Patent Grant
6965806
The present invention relates generally to process control systems within process plants and, more particularly, to the use of a coordinated expert system to aid asset utilization and optimization in a process plant.
Process control systems, like those used in chemical, petroleum or other processes, typically include one or more centralized or decentralized process controllers communicatively coupled to at least one host or operator workstation and to one or more process control and instrumentation devices, such as field devices, via analog, digital or combined analog/digital buses. Field devices, which may be, for example valves, valve positioners, switches, transmitters, and sensors (e.g., temperature, pressure and flow rate sensors), perform functions within the process such as opening or closing valves and measuring process parameters. The process controller receives signals indicative of process measurements or process variables made by or associated with the field devices and/or other information pertaining to the field devices, uses this information to implement a control routine and then generates control signals which are sent over one or more of the buses to the field devices to control the operation of the process. Information from the field devices and the controller is typically made available to one or more applications executed by an operator workstation to enable an operator to perform desired functions with respect to the process, such as viewing the current state of the process, modifying the operation of the process, etc.
While a typical process control system has many process control and instrumentation devices, such as valves, transmitters, sensors, etc. connected to one or more process controllers which execute software that controls these devices during the operation of the process, there are many other supporting devices which are also necessary for or related to process operation. These additional devices include, for example, power supply equipment, power generation and distribution equipment, rotating equipment such as turbines, etc., which are located at numerous places in a typical plant. While this additional equipment does not necessarily create or use process variables and, in many instances, is not controlled or even coupled to a process controller for the purpose of affecting the process operation, this equipment is nevertheless important to and ultimately necessary for proper operation of the process. In the past however, process controllers were not necessarily aware of these other devices or the process controllers simply assumed that these devices were operating properly when performing process control.
Still further, many process plants have other computers associated therewith which execute applications related to business functions or maintenance functions. For example, some plants include computers which execute applications associated with ordering raw materials, replacement parts or devices for the plant, applications related to forecasting sales and production needs, etc. Likewise, many process plants, and especially those which use smart field devices, include applications which are used to help monitor and maintain the devices within the plant regardless of whether these devices are process control and instrumentation devices or are other types of devices. For example, the Asset Management Solutions (AMS) application sold by Fisher-Rosemount Systems, Inc. enables communication with and stores data pertaining to field devices to ascertain and track the operating state of the field devices. An example of such a system is disclosed in U.S. Pat. No. 5,960,214 entitled “Integrated Communication Network for use in a Field Device Management System.” In some instances, the AMS application may be used to communicate with devices to change parameters within the device, to cause the device to run applications on itself, such as self calibration routines or self diagnostic routines, to obtain information about the status or health of the device, etc. This information may be stored and used by a maintenance person to monitor and maintain these devices. Likewise, there are other types of applications which are used to monitor other types of devices, such as rotating equipment and power generation and supply devices. These other applications are typically available to the maintenance persons and are used to monitor and maintain the devices within a process plant.
However, in the typical plant or process, the functions associated with the process control activities, the device and equipment maintenance and monitoring activities, and the business activities are separated, both in the location in which these activities take place and in the personnel who typically perform these activities. Furthermore, the different people involved in these different functions generally use different tools, such as different applications run on different computers to perform the different functions. In many instances, these different tools collect or use different types of data associated with or collected from the different devices within the process and are set up differently to collect the data they need. For example, process control operators who generally oversee the day to day operation of the process and who are primarily responsible for assuring the quality and continuity of the process operation typically affect the process by setting and changing set points within the process, tuning loops of the process, scheduling process operations such as batch operations, etc. These process control operators may use available tools for diagnosing and correcting process control problems within a process control system, including, for example, auto-tuners, loop analyzers, neural network systems, etc. Process control operators also receive process variable information from the process via one or more process controllers which provide information to the operators about the operation of the process, including alarms generated within the process. This information may be provided to the process control operator via a standard user interface.
Still further, it is currently known to provide an expert engine that uses process control variables and limited information about the operating condition of the control routines or function blocks or modules associated with process control routines to detect poorly operating loops and to provide information to an operator about suggested courses of action to correct the problem. Such an expert engine is disclosed in U.S. patent application Ser. No. 09/256,585 entitled “Diagnostics in a Process Control System,” which was filed on Feb. 22, 1999 and in U.S. patent application Ser. No. 09/499,445 entitled “Diagnostic Expert in a Process Control System,” which was filed on Feb. 7, 2000, both of which are hereby expressly incorporated by reference herein. Likewise, it is known to run control optimizers, such as real time optimizers, within a plant to optimize the control activities of the process plant. Such optimizers typically use complex models of the plant to predict how inputs may be changed to optimize operation of the plant with respect to some desired optimization variable such as, for example, profit.
On the other hand, maintenance personnel who are primarily responsible for assuring that the actual equipment within the process is operating efficiently and for repairing and replacing malfunctioning equipment, use tools such as maintenance interfaces, the AMS application discussed above, as well and many other diagnostic tools which provide information about operating states of the devices within the process. Maintenance persons also schedule maintenance activities which may require shut down of portions of the plant. For many newer types of process devices and equipment, generally called smart field devices, the devices themselves may include detection and diagnostic tools which automatically sense problems with the operation of the device and automatically report these problems to a maintenance person via a standard maintenance interface. For example, the AMS software reports device status and diagnostic information to the maintenance person and provides communication and other tools that enable the maintenance person to determine what is happening in devices and to access device information provided by devices. Typically, maintenance interfaces and maintenance personnel are located apart from process control operators, although this is not always the case. For example, in some process plants, process control operators may perform the duties of maintenance persons or vice versa, or the different people responsible for these functions may use the same interface.
Still further, persons responsible and applications used for business applications, such as ordering parts, supplies, raw materials, etc., making strategic business decisions such as choosing which products to manufacture, what variables to optimize within the plant, etc. are typically located in offices of the plant that are remote from both the process control interfaces and the maintenance interfaces. Likewise, managers or other persons may want to have access to certain information within the process plant from remote locations or from other computer systems associated with the process plant for use in overseeing the plant operation and in making long term strategic decisions.
