the controller 1510 or more specifically, by the Pareto optimizer 1512, according to some embodiments. In some embodiments, process 2900 is performed to determine various Pareto optimal values, or to determine a Pareto optimal solution and associated operating parameters that results in optimal tradeoff between infection risk and energy cost.

Process 2900 includes obtaining multiple sets of values of control decision variables, each set of values including a different combination of the control decision variables (step 2902), according to some embodiments. In some embodiments, the control decision variables include a minimum ventilation setpoint and a supply temperature setpoint. In some embodiments, the control decision variables include operating parameters or control decisions of the UV lights 706 or the AHU 704 (e.g., a fresh air intake fraction x). It should be understood that the control decision variables described herein are not limited to only two variables, and may include any number of variables. In some embodiments, each set of the values of the control decision variables is a unique combination of different values of the control decision variables. For example, graph 2502 of FIG. 25 shows different unique combinations of values of supply temperature setpoint and minimum ventilation setpoint. Similarly, regardless of a number of control decision variables, each set of the values of the control decision variables may be a unique combination, according to some embodiments.

Process 2900 includes performing a simulation for each set of the values of the control decision variables to determine sets of values of energy cost and infection risk (step 2904), according to some embodiments. In some embodiments, the simulation is performed using the dynamic models as generated by the model manager 816, or using the techniques of the dynamic model simulator 1632 (e.g., the specific model forms 1628, the generic model simulation 1630, etc.). In some embodiments, the simulations are performed by the controller 1510, or processing circuitry 802 thereof. In some embodiments, the simulations are performed for a future time horizon to generate predicted or simulated values for the energy cost and infection risk. In some embodiments, the simulations are performed for a previous or historical time period to determined values of the energy cost and infection risk for analysis (e.g., for comparison with actual historical data of the energy cost and infection risk). In some embodiments, each of the sets of values of control decisions (e.g., as obtained in step 2902) is used for a separate simulation to determine a corresponding set of performance variables (e.g., the values of energy cost and infection risk). In some embodiments, the simulations are performed subject to one or more constraints. In some embodiments, step 2904 includes performing steps 1002-1016 of process 1000. In some embodiments, step 2904 includes performing steps 1102-1116 of process 1100.

Process 2900 includes determining which of the sets of values of energy cost and infection risk are infeasible and which are feasible (step 2906), according to some embodiments. In some embodiments, step 2906 is performed by the Pareto optimizer 1512 or by the Pareto optimizer 1636. In some embodiments, step 2906 is performed using one or more constraints. The one or more constraints can be minimum or maximum allowable values of either of the energy cost and infection risk, according to some embodiments. For example, if one of the sets of values of energy cost and infection risk has an energy cost or energy consumption that exceeds a maximum allowable value of energy cost (e.g., exceeds a maximum threshold), then such a set of values of energy cost and infection risk, and consequently the corresponding sets of values of the control decision variables, may be considered infeasible, according to some embodiments. In some embodiments, the constraints are set based on capabilities of an HVAC system that the process 2900 is performed to optimize, user inputs, budgetary constraints, minimum infection risk reduction constraints, etc.

Process 2900 includes determining which of the feasible sets of values of energy cost and infection risk are Pareto optimal solutions (step 2908), according to some embodiments. In some embodiments, the step 2908 is performed by the Pareto optimizer 1512 or by the Pareto optimizer 1636. In some embodiments, the step 2908 is performed to determine which of the sets of values of energy cost and infection risk are Pareto optimal from the feasible sets of values of energy cost and infection risk. In some embodiments, process 2900 includes performing steps 2902-2908 iteratively to determine sets of decision variables. For example, the decision variables can be iteratively generated based on simulation results (e.g., by generating additional points that are likely to be Pareto optimal based on the results of step 2908).

Process 2900 includes determining, based on the Pareto optimal solutions, a minimum energy cost solution, a maximum disinfection solution, and an equal priority energy cost/disinfection solution, according to some embodiments. In some embodiments, step 2910 is performed the Pareto optimizer 1512 or by the Pareto optimizer 1636. In some embodiments, the minimum energy cost solution is the set of values of the energy cost and infection risk that are Pareto optimal, feasible, and also have a lowest value of the energy cost. In some embodiments, the maximum disinfection solution is selected from the set of values of the energy cost and infection risk that are feasible and Pareto optimal, and that has a lowest value of the infection risk. In some embodiments, the equal priority energy cost/disinfection solution is selected from the set of values of the energy cost and infection risk that are feasible and Pareto optimal, and that equally prioritizes energy cost and infection risk. For example, the energy cost/disinfection solution can be a point that is proximate an inflection of a curve that is fit to the sets of values of energy cost and infection risk (e.g., including the feasible and infeasible points, only the feasible points, only the Pareto optimal points, etc.).

Process 2900 includes providing one or more of the Pareto optimal solutions to a user via a display screen (step 2912), according to some embodiments. In some embodiments, step 2912 includes operating the display device 822 to display the Pareto optimal solutions to the user as different operational modes or available operating profiles. In some embodiments, step 2912 is performed by the display device 822 and the controller 1510. In some embodiments, step 2912 includes providing the Pareto optimal solutions and historical data (e.g., historical data of actually used control decisions and the resulting energy cost and infection risks). In some embodiments, step 2912 is optional. For example, if the user has already set a mode of operation (e.g., always use minimum infection risk settings, always use minimum energy cost solution, always use equal priority energy cost/disinfection solution, etc.), then step 2912 can be optional.

Process 2900 includes automatically selecting one of the Pareto optimal solutions or receiving a user input of a selected Pareto optimal solution (step 2914), according to some embodiments. In some embodiments, a user may select a setting for the controller 1510 to either automatically select one of the Pareto optimal solutions, or that the Pareto optimal solutions should be provided to the user for selection. In some embodiments, step 2914 is performed by the controller 1510 and the display device 822. For example, step 2914 can be performed by a user providing a selection of one of the Pareto optimal solutions (and therefore the corresponding control decisions) to be used by the HVAC system for operation, according to some embodiments. In some embodiments, step 2914 is performed automatically (e.g., if a user or administrator has selected a preferred mode of operation for the HVAC system) by the Pareto optimizer 1512 to select one of the Pareto optimal solutions and therefore the corresponding control decisions for operational use of the HVAC system.

Process 2900 includes operating equipment of an HVAC system according to the control decisions of the selected Pareto optimal solution (step 2916), according to some embodiments. In some embodiments, step 2916 includes operating the HVAC system 700. More specifically, step 2916 can include operating the UV lights 706, the AHU 704, etc., of the HVAC system 700 according to the control decisions of the selected Pareto optimal solution. Advantageously, using the control decisions of the selected Pareto optimal solution can facilitate optimal control of the HVAC system 700 in terms of risk reduction, energy consumption, or an equal priority between infection risk reduction and energy consumption or energy cost.

Pareto Optimal Analysis and Advisory

Referring again to FIGS. 15-28, the controller 1510 can be configured to perform any of the Pareto optimization techniques described herein to perform a historical analysis for the building 10 that the HVAC system 700 serves. For example, the controller 1510 can use modeling data 1618 and/or a data model 1602 that is based on historical data of the building 10, weather conditions, occupancy data, etc. In some embodiments, the controller 1510 is configured to perform the simulation and Pareto optimization techniques to determine different sets of values for the energy cost and infection risk, determine which of these sets are feasible, infeasible, Pareto optimal, etc., and compare the different Pareto optimal solutions to actual energy consumption (e.g., as read on a meter or other energy consumption sensor) and to estimated infection risks that are determined based on historical data of the building 10 or the HVAC system 700. In some embodiments, the analysis mode outputs 1634 are configured to determine potential advantages (e.g., missed energy cost or consumption opportunities) that could have been achieved if the HVAC system 700 had been operated according to a Pareto optimal solution over a previous time period. In some embodiments, the controller 1510 is configured to use newly obtained energy cost or energy consumption data and associated infection risk data (e.g., control objectives) and compare them to energy cost or energy consumption data and associated infection risk data of a previous time period (e.g., a same month from a year ago) to provide the user with information regarding improved efficiency of the HVAC system 700 resulting from operating the HVAC system 700 according to a Pareto optimal solution.

Parallel Analysis and Advisory Outputs

Referring again to FIG. 16, the controller 1510 is configured to output both analysis mode outputs 1634 and advisory mode outputs 1640 concurrently or simultaneously, according to some embodiments. In some embodiments, the controller 1510 is configured to output both the analysis mode outputs 1634 and the advisory mode outputs 1640 to a building administrator or a user as display data via the display device 822.

The analysis mode outputs 1634 can be analysis data based on historical and/or current BMS data, according to some embodiments. In some embodiments, the analysis mode outputs 1634 include energy consumption, infection risks, ventilation setpoints, etc., over a previous time period. In some embodiments, the analysis mode outputs 1634 includes sensor or meter data (e.g., of the energy consumption, energy cost, etc.), and one or more calculated values (e.g., the calculated infection risk as described using the techniques herein) for the HVAC system 700 or the building 10.

In some embodiments, the advisory mode outputs 1640 are the results of the Pareto optimizer 1636 for future or predicted time periods. In some embodiments, the advisory mode outputs 1640 include various predicted infection risks and energy costs for different values of minimum ventilation setpoint and supply temperature setpoint (e.g., different Pareto optimal solutions as described in greater detail above with reference to FIGS. 25-28). In some embodiments, the Pareto optimal solutions or operating points are presented as suggested operating points for a future time period.

In some embodiments, the analysis mode outputs 1634 (e.g., analysis results over a previous or historical time period) and the advisory mode outputs 1640 (e.g., suggested operating points for the HVAC system 700 such as different Pareto optimal points), are determined (e.g., detected, sensed, read from a meter, determined based on operating parameters of equipment of the HVAC system 700, calculated, etc.) simultaneously and presented to the user simultaneously. In this way, the controller 1510 can both “look backwards” and “look forwards” to analyze or assess previous operation of the HVAC system 700 and present analysis data, and to simultaneously determine suggested or simulated setpoints for future operation that minimize energy consumption or energy cost (e.g., a Pareto optimal solution that has lowest energy consumption or energy cost), minimize infection risk (e.g., a Pareto optimal solution that has lowest infection risk or highest disinfection), or an equal priority operating point between minimizing infection risk and minimizing energy consumption or energy cost, according to some embodiments.

In some embodiments, the previous or historical time period over which data is analyzed to determine the analysis mode outputs 1634 is a different length or time duration than the future time period for the advisory mode outputs 1640. For example, the previous or historical time period and the future time period can have different periodicities or the same periodicities. In one example, the previous or historical time period may be a previous 24 hours, while the future time period is a 1 hour time horizon, a 12 hour time horizon, etc. In some embodiments, the periodicities of the historical or previous time period and the future time period is user-selectable and can be adjusted by the user providing inputs to the controller 1510 via the display device 822. In some embodiments, the analysis mode outputs 1634 are for an hourly previous time period and the advisory mode outputs 1640 are for a future 24 hour period. In some embodiments, the analysis mode outputs 1634 include calculations of clean-air delivery and infection risk (e.g., based on BMS data obtained over the previous time period) such as minimum ventilation setpoint, supply temperature setpoint, infection risk, and energy cost. In some embodiments, the analysis mode outputs 1634 are determined based on sensor data and/or setpoints or operating parameters of the HVAC system 700 or equipment thereof (e.g., the AHU 704). In some embodiments, both the analysis mode outputs 1634 and the advisory mode outputs 1640 are determined based on common models, configurations, and data streams (e.g., the same data model 1602, the same modeling data 1618, the same dynamic model simulator 1632, etc., except using historical or previous data, and predicted or future data).

Analysis Mode Process

Referring particularly to FIG. 30, a process 3000 for determining analysis mode outputs is shown, according to some embodiments. Process 3000 can be performed for a previous or historical time period of the building 10 using BMS data obtained over the previous or historical time period, according to some embodiments. In some embodiments, process 3000 is performed by the controller 1510. In some embodiments, process 3000 includes steps 3002-3012 and can be performed simultaneously with process 5000 as described in greater detail below with reference to FIG. 50.

Process 3000 includes obtaining one or more input parameters of one or more zones of a building served by an HVAC system (step 3002), according to some embodiments. In some embodiments, the one or more input parameters are timeseries values of the input parameters obtained (e.g., via obtaining BMS data) over a previous or historical time period. In some embodiments, step 3002 includes obtaining values of setpoints of the equipment of the HVAC system (e.g., the HVAC system 300) over the previous or historical time period. In some embodiments, the values of setpoints include values of the minimum ventilation setpoint and/or supply temperature setpoint (e.g., values of decision variables). In some embodiments, the one or more input parameters include setpoints for temperature, humidity, etc., of the building 10 or various zones thereof. In some embodiments, step 3002 is performed the model generator 1610 and/or the timeseries resampler 1612. In some embodiments, the one or more input parameters are or include values of different infection parameters such as a number of infected individuals, a number of susceptible individuals, a number of infectious individuals, a volumetric breath rate of one individual, a disease quanta generation rate, etc.

Process 3000 includes generating an infection model for the one or more zones of the building served by the HVAC system based on the one or more input parameters (step 3004), according to some embodiments. In some embodiments, the infection model predicts a probability of infection as a function of at least one of the one or more input parameters. In some embodiments, the infection model is a Wells-Riley based model. In some embodiments, the infection model is any infection risk management standard, such as ASHRAE 241. In some embodiments, step 3004 is performed by the controller 1510, or more particularly, by the model manager 816 using any of the techniques described in greater detail above with reference to the model manager 816 (see e.g., FIG. 8).

