Patent Application
20210315170
The disclosure relates generally to the field of controlled-environment agriculture, and in particular to controlling latent and sensible loads to optimize the ratio of yielded product for each unit of energy consumed (e.g., as represented by yield (e.g., harvest weight) per kw-hour of energy) in a controlled growing environment.
During the twentieth century, agriculture slowly began to evolve from a conservative industry to a fast-moving high-tech industry in order to keep up with world food shortages, climate change and societal changes. Farming began to move away from manually-implemented agriculture techniques toward computer-implemented technologies. In the past, and in many cases still today, farmers only had one growing season to produce the crops that would determine their revenue and food production for the entire year. However, this is changing. With indoor growing as an option and with better access to data processing technologies, among other advanced techniques, the science of agriculture has become more agile. It is adapting and learning as new data is collected and insights are generated.
Advancements in technology are making it feasible to control the effects of nature with the advent of “controlled indoor agriculture” or “controlled-environment agriculture.” Improved efficiencies in space utilization, lighting, and a better understanding of hydroponics, aeroponics, crop cycles, and advancements in environmental control systems have allowed humans to better recreate environments conducive for agriculture crop growth with the goals of greater yields per square foot, better nutrition and lower cost.
Even with these advances, energy consumption and its accompanying cost remain some of the biggest challenges facing indoor farms. Sunlight is free, but indoor lighting, cooling, dehumidification and other climate control methods for large grow rooms can be expensive. Others have tried to reduce energy consumption by employing more energy-efficient lighting and smart energy management approaches, among other means.
Embodiments of the disclosure optimize yield and energy consumption to maximize efficiency. This is done, in part, by taking advantage of the net evapotranspirative cooling effect of dense plant growth environments to reduce the need for cooling, and by enabling the recycling of waste heat from sources such as grow lights. This shift in the energy balance allows for greater efficiency of energy per kg of yield compared to conventional indoor farms. Additionally, predictive modeling of both yield and energy consumption allows for optimization of both design and system operation in terms of cost and yield, under particular conditions.
Plants convert water and carbon dioxide into energy through photosynthesis. Photosynthesis and many other plant processes also require micronutrients which are transported through the plant tissue by water. Because the plant can only store a small portion of the water required to transport sufficient nutrients, plants must release water to their surroundings. One way in which a plant releases this water is through transpiration. During transpiration, water evaporates from small openings in the leaves called stomata, and then diffuses to the surrounding air. Diffusion is driven by a difference in vapor pressure between the leaf surface and the surrounding air. This pressure difference is refered to as the vapor pressure deficit (VPD), measured in kPa, and is one of the most critical parameters used to estimate transpiration rate. Other factors such as light intensity, CO2 concentration, nutrient concentration and others can also affect transpiration rate.
If the vapor pressure inside a plant is high (due to, e.g., hydration), water vapor will exit the stomata if the outside air has a lower vapor pressure. As leaf temperature increases at a given air temperature and relative humidity, transpiration rate via the stomata increases. Water is effectively evaporating within or at the surface of the plants' leaves, transferring energy while the phase changes from liquid to vapor. This phenomenon is known as “transpirative cooling.” Evaporation from wet surfaces within the indoor agricultural environment also contributes to cooling. The term “evapotranspirative” cooling refers to the cooling effect due to the phase change of liquid water to water vapor in the form of evaporation off of wet surfaces and transpiration from plants. For convenience “evapotranspiration” will be referred to herein interchangeably as “EVT.”
Evapotranspiration rate is one of the most important parameters to understand for optimization of the system because it both affects mechanical efficiency and is closely related to growth rate and thus yield. Embodiments of the disclosure employ an empirically based model of growth rate and transpiration rate as a function of various environmental parameters paired with a physics based model of mechanical efficiency as a function of environmental parameters and transpiration rate to optimize yield vs. energy consumption.
Conventional approaches focus on cooling the growing environment to offset the sensible heat created by heat sources such as lighting. One approach adopted by Desert Aire accounts for the sensible cooling effect created by evapotranspiration. See “Grow Room Load Determination,” Application Note 25, Desert Aire, March 2016 (“Desert Aire Note”). The Desert Aire approach takes EVT cooling as a given variable that changes as the plants grow, and adjusts HVAC equipment accordingly to account for EVT cooling. The Desert Aire Note discusses modulating sensible heat ratios in response to the grow room environment. The Desert Aire Note provides examples concerning the type of energy (e.g., sensible or latent) that must be removed from the room, demonstrating its focus on cooling the environment.