Because, for the most part, very different applications used to perform the different functions within a plant, e.g., process control operations, maintenance operations and business operations are separated, the different applications used for these different tasks are not integrated and, thus, do not share data or information. In fact, many plants only include some, but not all, of these different types of applications. Furthermore, even if all of the applications are located within a plant, because different personnel use these different applications and analysis tools and because these tools are generally located at different hardware locations within the plant, there is little if any flow of information from one functional area of the plant to another, even when this information may be useful to other functions within the plant. For example, a tool, such as a rotating equipment data analysis tool, may be used by a maintenance person to detect a poorly functioning power generator or piece of rotating equipment (based on non-process variable type data). This tool may detect a problem and alert the maintenance person that the device needs to be calibrated, repaired or replaced. However, the process control operator (either a human or a software expert) does not have the benefit of this information, even though the poorly operating device may be causing a problem that is affecting a loop or some other component which is being monitored by the process control operation. Likewise, the business person is not aware of this fact, even though the malfunctioning device may be critical to and may be preventing optimization of the plant in a manner that the business person may desire. Because the process control expert is unaware of a device problem which may be ultimately causing poor performance of a loop or unit in the process control system and because the process control operator or expert assumes that this equipment is operating perfectly, the process control expert may misdiagnose the problem it detects within the process control loop or may try to apply a tool, such as a loop tuner, which could never actually correct the problem. Likewise, the business person may make a business decision to run the plant in a manner that will not achieve the desired business effects (such as optimizing profits) because of the malfunctioning device.
Due to the abundance of data analysis and other detection and diagnostic tools available in the process control environment, there is a lot of information about the health and performance of devices available to the maintenance person which could be helpful to the process operator and the business persons. Similarly, there is a lot of information available to the process operator about the current operational status of the process control loops and other routines which may be helpful to the maintenance person or to the business person. Likewise, there is information generated by or used in the course of performing the business functions which could be helpful to the maintenance person or the process control operator in optimizing the operation of the process. However, in the past, because these functions were separated, the information generated or collected in one functional area was not used at all, or not used very well in other functional areas which led to an overall sub-optimal use of the assets within process plants.
A process control system uses data or information pertaining to the assets of a process plant from various devices within the process plant. The devices may include field devices such as valves, positioners, switches, multiplexers, transmitters, sensors, control systems, transceivers, variable speed drivers, actuators, I/O systems, two-, three-, or four-wire devices, etc., as well as various smart devices including devices according to a specific protocol such as Fieldbus, HART, PROFIBUS®, WORLDFIP®, Device-Net®, AS-Interface, and CAN. The devices may also include network communication devices such as TCP/IP protocol devices, Ethernet devices, and internet devices. The entities may further include field equipment such as power generation equipment, power distribution equipment, transformers, tanks, rotating equipment, measuring equipment, pumps, etc. The data and information of the process plant, such as control, process and maintenance data, is manipulated in a coordinated manner by tools, such as to create status information regarding the device. The status information may be utilized to perform better or more optimal maintenance activities, such as corrective measures pertaining to a problem with the device. For example, information or data may be collected by maintenance functions pertaining to the status of a device. This information may then used to determine the status of the device. The status information may be in the form of diagnostic information, maintenance requests, notifications of current or future problems, etc. If the status information indicates a problem with the device, a work order and/or parts order related to solving the problem may be automatically generated. The order may then be communicated to fulfill the order. In the case of a parts order, the communication may be with a supplier of the parts to be used in the repair. If the order is a work order, the appropriate maintenance personnel may be alerted. In other cases, such as with predicted future problems, the order may be scheduled such that it will be fulfilled prior to the predicted failure of the device. In addition, the status of the order may be tracked based on periodic reports regarding the order.
Referring now to
Still further, maintenance systems, such as computers executing the AMS application or any other device monitoring and communication applications may be connected to the process control systems 12 and 14 or to the individual devices therein to perform maintenance and monitoring activities. For example, a maintenance computer 18 may be connected to the controller 12B and/or to the devices 15 via any desired communication lines or networks (including wireless or handheld device networks) to communicate with and, in some instances, reconfigure or perform other maintenance activities on the devices 15. Similarly, maintenance applications such as the AMS application may be installed in and executed by one or more of the user interfaces 14A associated with the distributed process control system 14 to perform maintenance and monitoring functions, including data collection related to the operating status of the devices 16.
The process plant 10 also includes various rotating equipment 20, such as turbines, motors, etc. which are connected to a maintenance computer 22 via some permanent or temporary communication link (such as a bus, a wireless communication system or hand held devices which are connected to the equipment 20 to take readings and are then removed). The maintenance computer 22 may store and execute known monitoring and diagnostic applications 23 provided by, for example, CSi Systems or other any other known applications used to diagnose, monitor and optimize the operating state of the rotating equipment 20. Maintenance personnel usually use the applications 23 to maintain and oversee the performance of rotating equipment 20 in the plant 10, to determine problems with the rotating equipment 20 and to determine when and if the rotating equipment 20 must be repaired or replaced.
Similarly, a power generation and distribution system 24 having power generating and distribution equipment 25 associated with the plant 10 is connected via, for example, a bus, to another computer 26 which runs and oversees the operation of the power generating and distribution equipment 25 within the plant 10. The computer 26 may execute known power control and diagnostics applications 27 such a as those provided by, for example, Liebert and ASCO or other companies to control and maintain the power generation and distribution equipment 25.
In the past, the various process control systems 12 and 14 and the power generating and maintenance systems 22 and 26 have not been interconnected with each other in a manner that enables them to share data generated in or collected by each of these systems in a useful manner. As a result, each of the different functions such as the process control functions, power generation functions and rotating equipment functions have operated on the assumption that the other equipment within the plant which may be affected by or have an affect on that particular function is operating perfectly which, of course, is almost never the case. However, because the functions are so different and the equipment and personnel used to oversee these functions are different, there has been little or no meaningful data sharing between the different functional systems within the plant 10.
To overcome this problem, a computer system 30 is provided which is communicatively connected to the computers or interfaces associated with the various functional systems within the plant 10, including the process control functions 12 and 14, the maintenance functions such as those implemented in the computers 18, 14A, 22 and 26 and the business functions. In particular, the computer system 30 is communicatively connected to the traditional process control system 12 and to the maintenance interface 18 associated with that control system, is connected to the process control and/or maintenance interfaces 14A of the distributed process control system 14, is connected to the rotating equipment maintenance computer 22 and to the power generation and distribution computer 26, all via a bus 32. The bus 32 may use any desired or appropriate local area network (LAN) or wide area network (WAN) protocol to provide communications.