Process 3000 includes obtaining an occupancy profile for the one or more zones of the building (step 3006), according to some embodiments. In some embodiments, the occupancy profile includes a scheduled occupancy of each zone for the previous time period. In some embodiments, the occupancy profile is timeseries data indicating a number of occupants in each of the zones. In some embodiments, the occupancy profile is obtained from a scheduling service or from occupancy detectors (e.g., detectors at the door that read a number of occupants that enter the building 10 or that enter a specific zone). In some embodiments, step 3006 is performed by the controller 1510 or more specifically by the model manager 816. In some embodiments, the occupancy profile is generated based on zone scheduling and ASHRAE 90.1 standards.

Process 3000 includes performing a simulation of the infection model for the previous time period (step 3008), according to some embodiments. In some embodiments, the simulation is performed using the infection model generated in step 3004. In some embodiments, the simulation is performed for the previous time period using the one or more input parameters as obtained in step 3002. The simulation can be performed to determine infection risks, infection probability, disinfection magnitude, etc., resulting from the operation of the HVAC system over the previous time period, according to some embodiments. In some embodiments, step 3008 is performed by the optimization manager 812 or the dynamic model simulator 1632. In some embodiments, the simulation is performed for a 24 hour period preceding a current time in 15 minute timesteps.

Process 3000 includes determining one or more infection metrics based on outputs of the simulation (step 3010), according to some embodiments. In some embodiments, the one or more infection metrics include infection probability, infection probability reduction, disinfection amount, etc., of the zones of the building that is served by the HVAC system. In some embodiments, the infection metrics include ventilation rate of air for the zones, clean-air delivery rate to the zones, infection risk, and/or clean air score. In some embodiments, step 3010 is performed by the dynamic model simulator 1632, the analysis mode outputs 1634, or the results manager 818.

Process 3000 includes operating a display to provide the infection metrics of the previous time period to a user as analysis data (step 3012), according to some embodiments. In some embodiments, step 3012 is performed using the display device 822. In some embodiments, step 3012 includes operating the display device 822 to provide any of the ventilation rate of air for each of the zones, clean-air delivery rate to each of the zones, infection risk of each of the zones or of the overall building, and/or clean air score for each of the zones. In some embodiments, the infection metrics are for the previous time period.

Advisory Mode Process

Referring particularly to FIG. 31, a process 3100 for determining advisory mode outputs is shown, according to some embodiments. Process 3100 can be performed for a future or subsequent time period of the building 10, according to some embodiments. In some embodiments, process 3100 is performed simultaneously or concurrently with process 3000 so that the future time period is relative to a current time, and the previous or historical time period is also relative to the current time. Process 3100 can include steps 3102-3114 and can be performed by the controller 1510, according to some embodiments.

Process 3100 includes obtaining one or more input parameters of one or more zones of a building served by an HVAC system (step 3102), according to some embodiments. In some embodiments, step 3102 is similar to step 3002 as described in greater detail above with reference to FIG. 30 but is performed for a future time period. In some embodiments, step 3102 includes defining one or more input parameters for use and obtaining BMS data. In some embodiments, step 3102 is performed using the data model 1602 and the modeling data 1618. In some embodiments, step 3102 is performed by the timeseries resampler 1612.

Process 3100 includes performing a regression technique using operational data to determine key model parameters (step 3104), according to some embodiments. In some embodiments, step 3104 includes using artificial intelligence, machine learning, a neural network, etc., to train, adjust, calibrate, etc., different model parameters (e.g., parameters of an infection risk model). In some embodiments, step 3104 is performed using a predefined model or predefined model parameters (e.g., generic parameters) that is/are adjusted based on operational data obtained from the BMS of the HVAC system to improve an accuracy of simulations or predictions using the model.

Process 3100 includes generating an infection model for the one or more zones of the building served by the HVAC system based on the one or more input parameters (step 3106), according to some embodiments. In some embodiments, step 3106 is the same as or similar to step 3004 of process 3000. In some embodiments, step 3106 is performed using the model parameters that are updated, adjusted, calibrated, determined, etc., in step 3104 using the regression technique and operational data (e.g., actual data of the HVAC system obtained from a BMS).

Process 3100 includes obtaining an occupancy profile for the one or more zones of the building (step 3108), according to some embodiments. In some embodiments, step 3108 is the same as or similar to step 3006 of process 3000 but is performed for a future time period. In some embodiments, step 3108 includes generating the occupancy profile for the zones based on ASHRAE standards and zone scheduling.

Process 3100 includes performing a simulation of the infection model for a future time period (step 3110), according to some embodiments. In some embodiments, step 3110 is the same as or similar to step 3008 of process 3000 but performed for the future time period (e.g., a future time horizon). In some embodiments, the simulation performed in step 3110 is performed at a different periodicity (e.g., the duration of the future time period, and the time steps of the simulation performed at step 3110 are different than the duration of the previous time period and the time steps of the simulation performed at step 3008 of process 3000). In some embodiments, the simulation is performed to determine energy consumption or energy cost and corresponding infection risks for different decision variables (e.g., minimum ventilation setpoint, supply temperature setpoint, etc.).

Process 3100 includes determining one or more infection metrics based on outputs of the simulation (step 3112) and performing a Pareto optimization to determine different Pareto optimal operating points (step 3114), according to some embodiments. In some embodiments, step 3112 is the same as or similar to the step 3010 of process 3000. In some embodiments, step 3114 includes performing steps 2906-2910 of process 2900 as described in greater detail above with reference to FIG. 29. For example, the simulation performed in step 3110 can result in different combinations of energy cost or energy consumption and infection risk for different operating parameters (e.g., for different values of decision variables). The Pareto optimization is performed to determine different Pareto optimal points that may be provided in step 3116 as suggested operating points or as advised operating points, according to some embodiments. In some embodiments, step 3114 includes determining which of the Pareto optimal operating points result in lowest energy cost or energy consumption, lowest infection risk (e.g., highest disinfection), and an equal priority between energy cost or energy consumption and infection risk.

Process 3100 includes operating a display to provide the Pareto optimal points (or a subset thereof) to a user as advisory data (step 3116), according to some embodiments. In some embodiments, step 3116 includes providing the Pareto optimal points associated with lowest or minimal energy cost or energy consumption, lowest or minimal infection risk (e.g., maximum disinfection), and an equal priority Pareto optimal point between energy consumption or cost and infection risk. In some embodiments, the Pareto optimal points are provided as advisory or suggested operating conditions for the future time period.

Combined Analysis and Advisory Process

Referring particularly to FIG. 32 a process 3200 for performing both an infection metric analysis for a previous time period and a Pareto optimization for a future time period is shown, according to some embodiments. In some embodiments, process 3200 includes steps 3202-3210 and is performed by the controller 1510. In some embodiments, process 3200 illustrates the simultaneous performance of process 3000 and process 3100.

Process 3200 includes performing an infection risk and energy cost analysis of an HVAC system for a previous time period to determine analysis data (step 3202) and performing an infection risk and energy cost Pareto optimization of the HVAC system for a future time period to determine advisory data (step 3204), according to some embodiments. In some embodiments, performing step 3202 includes performing process 3000. In some embodiments, performing step 3204 includes performing process 3100. In some embodiments, steps 3202 and 3204 are performed at least partially simultaneously with each other.

Process 3200 includes operating a display to provide both the analysis data and the advisory data to a user (step 3206), according to some embodiments. In some embodiments, step 3206 includes performing steps 3012 and 3116 of processes 3000 and 3100, respectively. In some embodiments, step 3206 is performed by the display device 822. Providing both the analysis data and the advisory data can facilitate a temporal bi-directional informing of the user regarding past operation of the HVAC system (e.g., the HVAC system 700) and suggested future suggestions or advisory control decisions to achieve desired energy costs and infection reduction or acceptable infection risk levels.

Process 3200 includes automatically selecting control decisions or receiving a user input of a selected control decision (step 3208) and operating equipment of an HVAC system according to the control decisions (step 3210), according to some embodiments. In some embodiments, steps 3208 and 3210 are the same as or similar to steps 2914 and 2916 of process 2900.

User Interfaces

Referring particularly to FIGS. 33-35, different user interfaces 3300, 3400, and 3500 display the various outputs of the controller 1510 (e.g., the display data), according to some embodiments. In some embodiments, the user interfaces 3300, 3400, and 3500 are displayed on the display device 822 and presented to a user or a building administrator. The user interface 3300 shows display of the analysis mode outputs 1634, the user interface 3400 shows display of the advisory mode outputs 1640, and the user interface 3500 shows checklists for implementing one of the various options of the advisory mode outputs 1640.

Referring particularly to FIG. 33, the user interface 3300 includes an infectious disease risk score icon 3302, and an indoor air quality score icon 3304. In some embodiments, the infectious disease risk score 3302 is a scaled version of the infection risk for the previous time period. In some embodiments, the infectious disease risk score is a weighted average, a time-series average, etc., of the infection risks of zones of the building 10 over the previous time period. In some embodiments, the indoor air quality score icon 3304 displays a similarly aggregated, average, etc., score of the indoor air quality of the zones of the building 10 over the previous time period. In some embodiments, the values of the infectious disease risk score and the indoor air quality score are normalized values from ranging from 0 or 1 to 100. In some embodiments, the indoor air quality score icon 1604 and the infectious disease risk score icon 3302 are graphical icons that display a bar or a circle chart and a textual or numeric value of the indoor air quality score and the infectious disease risk score for the zones of the building 10 over the previous time period. In some embodiments, the indoor air quality score icon 1604 and the infectious disease risk score icon 3302 are color-coded based on their values. For example, if the indoor air quality score is between a first or normal range, then the color of the indoor air quality score icon 1604 may be yellow, according to some embodiments. In some embodiments, if the indoor air quality score is between a second range or less than a lower value of the first or normal range, this may indicate that the indoor air quality score is poor and the color of the indoor air quality score icon 3304 may be red. In some embodiments, if the indoor air quality score is between a third range or greater than a higher value of the first or normal range, this may indicate that the indoor air quality score is good and the color of the indoor air quality score 3304 may be green.

Referring still to FIG. 33, the user interface 3300 includes a list 3306 of one or more infectious disease high risk alerts, according to some embodiments. In some embodiments, the list 3306 includes different items 3308, each item corresponding to a different zone of the building 10 and an individual infection risk score associated with the different zones. In some embodiments, the items 3308 of the list 3306 are zone-specific and are determined based on the infectious disease risk score for each of the zones of the building 10. For example, if one of the zones has an associated infectious disease risk score that is below a threshold amount, then that zone may be added with the associated infectious disease risk score to the list 3306 as one of the items 3308.

Referring still to FIG. 33, the user interface 3300 also includes a list 3310 of one or more low indoor air quality alerts, according to some embodiments. In some embodiments, the list 3310 includes different items 3312, each item corresponding to a different zone of the building 10 and an individual indoor air quality associated with the different zones. In some embodiments, the items 3312 of the list 3310 are zone-specific and are determined based on the indoor air quality for each of the zones of the building 10. For example, if one of the zones has an associated indoor air quality that is below a threshold amount, then that zone may be added with the associated indoor air quality to the list 3310 as one of the items 3312.

Referring still to FIG. 33, the user interface 3300 includes a list 3314 of each of the zones of the building 10 (e.g., organized by zone type, floor of the building 10, etc.). Each of the items of the list 3314 includes an indication of the zone or floor, an associated infectious disease risk score for the zone or floor, a number of infectious disease risk alerts for the zone or floor, an indoor air quality score for the zone or floor, an indoor air quality trend (e.g., a 30 day trend), a number of indoor air quality alerts, and/or an energy spend versus budget (e.g., for 30 days).

Referring to FIG. 34, the user interface 3400 includes different widgets 3402-3408 indicating the results of the Pareto optimization (e.g., the advisory mode outputs 1640) as described in greater detail above with reference to FIGS. 15-32. Specifically, the user interface 3400 includes a current operational state widget 3402 illustrating current energy costs and associated infectious disease risk score with additional air flow, comfort, UV disinfection, and filtration specifics, according to some embodiments. The user interface 3400 includes a widget 3404 illustrating a first option, namely, the Pareto optimal result for optimizing the infectious disease risk score (e.g., minimizing infection risk or infection probability such as the maximum disinfection point 2518), a widget 3406 illustrating a second option, namely, the Pareto optimal result for equal priority between disinfection and energy consumption (e.g., the equal priority point 2520), and a widget 3408 illustrating a third option, namely, the Pareto optimal result for operating with minimum energy cost (e.g., the minimum energy consumption point 2522). Each of the widgets 3404-3408 include graphical and/or textual information regarding a corresponding infectious disease risk score, an energy cost per a time period (e.g., a monthly time period), air flow parameters, required operational adjustments, optional design adjustments, etc., for each of the options. In some embodiments, the user or building administrator may select one of the options by selecting one of the widgets 3404-3408.

Referring particularly to FIG. 35, the user interface 3500 illustrates different operational adjustments for the HVAC system 800 that the building administrator should implement in order to configure the HVAC system 800 to perform the selected option, according to some embodiments. The user interface 3500 includes widgets 3502a-3502d, each of which illustrate a next step that should be performed to implement the selected option, according to some embodiments. In some embodiments, each widget 3502 includes a button 3504 which, when selected, navigates the user to a command and control panel where the user or building administrator can perform the specific operational adjustment (e.g., adjusting the supply temperature setpoint). Each widget 3502 also includes a button 3506 which, when selected, marks the task associated with the widget 3502 as completed, according to some embodiments.