Like Desert Aire, other conventional HVAC systems in indoor farms treat transpiration as an input, requiring conditioning of other factors to maintain a healthy grow room environment. This leads to equipment that is sized for maximum and worst-case load conditions, which requires flexibility to handle varying loads (such as modulating hot gas reheat in the Desert Aire unit). These requirements make traditional controlled-environment HVAC in indoor farms very capital intensive.
There are two primary loads to the grow space: sensible (energy in the form of heat), e.g., from lights, and latent (energy in the form of water) from evaporation off of wet surfaces and transpiration of the plants.
The heating load is the amount of heat energy that must be added to the grow room space to maintain a target air temperature. This is often considered the heat loss calculation as it calculates the level of heat that must be added to offset the loss through, e.g., the building's ceilings and walls. This heat loss is especially significant in the winter when the outside temperature is much lower than the inside temperature.
The cooling load is the amount of heat energy that needs to be removed to attain the target indoor temperature. This is the heat gain calculation with heat from lighting systems being the most significant portion of this load, plus some contributions from solar gain and the building envelope.
Embodiments of the disclosure provide a control system for controlling one or more environmental conditions to control latent and sensible loads in a controlled agricultural environment (e.g., a grow space such as a fully or partially enclosed grow chamber). The control system may control a sensible load in the grow chamber to provide heat to at least partially offset the latent load. According to embodiments of the disclosure, the density of plant receptacles in the grow chamber is such that, when plants are held in the plant receptacles, evapotranspiration contributes to the latent load so that the latent load exceeds a sensible load, resulting in a net evapotranspirative cooling effect. This shift in the energy balance allows for greater energy savings, as compared with conventional indoor farms that focus on removing heat from the growth environment.
According to embodiments of the disclosure, the sensible load, the latent load or both are controlled to achieve at least one desired condition. Controlling the one or more environmental conditions may comprise setting the one or more environmental conditions to one or more environmental setpoints that are determined using a physics based model. The one or more environmental setpoints may additionally be determined using an empirically based model.
The desired condition may be a desired ambient temperature or other setpoint, a desired energy consumption, a desired productivity of plant matter, or a desired ratio of plant product yield to energy use. The control system may employ machine learning to achieve the desired condition. Embodiments of the disclosure may increase or decrease transpiration to, respectively, decrease or increase a sensible cooling load.
The control system may control evapotranspiration to regulate the latent load by controlling at least one of irrigation, CO2 concentration, or the supply of chemicals (e.g., hormones) that regulate transpiration. The control system may control evapotranspiration by controlling at least one of temperature, relative humidity, vapor pressure deficit, light intensity, light wavelength (including electromagnetic radiation wavelength in the visible range and in the non-visible range, such as ultraviolet and infrared), light duration, or air velocity. The control system may control evapotranspiration by varying lighting based on daytime or nighttime condition. Any factor subject to such control, such as those enumerated above (e.g., irrigation, temperature, lighting) or any other that is characteristic of the grow room environment may be referred to herein as an “environmental condition.”
The control system may control the latent load within a control volume to achieve a desired condition within the control volume. The control volume may include lighting and the plant receptacles. The control system may receive sensor signals representing characteristics of at least one plant or plant environment in the control volume or signals that can be extrapolated to estimate the plant characteristics or environment. Data from the sensors may be used to construct an empirically based model with outputs of transpiration rate and yield.
The control system may include a dehumidification subsystem for dehumidifying the grow chamber and sensible conditioning equipment for sensibly heating and cooling the grow chamber. The control system may employ waste heat to warm the air or evaporate moisture in the grow chamber. Lighting in the agricultural environment may provide the waste heat. The control system may control cooling of the lighting to control the waste heat. More particularly, the control system may include fluid-cooled lighting in the grow chamber as well as a dehumidifier and a heat exchanger. The heat exchanger may employ waste heat from the lighting in the grow chamber or from components within an air and fluid conditioning system to heat air output from the dehumidifier and provide the heated air to the agricultural environment.
The present description is made with reference to the accompanying drawings, in which various example embodiments are shown. However, many different example embodiments may be used, and thus the description should not be construed as limited to the example embodiments set forth herein. Rather, these example embodiments are provided so that this disclosure will be thorough and complete. Various modifications to the exemplary embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the disclosure. Thus, this disclosure is not intended to be limited to the disclosed embodiments, but is to be accorded the widest scope consistent with the claims and the principles and features disclosed herein.