As illustrated in
In one embodiment, the communications over the bus 32 occur using the XML protocol. Here, data from each of the computers 12A, 18, 14A, 22, 26, 35, 36, etc. is wrapped in an XML wrapper and is sent to an XML data server which may be located in, for example, the computer 30. Because XML is a descriptive language, the server can process any type of data. At the server, if necessary, the data is encapsulated with to a new XML wrapper, i.e., this data is mapped from one XML schema to one or more other XML schemas which are created for each of the receiving applications. Thus, each data originator can wrap its data using a schema understood or convenient for that device or application, and each receiving application can receive the data in a different schema used for or understood by the receiving application. The server is configured to map one schema to another schema depending on the source and destination(s) of the data. If desired, the server may also perform certain data processing functions or other functions based on the receipt of data. The mapping and processing function rules are set up and stored in the server prior to operation of the system described herein. In this manner, data may be sent from any one application to one or more other applications.
Generally speaking, the computer 30 stores and executes an asset utilization expert 50 that collects data and other information generated by the process control systems 12 and 14, the maintenance systems 18, 22 and 26 and the business systems 35 and 36 as well as information generated by data analysis tools executed in each of these systems. The asset utilization expert 50 may be based on, for example, the OZ expert system currently provided by NEXUS. However, the asset utilization expert 50 may be any other desired type of expert system including, for example, any type of data mining system. Importantly, the asset utilization expert 50 operates as a data and information clearinghouse in the process plant 10 and is able to coordinate the distribution of data or information from one functional area, such as the maintenance area, to other functional areas, such as the process control or the business functional areas. The asset utilization expert 50 may also use the collected data to generate new information or data which can be distributed to one or more of the computer systems associated with the different functions within the plant 10. Still further, the asset utilization expert 50 may execute or oversee the execution of other applications that use the collected data to generate new types of data to be used within the process plant 10.
In particular, the asset utilization expert 50 may include or execute index generation software 51 that creates indices associated with devices, like process control and instrumentation devices, power generation devices, rotating equipment, units, areas, etc, or that are associated with process control entities, like loops, etc. within the plant 10. These indices can then be provided to the process control applications to help optimize process control and can be provided to the business software or business applications to provide the business persons more complete or understandable information associated with the operation of the plant 10. The asset utilization expert 50 can also provide maintenance data (such as device status information) and business data (such as data associated with scheduled orders, timeframes, etc.) to a control expert 52 associated with, for example, the process control system 14 to help an operator perform control activities such as optimizing control. The control expert 52 may be located in, for example, the user interface 14A or any other computer associated with the control system 14 or within the computer 30 if desired.
In one embodiment, the control expert 52 may be, for example, the control expert described in U.S. patent application Ser. Nos. 09/256,585 and 09/499,445 identified above. However, these control experts may additionally incorporate and use data related to the status of devices or other hardware within the process plant 10 in the decision making performed by these control experts. In particular, in the past, the software control experts generally only used process variable data and some limited device status data to make decisions or recommendations to the process operator. With the communication provided by the asset utilization expert 50, especially that related to device status information such as that provided by the computer systems 18, 14A, 22 and 26 and the data analysis tools implemented thereon, the control expert 52 can receive and incorporate device status information such as health, performance, utilization and variability information into its decision making along with process variable information.
Additionally, the asset utilization expert 50 can provide information pertaining to states of devices and the operation of the control activities within the plant 10 to the business systems 35 and 36 where, for example, a work order generation application or program 54 can automatically generate work orders and order parts based on detected problems within the plant 10 or where supplies can be ordered based on work being performed. Similarly, changes in the control system detected by the asset utilization expert 50 may cause the business systems 35 or 36 to run applications that perform scheduling and supply orders using, for example, the program 54. In the same manner, changes in customer orders etc. can be entered into the business systems 35 or 36 and this data can be sent to the asset utilization expert 50 and sent to the control routines or control expert 52 to cause changes in the control to, for example, begin making the newly ordered products or to implement the changes made in the business systems 35 and 36. Of course, if desired, each computer system connected to the bus 32 may have an application therein that functions to obtain the appropriate data from the other applications within the computer and to sending this data to, for example, the asset utilization expert 50.
Additionally, the asset utilization expert 50 can send information to one or more optimizers 55 within the plant 10. For example, a control optimizer 55 can be located in the computer 14A and can run one or more control optimization routines 55A, 55B, etc. Additionally or alternatively, optimizer routines 55 could be stored in and executed by the computer 30 or any other computer, and the data necessary therefore could be sent by the asset utilization expert 50. If desired, the plant 10 may also include models 56 that model certain aspects of the plant 10 and these models 56 can be executed by the asset utilization expert 50 or a control or other expert such as the control expert 52 to perform modeling functions, the purpose of which will be described in more detail herein. Generally speaking, however, the models 56 can be used to determine device, area, unit, loop, etc. parameters, to detect faulty sensors or other faulty equipment, as part of optimizer routines 55, to generate indices such as performance and utilization indices for use in the plant 10, to perform performance or condition monitoring, as well as for many other uses. The models 56 may be models such as those created by and sold by MDC Technology located in Teeside, England or may be any other desired types of models. There are, of course, many other applications that can be provided within the plant 10 and that can use the data from the asset utilization expert 50 and the system described herein is not limited to the applications specifically mentioned herein. Overall, however, the asset utilization expert 50 helps to optimize the use of all of the assets within the plant 10 by enabling the sharing of data and coordination of assets between all of the functional areas of the plant 10.