Sustainability Metric Techniques

Referring to FIGS. 36-42, various techniques for performing the Pareto optimization based on a sustainability metric are shown, according to some embodiments. In some embodiments, the sustainability metric can be introduced into the Pareto optimization as a third control objective such that the Pareto optimization considers energy cost, infection risk, and the sustainability metric. In some embodiments, the sustainability metric may replace one of the control objectives discussed above (e.g., replacing energy cost or infection risk) such that the Pareto optimization considers energy cost and the sustainability metric (as shown in FIGS. 38 and 42) or considers infection risk and the sustainability metric (as shown in FIG. 36).

The sustainability metric may include any of a variety of metrics that quantify the performance of a building, campus, or organization with respect to energy sustainability or environmental sustainability. Some examples of sustainability metrics include carbon dioxide (CO2) related metrics (i.e., carbon equivalents) such as carbon emissions, carbon footprint, carbon credits, carbon offsets, and the like. Other examples of sustainability metrics include greenhouse gas emissions (e.g., methane, nitrous oxide, fluorinated gases, etc.), water usage, water pollution, waste generation, ecological footprint, resource consumption, or any other metric that can be used to quantify sustainable building operations. In some embodiments, sustainability metrics can be expressed on a per unit basis such as carbon per number of widgets produced, carbon per volume of product produced, carbon per meals served, carbon per patients treated, carbon per experiments run, carbon per sales revenue, carbon per items shipped, carbon per emails sent, carbon per unit of data processed, carbon per occupant, carbon per occupied room, carbon per normalized utilization value, etc. In some embodiments, sustainability metrics can be generated on an enterprise-wide basis (e.g., one value for the whole enterprise), on a building-by-building basis, on a campus-by-campus basis, by business unit/department, by building system or subsystem (e.g., HVAC, lighting, security, etc.), by control loop (e.g., chiller control loop, AHU control loop, waterside control loop, airside control loop, etc.), by building space (e.g., per room or floor,) or by any other division or aggregation. Sustainability metrics can be calculated or generated based on actual or historical building operations or predicted for future building operations using one or more predictive models.

The techniques described herein with reference to FIGS. 36-39 can be performed or implemented by the controller 1510, the functionality of the controller 1510 as shown in the diagram 1600, etc., with the one or more sustainability metrics used in the Pareto optimization as an additional parameter (e.g., an additional degree of freedom for the optimization to thereby result in points that define a Pareto optimal surface), in place of the energy cost or consumption, in place of infection risk, or as a post-processing calculation to determine a sustainability metric or carbon equivalence for the various proposed solutions or operating schedules. Advantageously, the sustainability techniques described herein can be used for at least one of (i) an operational optimization to minimize or to inform a building administrator regarding sustainable building operation, or (ii) a design optimization to determine infrastructure (e.g., building equipment, HVAC equipment, etc.) that results in a cost-effective and sustainable (e.g., reduced carbon emissions) HVAC system infrastructure.

It should be understood that the sustainability metric described herein provides an estimation of how sustainable or environmentally friendly a proposed solution is (e.g., in terms of carbon emissions). Therefore, “high” or “maximum” values of the sustainability metric indicate an increased amount of carbon emissions and decreased or minimal environmentally friendliness of the solution, and “low” values of the sustainability metric indicate a decreased amount of carbon emissions and increased or maximal environmentally friendliness of the solution. It should further be understood that while FIGS. 36-41 described herein show infection risk being used as one of the optimization objectives (e.g., one of the parameters that is calculated by an objective function), the optimization objectives can also be a combination of the sustainability metric and an energy consumption or energy cost (e.g., as shown in FIG. 42).

Referring particularly to FIG. 36, a diagram 3600 shows the graph 2502 of different decision variables (e.g., supply temperature setpoint and minimum ventilation setpoint), and a graph 3604 that shows corresponding objective function values, shown as points 3608, for each of the different combinations of decision variables. The decision variables 2506 as shown in the graph 2502 of FIG. 36 may be the same as the decision variables 2506 as shown in the graph 2502 of FIG. 25 or any other decision variables that can be adjusted to influence the objective function values. In the embodiment shown in FIG. 36, the objective function values are infection risk and a sustainability metric. In other embodiments, the objective function values may be the sustainability metric and energy cost (described in greater detail with reference to FIG. 42). It is contemplated that any set of objective function values can be used and the teachings of the present disclosure are not limited to the specific examples provided herein. The sustainability metric can be an estimated amount of CO2 emissions that are predicted to occur due to operation according to the minimum ventilation setpoint and supply temperature setpoint, or any other sustainability metric as described in detail above.

The objective function and corresponding predictive models used to define the sustainability metric and the infection risk can be similar to the objective function and predictive models described above with reference to FIGS. 15-32, with the exception that the sustainability metric is calculated/predicted in place of energy cost or consumption (e.g., as shown in the graph 3604). For each combination of the decision variables 2506 (i.e., for each of the points 2506a-25061), the controller 1510 may perform a simulation to determine corresponding values 3608a-36081 of the control objectives, as described in greater detail above with reference to FIGS. 25-27 and FIGS. 29-32. In this way, the steps of the Pareto optimization as described in greater detail above with reference to FIGS. 25-27 can be performed (e.g., by the controller 1510) using an objective function that predicts a sustainability metric and an infection risk.

In some embodiments, the value of the sustainability metric is calculated or predicted directly using one or more predictive models that define a relationship between the sustainability metric and building control decisions. For example, a predictive model may define the amount of carbon emissions as a function of operating decisions for building equipment (e.g., equipment on/off decisions, operating setpoints, etc.) over the duration of the optimization period. In other embodiments, the value of the sustainability metric can be calculated based on a corresponding amount of energy consumption. In this scenario, the objective function may be the same as the objective function previously described (e.g., an objective function that predicts energy cost or energy consumption as a function of the supply temperature setpoint and the minimum ventilation setpoint), and the resulting value of the objective function (e.g., energy consumption or cost) can be converted to a value of the sustainability metric (e.g., the carbon equivalence) using a conversion relationship. In some embodiments, the conversion relationship is a linear relationship that is used to map energy cost or energy consumption (e.g., for the HVAC system 100) to carbon emissions equivalence or any other sustainability metric.

Referring particularly to FIG. 37, a graph 3700 shows a relationship between energy cost and carbon equivalence for an operational implementation of the Pareto optimization, according to some embodiments. The graph 3700 includes points 3704 and a trendline 3702 that represents the relationship between the energy cost and the carbon equivalent. In some embodiments, the trendline 3702 is a linear relationship (e.g., y=mx+b) that can be used to convert between energy cost and carbon equivalent in either direction (e.g., to estimate an energy cost based on carbon emissions, or to estimate carbon emissions or carbon equivalent based on energy cost). In some embodiments, the relationship shown in FIG. 37 is used to convert an output of the objective function from energy consumption to the carbon equivalent prior to performing the Pareto optimization described in greater detail above with reference to FIGS. 25-27, in process 2900, process 3100, etc. In some embodiments, the relationship shown in FIG. 37 is used when the controller 1510 is implemented in an on-line mode for an operational optimization.

Referring particularly to FIG. 38, a graph 3800 shows a relationship between total cost (e.g., a sum of capital expenses for purchase and installation of equipment, and energy cost associated with operating the equipment) and carbon equivalent is shown, according to some embodiments. In some embodiments, the relationship shown in FIG. 38 is used for an off-line implementation of the controller 1510 (e.g., when the functionality of the controller is implemented as a design tool or used to determine what equipment should be purchased). The graph 3800 includes points 3804 and a curve fit 3802 that illustrates the relationship between the total cost and the carbon equivalent (e.g., a curve fit approximation of the points 3804). In some embodiments, the relationship between total cost and the carbon equivalent is used by the controller 1510 when the capital cost is affected by a decision variable or depends on a decision variable (e.g., when the controller 1510 is used to determine what equipment should be purchased or used, or to provide a recommendation to a building administrator regarding what equipment should be purchased). In some embodiments, the relationship as shown in FIG. 38 illustrates that as the carbon equivalent increases, a significant tradeoff between carbon emissions and total cost may occur (e.g., the total cost may be significantly reduced but cause increasingly higher carbon equivalents).

Referring to FIG. 39, the diagram 1600 that illustrates the functionality of the controller 1510 (see e.g., FIG. 16, above) is shown as diagram 3900, modified to account for sustainability instead of energy cost or energy consumption, according to some embodiments. In some embodiments, the components of the diagram 3900 are the same as the components of the diagram 1600 but with an additional sustainability conversion 3902 that is performed either (i) using an output of the dynamic model simulator 1632 to convert the resulting objective function values into a sustainability metric (e.g., CO2 emission or carbon equivalent), (ii) using an output of the Pareto optimizer 1636 to convert the optimized values to a sustainability metric (e.g., CO2 emission or carbon equivalent) so that at least one of the optimized values in terms of energy cost or energy consumption or the carbon equivalent is provided to the user as the advisory mode outputs 1640 or the analysis mode outputs 1634, or (iii) using a different generic model simulation 1630 that predicts the sustainability metric as a function of one or more decision variables, or incorporates a conversion between energy consumption or cost and the sustainability metric.

As shown in FIG. 39, the sustainability conversion 3902 can be implemented using an output of the dynamic model simulator 1632 to convert the objective function values from being in terms of energy cost or energy consumption to being in terms of the sustainability metric. In some embodiments, the sustainability conversion 3902 is configured to receive values of the objective function from the dynamic model simulator 1632 (e.g., values of energy cost or energy consumption) and determine a corresponding or equivalent value of a sustainability metric such as carbon or CO2 emissions. In some embodiments, the sustainability conversion 3902 or any of the other sustainability conversions described herein is/are implemented using the linear relationship shown in FIG. 37 or using the curve-fit relationship as shown in FIG. 38. In some embodiments, the linear relationship shown in FIG. 37 is implemented when the controller 1510 operates to perform operational optimizations, and the curve-fit relationship as shown in FIG. 38 is implemented when the controller 1510 operates to perform building design and operational optimization. In some embodiments, the sustainability conversion 3902 is implemented using outputs of the Pareto optimizer 1636, or as part of a post-process when providing the analysis mode outputs 1634 or the advisory mode outputs 1640 to the user.

Referring still FIGS. 39, 8-9, and 15, the functionality of the sustainability conversion 3902 can be performed in an on-demand manner in response to a user input, according to some embodiments. In some embodiments, the user can provide an input to the controller 1510 to transition use of the sustainability conversion 3902 or to toggle the controller 1510 from operating using an objective function that predicts energy consumption or energy cost, and an objective function that predicts the sustainability metric. In this way, the dynamic model simulator 1632 can be transitioned between performing a simulation for energy consumption or energy cost (e.g., the objective function values) and performing a simulation for a sustainability metric. In some embodiments, the functionality of the sustainability conversion 3902 is performed as a display feature. For example, the Pareto optimal values may be provided to the user (e.g., via the display device 822, a user interface, a display screen, etc.) in terms of energy consumption or energy cost, and can be toggled (e.g., in response to a user input) to the sustainability metric (e.g., using the linear relationship shown in FIG. 37) in response to a user input. In this way, the user can use either energy consumption, energy cost, or the sustainability metric to determine which control decisions to select. In some embodiments, the controller 1510 operates to provide both the sustainability metric and the energy cost or consumption to the user (e.g., via the display device 822) without requiring a user input.

Referring to FIGS. 25 and 36, all of the infection risk, the energy cost or consumption, and the sustainability metric can be used to determine objective function values based on one or more decision variables, according to some embodiments. Specifically, while FIGS. 25 and 36 show only two objective function values (e.g., predicted energy cost/consumption and predicted infection risk in FIG. 25, and predicted sustainability metric and predicted infection risk in FIG. 36), the objective function values may be combined so that more than two objective function values are used to assess Pareto optimality of the different combinations of decision variables. In some embodiments, the sustainability metric, the energy cost or consumption, and the infection risk are calculated using several different objective functions to determine a surface of points (e.g., each point having a value of the sustainability metric, the infection risk, and the energy cost or consumption). The subsequent steps of determining feasible points, and the various Pareto optimal points can be performed by the controller 1510 similarly as described above but using the surface graph. In some embodiments, the equal priority point is an equal priority between the sustainability metric, the infection risk, and the energy cost or consumption, and may be an inflection point of the surface of feasible points that is identified by the controller 1510.

Referring particularly to FIG. 40, a process 4000 shows the process 2900 modified to use the sustainability metric in place of the energy cost or energy consumption, according to some embodiments. In some embodiments, the process 4000 is the same as the process 2900 but includes an additional step 4002 performed using the outputs of the simulation performed in step 2904. Process 4000 includes converting values of the energy cost and infection risk to values of a sustainability metric and infection risk (step 4002), according to some embodiments. In some embodiments, step 4002 includes using either of the relationships shown in FIGS. 37 and 38 to convert the energy cost or energy consumption resulting from the simulations into a corresponding value of carbon emissions (e.g., the sustainability metric).