System Mechanics
This section first discusses loads within a grow space and then discusses methodologies used to handle those loads to maintain the environment at desired conditions, according to embodiments of the disclosure. Embodiments of the disclosure treat the problem differently than conventional systems by leveraging the transient nature of the loads implicit to the abundance of biological material within the conditioned space. According to embodiments of the disclosure, because the implication of various operational strategies on growth rate and energy consumption is non-linear, a unique empirical model is utilized along with a physics based mechanical efficiency model to optimize the system.
Loads
In standard HVAC design, the peak load defines the maximum amount of heating, cooling and dehumidification that is required to maintain the conditions within a space to achieve desired environmental setpoints (e.g., temperature, relative humidity) under a worst case scenario. The term “load” is used herein rather than “peak load” because it is used to describe the amount of heating, cooling, and dehumidification that is required at a given time and the load may vary significantly based on changes in system operation. When computing the load there are two primary components that are used to describe the amount of heating, cooling and dehumidification. These components are sensible and latent loads.
The sensible load refers to the amount of heat energy that needs to be added or removed from a space in order to maintain a desired setpoint temperature. Examples of factors that contribute to the sensible load are heat from lights, heat from mechanical equipment, and heat from infiltration of air from the surroundings.
The latent load refers to the amount of energy in the form of water that must be removed or added to a space in order to maintain a desired relative or absolute humidity setpoint. An example of a latent load is evaporation from wet surfaces and transpiration of the plants. Here, a control volume 100, as shown in
Associated with each unit surface area of leaf is a transpiration rate. Transpiration is the process by which plants release water to the air through stomates in the leaf surface. As the liquid water within the plant is converted to water vapor, it absorbs a large amount of energy. This has a cooling effect on the air within the space. Transpiration rate as well as surface area will vary based on age of the plant, variety, treatment, and environmental conditions.
Depending on the magnitude of each component vector, the slope of the resultant (solid) line will change.
Systems according to embodiments of the disclosure maintain a plant density such that the result of sensible and latent loads on the room has a cooling or heating effect that may be premeditated by system design and predictive modeling. The density may be expressed as the leaf surface area per unit volume (which may be expressed in cm2/m3 or m2/m3, for example). The density of the plants is arranged such that within a control volume, the latent-sensible ratio
(T*A*L+E*L)/S is optimized for energy efficient operation,
where:
T=Transpiration rate for a unit area of leaf (kg/(m2*hr))
T is a function of crop type, time since seeding, environmental conditions, nutrient treatment, light intensity, and other known factors, and may be predicted by a predictive model, such as the empirically based model described elsewhere herein
A=Leaf area within the control volume (m2)
L=Latent heat of vaporization of water (kw/kg)
E=Rate of evaporation within the control volume (kg/hr)
S=Rate of sensible heat transfer to the space (kw/hr)
The resulting parameter is dimensionless for the one-hour period.
According to embodiments of the disclosure, a controller 620 (described below) may optimize the latent-sensible ratio through a variety of optimization schemes, including machine learning, using, as inputs/features, crop type, time since seeding, environmental conditions (e.g., temperature, relative humidity), nutrient treatment, light intensity, or other known factors, as well as those represented by the variables A, L, E, and S above, with the objective of improving plant yield per unit energy consumption, flavor or other parameters, alone or in combination.
As an example for a one cubic meter control volume, the controller 620 (described below) may achieve evapotranspirative cooling by optimizing parameters to take on the following values:
T=0.0088 kg/(m2*hr) and A=3.42 (m2) for:
L=0.67
7 kw/hr
E=0.0044 (kg/hr)
S=0.195 (kw/hr)
Because transpiration rate is variable, a controller (such as controller 620) controlling appropriate equipment to condition the environment of the grow space may manipulate the direction of the resultant arrow on the psychrometric chart, according to embodiments of the disclosure. By taking advantage of evapotranspirative cooling in the grow space, embodiments of the disclosure may employ the sensible heat added to the space in a beneficial, rather than detrimental, manner with respect to operational efficiency of the system.
In most conventional systems, the resultant arrow of sensible and latent loads is tilted towards the right as in
In the example of
Conditioning System
According to embodiments of the disclosure, the conditioning system is designed to respond to and handle loads in the grow space. Embodiments of the disclosure recognize and make use of the flexibility of the latent:sensible load ratios. Referring to
Conventional HVAC systems treat sensible and latent loads as a fixed input to sizing and energy use calculations. In many ways this is a holdover from traditional building design where the controlled environment is designed for human comfort. In reality, plants are much more resilient and flexible than human comfort standards. The traditional methodology of defining worst case loads and sizing for such loads leads to oversized, expensive equipment designed for a very specific operational condition. By recognizing that the loads can be manipulated, there is a large, unrecognized potential to cut capital as well as operational costs.