Also, generally speaking, one or more user interface routines 58 can be stored in and executed by one or more of the computers within the plant 10. For example, the computer 30, the user interface 14A, the business system computer 35 or any other computer may run a user interface routine 58. Each user interface routine 58 can receive or subscribe to information from the asset utilization expert 50 and either the same or different sets of data may be sent to each of the user interface routines 58. Any one of the user interface routines 58 can provide different types of information using different screens to different users. For example, one of the user interface routines 58 may provide a screen or set of screens to a control operator or to a business person to enable that person to set constraints or to choose optimization variables for use in a standard control routine or in a control optimizer routine. The user interface routine 58 may provide a control guidance tool that enables a user to view the indices created by the index generation software 51 in some coordinated manner. This operator guidance tool may also enable the operator or any other person to obtain information about the states of devices, control loops, units, etc. and to easily see the information related to the problems with these entities, as that information has been detected by other software within the process plant 10. The user interface routine 58 may also provide performance monitoring screens using performance monitoring data provided by or generated by the tools 23 and 27, the maintenance programs such as the AMS application or any other maintenance programs, or as generated by the models in conjunction with the asset utilization expert 50. Of course, the user interface routine 58 may provide any user access to and enable the user to change preferences or other variables used in any or all functional areas of the plant 10.
Referring now to
Also, the asset utilization expert 50 receives information from (and may actually execute) data analysis tools such as typical maintenance data analysis tools which are currently provided today, performance tracking tools, such as those associated with devices, as well as performance tracking tools for process control systems like that described in U.S. patent application Ser. Nos. 09/256,585 and 09/499,445 identified above. The data analysis tools may also include, for example, a root cause application which detects root causes of certain types of problems, event detection such as that described in U.S. Pat. No. 6,017,143, regulatory loop diagnostics such as that disclosed in U.S. patent application Ser. No. 09/303,869 (filed May 3, 1999), which is hereby expressly incorporated by reference herein, impulse lines plugging detection applications, such as that described in U.S. patent application Ser. No. 09/257,896 (filed Feb. 25, 1999), which is hereby expressly incorporated by reference herein, other plugged line detection applications, device status applications, device configuration applications and maintenance applications, device storage, historian and information display tools, such as AMS, Explorer applications and audit trail applications. Still further, the expert 50 can receive data and any information from process control data analysis tools such as the advanced control expert 52, model predictive control process routines such as those described in U.S. patent application Ser. Nos. 09/593,327 (filed Jun. 14, 2000) and 09/412,078 (filed Oct. 4, 1999), which are hereby expressly incorporated by reference herein, tuning routines, fuzzy logic control routines and neural network control routines, as well as from virtual sensors such as that described in U.S. Pat. No. 5,680,409, which may be provided within the process control system 10. Still further, the asset utilization expert 50 may receive information from data analysis tools related to rotating equipment such as on-line vibration, RF wireless sensors and hand-held data collection units, oil analysis associated with rotating equipment, thermography, ultra-sonic systems and laser alignment and balancing systems, all of which may be related to detecting problems or the status of rotating equipment within the process plant 10. These tools are currently known in the art and so will not be described further herein. Still further, the asset utilization expert 50 may receive data related to power management and power equipment and supplies such as the applications 23 and 27 of
In one embodiment, the asset utilization expert 50 executes or oversees the execution of mathematical software models 56 of some or all of the equipment within the plant 10, such as device models, loops models, unit models, area models, etc., which are run by, for example, the computer 30 or any other desired computer within process plant 10. The asset utilization expert 50 may use the data developed by or associated with these models for a number of reasons. Some of this data (or the models themselves) may be used to provide virtual sensors within the plant 10. Some of this data, or the models themselves, may be used to implement predictive control or real time optimal control within the plant 10. Some of the data generated by the models 56 may be used by the index generation routine 51 to generate indices which are used in other applications, such as business and process control applications. The use of the models 56 for these and other purposes will be described in more detail below.
The asset utilization expert 50 receives data as it is generated or at certain periodic times over, for example, the bus 32 or other any communication network within the process plant 10. Thereafter, periodically or as needed, the asset utilization expert 50 redistributes the data to other applications or uses that data to generate and provide other information useful in different aspects of the control or operation of the process plant 10 to other function systems within the plant 10. In particular, the asset utilization expert 50 may supply data to cause the index generation routine 51 to create a series of composite indices such as a performance index, a utilization index, a health index and a variability index associated with one or more of the devices, units, loops, areas, or other entities within the process plant 10. The generation and use of these indices will also be discussed in more detail herein.
The asset utilization expert 50 may also provide data to and receive data from control routines 62 which may be located in process controllers or interfaces associated with those controllers, optimizers 55, business applications 63, maintenance applications 66, etc.
Furthermore, a control expert 65 (which may include a predictive process controller), which in the past simply assumed that the devices it was controlling either worked properly or not at all, can receive information from the asset utilization expert 50 related to the status or health of the devices it is controlling, such as the utilization, variability, health or performance indices mentioned above or other information related to the operating status of devices, loops, etc. which can be taken into account when trying to control a process. The predictive controller 65, as well as the optimizers 55 may provide additional information and data to user interface routines 58. The predictive controller 65 or optimizer 55 may use the status information pertaining to actual current status of the devices in the network, as well as take into account goals and future needs such as those identified by business solution software provided from the asset utilization expert 50 as defined by, for example, business applications 63, to optimize control based on predictions within the control system.
Still further, the asset utilization expert 50 may provide data to and receive data from enterprise resource planning tools such as those typically used in business solutions or business computers 35 and 36. These applications may include production planning tools which control production planning, material resource planning, the work order generation tool 54 which automatically generates part orders, work orders, or supply orders for use in the business applications, etc. Of course, the part order, work order and supply order generation may be completed automatically based on information from the asset utilization expert 50, which decreases the time required to recognize that an asset needs to be fixed as well as the time is takes to receive the parts necessary to provide corrective action with respect to maintenance issues.
The asset utilization expert 50 may also provide information to the maintenance system applications 66, which not only alert maintenance people to problems immediately, but also take corrective measures such as ordering parts, etc. which will be needed to correct a problem. Still further, new models 68 may be generated using types of information that are available to the asset utilization expert 50 but that were previously unavailable to any single system. Of course, it will be understood from
Moreover, one or more coordinated user interface routines 58 may communicate with the asset utilization expert 50 as well as any other applications within the plant 10 to provide help and visualization to operators, maintenance persons, business persons, etc. The operators and other users may use the coordinated user interface routines 58 to perform or to implement predictive control, change settings of the plant 10, view help within the plant 10, or perform any other activities related to the information provided by the asset utilization expert 50. As discussed above, the user interface routines 58 may include an operator guidance tool that receives information from the predictive controller 65 as well as information related to the indices, which can be used by an operator or other user to help perform many functions such as viewing the status of a process or devices within the process, to guide the predictive controller 65 or to perform predictive or optimized control. Still further, the user interface routines 58 may be used to view data or to obtain data from any of the tools in the other parts of the process plant 10 via, for example, the asset utilization expert 50. For example, managers may want to know what is happening in the process or may need high level information related to the process plant 10 to make strategic plans.