Process 4000 also includes modified steps 4004, 4006, and 4008, and steps 2912-2916 of process 2900, according to some embodiments. In some embodiments, the modified step 4004 is the same as the step 2906 but is performed based on the sustainability metric instead of the energy cost. Similarly, step 4006 can be the same as or similar to the step 2908 but performed to determine Pareto optimal solutions based on the sustainability metric. Finally, step 4008 can be the same as the step 2910 of the process 2900 but performed to determined various of the Pareto optimal points (e.g., a minimum sustainability metric solution, a maximum disinfection solution, and an equal priority sustainability/disinfection solution), according to some embodiments. In some embodiments, the step 4002 is incorporated in the step 2904, or the simulation is performed to determine sets of values of the sustainability metric and the infection risk directly. In such an implementation, step 4002 can be performed to determine energy cost or energy consumption based on the determined values of the sustainability metric.

Referring particularly to FIG. 41, a process 4100 for performing a Pareto optimization while accounting for the sustainability metric of an HVAC system, the energy cost, and an infection risk is shown, according to some embodiments. Process 4100 includes steps 4102-4116 and can be the same as or similar to the steps 2902-2916 of process 2900. Process 4100 differs from the process 2900 in that the process 4100 is performed for three objective function values (the energy cost, the sustainability metric, and the infection risk) instead of only two objective function values (the energy cost and the infection risk).

Particularly, process 4100 includes performing a simulation for each set of the values of the control decision variables to determine sets of values of energy cost, a sustainability metric, and infection risk (step 4104), according to some embodiments. In some embodiments, step 4104 is performed by the controller 1510 similarly to step 2904 but for several objective function values (e.g., optimization objectives). In some embodiments, step 4104 is performed using multiple objective functions, namely, an objective function that estimates energy cost based on the control decision variables, an objective function that estimates a sustainability metric based on the control decision variables, and an objective function that estimates or predicts infection risk based on the control decision variables. In some embodiments, the objective function used to predict the sustainability metric is the same as the objective function used to predict the energy cost but with an additional conversion factor or function to convert the energy cost to the sustainability metric.

Process 4100 includes determining which of the sets of values of energy cost, the sustainability metric, and the infection risk are infeasible and which are feasible (step 4106), according to some embodiments. In some embodiments, the sets of values of energy cost, sustainability metric, and the infection risk are used to construct a surface or 3-d plot. In some embodiments, the step 4106 includes comparing various values of the control decision variables, or values of any of the energy cost, the sustainability metric, or the infection risk to constraints (e.g., user-specified constraints, system operating constraints, etc.) to determine which of the sets of values of the energy cost, the sustainability metric, or the infection risk are feasible or infeasible.

Process 4100 includes determining which of the feasible sets of values of energy cost, the sustainability metric, and the infection risk are Pareto optimal solutions (step 4108), according to some embodiments. In some embodiments, step 4108 is performed similarly to step 2908 of process 2900 but also accounting for the sustainability metric. In some embodiments, the Pareto optimal solutions are curves that define multiple Pareto optimal solutions. Process 4100 includes determining, based on the Pareto optimal solutions, a minimum energy cost solution, a maximum disinfection solution, a minimum sustainability metric solution, and an equal priority solution (step 4110), according to some embodiments. In some embodiments, step 4110 is similar to step 2910 but also accounts for the sustainability metric. In some embodiments, the Pareto optimal solutions are curves that define multiple Pareto optimal solutions along the surface graph in terms of the minimum energy cost solution, the maximum disinfection solution, the minimum sustainability metric solution, and the equal priority solution. In some embodiments, the equal priority solution is an equal priority between the energy cost, the infection risk, and the sustainability metric.

Process 4100 includes steps 4112-4116 that are the same as or similar to steps 2912-2916 but also including display and accordingly user selection of options that take into account the sustainability metric. In some embodiments, only one of the sustainability metric or the energy cost is displayed to the user via the display screen, and the user may toggle between the sustainability metric and the energy cost for the different proposed solutions to facilitate proper selection.

Referring to FIG. 42, another diagram 4200 shows a graph 4202 of different decision variables, and a graph 4204 of corresponding objective function values for each of the different decision variables. The graph 4202 can be the same as or similar to the graph 2502 as described in greater detail above with reference to FIG. 25. As shown in FIG. 42, the graph 4202 shows different combinations for a first decision variable (the Y-axis) and a second decision variable (the X-axis) for the HVAC system 700, illustrated as points 4206, according to some embodiments. In some embodiments, the first decision variable and the second decision variable are minimum temperature setpoint and minimum ventilation setpoint for the HVAC system 700. It should be understood that the decision variables 4206 are not limited to the minimum temperature setpoint and the minimum ventilation setpoint and may be any other setpoints, operating parameters, etc., such as temperature setpoints, humidity setpoints, comfort parameters, HVAC operating parameters, etc. The points 4206 are shown to include points 4206a-42061, each point corresponding to a different pair of objective function values.

In some embodiments, a simulation is performed using one or more dynamic models to determine one or more corresponding values of Pareto optimization objectives (e.g., values of the sustainability metric and the energy cost) for each of the decision variables, represented by the points 4206. In some embodiments, the corresponding values of the Pareto optimization objectives are shown as points 4208, including points 4208a-42081. Each of the points 4208a-42081 correspond to one of the points 4206a-42061. In some embodiments, the one or more dynamic models include models that predict energy cost and/or the sustainability metric as a function of both the first decision variable and the second decision variable. In some embodiments, the points 4208 are determined by the dynamic model simulator 1632 based on the different values of the first decision variable and the second decision variable using one or more dynamic models.

The points 4208 can be used by the Pareto optimizer 1636 or the Pareto optimizer 1512 to determine which of the points 4208 are feasible and in-feasible, and to further determine which of the feasible points 4208 are Pareto optimal points in terms of sustainability (e.g., a Pareto optimal point having a lowest value of the sustainability metric), energy cost (e.g., a Pareto optimal point having a lowest value of energy cost), or an equal priority Pareto point that optimizes both the sustainability metric and the energy cost equally, according to some embodiments. In some embodiments, the sustainability metric accounts for operational carbon emissions (e.g., if the Pareto optimization is performed in the context of an operational tool) or accounts for carbon emissions resulting from both operation of the HVAC system 700 and/or installing equipment in the HVAC system 700 (e.g., if the Pareto optimization is performed in the context of a design tool). In some embodiments, the points 4208 are used for selection or determination of the various Pareto optimal points. These Pareto optimal points may be automatically selected for use by the HVAC system 700 or may be presented to a building administrator for selection thereof. Each of the Pareto optimal points corresponds to different values of the decision variables or schedules of decision variables over a time period (e.g., setpoints over a future time period) and selection of one of the Pareto optimal points of the points 4208 results in the selection of the corresponding decision variables or schedules of the decision variables.

Once a particular Pareto optimal point of the Pareto optimal points 4208 are selected, the controller 1510 or the controller 710 generates control signals for the HVAC system 700 to operate the HVAC system 700 according to the selected Pareto optimal point (e.g., according to the corresponding combination of decision variables or the corresponding schedule of the decision variables over a time horizon such as a future time horizon).

Mold Risk

Referring now to FIGS. 43-54, various techniques for quantifying mold risk in a building and controlling building conditions (e.g., temperature, humidity, air quality, etc.) based on mold risk are shown, according to some embodiments. The quantified mold risk (e.g., a mold risk metric or mold risk score) can be used by any of the systems and methods described herein as a control objective (e.g., minimize mold risk) either alone or in combination with any other control objectives. In some embodiments, the mold risk can be introduced into the Pareto optimization of the present disclosure in combination with any of the other control objectives described herein. For example, in some embodiments the mold risk metric can be used in combination with an energy metric (e.g., energy consumption, energy cost, etc.) in a 2-objective Pareto optimization, as shown in FIGS. 44-47. In some embodiments, the mold risk metric can be used in combination with both an energy metric and an infection risk metric in a 3-objective Pareto optimization. In some embodiments, the mold risk metric can be used in combination with an energy metric, a sustainability metric, an infection risk metric, and/or any other control objective which may be of interest (e.g., occupant comfort, occupant productivity, etc.). In various embodiments, users may use the Pareto optimization to make trade-offs between control objectives based on user preferences or tolerances, even if the user preferences do not align with a Pareto-optimal solution. For example, if the presence of mold is unacceptable to a user, the user may adjust the control objectives such that mold risk is at a lowest possible value, regardless of a corresponding energy cost or consumption. A lower or lowest mold risk value may correspond to an increased energy cost or consumption or other control parameter (e.g., an increased infection risk, etc.)

In some embodiments, the building control system switches between various control strategies depending on time of day or whether occupants are present in the building. For example, the building control system may use a 3-objective control strategy (e.g., a 3-objective Pareto optimization) that considers mold risk, energy, and infection risk when occupants are present in the building or during normal occupancy hours, but switches to a 2-objective control strategy (e.g., a 2-objective Pareto optimization) that considers mold risk and energy (without considering infection risk) when occupants are not present in the building or during normal unoccupied hours. Advantageously, such an approach may allow the building control system to protect against mold growth continuously, whereas infection risk may only be considered for time periods when occupants are present.

Mold risk can be quantified and predicted using similar techniques as infection risk as described above. For example, the building system can use a predictive mold risk model that predicts mold risk as a function of building conditions (e.g., temperature, humidity, dew point, mold spore concentration, etc.) that affect the likelihood of mold growth. In some embodiments, the predictive mold risk model also or alternatively predicts mold risk based on the types of materials exposed to air in the building. For example, carpeted or porous materials or surfaces in the building zone (e.g., walls, floors, objects, etc.) may be more susceptible to mold growth, whereas tiled or hard surfaces may be less susceptible to mold growth. Each building zone may be associated with a particular type of material or set of materials in the predictive mold risk model, which may act as a multiplier or additional factor that affects the mold risk based on building conditions. Mold risk may not be an absolute threshold indicating mold growth, but rather may be an indicator of a relative risk of mold growth. Predictive temperature models, predictive humidity models, predictive air quality models and/or other types of predictive models can be used to predict the building conditions as a function of control decisions over a given time period (e.g., temperature setpoints, humidity setpoints, air filtration setpoints, etc.). The building system can consider the mold risk in combination with other control objectives and allow the user to select between various control strategies that prioritize minimizing mold risk vs. other control objectives such as energy consumption/cost, infection risk, sustainability, etc. These and other features relating to mold risk are described in detail below.

The control objective of mold risk may be rooted in protecting the building against mold growth. Protecting the building against mold growth may indirectly protects the building occupants. The control objective of infection risk may be to determine or ensure that the indoor air that occupants actively breathe is suitably healthy and free of pathogens. Because mold can grow at any time, mitigating mold risk may be irrespective of the building's occupancy. Accordingly, mitigating mold risk may be applicable at all times (e.g., 24 hours a day, 7 days a week). Conversely, the control objective of infection risk may apply only when the building is occupied (e.g., during business hours). Controlling equipment to minimize the risk of mold growth may impact how equipment is controlled during both occupied time periods and unoccupied time periods.

Mold growth can be caused by a variety of factors. For example, one or both of fenestration and natural ventilation may introduce outdoor air to the inside of the building. Outdoor air may be introduced through, for example, leaks in the building envelope (i.e., fenestration) or intentionally through open windows and/or doors (i.e., natural ventilation). Fenestration and/or natural ventilation may be common in older buildings that have less tight envelopes than newer buildings. The introduction of unconditioned outdoor air to the indoor environment can result in excessive indoor humidity. Excessive indoor humidity may cause an indoor dew point to rise towards, and subsequently up to, outdoor dew point levels. If any surface indoors is cooler than the indoor air dew point (for example, an indoor-facing surface of a poorly insulated exterior wall), water may condense on the surface and create an environment suitable for mold growth.

Another factor that may affect mold growth in some cases is unoccupied setpoint setbacks. Unoccupied setback describes a practice that “sets back” the temperature setpoints (e.g., increasing cooling setpoints and decreasing heating setpoints) during unoccupied periods. Having no occupants in the building may mean that thermal comfort does not need to be maintained. The “widened” unoccupied temperature setpoints may decrease the runtime of the HVAC system, with the intention of reducing the energy consumption and/or cost. Setting back temperature setpoints may save energy. However, the reduced equipment runtime may mean that the equipment is no longer able to dehumidify the indoor spaces. Indoor moisture levels can increase during unoccupied times when there is fenestration or a moisture-generating process.

Another factor that may affect mold growth in some cases is pre-occupancy purges. Pre-occupancy purge is a control sequence used to purge “stale” indoor air (e.g., air that has been circulating in the space for a certain period of time) from a space prior to occupancy. The stale air may potentially be polluted and/or contaminated. During a pre-occupancy purge, the stale air may be replaced with fresh outdoor air. The goal of pre-occupancy purge may be to ensure good indoor air quality (e.g., air quality meeting certain metrics) when people begin occupying the space.

While pre-occupancy purge may help improve indoor air quality, it may potentially introduce excessive humidity into indoor spaces. Excessive humidity may be introduced during a pre-occupancy purge if outdoor conditions are humid and the ventilation air is not sufficiently conditioned. For example, a building may be in a hot and humid climate, with night-time lows of 65 degF & 95% relative humidity. The building may require cooling all the time, with occupied cooling setpoints around 74 degF and unoccupied cooling setpoints around 83 degF. The HVAC may be scheduled to transition from unoccupied-to-occupied at 6:00 am, with a 1-hour pre-occupancy purge scheduled to start at 5:00 am. When the purge begins, the indoor space may need cooling, and the outdoor air may be suitable for “free-cooling” (e.g., an economizer mode) with no need for mechanical cooling. The equipment may introduce a high volume of outdoor air at 65 degF and 95% relative humidity into the space without dehumidifying it. This may ultimately raise indoor humidity levels beyond design.