The level of control and flexibility built into the system can be optimized for plant health and efficiency. A controlled vapor-pressure-deficit, among other variables, can maintain nutrient uptake and effectively grow the same volume of crop in a small volume of air with lower energy consumption.
According to embodiments of the disclosure, the controller 620 utilizes two model components to determine environmental conditions that optimize the system (e.g., grow space 600, including lighting 608, CO2 supply 611, irrigation system 609, and conditioning system 602) for yield vs. energy consumption. For the sake of convenience, this disclosure will refer to the controller 620 as performing this function, but those skilled in the art will recognize that in other embodiments one or more other computing devices may perform the same function, and provide their determination to the controller 620 to control the conditioning system 602, among other things. The two model components share many inputs, including, e.g., temperature, relative humidity, light intensity, light spectrum, CO2 concentration, and mechanical ventilation rate.
The first component is an empirically based model for predicting yield and transpiration rate as a function of many parameters. The output transpiration rate is used as an input to the second model component, which is a physics based model used to predict energy consumption of the system.
According to embodiments of the disclosure, the first, empiricially based model component uses data collected via a sensor network to establish numerical relationships between a fixed number of environmental and plant parameters and yield and transpiration rate. Environmental and plant parameters may include temperature, relative humidity, light intensity and spectrum, CO2 concentration, plant variety, plant age, nutrient concentrations, and others. In general, these numerical relationships are determined by systematically varying a single parameter or multiple parameters while observing the effect on the output parameters. According to embodiments of the disclosure, the controller 620 applies to this empirical data predictive techniques, such as a multivariate regression model or machine learning, to relate all the parameters to yield and transpiration rate.
According to embodiments of the disclosure, the controller 620 uses the physics-based model component to predict the efficiency (amount of work performed/energy consumed) of the mechanical system (e.g., conditioning system 602, lighting 608, CO2 supply 611, irrigation system 609) given various operating conditions such as temperature, relative humidity, VPD, CO2 concentration, mechanical ventilation rate, light intensity and transpiration rate. From efficiency, the total energy consumed is determined for use in determining the desired yield-to-energy (Y/E) ratio.
According to embodiments of the disclosure, the physics based model determines total system efficiency (amount of work performed/energy consumption) using the predicted transpiration rate, conventional psychrometric equations as defined in ASHRAE and as reflected in psychrometric charts such as those in
According to embodiments of the disclosure, the physics based model employs the Heat Balance Method such as that outlined in J. Spitler, Load Calculation Applications Manual, second edition, I-P edition, ASHRAE (2014) (the “ASHRAE manual”), incorporated by reference in its entirety herein. This manual is an in-depth, application-oriented reference that provides a clear understanding of state of the art heating and cooling load calculation methods plus the tools and resources needed to impement them in practice. Although one skilled in the art would know how to compute system efficiency with the above inputs, embodiments of the disclosure also incorporate predicted transpiration rate into the efficiency computation as at least part of the latent load.
The physics based model employing the Heat Balance Method, as depicted in the simplified diagram of
In this model, the controller 620 identifies and sums the primary energy and mass (e.g., water) sources and sinks, and uses the sum to estimate the amount of work (e.g., the total of sensible and latent loads) that the conditioning system 602 must do to maintain the setpoint conditions (i.e., the amount of work performed in the numerator of the total system efficiency computed above). In
T
return_air
=T
supply
a+total_load =Tsupply_air+total_heat/(mass_flow_rate*heat_capacity)
Values derived from the the Heat Balance Method provides conditions to use with the mechanical equipment operating curves mentioned above in order to determine system efficiency. Using the system efficiency, energy consumption at any given point (instantaneous energy) during a grow cycle (length of time it takes a plant to mature to harvest weight) may be computed for given environmental conditions and estimated transpiration rate for the age of the subject crop in the grow space, according to embodiments of the disclosure. According to embodiments of the disclosure, the instantaneous energy is integrated over the full length of the grow cycle to determine total energy consumption (the variable E used in computing the ratio of yield to energy consumption). The unique aspect of this model is that transpiration rate is treated as a controllable variable that is determined based on the empirical model. The physics based model component allows for a variable transpiration input which can be used for yield and energy optimization.