As mentioned above, the asset utilization expert 50 can execute or oversee the execution of one or more mathematical or software models 56 that model the operation of a particular plant or entities within the plant, such as devices, units, loops, areas, etc. These models may be hardware models or they may be process control models. In one embodiment, to generate these models, a modeling expert divides the plant into component hardware and/or process control parts and provides a model for the different component parts at any desired level of abstraction. For example, the model for a plant is implemented in software and is made up of or may include a set of hierarchically related, interconnected models for the different areas of the plant. Similarly, the model for any plant area may be made up of individual models for the different units within the plant with interconnections between the inputs and outputs of these units. Likewise, units may be made up of interconnected device models, and so on. Of course, area models may have device models interconnected with unit models, loop models, etc. In this example model hierarchy, the inputs and outputs of models for the lower level entities, such as devices, may be interconnected to produce models for higher level entities, such as units, the inputs and outputs of which may be interconnected to create still higher level models, such as area models, and so on. The way in which the different models are combined or interconnected will, of course depend on the plant being modeled. While a single, complete mathematical model for the whole plant could be used, it is believed that providing different and independent component models for different portions of or entities within the plant, such as areas, units, loops, devices, etc. and interconnecting these different models to form larger models is useful for a number of reasons. Furthermore, it is desirable to use component models that can be run independently of one another as well as together with other component models as part of a larger model.
While highly mathematically accurate or theoretical models (such as third or fourth order models) may be used for the entire plant or for any or all of the component models, the individual models need not necessarily be as mathematically accurate as possible and could be, for example, first or second order models or other types of models. These simpler models can generally be executed more quickly in software and can be made more accurate by matching the inputs and outputs of the models with actual measurements of inputs and outputs made within the plant in a manner described herein. In other words, the individual models may be tuned or tweaked to accurately model the plant or the entities within the plant based on actual feedback from the plant.
The use of hierarchical software models will now be described with respect to
Referring now to
As noted above, the component model for the distillation process 90 may be executed as part of a model for the area 82 or may be executed separately and apart from any other models. In particular, the input 103 to the distillation column 100 and/or the outputs C2H4 and C2H6 can actually be measured and these measurements may be used within the model of the distillation process 90 in a number of ways as described below. In one embodiment, the inputs and outputs of the model of the distillation process 90 may be measured and used to determine other factors or parameters associated with the model of the distillation process 90 (such as the distillation column efficiency, etc.) to force the model of the distillation process 90 to more accurately match the operation of the actual distillation column within the plant 10. The model of the distillation process 90 may then be used with the calculated parameters, as part of a larger model, such as an area or plant model. Alternatively or additionally, the model of the distillation process 90 with the calculated parameters may be used to determine virtual sensor measurements or to determine if actual sensor measurements within the plant 10 are in error. The model of the distillation process 90 with the determined parameters may also be used to perform control or asset utilization optimization studies, etc. Still further, component models may be used to detect and isolate developing problems in the plant 10 or to see how changes to the plant 10 might affect the selection of optimization parameters for the plant 10.
If desired, any particular model or component model may be executed to determine the values of the parameters associated with that model. Some or all of these parameters such as efficiency parameters may mean something to an engineer within the context of the model but are generally unmeasurable within the plant 10. More particularly, a component model may be generally mathematically described by the equation Y=F(X, P), wherein the outputs Y of the model are a function of the inputs X and a set of model parameters P. In the example of the distillation column model of the distillation process 90 of
Similarly, a regression analysis can be performed to determine if sensors making measurements within the process 10 are drifting or have some other error associated therewith. Here, the same data or potentially different data pertaining to the measured inputs and outputs of the entity being modeled are collected and stored by, for example, the asset utilization expert 50. In this case, the model can be generally mathematically expressed as Y+dY=F(X+dX, P), wherein dY are the errors associated with the measurements of the outputs Y, and dX are the errors associated with measurements of the inputs X. Of course, these errors could be any types of errors, such as bias, drift, or non-linear errors and the model may recognize that the inputs X and outputs Y may have different kinds of errors associated therewith, with the different kinds of possible errors having different mathematical relationships to the actual measured values. In any event, a regression analysis, such as a maximum likelihood, least squares or any other regression analysis may be performed using the model with the measured inputs and outputs to determine a best fit for the unknown sensor errors dY and dX. Here, the model parameters P may be based on the parameters P calculated using a previous regression analysis for the model, or may be treated as further unknowns and may be determined in conjunction with this regression analysis. Of course, as the number of unknowns used within regression analysis increases, the amount of data required increases and the longer it takes to run the regression analysis. Furthermore, if desired, the regression analysis for determining the model parameters and the regression analysis for determining the sensor errors may be run independently and, if desired, at different periodic rates. This different periodicity may be beneficial when, for example, the time frame over which measurable sensor errors are likely to occur is much different, either greater than or less than, the time frame over which changes in the model parameters are likely to occur.
In any event, using these component models, the asset utilization expert 50 can perform asset performance monitoring by plotting the values of the determined model parameter(s) (and/or model inputs and outputs) versus time. Still further, the asset utilization expert 50 can detect potentially faulty sensors by comparing the determined sensor errors dY and dX to thresholds. If one or more of the sensors appears to have a high or an otherwise unacceptable error associated therewith, the asset utilization expert 50 can notify a maintenance person and/or a process control operator of the faulty sensor.
It will be understood from this discussion that the component models may be executed independently for different purposes at different times and, in many cases, may be executed periodically to perform the above described performance monitoring activities. Of course, the asset utilization expert 50 can control the execution of the appropriate models for the appropriate purposes and use the results of these models for asset performance monitoring and optimization. It will be understood that the same model may be run by the asset utilization expert 50 for different purposes and for calculating different parameters or variables associated with the model.