Another factor that may affect mold growth or risk in some cases is changing the temperature of the air in a building or space of the building. Mold may form when there is moisture on surfaces. If a building does not have any leaks, moisture may form when surfaces are below the dew point of the air in the building. Thus, reducing mold risk may be achieved by reducing a moisture content of the air, which consequently reduces the dew point of the air. Therefore, surfaces may generate condensation when the surfaces are cold. A cold surface may be unlikely in a building or indoor space. An alternative approach may be increasing the temperature of the air (e.g., by heating the air), which may result in heating surfaces. For example, if the dew point in the building or space is 70 degF, the building or space may be operated at a room temperature of 85 degF, thus keeping surfaces above the dew point temperature. A building operated at a high temperature, such as 85 degF, may be uncomfortable for an occupant to be in, but may be acceptable when the building is unoccupied.

Referring now to FIG. 43, several plots 4310, 4320, and 4330 (collectively plots 4300) illustrating the effects of a given control strategy on the factors that impact mold growth are shown for an example building. Plot 4310 shows temperature (T), plot 4320 shows relative humidity (RH), and plot 4330 shows dew point (T_dp) timeseries data for indoor conditions, indicated by lines 4340, and outdoor conditions indicated by lines 4345. The figure may represent, for example, the above parameters at a school in early September.

The gray background 4350 indicates time periods during which the HVAC is scheduled for occupied control. Occupied control may mean, for example, providing outdoor ventilation and controlling to occupied temperature setpoint. For example, the HVAC may be scheduled for occupied control during the hours of 5:30 am to 9:30 pm.

The white background 4355 indicates time periods during which the HVAC systems appears to be off. The off time periods may be either explicitly scheduled as off or scheduled for unoccupied control and remain idle for the duration of the time period. The HVAC systems may be off, for example, between the hours of 9:30 pm and 5:30 am.

When the HVAC transitions to occupied control, the indoor dry bulb and dew point may drop, as indicated by time period 4360. The drop may indicate that the system is in cooling and thus providing sufficient dehumidification to the space. When the HVAC is off (either scheduled as off or remaining idle due to unoccupied control), the indoor dew point may increase towards outdoor levels, as indicated by line 4340 on plot 4330 nearing line 4345 in period 4370. Period 4370 thus indicates that some outdoor air is leaking indoors. The indoor dew point may stay at a level (e.g., at or nearing the outdoor dew point level) until the following time period (e.g., the following morning) when the HVAC is operational. The building may therefore spend a prolonged period of time at elevated dew points. Thus, any surface cooler than the elevated dew point may cause water vapor to condense out of the air, creating an environment conducive of mold growth. In some embodiments, the HVAC system may not dehumidify the indoor space when an outdoor temperature and outdoor dew point is low (i.e., not hot and/or humid). Time period 4380 of plot 4330 indicates that the indoor conditions and the outdoor conditions are similar, as indicated by line 4340 and line 4345 following the same path.

The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) provides a standard that relates to preventing or mitigating mold growth. For example, ASHRAE Standard 62.1 provides, in Section 5.12, that systems that cool by mechanical means or indirect evaporation shall be designed to limit the indoor humidity to a maximum dew point of 60° F. (15° C.) during both occupied and unoccupied hours whenever the outdoor air dew point is above 60° F. (15° C.). The dew-point limit shall not be exceeded when system performance is analyzed with outdoor air at the dehumidification design condition (that is, design dew point and mean coincident dry-bulb temperatures) and with the space interior loads (both sensible and latent) at cooling design values and space solar loads at zero. Exceptions to 5.12 include (1) spaces equipped with materials, assemblies, coatings, and furnishings that resist microbial growth and that are not damaged by continuously high indoor air humidity and (2) during overnight unoccupied periods not exceeding 12 hours, the 60° F. (15° C.) dew-point limit shall not apply, provided that indoor relative humidity does not exceed 65% at any time during those hours.

Examples of spaces that are potentially exempted by Exception 1 may be shower rooms, swimming pool enclosures, kitchens, spa rooms, or semicooled warehouse spaces that contain stored contents that are not damaged by continuously high indoor air humidity or microbial growth. This requirement reduces the risk of microbial growth in buildings and their interstitial spaces because it limits the mass of indoor water vapor that can condense or be absorbed into mechanically cooled surfaces. The dew-point limit is explicitly extended to unoccupied hours because of the extensive public record of mold growth in schools, apartments, dormitories, and public buildings that are intermittently cooled during unoccupied hours when the outdoor air dew point is above 60° F. (15° C.).

While ASHRAE standards provide some guidance on the types of building conditions that affect mold growth and threshold levels for those conditions, ASHRAE guidance falls short on how mold risk can be incorporated into a more comprehensive control strategy that also considers other control objectives such as energy consumption/cost, occupant comfort, infection risk, sustainability, and/or other control objectives which may be of interest. Accordingly, it is an objective of the present disclosure to incorporate mold risk in an optimal manner in combination with other control objectives.

Mold Risk Pareto Optimization

In various embodiments, a control objective for a Pareto optimization may be mold risk. Mold risk may be a function of indoor air temperature and/or indoor humidity. Temperature, humidity, and mold risk control objectives may be modeled using one or more predictive models. The controller 1510, described in greater detail above with respect to FIG. 15, may be configured to receive control objectives provided to the Pareto optimizer 1512 that include energy cost or energy consumption and mold risk. The Pareto optimizer 1512 may be configured to use optimization results from the optimization manager 812 and perform a Pareto optimization to determine feasible and infeasible operating points, and to determine, from the feasible operating point, which is the Pareto optimal point. In some embodiments, the optimization results provided to the Pareto optimizer 1512 are or include values of the objective function. Values of the objective function may be referred to as control objectives. For example, the values of the objective function can include values of two or more variables of interest. In some embodiments, the control objectives can include energy cost and mold risk for an associated pair of decision variables such as humidity setpoint and supply temperature setpoint. In some embodiments, both the values of the decision variables and the control objectives are provided to the Pareto optimizer 1512 for use in determining what values of the decision variables should be used to achieve the Pareto optimal values of the control objectives.

Mold risk can be quantified, predicted, and handled using the same framework described throughout the present disclosure with respect to infection risk and other control objectives. For example, in various embodiments, the diagram 1600 illustrating the functionality of the controller 1510 may include mold risk parameters when the control objectives of the objective function include mold risk. The functionality of the controller 1510 may be the same or similar to the functionalities described above with respect to FIGS. 15 and 16 with respect to calculating or predicting mold risk. For example, the controller 1510 may use predictive temperature and humidity models to predict these building conditions as a function of control decisions for building equipment and then predict mold risk as a function of the predicted temperature and humidity conditions of the building.

In various embodiments, the controller 1510 may determine a temperature model for use in an optimization problem. The temperature model can be generated or determined by model manager 816 for use in an optimization problem. In some embodiments, the temperature model is:

$ρ c V_{k} (\frac{d T_{k}}{dt}) = ρ c f_{k} (T_{0} - T_{k}) + Q_{k} (T_{k})$

where ρ is a mass density of air, c is a heat capacity of air, V_kis a volume of the kth zone, f_kis a volumetric flow of air into the kth zone, T₀is the temperature of air output by the AHU, T_kis the temperature of the kth zone, and Q_kis the heat load on the kth zone. The temperature model generation can be performed by model manager 816 as described in greater detail above with reference to FIGS. 8-9.

Determining mold risk may include determining a humidity model for each of the multiple zones to predict a humidity of the corresponding zone based on one or more conditions or parameters of the corresponding zone, according to some embodiments. Generating a humidity model may be similar to generating the temperature model. In some embodiments, the humidity model is:

$ρ V_{k} (\frac{d ω_{k}}{dt}) = ρ f (ω_{0} - T_{0}) + w_{k}$

for a kth zone 206, where ω is a humidity ratio expressed in pounds H2O/pounds dry air. In some embodiments, generating the humidity model is performed by model manager 816 as described in greater detail above with reference to FIGS. 8-9.

A mold risk model may be determined for each of the multiple zones to predict a mold risk of the corresponding zone based on one or more conditions or parameters of the corresponding zone, according to some embodiments. In some embodiments, the mold risk model is similar to and/or based on the humidity model above or the temperature model above. In some embodiments, the mold risk model predicts mold risk as a function of the predicted temperature values T_kand the predicted humidity values ω_kgenerated using the predictive temperature and humidity models, respectively. For example, the mold risk model can define mold risk as a function of temperature and humidity as follows:

$M_{k} = f (T_{k}, ω_{k})$

according to some embodiments, where M is the mold risk defined as a function of temperature and humidity at time k. In some embodiments, generating the mold risk model is performed by model manager 816.

Pareto Optimization Techniques

Referring particularly to FIGS. 44-47, the Pareto optimizer 1636 or the Pareto optimizer 1512 are configured to perform various Pareto optimization techniques as described herein to determine Pareto optimization results, according to some embodiments. It should be understood that while the techniques described herein with reference to FIGS. 44-47 are described as being performed by the Pareto optimizer 1636, the techniques can also be performed by the Pareto optimizer 1512 or processing circuitry 802 thereof.

Referring particularly to FIG. 44, a diagram 4400 shows a graph 4402 of different decision variables, and a graph 4404 of corresponding control objectives for each of the different decision variables. The graph 4402 shows different combinations for a supply temperature setpoint and a humidity setpoint (shown on the Y and X axes, respectively) for the HVAC system 700, according to some embodiments. It should be understood that only two decision variables are shown for ease of explanation, and that any number of decision variables may be used. Different values and combinations of both the decision variables are represented in FIG. 44 as points 4406. For example, points 4406a-4406d have a same value for the humidity setpoint decision variables but different values of the supply temperature setpoint decision variable. Similarly, points 4406e-4406h have the same value for the humidity setpoint decision (different than the value of the humidity setpoint decision variable for points 4406a-4406d) but different values of the supply temperature setpoint decision variable. Points 4406i-44061 likewise have the same value of the humidity setpoint decision variable (different than the values of the humidity setpoint decision variable for points 4406a-4406d and 4406e-4406h) but different values of the supply temperature setpoint decision variable. In some embodiments, the values of the decision variables are a fixed set (e.g., generated as a grid using minimum and maximum allowed values for each of the multiple decision variables) or are generated iteratively based on simulation results (e.g., by adding additional points that are likely to be Pareto optimal based on simulation results of proximate points).

In some embodiments, a simulation is performed to determine corresponding energy cost and mold risk for each of the different points 4406. The corresponding energy cost and mold risk are shown as points 4408 in graph 4404. In some embodiments, points 4408a-44081 of graph 4404 correspond to points 4406a-44061 of graph 4402. For example, point 4408a illustrates the corresponding energy cost and mold risk for the values of the humidity setpoint and the supply temperature of the point 4406a. Likewise, points 4408b-44081 illustrate the various corresponding energy costs and mold risks for each of the humidity setpoint and supply temperature setpoint values as represented by points 4406b-44061. In some embodiments, each of the points 4406a-44061 and the corresponding points 4408a-44081 correspond to a simulation performed by the optimization manager 812, or the dynamic model simulator 1632. For example, the dynamic model simulator 1632 may perform a simulation for each of the sets of values of the supply temperature setpoint and the humidity setpoint (e.g., the decision variables) and output values of the energy cost and mold risk for each simulation (shown as points 4408). It should be understood that while FIG. 44 shows only two objectives of the Pareto optimization (e.g., energy cost and mold risk), the Pareto optimization may have any number of optimization objectives (e.g., more than two, etc.).

Referring particularly to FIG. 45, a diagram 4500 shows the graph 4404 and a graph 4410, according to some embodiments. The graph 4404 shows the points 4408 that illustrate the various combinations of energy cost and mold risk for the corresponding values of the decision variables (the supply temperature setpoint and the humidity setpoint shown in graph 4402 in FIG. 44). In some embodiments, the graph 4410 illustrates groups 4412 that are include the points 4408 grouped according to feasibility, and further group according to Pareto optimality. Specifically, groups 4412 includes a first group of points 4412a (e.g., points 4408i and 4408e), a second group of points 4412b (e.g., points 4408f, 4408a, 4408b, 4408c, and 4408d), and a third group of points 4412c (e.g., points 4408j, 4408k, 44081, 4408g, and 4408h).

The first group of points 4412a are points that are infeasible, unfeasible, or non-feasible. In some embodiments, the Pareto optimizer 1636 is configured to determine or identify which of the points 4408 are infeasible and group such combinations of energy cost and mold risk as infeasible solutions. In some embodiments, the Pareto optimizer 1636 is configured to use threshold energy costs/consumption or mold risks, and if some of the points 4408 are greater than a maximum allowable energy cost/consumption or mold risk, or less than a minimum allowable mold risk or energy cost, the Pareto optimizer 1636 can determine that such points are infeasible and group them accordingly as the first group of points 4412a. In some embodiments, the maximum or minimum allowable energy cost/consumption or mold risk values used by the Pareto optimizer 1636 are user inputs, values set by legal regulations, or values determined based on abilities of the HVAC system 700. In some embodiments, the second group of points 4412b and the third group of points 4412c are feasible solutions.