According to embodiments of the disclosure, the controller 620 uses the two model components to estimate the impact on energy and yield of various operational strategies.
Block 904 represents the empirical model with outputs of yield and transpiration rate. Block 906 represents the physics based mechanical system model, which receives as inputs transpiration rate from block 904 and the parameters from block 902. According to embodiments of the disclosure, the controller 620 employs the physics based mechanical system model (906) to predict energy consumption, as described elsewhere herein. The controller 620 uses the predicted energy consumption along with the yield predicted by the empirical model component 904 to determine the desired condition of the ratio of yield/energy consumption (e.g, in kg/kwh) (908).
Determining temperature and relative humidity setpoints is one example of how the two-component model can be used for optimization. In an example, consider a two dimensional parametric analysis of various temperature and relative humidity setpoints. In a prophetic example, the result of this analysis may be the yield divided by the energy (kg/kw-hour) over the course of a 10 day grow cycle where temperature is varied by increments of 1 degree from 18 to 40 degrees C. and relative humidity is varied by 1% from 55% to 85%. In this example, the controller 620 may determine that the kg/kw output is maximized at 22 degrees and 80% relative humidity.
This type of analysis is valuable because it allows operators to quantify the tradeoffs of various scenarios which are not immediately apparent. For example, in the above scenario where temperature is varied, the mechanical equipment may be more efficient at a higher temperature whereas yield is higher at a lower temperature. The controller 620 may also determine that there is a higher transpiration rate at higher temperatures and thus an increased dehumidification load which leads to increased energy consumption. From a purely mechanical standpoint, one would expect energy consumption to be reduced when operating at a higher temperature when cooling. For example, refrigeration equipment requires more energy per unit of work output at colder temperatures because there is less heat per unit of air, but when coupled with the empirical model, the model predicts that the system is optimized at a temperature setpoint where the mechanical equipment is actually running less efficiently.
In this example, using the empirical model 904, the controller 620 determines that Region B represents the region in which yield is greater than 90% of the maximum yield (maximum represented by point E) under the same conditions. As noted, points D and E represent minimum energy consumption and maximum yield. However, because those points do not overlap, the ideal, maximum ratio of yield/energy cannot be achieved, in this example. Therefore, the inventors determined to establish regions around the ideal setpoints to generate an overlap region in which the optimum Y/E ratio could be found within real-world constraints.
Note that the system operator may vary the allowable percentages above based upon the cost of energy and the profit from yield (e.g., harvest weight) as dictated by market conditions. An objective is the highest profit from yield per unit cost of energy. As an example, if energy costs were to increase, the allowable percentage from minimum cost would decrease because the cost factor would be more critical.
Using the boundary values of Regions A and B, the controller 620 identifies that the resultant overlapping window C is the region in which the system can operate and achieve a desired, optimum yield to energy ratio within real-world constraints. In this example, Region C represents acceptable temperature and relative humidity environmental conditions that are predicted to achieve energy consumption of less than 110% of the minimum energy use and a yield greater than 90% of the maximum yield in the example above, i.e., each environmental condition falls between lower and upper thresholds of acceptable environmental parameter values that achieve those objectives.
According to embodiments of the disclosure, the controller 620 selects particular environmental conditions from within the range of acceptable setpoints (as represented by the overlap region C in this example) to determine target setpoints to be applied through the environmental conditioning system 602 to the grow space 600. It is desired that both supply and return conditions fall within Region C. Referring to
To operate most efficiently in this example, it is desired that both supply and return conditions fall within Region C. Maximizing the magnitude (length) of load line 1- 2 minimizes the rate of air exchange and energy demand of the supply fan necessary to maintain supply and return within Region C. Thus, according to embodiments of the disclosure, the controller 620 selects the supply and return temperature and relative humidity setpoints as the endpoints of the longest load line 1-2 (which is at an angle defined by the ratio of the latent load to the sensible load) that fits within Region C. The longest line represents taking the greatest advantage of the evapotranspirative cooling effect (because its endpoint is at the coolest point in the x direction within region C) and specifies operating conditions that are predicted to hit the desired ratio of yield to energy consumption.