As noted above, the parameters, inputs, outputs or other variables associated with any particular model may be stored and tracked to provide performance monitoring for a device, a unit, a loop, an area or any other entity of a process or a plant. If desired, two or more of these variables may be tracked or monitored together to provide a multi-dimensional plot or measure of the performance of the entity. As part of this performance modeling, the location of the parameters or other variables within this multi-dimensional plot may be compared to thresholds to see if the entity, as defined by the coordinated parameters being monitored, is within a desired or acceptable region or is, instead, outside of that region. In this manner, the performance of an entity may be based on one or more parameters or other variables associated with that entity.
It will be understood that the asset utilization expert 50 can monitor one or more entities using the monitoring technique described above based on model parameters or other model variables and can report the operating states or performance measures of these entities to any other desired persons, functions or applications within the process plant 10, such as to a process control expert system, a maintenance person, a business application, a user interface routine 58, etc. Of course, it will also be understood that the asset utilization expert 50 may perform performance or condition monitoring on any desired entity, based on one, two, three or any other desired number of parameters or variables for each entity. The identity and number of variables or parameters to be used in this performance monitoring will generally be determined by an expert familiar with the process and will be based on the type of entity being monitored.
If desired, the asset utilization expert 50 may also define a performance index or plot by comparing one or more of the parameters determined by the models as described above with the same parameter determined by the model run in accordance with the design parameters of the entity being modeled. In particular, the asset utilization expert 50 may execute a model using the design parameters of the entity within the plant 10 to which the model pertains to determine what the designed performance of the entity would be if it was operating according to the current state of the process and using the actual inputs to the entity as measured within the plant 10. This design performance can then be compared to the actual performance of the entity as determined by the component model for that entity or as determined by the measured inputs and outputs of the entity to generate a measure of the performance of the entity.
Thus, for example, the efficiency of an entity may be determined using the component model which estimates the parameters of the entity (one of which may be efficiency) based on the regression analysis described above. At the same time, a model of the entity may be run using the parameters that would result according to the design criteria of the entity, but based on the actual inputs to and/or outputs from the entity. Thus, for example, if a different raw material was being input to the entity, the design model would be run using the efficiency that would result from the change in raw materials. The performance of the entity in both cases may be compared to determine a performance index which indicates how far from the possible or designed operation the actual entity is operating. This performance index can then be reported to and used by other applications or users of the system, such as a process control, a maintenance or a business person or application.
The component models 56 may also be used to perform process optimization. In particular, the asset utilization expert 50 may use one or more of the optimization routines 55 which execute the individual component models to optimize the operation of the plant in terms of some optimization criteria provided by, for example, a process control operator or a business person via a business application. The optimizer 55 can be a real time optimizer which operates in real time to optimize the plant 10 based on the actual state of the plant 10 at that time. Alternatively or additionally, an optimizer 55 may determine changes to be made to the plant 10, such as bringing certain devices or units back on line, that will provide the greatest optimization of the plant 10. Of course, other types of optimization routines 55 may be executed instead of or in addition to those mentioned here.
In one embodiment, the RTO+ real time optimization routine, which is provided by MDC Inc., may be used as a real time optimizer and may be executed at various or periodic times during operation of the plant 10, such as every 3-5 minutes, every 10-15 minutes, every hour, etc. Of course, other known or later developed optimizer routines which perform optimization less frequently, such as every 3 or 4 hours or every 3 to 4 days could be used instead.
The RTO+ optimizer implements three general phases each time it is executed to perform real time optimization. The RTO+ optimization routine first executes an input phase during which the routine checks to determine whether the variables that were previously indicated at the design of the optimizer as being variables which could be manipulated by the optimizer to perform optimization, such as set points or other inputs of various devices, units, etc., can actually be manipulated at the current time. This information may be available to the optimizer from the asset utilization expert 50 which obtains this information from the process control system and stores this information within any desired database. Thus, during the input phase, the optimizer actually determines, based on the data provided to it from the asset utilization expert 50, whether each of the possible manipulated inputs is still available to be changed because, in many instances, one or more of the potential manipulated inputs may not be available for change due to the fact that, for example, a device which provides that input is not operating or has been taken off-line or the device is being run in a mode other than the designed mode which prevents the controller from changing inputs to the device.
As part of the input phase, the real time optimizer may also determine if the variables that were supposed to change during the last run of the optimizer were actually changed to and reached the suggested or calculated values from the last run of the optimizer, i.e., the values to which they were supposed to be changed. If, for whatever reason, a variable that was supposed to change to a particular value did not reach that value, the optimizer recognizes that there is a problem that is preventing the change from occurring and, effectively, removes the option of changing this variable to that value during the next run of the optimizer. Detecting a failure of a variable to reach a value which it should have theoretically reached may also cause the optimizer to indicate to an operator that there may be a problem within the system that needs to be addressed.
Next, during the input phase, the optimizer performs a quick execution of each of the individual component models that make up the entire model using, for example, the actual inputs and outputs measured from the plant 10. The calculated outputs of each component model are then reviewed to see if there is any problem with any particular component model that will prevent the entire model from running accurately. Here, the optimizer may use the actual measured inputs for each entity (which have previously stored) to see if the individual component parts of the model work for these actual inputs to produce realistic outputs.
Assuming that each of the component models can be executed, the optimizer may look for discrepancies in the models which may effect the ability of the optimizer to optimize. For example, the optimizer may determine if a measurement made by a real device is the same as predicted by the component model using the actual inputs to the device. If the model (using the most recently calculated model parameters) predicts an output that deviates from the actual measurement of the output, such as if the model predicts a flow rate of 18 and a flow rate meter reads 20, then the optimizer may reset a constraint associated with the flow rate at two below the previously defined constraint. Thus, if the constraint associated with that flow rate was originally set at 25, the optimizer may use a constraint of 23 because the optimizer recognizes that its model is in error by two with respect to this variable. Of course, the optimizer may look for other inconsistencies or deviations between models and the actual measurements of the plant to reset, update or tweak constraints or other variables within the optimization routine.