In some embodiments, the Pareto optimizer 1636 is configured to perform a Pareto optimization based on the points 4408 to determine which of the points 4408 are Pareto optimal. A Pareto optimal point is a point where neither the energy cost/consumption or the mold risk can be reduced without causing a corresponding increase in the mold risk or the energy cost/consumption. In the example shown in FIG. 45, the points 4408j, 4408k, 44081, 4408g, and 4408h are Pareto optimal points, and the Pareto optimizer 1636 is configured to classify these points as such, thereby defining the third group of points 4412c, according to some embodiments. In some embodiments, the Pareto optimizer 1636 is configured to determine that points which are feasible but are not Pareto optimal (e.g., points 4408f, 4408a, 4408b, 4408c, and 4408d) should define the second group of points 4412b. In this way, the Pareto optimizer 1636 can define several groups of the points 4408 (e.g., groups of various solutions to be considered): (i) infeasible points, (ii) feasible points that are Pareto optimal, and (iii) feasible points that are not Pareto optimal.

Referring particularly to FIG. 46, the Pareto optimizer 1636 can further identify which of the Pareto optimal points, shown as the third group of points 4412c in graph 4410, result in minimal mold risk, minimum energy consumption or cost, and an equal priority between mold risk and energy consumption or cost, according to some embodiments. In some embodiments, the Pareto optimizer 1636 is configured to determine which of the Pareto optimal points (i.e., the points 4408 of the third group of points 4412c) have a minimum mold risk, a minimum energy consumption or cost, and an equal priority between energy consumption or cost and mold risk. Specifically, in the example shown in FIGS. 44-46, the point 4408j is a Pareto optimal point that has a lowest mold risk, and therefore the Pareto optimizer 1636 identifies point 4408j as a minimal mold risk point 4418. Similarly, in the example shown in FIGS. 44-46, the Pareto optimizer 1636 determines that the point 4408h is a Pareto optimal point associated with minimum energy consumption or cost, and therefore the Pareto optimizer 1636 identifies the point 4408h as a minimum energy consumption/cost point 4422. Finally, in the example shown in FIGS. 44-46, the Pareto optimizer 1636 determines that the point 44081 is a Pareto optimal point that results in an equal priority between the mold risk and the energy consumption, shown as equal priority point 4420.

In some embodiments, the minimal mold risk point 4418 (e.g., point 4408j), the minimum energy consumption (or energy costs) point 4422 (e.g., point 4408h), and the equal priority point 4420 (e.g., point 44081) are the Pareto optimization results. In some embodiments, the Pareto optimization results also include the energy cost, mold risk, as well as the humidity setpoint, and the supply temperature setpoint for each of the minimal mold risk point 4418, the minimum energy consumption point 4422, and the equal priority point 4420. In some embodiments, the Pareto optimization results are provided to the user via the display device 822 for selection. For example, the display device 822 can provide different recommended operating possibilities such as a minimal mold risk operating possibility, a minimum energy consumption operating possibility, and an equal energy and mold risk. In some embodiments, the user may select one of the different operating recommendations, and provide the selection to the controller 1510 for use in operating the HVAC system 700 according to the selected operating possibility.

Referring particularly to FIG. 47, a diagram 4700 shows another example of a graph 4702 and a graph 4704 illustrating the functionality of the Pareto optimizer 1636, according to some embodiments. Graph 4702 includes points 4722 that illustrate different combinations of humidity and supply temperature setpoint (e.g., the decision variables) usable by the dynamic model simulator 1632 or the optimization manager 812 for performing a simulation to determine corresponding mold risk and energy costs or consumption (shown as points 4712 in graph 4704), according to some embodiments. In some embodiments, the points 4712 illustrated in the graph 4704 are mapped to the points 4722. For example, the simulation can be performed for each of the points 4722 to determine the corresponding mold risk and energy cost or energy consumption, as represented by points 4712 in graph 4704.

In the example shown in FIG. 47, the Pareto optimizer 1636 identifies a group 4720 of infeasible points, according to some embodiments. The infeasible points may be points that cannot be achieved due to constraints and operational ability of the HVAC system 700. In various embodiments, if one or more infeasible points correspond to levels of mold risk deemed to be acceptable by a user, the HVAC system 700 or other component of the BMS may be or include a mode or tool to recommend equipment upgrades. For example, if the building is not able to operate at a desired mold risk level, the HVAC system 700 or BMS may recommend, for example, energy recovery ventilators, dedicated outdoor air systems, reheat coils, etc. such that the building can operate at the desired mold risk level. In some embodiments, the Pareto optimizer 1636 further identifies which of the points 4712 are Pareto optimal points, and determines which of the Pareto optimal points are associated with lowest mold risk, lowest energy costs or energy consumption, and a balanced priority point where energy costs and mold risk are equally prioritized. In the example shown in FIG. 47, the Pareto optimizer 1636 identifies that a point 4718 of graph 4704 is the Pareto optimal point that results in lowest mold risk, which corresponds to point 4710 of graph 4702. Similarly, the Pareto optimizer 1636 may determine that the point 4714 is the Pareto optimal point that results in lowest energy cost, which corresponds to the point 4706 in graph 4702, according to some embodiments. Finally, the Pareto optimizer 1636 can determine that a point 4716 of graph 4704 is a Pareto optimal point that results in a solution with equal priority between the mold risk and the energy cost, which corresponds to point 4708 of graph 4702, according to some embodiments.

In some embodiments, the Pareto optimizer 1636 is configured to provide all of the Pareto optimal points to the user via the display device 822. In some embodiments, the Pareto optimizer 1636 or the Pareto optimizer 1512 is configured to provide the Pareto optimal points as the Pareto optimal results to the results manager 818 for use in display to the user. In some embodiments, the Pareto optimizer 1512 or the Pareto optimizer 1636 automatically selects one of the Pareto optimal points for use and provides the Pareto optimal points and its associated control decisions (e.g., the humidity and supply temperature setpoints) to the control signal generator 808.

Pareto Optimization Process

Referring particularly to FIG. 48, a process 4800 for performing a Pareto optimization to determine operation of a building HVAC system is shown, according to some embodiments. Process 4800 includes steps 4802-4816 and can be performed by the controller 1510 or more specifically, by the Pareto optimizer 1512, according to some embodiments. In some embodiments, process 4800 is performed to determine various Pareto optimal values, or to determine a Pareto optimal solution and associated operating parameters that results in optimal tradeoff between mold risk and energy cost.

Process 4800 includes obtaining multiple sets of values of control decision variables, each set of values including a different combination of the control decision variables (step 4802), according to some embodiments. In some embodiments, the control decision variables include a humidity setpoint and a supply temperature setpoint. In some embodiments, the control decision variables include operating parameters or control decisions of the UV lights 706 or the AHU 704 (e.g., a fresh air intake fraction x). It should be understood that the control decision variables described herein are not limited to only two variables, and may include any number of variables. In some embodiments, each set of the values of the control decision variables is a unique combination of different values of the control decision variables. For example, graph 4402 of FIG. 44 shows different unique combinations of values of supply temperature setpoint and humidity setpoint. Similarly, regardless of a number of control decision variables, each set of the values of the control decision variables may be a unique combination, according to some embodiments.

Process 4800 includes performing a simulation for each set of the values of the control decision variables to determine sets of values of energy cost or consumption and mold risk (step 4804), according to some embodiments. In some embodiments, the simulation is performed using the dynamic models as generated by the model manager 816, or using the techniques of the dynamic model simulator 1632 (e.g., the specific model forms 1628, the generic model simulation 1630, etc.). In some embodiments, the simulations are performed by the controller 1510, or processing circuitry 802 thereof. In some embodiments, the simulations are performed for a future time horizon to generate predicted or simulated values for the energy cost and mold risk. In some embodiments, the simulations are performed for a previous or historical time period to determined values of the energy cost and mold risk for analysis (e.g., for comparison with actual historical data of the energy cost and mold risk). In some embodiments, each of the sets of values of control decisions (e.g., as obtained in step 4802) is used for a separate simulation to determine a corresponding set of performance variables (e.g., the values of energy cost and mold risk). In some embodiments, the simulations are performed subject to one or more constraints. In some embodiments, step 4804 includes performing steps 1002-1016 of process 1000. In some embodiments, step 4804 includes performing steps 1102-1116 of process 1100. In various embodiments, steps of processes 1000 and 1100 with reference to infection risk or related terms (e.g., infectious quanta, infectious disease risk, etc.) may be performed for mold risk and mold-risk related terms and models. For example, step 4804 may include performing steps 1002-1016 and/or steps 1102-1116. However, a mold risk quanta model or mold risk probability constraint may be generated, determined, and/or used instead of an infectious quanta model or infection risk probability constraint, respectively.

Process 4800 includes determining which of the sets of values of energy cost or consumption and mold risk are infeasible and which are feasible (step 4806), according to some embodiments. In some embodiments, step 4806 is performed by the Pareto optimizer 1512 or by the Pareto optimizer 1636. In some embodiments, step 4806 is performed using one or more constraints. The one or more constraints can be minimum or maximum allowable values of either of the energy cost and mold risk, according to some embodiments. For example, if one of the sets of values of energy cost and mold risk has an energy cost or energy consumption that exceeds a maximum allowable value of energy cost (e.g., exceeds a maximum threshold), then such a set of values of energy cost and mold risk, and consequently the corresponding sets of values of the control decision variables, may be considered infeasible, according to some embodiments. In some embodiments, the constraints are set based on capabilities of an HVAC system that the process 4800 is performed to optimize, user inputs, budgetary constraints, minimum mold risk reduction constraints, etc. In various embodiments, one or more sets of control objective values may correspond to levels of mold risk deemed to be acceptable by a user, but may be unable to be attained by the building. For example, an occupant may want a mold risk level below a certain value, but the operating point may be infeasible. The HVAC system 700 or other component of the BMS may be or include a mode or tool to recommend equipment upgrades so that the mold risk level can be met. For example, if the building is not able to operate at a desired mold risk level, the HVAC system 700 or BMS may recommend, for example, energy recovery ventilators, dedicated outdoor air systems, reheat coils, etc. such that the building can operate at the desired mold risk level.

Process 4800 includes determining which of the feasible sets of values of energy cost/consumption and mold risk are Pareto optimal solutions (step 4808), according to some embodiments. In some embodiments, the step 4808 is performed by the Pareto optimizer 1512 or by the Pareto optimizer 1636. In some embodiments, the step 4808 is performed to determine which of the sets of values of energy cost and mold risk are Pareto optimal from the feasible sets of values of energy cost and mold risk. In some embodiments, process 4800 includes performing steps 4802-4808 iteratively to determine sets of decision variables. For example, the decision variables can be iteratively generated based on simulation results (e.g., by generating additional points that are likely to be Pareto optimal based on the results of step 4808).

Process 4800 includes determining, based on the Pareto optimal solutions, a minimum energy cost solution, a minimum mold risk solution, and an equal priority energy cost/mold risk solution, according to some embodiments. In some embodiments, step 4810 is performed the Pareto optimizer 1512 or by the Pareto optimizer 1636. In some embodiments, the minimum energy cost solution is the set of values of the energy cost and mold risk that are Pareto optimal, feasible, and also have a lowest value of the energy cost. In some embodiments, the minimal mold risk solution is selected from the set of values of the energy cost and mold risk that are feasible and Pareto optimal, and that has a lowest value of the mold risk. In some embodiments, the equal priority energy cost/mold risk solution is selected from the set of values of the energy cost and mold risk that are feasible and Pareto optimal, and that equally prioritizes energy cost and mold risk. For example, the energy cost/mold risk solution can be a point that is proximate an inflection of a curve that is fit to the sets of values of energy cost and mold risk (e.g., including the feasible and infeasible points, only the feasible points, only the Pareto optimal points, etc.).

Process 4800 includes providing one or more of the Pareto optimal solutions to a user via a display screen (step 4812), according to some embodiments. In some embodiments, step 4812 includes operating the display device 822 to display the Pareto optimal solutions to the user as different operational modes or available operating profiles. In some embodiments, step 4812 is performed by the display device 822 and the controller 1510. In some embodiments, step 4812 includes providing the Pareto optimal solutions and historical data (e.g., historical data of actually used control decisions and the resulting energy cost and mold risks). In some embodiments, step 4812 is optional. For example, if the user has already set a mode of operation (e.g., always use minimum mold risk settings, always use minimum energy cost/consumption solution, always use equal priority energy cost/mold risk solution, etc.), then step 4812 can be optional.

Process 4800 includes automatically selecting one of the Pareto optimal solutions or receiving a user input of a selected Pareto optimal solution (step 4814), according to some embodiments. In some embodiments, a user may select a setting for the controller 1510 to either automatically select one of the Pareto optimal solutions, or that the Pareto optimal solutions should be provided to the user for selection. In some embodiments, step 4814 is performed by the controller 1510 and the display device 822. For example, step 4814 can be performed by a user providing a selection of one of the Pareto optimal solutions (and therefore the corresponding control decisions) to be used by the HVAC system for operation, according to some embodiments. In some embodiments, step 4814 is performed automatically (e.g., if a user or administrator has selected a preferred mode of operation for the HVAC system) by the Pareto optimizer 1512 to select one of the Pareto optimal solutions and therefore the corresponding control decisions for operational use of the HVAC system.