Benefits of HVAC systems designed according to embodiments of the disclosure include:
Control of Sensible Loads
Embodiments of the disclosure adjust the sensible load by altering light intensity or cooling the lights (see, e.g., U.S. Patent Application Pub. No. US 2017/0146226, filed Nov. 15, 2016, assigned to the assignee of the present invention and incorporated by reference herein in its entirety). For example, the conditioning system may adjust the flow rate through water-cooled light fixtures to affect whether heat generated at the fixture is rejected to the grow space air or whether it is removed via the water.
According to embodiments of the disclosure, the controller 620 may control the appropriate equipment (e.g., in environmental conditioning system 602) to regulate the sensible load in the grow space by controlling variables such as:
Control of Latent Loads
According to embodiments of the disclosure, the controller 620 may control the appropriate equipment (e.g., in environmental conditioning system 602) to regulate the evapotranspiration rate in the grow space by controlling variables such as:
According to embodiments of the disclosure, the environmental control system equipment is specified to remove an equivalent or larger amount of water than occurs during worst case transpiration and evaporation scenarios for both night and day. Worst case transpiration may be defined by the amount of water that needs to be removed from the air when plants are at the most dense phase of their life cycle (usually just before harvest when they have the most leaf surface area per unit volume), are at peak transpiration of their growth cycle (transpiring the most due to plant processes), and conditions are such that the vapor pressure deficit and other contributing factors promote evapotranspiration. The combination of these conditions results in the highest rate of evapotranspiration expected in the system.
Control of sensible and latent loads, and optimizing systems for them, is an advantage of embodiments of the disclosure.
An irrigation pump 609 circulates water and nutrients through the plant receptacle 604. Carbon dioxide supply equipment 611 provides carbon dioxide to the plants. The irrigation pump 609 and carbon dioxide supply equipment 611 may be considered as part of the conditioning system 602, according to embodiments of the disclosure.
According to embodiments of the disclosure, the conditioning system 602 includes a dehumidifier 610, a fluid (e.g., water) conditioning system 612, and a heating coil 614 in heat exchanger 615. (The lighting 608, heating coil 614 and other heating and cooling elements that sensibly heat or cool the grow space may be considered to be sensible conditioning equipment.) The dehumidifier 610 receives from the grow space 600 return air A, having a temperature and relative humidity that depends on the plant transpiration rate and rate of evaporation from wet surfaces in the environment 600. The conditioning system 602 provides supply air B, having a temperature and relative humidity that is controlled to meet set points for desired operating conditions of the plants in the environment 600.
The fluid conditioning system 612 receives return fluid C from the fluid-cooled light fixture 608. The fluid conditioning system 612 can control the fluid temperature by varying the fluid flow rate through the light fixtures 608. The fluid conditioning system 612 supplies to the fluid-cooled light fixture 608 a supply fluid D, having a temperature that is controlled to meet sensible load set points for desired operating conditions of the plants in the environment 600.
Waste heat from the fluid passing through fluid conditioning system 612 may be provided to the heating coil 614 in the heat exchanger 615 to heat air E that is output from the dehumidifier 610. The air heated by the coil 614 is output as heated air B to the grow space 600.
A controller 620 may control all the elements of the conditioning system 602, according to embodiments of the disclosure. The controller 620 may receive sensed parameters from sensors distributed throughout the plant growing environment 600 and the air and water conditioning system 602, according to embodiments of the disclosure. Such sensors may include, for example, sensors that measure temperature, humidity, soil moisture, plant characteristics (e.g., size, shape, color), and irrigation flow rate. The controller 620 may also receive operating settings for those same parameters as well as others. The controller 620 may use the sensed parameters as feedback to instruct the conditioning system 602 to control environmental treatments (e.g., temperature, humidity) of the plant growing environment 600, according to embodiments of the disclosure. The controller 620 may employ machine learning or other predictive methods to adjust the treatments to achieve a desired objective relating to parameters such as ambient environmental conditions (e.g., temperature), energy usage, productivity, or plant product yield to energy use.
According to embodiments of the disclosure, the controller 620 may control evapotranspiration by controlling the following factors, whether alone or in any combination: irrigation, CO2 concentration, temperature, relative humidity, vapor pressure deficit, light intensity (e.g., based on daytime or nighttime condition), light wavelength (including electromagnetic radiation wavelength in the visible range and in the non-visible range, such as ultraviolet and infrared), light duration, light modulation (e.g., pulse width modulation), or air velocity, or by varying the supply of chemicals (e.g., hormones) that regulate transpiration. By increasing evapotranspiration, the controller 620 may cause a decrease in the sensible cooling load.