In the next phase, known generally as the optimization phase, the optimizer runs the individual models in a predetermined order using the outputs from one component model as inputs to one or more of the other component models making up the entire model. Using the entire model, the constraints provided by the user and the new constraints determined by the input phase, as well as the optimization criteria, the optimizer determines the changes to be made to the input or manipulated variables which have been detected as currently being capable of manipulation which will optimize the plant over the time window in which the optimizer runs. This time window may be 3-4 minutes, 3-4 hours, etc., and is generally the periodic rate at which the optimizer runs. The use of optimization software is well known and any desired optimization software for this purpose could be used. Thus, in one example, a refining plant may be optimized to operate a C2 cracker to make as much profit as possible based on the possible output products that can be produced by the C2 cracker as determined by the current prices and lots of production associated with each of those possible outputs. A constraint, however, may be that the C2 cracker has to produce a certain amount of one product because there is a business contract to provide that amount of that product which must be fulfilled no matter what the current price of that product happens to be. In another example, an optimization criteria may be to maximize the use of one particular raw material because the cost of keeping excess amounts of that raw material at the plant overshadows other costs or price concerns, such as products that have the highest current price.
It will be seen that the determination of the optimization criteria, which is typically performed by a business person or a business application, is very critical to the operation of the optimizer and thus, ultimately, to the operation of the plant 10. As a result, the asset utilization expert 50 may provide the business person, via the user interface routines 58, a systematic set of choices of what the optimization criteria will be at any particular time and provides the choices made by the operator or any other user to the optimization routine. In fact, there are many optimization variables that can be selected and the choice of these different criteria may be provided to the operator or business person via the user interface to allow the operator or business person to choose different optimization criteria in any desired manner.
Next, the optimization routine enters an output phase in which implementation of the results of the optimizer may be accomplished. In particular, after computing the suggested changes to the manipulated variables, the optimizer may determine if the manipulated variables or inputs to be changed are still available because one or more of the devices which were available at the time the optimizer began the optimization phase may have gone off-line or may have otherwise become unavailable, which will prevent the suggested changes in the input variables from being implemented. However, if all of the manipulated variables to be changed can still be changed, the suggested changes may be provided to an operator via, for example, a user interface (e.g., a graphical user interface). The operator may be able to simply press a button and have the changes to the manipulated variables initiated or downloaded to the process control routine automatically, such as changing set points, etc. in a matter determined by the optimizer. In another embodiment or in later stages of operation, for example, when the process is running properly, the optimizer may automatically implement the suggested changes if the operator does not prevent the instantiation of the these changes within a particular time window. Thus, the output of the optimizer may be used every time the optimizer executes unless the operator intervenes to prevent the changes from the optimizer from being used. As part of this operation, one or more of the user interface routines 58 may provide a screen to the operator indicating the suggested changes to be made and a button or bar which the operator uses to install the changes or to prevent the changes from being installed. If, in one embodiment, the user pushes a button to install the changes, all of the changes are sent to the appropriate controllers where they are checked for limits and then implemented.
The real time optimizer, such as that described above, may be executed relatively frequently compared to most optimizers for a number of reasons. First, the real time optimizer uses fitness for purpose component models which typically run much faster than the highly theoretical models that are typically used to design a plant. These component models are accurate, however, because one or more parameters of the models are being tweaked or altered (using the regression analysis described above) based on actual inputs and outputs of the plant 10. That is, the real time optimizer uses the model parameters provided by the last run of the regression analysis on the model to reconcile the model with the actual operation of the plant 10. Still further, the optimizer described above is faster than conventional optimizers because it uses restricted optimization steps. That is, the optimizer only tries to optimize over the time period between individual runs of the optimizer, which reduces the amount of processing performed by the optimization routine. Still further, the real time optimizer may be configured to recognize the occurrence of one or more significant events, which may trigger the real time optimizer to restart as these events may make previous recommendations unfeasible or undesirable.
While the use of a closed loop, real time optimizer has been discussed above, other types of optimizers 55 which use the same or different component models could also be executed by the asset utilization expert 50 in conjunction with or separately from the real time optimizer. These other optimizers could be executed less frequently and for other reasons. For example, a wideband optimizer may be used to look at or determine where the final optimal operating point of a process may ultimately be, even though the real time optimizer may not be able to drive the plant 10 to that point for some time. This wideband optimizer may enable the business person to make long term predictions about the plant 10 or may enable the operator to determine if the operation of the plant 10 is headed toward a desirable region. If the wideband optimizer determines that the ultimately achievable optimization point is still not acceptable, the operator may decide to change the configuration or other operating parameters of the plant 10.
Other optimizers, such as selection optimizers, may determine whether changes in the process configuration that need to be performed by the operator or maintenance person could better optimize the process. For example, in some cases, a selection optimizer may recognize that certain units or other manipulated inputs that are supposed to be available to the real time optimizer are no longer available for some reason, e.g., the devices associated with these inputs are down or are off-line. The selection optimizer runs one or more optimization tests assuming that one or more of these devices, units, etc. are available, to determine how much better the plant 10 would operate (i.e., how much more optimal the plant 10 could be with respect to some optimization criteria) if these entities were put back into operation. This optimizer may, for example, tell the operator or business person how much more money the plant 10 could make by getting certain units or devices online, or could tell the operator or business person which devices or units to focus on getting back into operation first. Such a selection optimizer may also try to optimize by turning particular pumps or valves on or off, by replacing other devices that are running in sub-optimal modes, etc. to determine which critical changes to the process or to the assets thereof could be made to make the process more profitable or optimal. The selection optimizer might use inputs from the operator or business person and/or might use other branch and bound techniques which are common in data processing or data mining routines to select the way in which to adjust the optimization variables. Other selection techniques could also be used, such as providing the selection optimizer with a series of rules to apply in a particular order to determine how to change the process or to determine which changes to the process, if implemented, would provide improvements or the most improvement to the plant 10.
As a result of the above discussion, it can be seen that the use of models provides many new types of data or information for the business applications, process control applications and asset maintenance and monitoring applications. In particular, the models can be used to perform performance monitoring and to produce a performance index which indicates the relative performance of a device, unit, area, etc. within a plant. This performance index may be a measure of the performance of an entity with respect to the possible performance of that entity. Furthermore, while device and unit models have been discussed above, similar models could be made and executed for process control entities, such as loops, units, etc. to provide performance measures and optimization criteria for these types of entities as well. Also, as indicated above, models may, in some cases, be used to measure or indicate the health of certain devices or other entities and to provide a health index indicative of these entities. For example, the error measurements of certain input and output sensors as determined by the regression analysis used on certain models may be used as or converted into an indication of the health of those devices. Also, other information not otherwise available to the process controller, such as model parameters and virtual sensor measurements based on the models could be provided to the process controllers or to the business persons for use in numerous manners.