Process 4800 includes operating equipment of an HVAC system according to the control decisions of the selected Pareto optimal solution (step 4816), according to some embodiments. In some embodiments, step 4816 includes operating the HVAC system 700. More specifically, step 4816 can include operating the UV lights 706, the AHU 704, etc., of the HVAC system 700 according to the control decisions of the selected Pareto optimal solution. Advantageously, using the control decisions of the selected Pareto optimal solution can facilitate optimal control of the HVAC system 700 in terms of risk reduction, energy consumption, or an equal priority between mold risk reduction and energy consumption or energy cost. If the building is not able to operate at a desired mold risk level, the HVAC system 700 or BMS may be or include a mode or tool to recommend equipment upgrades so that the mold risk level can be met. Examples of equipment upgrades include, but are not limited to, energy recovery ventilators, dedicated outdoor air systems, and reheat coils.

Pareto Optimal Analysis and Advisory

Referring again to FIGS. 15, 25 and 44-47, the controller 1510 can be configured to perform any of the Pareto optimization techniques described herein to perform a historical analysis for the building 10 that the HVAC system 700 serves. For example, the controller 1510 can use modeling data 1618 and/or a data model 1602 that is based on historical data of the building 10, weather conditions, occupancy data, etc. In some embodiments, the controller 1510 is configured to perform the simulation and Pareto optimization techniques to determine different sets of values for the energy cost and mold risk, determine which of these sets are feasible, infeasible, Pareto optimal, etc., and compare the different Pareto optimal solutions to actual energy consumption (e.g., as read on a meter or other energy consumption sensor) and to estimated mold risks that are determined based on historical data of the building 10 or the HVAC system 700. In some embodiments, the analysis mode outputs 1634 are configured to determine potential advantages (e.g., missed energy cost or consumption opportunities) that could have been achieved if the HVAC system 700 had been operated according to a Pareto optimal solution over a previous time period. In some embodiments, the controller 1510 is configured to use newly obtained energy cost or energy consumption data and associated mold risk data (e.g., control objectives) and compare them to energy cost or energy consumption data and associated mold risk data of a previous time period (e.g., a same month from a year ago) to provide the user with information regarding improved efficiency of the HVAC system 700 resulting from operating the HVAC system 700 according to a Pareto optimal solution.

Parallel Analysis and Advisory Outputs

The analysis mode outputs 1634 can be analysis data based on historical and/or current BMS data, according to some embodiments. In some embodiments, the analysis mode outputs 1634 include energy consumption, mold risks, ventilation setpoints, etc., over a previous time period. In some embodiments, the analysis mode outputs 1634 includes sensor or meter data (e.g., of the energy consumption, energy cost, etc.), and one or more calculated values (e.g., the calculated mold risk as described using the techniques herein) for the HVAC system 700 or the building 10.

In some embodiments, the advisory mode outputs 1640 are the results of the Pareto optimizer 1636 for future or predicted time periods. In some embodiments, the advisory mode outputs 1640 include various predicted mold risks and energy costs for different values of humidity setpoint and supply temperature setpoint (e.g., different Pareto optimal solutions as described in greater detail above with reference to FIGS. 44-47). In some embodiments, the Pareto optimal solutions or operating points are presented as suggested operating points for a future time period.

In some embodiments, the analysis mode outputs 1634 (e.g., analysis results over a previous or historical time period) and the advisory mode outputs 1640 (e.g., suggested operating points for the HVAC system 700 such as different Pareto optimal points), are determined (e.g., detected, sensed, read from a meter, determined based on operating parameters of equipment of the HVAC system 700, calculated, etc.) simultaneously and presented to the user simultaneously. In this way, the controller 1510 can both “look backwards” and “look forwards” to analyze or assess previous operation of the HVAC system 700 and present analysis data, and to simultaneously determine suggested or simulated setpoints for future operation that minimize energy consumption or energy cost (e.g., a Pareto optimal solution that has lowest energy consumption or energy cost), minimize mold risk (e.g., a Pareto optimal solution that has lowest mold risk), or an equal priority operating point between minimizing mold risk and minimizing energy consumption or energy cost, according to some embodiments.

In some embodiments, the previous or historical time period over which data is analyzed to determine the analysis mode outputs 1634 is a different length or time duration than the future time period for the advisory mode outputs 1640. For example, the previous or historical time period and the future time period can have different periodicities or the same periodicities. In one example, the previous or historical time period may be a previous 24 hours, while the future time period is a 1 hour time horizon, a 12 hour time horizon, etc. In some embodiments, the periodicities of the historical or previous time period and the future time period is user-selectable and can be adjusted by the user providing inputs to the controller 1510 via the display device 822. In some embodiments, the analysis mode outputs 1634 are for an hourly previous time period and the advisory mode outputs 1640 are for a future 24 hour period. In some embodiments, the analysis mode outputs 1634 include calculations of clean-air delivery and mold risk (e.g., based on BMS data obtained over the previous time period) such as humidity setpoint, supply temperature setpoint, mold risk, and energy cost or consumption. In some embodiments, the analysis mode outputs 1634 are determined based on sensor data and/or setpoints or operating parameters of the HVAC system 700 or equipment thereof (e.g., the AHU 704). In some embodiments, both the analysis mode outputs 1634 and the advisory mode outputs 1640 are determined based on common models, configurations, and data streams (e.g., the same data model 1602, the same modeling data 1618, the same dynamic model simulator 1632, etc., except using historical or previous data, and predicted or future data).

Analysis Mode Process

Referring particularly to FIG. 49, a process 4900 for determining analysis mode outputs is shown, according to some embodiments. Process 4900 can be performed for a previous or historical time period of the building 10 using BMS data obtained over the previous or historical time period, according to some embodiments. In some embodiments, process 4900 is performed by the controller 1510. In some embodiments, process 4900 includes steps 4902-4912 and can be performed simultaneously with process 3100 as described in greater detail below with reference to FIG. 31.

Process 4900 includes obtaining one or more input parameters of one or more zones of a building served by an HVAC system (step 4902), according to some embodiments. In some embodiments, the one or more input parameters are timeseries values of the input parameters obtained (e.g., via obtaining BMS data) over a previous or historical time period. In some embodiments, step 4902 includes obtaining values of setpoints of the equipment of the HVAC system (e.g., the HVAC system 700) over the previous or historical time period. In some embodiments, the values of setpoints include values of the humidity setpoint and/or supply temperature setpoint (e.g., values of decision variables). In some embodiments, the one or more input parameters include setpoints for temperature, humidity, etc., of the building 10 or various zones thereof. In some embodiments, step 4902 is performed the model generator 1610 and/or the timeseries resampler 1612. In some embodiments, the one or more input parameters are or include values of different mold parameters.

Process 4900 includes generating a mold model for the one or more zones of the building served by the HVAC system based on the one or more input parameters (step 4904), according to some embodiments. In some embodiments, the mold model predicts a probability of mold as a function of at least one of the one or more input parameters. In some embodiments, step 4904 is performed by the controller 1510, or more particularly, by the model manager 816 using any of the techniques described in greater detail above with reference to the model manager 816 (see e.g., FIG. 8).

Process 4900 includes obtaining an occupancy profile for the one or more zones of the building (step 4906), according to some embodiments. In some embodiments, the occupancy profile includes a scheduled occupancy of each zone for the previous time period. In some embodiments, the occupancy profile is timeseries data indicating a number of occupants in each of the zones. In some embodiments, the occupancy profile is obtained from a scheduling service or from occupancy detectors (e.g., detectors at the door that read a number of occupants that enter the building 10 or that enter a specific zone). In some embodiments, step 4906 is performed by the controller 1510 or more specifically by the model manager 816. In some embodiments, the occupancy profile is generated based on zone scheduling and ASHRAE 90.1 standards.

Process 4900 includes performing a simulation of the mold model for the previous time period (step 4908), according to some embodiments. In some embodiments, the simulation is performed using the mold model generated in step 4904. In some embodiments, the simulation is performed for the previous time period using the one or more input parameters as obtained in step 4902. The simulation can be performed to determine mold risks, mold probability, molding magnitude, etc., resulting from the operation of the HVAC system over the previous time period, according to some embodiments. In some embodiments, step 4908 is performed by the optimization manager 812 or the dynamic model simulator 1632. In some embodiments, the simulation is performed for a 24 hour period preceding a current time in 15 minute timesteps.

Process 4900 includes determining one or more mold metrics based on outputs of the simulation (step 4910), according to some embodiments. In some embodiments, the one or more mold metrics include mold probability, mold probability reduction, mold reduction amount, room temperature history, room materials (e.g., is the floor tiled or carpeted), etc., of the zones of the building that is served by the HVAC system. In some embodiments, the mold metrics include ventilation rate of air for the zones, clean-air delivery rate to the zones, mold risk, and/or clean air score. In some embodiments, step 4910 is performed by the dynamic model simulator 1632, the analysis mode outputs 1634, or the results manager 818.

Process 4900 includes operating a display to provide the infection metrics of the previous time period to a user as analysis data (step 4912), according to some embodiments. In some embodiments, step 4912 is performed using the display device 822. In some embodiments, step 4912 includes operating the display device 822 to provide any of the ventilation rate of air for each of the zones, clean-air delivery rate to each of the zones, mold risk of each of the zones or of the overall building, and/or clean air score for each of the zones. In some embodiments, the mold metrics are for the previous time period.

Advisory Mode Process

Referring particularly to FIG. 50, a process 5000 for determining advisory mode outputs is shown, according to some embodiments. Process 5000 can be performed for a future or subsequent time period of the building 10, according to some embodiments. In some embodiments, process 5000 is performed simultaneously or concurrently with process 4900 so that the future time period is relative to a current time, and the previous or historical time period is also relative to the current time. Process 5000 can include steps 5002-5014 and can be performed by the controller 1510, according to some embodiments.

Process 5000 includes obtaining one or more input parameters of one or more zones of a building served by an HVAC system (step 5002), according to some embodiments. In some embodiments, step 5002 is similar to step 4902 as described in greater detail above with reference to FIG. 49 but is performed for a future time period. In some embodiments, step 5002 includes defining one or more input parameters for use and obtaining BMS data. In some embodiments, step 5002 is performed using the data model 1602 and the modeling data 1618. In some embodiments, step 5002 is performed by the timeseries resampler 1612.

Process 5000 includes performing a regression technique using operational data to determine key model parameters (step 5004), according to some embodiments. In some embodiments, step 5004 includes using artificial intelligence, machine learning, a neural network, etc., to train, adjust, calibrate, etc., different model parameters (e.g., parameters of an mold risk model). In some embodiments, step 5004 is performed using a predefined model or predefined model parameters (e.g., generic parameters) that is/are adjusted based on operational data obtained from the BMS of the HVAC system to improve an accuracy of simulations or predictions using the model.

Process 5000 includes generating a mold model for the one or more zones of the building served by the HVAC system based on the one or more input parameters (step 5006), according to some embodiments. In some embodiments, step 5006 is the same as or similar to step 4904 of process 4900. In some embodiments, step 5006 is performed using the model parameters that are updated, adjusted, calibrated, determined, etc., in step 5004 using the regression technique and operational data (e.g., actual data of the HVAC system obtained from a BMS).

Process 5000 includes obtaining an occupancy profile for the one or more zones of the building (step 5008), according to some embodiments. In some embodiments, step 5008 is the same as or similar to step 4906 of process 4900 but is performed for a future time period. In some embodiments, step 5008 includes generating the occupancy profile for the zones based on ASHRAE standards and zone scheduling.

Process 5000 includes performing a simulation of the mold model for a future time period (step 5010), according to some embodiments. In some embodiments, step 5010 is the same as or similar to step 4908 of process 4900 but performed for the future time period (e.g., a future time horizon). In some embodiments, the simulation performed in step 5010 is performed at a different periodicity (e.g., the duration of the future time period, and the time steps of the simulation performed at step 5010 are different than the duration of the previous time period and the time steps of the simulation performed at step 4908 of process 4900). In some embodiments, the simulation is performed to determine energy consumption or energy cost and corresponding mold risks for different decision variables (e.g., humidity setpoint, supply temperature setpoint, etc.).

Process 5000 includes determining one or more mold metrics based on outputs of the simulation (step 5012) and performing a Pareto optimization to determine different Pareto optimal operating points (step 5014), according to some embodiments. In some embodiments, step 5012 is the same as or similar to the step 4910 of process 4900. In some embodiments, step 5014 includes performing steps 4806-4810 of process 4800 as described in greater detail above with reference to FIG. 48. For example, the simulation performed in step 5010 can result in different combinations of energy cost or energy consumption and mold risk for different operating parameters (e.g., for different values of decision variables). The Pareto optimization is performed to determine different Pareto optimal points that may be provided in step 5016 as suggested operating points or as advised operating points, according to some embodiments. In some embodiments, step 5014 includes determining which of the Pareto optimal operating points result in lowest energy cost or energy consumption, lowest mold risk, and an equal priority between energy cost or energy consumption and mold risk.

Process 5000 includes operating a display to provide the Pareto optimal points (or a subset thereof) to a user as advisory data (step 5016), according to some embodiments. In some embodiments, step 5016 includes providing the Pareto optimal points associated with lowest or minimal energy cost or energy consumption, lowest or minimal mold risk, and an equal priority Pareto optimal point between energy consumption or cost and mold risk. In some embodiments, the Pareto optimal points are provided as advisory or suggested operating conditions for the future time period.

Combined Analysis and Advisory Process

Referring particularly to FIG. 51 a process 5100 for performing both an infection metric analysis for a previous time period and a Pareto optimization for a future time period is shown, according to some embodiments. In some embodiments, process 5100 includes steps 5102-5110 and is performed by the controller 1510. In some embodiments, process 5100 illustrates the simultaneous performance of process 4900 and process 5000.