According to embodiments of the disclosure, the combination of mechanical equipment that is specified to remove water from the air and maintain the desired air temperature can vary widely, but there are some primary characteristics that they should typically include: (a) the equipment should be specified to remove an equivalent or larger amount of water than occurs during worst case transpiration and evaporation scenarios for both night and day; and (b) when heat is required to either warm the air or evaporate moisture, waste heat as a byproduct from other components in the system should be used. Waste heat may come from within the grow space (e.g., from lighting) or from the environmental control system 602 (e.g., from a compressor).
Referring to
The environmental control system 602 may employ Direct-Exchange (“DX”) equipment. DX equipment uses a vapor compression cycle to condition the air to a desired condition. At the evaporator of the DX equipment, the air is cooled to its saturation point and moisture condenses out, dehumidifying the air. Electronically controlled expansion valves and modulating hot-gas reheat are two features used with DX systems to incorporate the desired amount of heat back into the airstream, after dehumidification. Any heat that is not put back into the air will eventually be rejected, or sent to other energy recovery devices. Heat rejection can occur in a number of ways not limited to condensing units. Multiple rows of coils, variable speed fans, multiple vapor-compression circuits, air-side economization, and air bypasses are some of many ways that the DX unit can be configured to condition specific and varying loads with energy savings within one DX unit.
The environmental control system 602 may employ chilled water or chilled fluid air handling units that use a working fluid to cool, dehumidify, or heat the air to a desired condition. Chilled fluid can be provided by, but is not limited to, a chiller or fluid cooler. Heat-recovery chillers are one way that energy used in the cooling process can be re-used to heat the airstream back up to the desired condition. Chillers can be used in conjunction with other equipment that provides or removes heat, according to the embodiments of the disclosure. Boilers, solar heaters, heat pumps, the utilization of heat from lighting, and the utilization of heat from other equipment are just a few of the other ways that heat can be added to the airstream.
The environmental control system 602 may employ desiccant dehumidification either standing alone or in conjunction with other air-handling equipment to achieve a desired air condition with systems-level energy savings. Desiccants make use of a chemical that adsorbs moisture from the airstream. The chemical, or desiccants, used in desiccant dehumidifiers are able, when heated, to release the amount of moisture that was adsorbed to another fluid stream. In this process, heat is required to recharge the desiccants rather than to reheat air that has been cooled below the dew point. Although these systems are both energy efficient and have the ability to supply air at a low humidity, the capital cost may be high. One of several efficient applications of desiccant dehumidification is further dehumidifying air that has already been cooled by mechanical cooling equipment, prior to heating the air back up for delivery to the controlled environment.
The environmental control system 602 may employ energy wheels, enthalpy wheels or other air-to-air heat exchangers to further improve energy efficiency and recovery in air-handling units. A wrap-around heat pipe dehumidifier exchanges sensible heat of the outgoing airstream with the incoming airstream, with a cooling coil in-between. In this configuration, air that passes over the coil exchanges heat with incoming air, pre-cooling the incoming air and heating the outgoing air; the net effect of this is less cooling energy required at the coil. Total energy wheels and enthalpy wheels are just some of the heat exchange equipment that can be used in system-level optimization for energy and cost.
Carbon dioxide supply and control is also a component of the design. Control of carbon dioxide levels in the grow space affects evapotranspiration rate and components of plant growth. Carbon dioxide control, with other control variables such as, but not limited to, light intensity, vapor-pressure deficit, fan speed and airflow velocity control, and nutrient supply, are ways that the conditioning system 602 can control latent load for a controlled agriculture environment.
Machine Learning
Embodiments of the disclosure may apply machine learning (“ML”) techniques to learn the relationship between the given parameters (e.g., environmental conditions such as temperature, humidity) and observed outcomes (e.g., experimental data concerning yield and energy consumption). In this framework, embodiments may use standard ML models, e.g. Decision Trees, to determine feature importance. In general, machine learning may be described as the optimization of performance criteria, e.g., parameters, techniques or other features, in the performance of an informational task (such as classification or regression) using a limited number of examples of labeled data, and then performing the same task on unknown data. In supervised machine learning such as an approach employing linear regression, the machine (e.g., a computing device) learns, for example, by identifying patterns, categories, statistical relationships, or other attributes exhibited by training data. The result of the learning is then used to predict whether new data will exhibit the same patterns, categories, statistical relationships or other attributes.
Embodiments of this disclosure may employ unsupervised machine learning. Alternatively, some embodiments may employ semi-supervised machine learning, using a small amount of labeled data and a large amount of unlabeled data. Embodiments may also employ feature selection to select the subset of the most relevant features to optimize performance of the machine learning model. Depending upon the type of machine learning approach selected, as alternatives or in addition to linear regression, embodiments may employ for example, logistic regression, neural networks, support vector machines (SVMs), decision trees, hidden Markov models, Bayesian networks, Gram Schmidt, reinforcement-based learning, cluster-based learning including hierarchical clustering, genetic algorithms, and any other suitable learning machines known in the art. In particular, embodiments may employ logistic regression to provide probabilities of classification along with the classifications themselves.
Embodiments may employ graphics processing unit (GPU) or Tensor processing units (TPU) accelerated architectures that have found increasing popularity in performing machine learning tasks, particularly in the form known as deep neural networks (DNN). Embodiments of the disclosure may employ GPU-based machine learning, such as that described in GPU-Based Deep Learning Inference: A Performance and Power Analysis, NVidia Whitepaper, November 2015, Dahl, et al., which is incorporated by reference in its entirety herein.
Computer System Implementation
Program code may be stored in non-transitory media such as persistent storage in secondary memory 810 or main memory 808 or both. Main memory 808 may include volatile memory such as random access memory (RAM) or non-volatile memory such as read only memory (ROM), as well as different levels of cache memory for faster access to instructions and data. Secondary memory may include persistent storage such as solid state drives, hard disk drives or optical disks. One or more processors 804 reads program code from one or more non-transitory media and executes the code to enable the computer system to accomplish the methods performed by the embodiments herein. Those skilled in the art will understand that the processor(s) may ingest source code, and interpret or compile the source code into machine code that is understandable at the hardware gate level of the processor(s) 804. The processor(s) 804 may include graphics processing units (GPUs) for handling computationally intensive tasks.
The processor(s) 804 may communicate with external networks via one or more communications interfaces 807, such as a network interface card, WiFi transceiver, etc. A bus 805 communicatively couples the I/O subsystem 802, the processor(s) 804, peripheral devices 806, communications interfaces 807, memory 808, and persistent storage 810. Embodiments of the disclosure are not limited to this representative architecture. Alternative embodiments may employ different arrangements and types of components, e.g., separate buses for input-output components and memory subsystems.
Those skilled in the art will understand that some or all of the elements of embodiments of the disclosure, and their accompanying operations, may be implemented wholly or partially by one or more computer systems including one or more processors and one or more memory systems like those of computer system 800. In particular, the elements of automated systems or devices described herein may be computer-implemented. Some elements and functionality may be implemented locally and others may be implemented in a distributed fashion over a network through different servers, e.g., in client-server fashion, for example.
Although the disclosure may not expressly disclose that some embodiments or features described herein may be combined with other embodiments or features described herein, this disclosure should be read to describe any such combinations that would be practicable by one of ordinary skill in the art. Unless otherwise indicated herein, the term “include” shall mean “include, without limitation,” and the term “or” shall mean non-exclusive “or” in the manner of “and/or.”
Those skilled in the art will recognize that, in some embodiments, some of the operations described herein may be performed by human implementation, or through a combination of automated and manual means. When an operation is not fully automated, appropriate components of embodiments of the disclosure may, for example, receive the results of human performance of the operations rather than generate results through its own operational capabilities.
All references, articles, publications, patents, patent publications, and patent applications cited herein are incorporated by reference in their entireties for all purposes to the extent they are not inconsistent with embodiments of the disclosure expressly described herein. However, mention of any reference, article, publication, patent, patent publication, and patent application cited herein is not, and should not be taken as an acknowledgment or any form of suggestion that they constitute valid prior art or form part of the common general knowledge in any country in the world, or that they are disclose essential matter.
In the claims below, a claim n reciting “any one of the preceding claims starting with claim x,” shall refer to any one of the claims starting with claim x and ending with the immediately preceding claim (claim n-1). For example, claim 35 reciting “The system of any one of the preceding claims starting with claim 28” refers to the system of any one of claims 28-34.
This application claims priority to U.S. Application No. 62/742,751, filed Oct. 8, 2018, and incorporated by reference in its entirety herein. This application is related to U.S. Patent Application Pub. No. 2018/0014486, filed Sep. 28, 2016, and U.S. Patent Application Pub. No. US 2017/0146226, filed Nov. 15, 2016, both assigned to the assignee of the present disclosure and incorporated by reference in their entirety herein.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US2019/055064 | 10/7/2019 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 62742751 | Oct 2018 | US |