Process 5100 includes performing a mold risk and energy cost or consumption analysis of an HVAC system for a previous time period to determine analysis data (step 5102) and performing a mold risk and energy cost Pareto optimization of the HVAC system for a future time period to determine advisory data (step 5104), according to some embodiments. In some embodiments, performing step 5102 includes performing process 4900. In some embodiments, performing step 5104 includes performing process 5000. In some embodiments, steps 5102 and 5104 are performed at least partially simultaneously with each other.

Process 5100 includes operating a display to provide both the analysis data and the advisory data to a user (step 5106), according to some embodiments. In some embodiments, step 5106 includes performing steps 4912 and 5016 of processes 4900 and 5000, respectively. In some embodiments, step 5106 is performed by the display device 822. Providing both the analysis data and the advisory data can facilitate a temporal bi-directional informing of the user regarding past operation of the HVAC system (e.g., the HVAC system 700) and suggested future suggestions or advisory control decisions to achieve desired energy costs and mold reduction or acceptable mold risk levels.

Process 5100 includes automatically selecting control decisions or receiving a user input of a selected control decision (step 5108) and operating equipment of an HVAC system according to the control decisions (step 5110), according to some embodiments. In some embodiments, steps 5108 and 5110 are the same as or similar to steps 4814 and 4816 of process 4800.

User Interfaces

Referring particularly to FIGS. 52 and 53, different user interfaces 5200 and 5300 display the various outputs of the controller 1510 (e.g., the display data), according to some embodiments. In some embodiments, the user interface 5200 and 5300 are displayed on the display device 822 and presented to a user or a building administrator. The user interface 5200 shows display of the analysis mode outputs 1634 and the user interface 5300 shows display of the advisory mode outputs 1640. In various embodiments, and the user interface 3500 of FIG. 35 shows checklists for implementing one of the various options of the advisory mode outputs 1640 and may show checklists relating to mold risk and advisory mode outputs relating to mold risk.

Referring particularly to FIG. 52, the user interface 5200 includes a mold risk score icon 5202, and an indoor air quality score icon 5204. In some embodiments, the mold risk score 5202 is a scaled version of the mold risk for the previous time period. In some embodiments, the mold risk score is a weighted average, a time-series average, etc., of the mold risks of zones of the building 10 over the previous time period. In some embodiments, the indoor air quality score icon 5204 displays a similarly aggregated, average, etc., score of the indoor air quality of the zones of the building 10 over the previous time period. In some embodiments, the values of the mold risk score and the indoor air quality score are normalized values from ranging from 0 or 1 to 100. In some embodiments, the indoor air quality score icon 1604 and the mold risk score icon 5202 are graphical icons that display a bar or a circle chart and a textual or numeric value of the indoor air quality score and the mold risk score for the zones of the building 10 over the previous time period. In some embodiments, the indoor air quality score icon 1604 and the mold risk score icon 5202 are color-coded based on their values. For example, if the indoor air quality score is between a first or normal range, then the color of the indoor air quality score icon 1604 may be yellow, according to some embodiments. In some embodiments, if the indoor air quality score is between a second range or less than a lower value of the first or normal range, this may indicate that the indoor air quality score is poor and the color of the indoor air quality score icon 5204 may be red. In some embodiments, if the indoor air quality score is between a third range or greater than a higher value of the first or normal range, this may indicate that the indoor air quality score is good and the color of the indoor air quality score 5204 may be green.

Referring still to FIG. 52, the user interface 5200 includes a list 5206 of one or more high mold risk alerts, according to some embodiments. In some embodiments, the list 5206 includes different items 5208, each item corresponding to a different zone of the building 10 and an individual mold risk score associated with the different zones. In some embodiments, the items 5208 of the list 5206 are zone-specific and are determined based on the mold risk score for each of the zones of the building 10. For example, if one of the zones has an associated mold risk score that is above a threshold amount, then that zone may be added with the associated mold risk score to the list 5206 as one of the items 5208.

Referring still to FIG. 52, the user interface 5200 also includes a list 5210 of one or more low indoor air quality alerts, according to some embodiments. In some embodiments, the list 5210 includes different items 5212, each item corresponding to a different zone of the building 10 and an individual indoor air quality associated with the different zones. In some embodiments, the items 5212 of the list 5210 are zone-specific and are determined based on the indoor air quality for each of the zones of the building 10. For example, if one of the zones has an associated indoor air quality that is below a threshold amount, then that zone may be added with the associated indoor air quality to the list 5210 as one of the items 5212.

Referring still to FIG. 52, the user interface 5200 includes a list 5214 of each of the zones of the building 10 (e.g., organized by zone type, floor of the building 10, etc.). Each of the items of the list 5214 includes an indication of the zone or floor, an associated mold risk score for the zone or floor, a number of mold risk alerts for the zone or floor, an indoor air quality score for the zone or floor, an indoor air quality trend (e.g., a 30 day trend), a number of indoor air quality alerts, and/or an energy spend versus budget (e.g., for 30 days).

Referring to FIG. 53, the user interface 5300 includes different widgets 5302-5308 indicating the results of the Pareto optimization (e.g., the advisory mode outputs 1640) as described in greater detail above with reference to FIGS. 15, 16 and 44-51. Specifically, the user interface 5300 includes a current operational state widget 5302 illustrating current energy costs and associated mold risk score with additional air flow, comfort, UV disinfection, and filtration specifics, according to some embodiments. The user interface 5300 includes a widget 5304 illustrating a first option, namely, the Pareto optimal result for optimizing the mold risk score (e.g., minimizing mold risk or mold probability such as the minimal mold risk point 4418), a widget 5306 illustrating a second option, namely, the Pareto optimal result for equal priority between mold risk and energy consumption or cost (e.g., the equal priority point 4420), and a widget 5308 illustrating a third option, namely, the Pareto optimal result for operating with minimum energy cost (e.g., the minimum energy consumption point 4422). Each of the widgets 5304-5308 include graphical and/or textual information regarding a corresponding mold risk score, an energy cost per a time period (e.g., a monthly time period), air flow parameters, required operational adjustments, optional design adjustments, etc., for each of the options. In some embodiments, the user or building administrator may select one of the options by selecting one of the widgets 5304-5308.

Referring now to FIG. 54, a process 5400 is shown for determining when to perform a two-control objective Pareto optimization or a three-control objective Pareto optimization. In various embodiments, monitoring infection risk of a building and air that occupants breathe in the building may be applicable when the building is occupied. However, mold may be of concern regardless of building occupancy because mold can grow at any time. Therefore, the controller 1510 may perform a three-control objective Pareto optimization when a building is occupied, to consider energy cost or consumption, mold risk, and infection risk. The controller 1510 may perform a two-control objective Pareto optimization when a building is not occupied, to consider energy cost or consumption and mold risk. The two- or three-control objective Pareto optimization to make trade-offs between control objectives based on user preferences or tolerances, even if the user preferences do not align with a Pareto-optimal solution. For example, if the presence of mold is unacceptable to a user, the user may adjust the control objectives such that mold risk is at a lowest possible value, regardless of a corresponding energy cost or consumption. A lower or lowest mold risk value may correspond to an increased energy cost or consumption and/or other control parameter (e.g., an increased infection risk, etc.)

At step 5410, an occupancy schedule of the building is received. The occupancy schedule may be received by the controller 1510 or another component of the building management system. The schedule may indicate when the building, a floor of the building, a zone of the building, etc. will be occupied and when the building will not be occupied. The schedule may also include a quantity of occupants at various times. The schedule may be an actual schedule (e.g., a timetable indicating times during each day of the week that the building is occupied). The schedule may be a predicted number of occupants of the building on various days of the week and at various times.

At step 5420, the controller 1510 may perform a two-control objective Pareto optimization responsive to determining that the building is not occupied. In various embodiments, the Pareto optimizer 1636 or the Pareto optimizer 1512 may perform the Pareto optimization as described above. The two-control objective Pareto optimization may consider energy cost or consumption and mold risk as control objectives. In various embodiments, the two-control objective Pareto optimization may be performed prior to a period of no occupancy for the building. The results of the two-control objective Pareto optimization may be stored in memory and implemented during future instances of periods of no occupancy in the building.

At step 5430, the controller 1510 may perform a three-control objective Pareto optimization responsive to determining that the building is occupied. In various embodiments, the Pareto optimizer 1636 or the Pareto optimizer 1512 may perform the Pareto optimization as described above. The the-control objective Pareto optimization may consider energy cost or consumption, mold risk, and infection risk as control objectives. In various embodiments, the three-control objective Pareto optimization may be performed prior to a period of building occupancy. The three-control objective Pareto optimization may be performed with steps similar to those of process 4100 of FIG. 41. The results of the three-control objective Pareto optimization may be stored in memory and implemented during future instances of periods of occupancy in the building.

Configuration of Exemplary Embodiments

Although the figures show a specific order of method steps, the order of the steps may differ from what is depicted. Also two or more steps can be performed concurrently or with partial concurrence. Such variation will depend on the software and hardware systems chosen and on designer choice. All such variations are within the scope of the disclosure. Likewise, software implementations could be accomplished with standard programming techniques with rule based logic and other logic to accomplish the various connection steps, calculation steps, processing steps, comparison steps, and decision steps.

The construction and arrangement of the systems and methods as shown in the various exemplary embodiments are illustrative only. Although only a few embodiments have been described in detail in this disclosure, many modifications are possible (e.g., variations in sizes, dimensions, structures, shapes and proportions of the various elements, values of parameters, mounting arrangements, use of materials, colors, orientations, etc.). For example, the position of elements can be reversed or otherwise varied and the nature or number of discrete elements or positions can be altered or varied. Accordingly, all such modifications are intended to be included within the scope of the present disclosure. The order or sequence of any process or method steps can be varied or re-sequenced according to alternative embodiments. Other substitutions, modifications, changes, and omissions can be made in the design, operating conditions and arrangement of the exemplary embodiments without departing from the scope of the present disclosure.

As used herein, the term “circuit” may include hardware structured to execute the functions described herein. In some embodiments, each respective “circuit” may include machine-readable media for configuring the hardware to execute the functions described herein. The circuit may be embodied as one or more circuitry components including, but not limited to, processing circuitry, network interfaces, peripheral devices, input devices, output devices, sensors, etc. In some embodiments, a circuit may take the form of one or more analog circuits, electronic circuits (e.g., integrated circuits (IC), discrete circuits, system on a chip (SOCs) circuits, etc.), telecommunication circuits, hybrid circuits, and any other type of “circuit.” In this regard, the “circuit” may include any type of component for accomplishing or facilitating achievement of the operations described herein. For example, a circuit as described herein may include one or more transistors, logic gates (e.g., NAND, AND, NOR, OR, XOR, NOT, XNOR, etc.), resistors, multiplexers, registers, capacitors, inductors, diodes, wiring, and so on).

The “circuit” may also include one or more processors communicably coupled to one or more memory or memory devices. In this regard, the one or more processors may execute instructions stored in the memory or may execute instructions otherwise accessible to the one or more processors. In some embodiments, the one or more processors may be embodied in various ways. The one or more processors may be constructed in a manner sufficient to perform at least the operations described herein. In some embodiments, the one or more processors may be shared by multiple circuits (e.g., circuit A and circuit B may comprise or otherwise share the same processor which, in some example embodiments, may execute instructions stored, or otherwise accessed, via different areas of memory). Alternatively or additionally, the one or more processors may be structured to perform or otherwise execute certain operations independent of one or more co-processors. In other example embodiments, two or more processors may be coupled via a bus to enable independent, parallel, pipelined, or multi-threaded instruction execution. Each processor may be implemented as one or more general-purpose processors, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), digital signal processors (DSPs), or other suitable electronic data processing components structured to execute instructions provided by memory. The one or more processors may take the form of a single core processor, multi-core processor (e.g., a dual core processor, triple core processor, quad core processor, etc.), microprocessor, etc. In some embodiments, the one or more processors may be external to the apparatus, for example the one or more processors may be a remote processor (e.g., a cloud based processor). Alternatively or additionally, the one or more processors may be internal and/or local to the apparatus. In this regard, a given circuit or components thereof may be disposed locally (e.g., as part of a local server, a local computing system, etc.) or remotely (e.g., as part of a remote server such as a cloud based server). To that end, a “circuit” as described herein may include components that are distributed across one or more locations.

The present disclosure contemplates methods, systems and program products on any machine-readable media for accomplishing various operations. The embodiments of the present disclosure can be implemented using existing computer processors, or by a special purpose computer processor for an appropriate system, incorporated for this or another purpose, or by a hardwired system. Embodiments within the scope of the present disclosure include program products comprising machine-readable media for carrying or having machine-executable instructions or data structures stored thereon. Such machine-readable media can be any available media that can be accessed by a general purpose or special purpose computer or other machine with a processor. By way of example, such machine-readable media can comprise RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to carry or store desired program code in the form of machine-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer or other machine with a processor. Combinations of the above are also included within the scope of machine-readable media. Machine-executable instructions include, for example, instructions and data which cause a general purpose computer, special purpose computer, or special purpose processing machines to perform a certain function or group of functions.

	Number	Date	Country
	63621946	Jan 2024	US
	63621947	Jan 2024	US

BUILDING SYSTEM WITH AIR QUALITY CONTROL BASED ON ENERGY, MOLD RISK, AND PRODUCTIVITY MODELING

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

International Classifications

Abstract

Description

Claims

CROSS-REFERENCE TO RELATED PATENT APPLICATIONS

Provisional Applications (2)