Besides performance and health indices, the asset utilization expert 50 can assist the index generation routine 51 in creating other types of indices such as a utilization index and a variability index. A variability index indicates how much some signal into or out of, or some other parameter associated with a device, loop, unit, etc. varies as compared to how much this signal or parameter is expected to vary. The data needed to create this variability index may be collected by the asset utilization expert 50 and provided to the index generation routine 51 at any desired or convenient times. Of course, the normal amount of variation of a signal or parameter may be set by a manufacturer, engineer, operator or maintenance person familiar with the entity or may be based on a statistical measure (such as an average, standard deviation, etc.) associated with that or other similar entities within the plant and this normal or expected variation may be stored by or updated within the index generation routine 51.
The utilization index, in one form or another, tracks or reflects the utilization of individual devices, units, loops, or other entities and may provide some indication as to whether these entities are being over utilized or under utilized based on previously determined bench marks or operational goals. A utilization index can be generated based on measured uses of the actual device. For example, a device may be measured as to how often it is being used within a process or is being left idle and this index may be compared to a desired utilization for that entity to determine if the entity is being over or under utilized. The utilization index might identify devices, units, loops, etc. which are not being utilized as often as they could be or should be or, on the other hand, which are being utilized too much and, thus, are being over used.
In some instances, a utilization index might be determined based on the business decisions made regarding the appropriate or desired use of a particular device. For example, many processing or refinery plants use furnaces which are subject to the problem of coking and, as a result, must be periodically cleaned. Generally speaking, coking occurs in furnaces that are used in petroleum industries where it is desirable to rapidly heat a petroleum based fluid in a short amount of time. Such a furnace might have an input tube which travels through the interior of the furnace so that fluid traveling within the tube is rapidly heated to a high temperature, such as 1400° F. In some cases, to get the fluid at the center of the tube to such a high temperature, it may be necessary to heat the outside of the tube to about 1700° F. However, because the fluid traveling through the tube is petroleum based, some of the material forms into carbon or other kinds of viscous or particulate material which is deposited on the inside surfaces of the tubes which, in turn, reduces the heat transfer efficiency of the furnace over time. The deposit of this material is generally called coking. The more fluid through the tube, the more coking that occurs and, thus, the lower the efficiency of the furnace. At some point, the efficiency of the furnace becomes so low that the furnace must be taken off-line and cleaned. This cleaning process is time consuming and requires a great deal of manpower and resources, including steam cleaning equipment. As a result, the cleaning of these furnaces must generally be scheduled in advance to assure that the manpower and necessary equipment is available.
A business person or process control operator, knowing the fact that particular furnaces within the plant 10 will experience coking, may attempt to obtain the most efficient use of those furnaces by specifying how coking is to occur within each furnace over time. Thus, for example, a business person might decide to operate a particular furnace for a period of 15 days and wants to get the most use out of that furnace in that time period by running the furnace to its maximum coking level at the end of the 15 day period. If the furnace reaches its maximum coking level too soon, i.e., before the end of the 15 day period, the furnace must be taken off-line but still cannot be cleaned until the end of the 15 day period because that is when the cleaning manpower and machinery becomes available (i.e., has been scheduled). This early removal of the furnace may adversely affect the process operation. However, if the furnace does not reach its maximum allowable coking level by the end of the 15 day period, the furnace must still be taken off-line and cleaned because, again, that is the only time that the manpower and machinery is available. However, the furnace has been underused which means that an opportunity has been lost, because it effectively costs just as much to clean an underutilized furnace as it does to clean a fully utilized furnace. Presumably, other furnaces were being overused and subject to more coking than necessary based on the under use of this furnace.
To plan the best utilization of the furnace over a given time period, such as the 15 day period discussed above, the operator or business person may draw a fiducial line which defines the desired amount of coking for the furnace over time. An exemplary fiducial line 200 in a plot of coking of a furnace versus time is illustrated in FIG. 6. The fiducial line 200 specifies how the business person or operator desires to use the furnace with respect to coking over a period of 15 days (after which time this furnace will be cleaned). The coking of the furnace is directly related to the heat transfer efficiency of the furnace. As can be see from
Unfortunately, it is impossible, due to the high temperatures within the furnace, to measure the coking in the furnace at any particular time. However, a measure of coking within the furnace may be made, instead, with the use of one or more models for the furnace constructed according to the principles discussed above. Here, coking may be a parameter of the model which is determined from other variables within the model, inputs to and outputs from the model or using a regression analysis as described above. The asset utilization expert 50 may then run this model periodically to obtain, as a virtual senor measurement or as a model parameter, the amount of coking in the furnace. The equations for this determination are well known in the art and will not be discussed herein. Furthermore, if desired, an optimizer may use the coking value defined by the fiducial line 200 of
Now, in the example of
Based on this utilization index, the business person or operator may recognize that there is an opportunity being missed to use the furnace more because, after the period of 15 days, the furnace will be shut down anyway whether or not the maximum amount of coking has occurred within the furnace. To get the coking parameter back on to the fiducial line 200 within a certain period of time such as, for example, by day 7 (which is marked as point 204 in FIG. 6), and to get the best utilization out of the furnace, the operator or the asset utilization expert 50 may define a line between the points 202 and 204. The slope of this line defines a rate of coking which is greater than the rate of coking allowed by the fiducial line 200 between days 6 and 7. The operator or optimization routine may then run the furnace more to obtain the rate of coking defined by the slope of the line between the points 202 and 204. This higher rate of coking may, for example, be used as a constraint, or even as an optimization parameter, within an optimization routine which controls the operation of the plant 10 or of the furnace. For example, and as is known, the amount or rate of coking can be adjusted by adjusting any of a number of parameters associated with the furnace, such as the speed that material is provided through the furnace (the faster the speed, the less coking that takes place), the exit temperature of the material out of the furnace (the higher the exit temperature, the more coking that takes place) and the amount of steam injected into the fluid sent through the furnace (typically, the more steam that is used, the less coking that takes place). The optimizer may use the desired rate of coking to adjust one or more of these parameters to obtain the new coking rate. In any event, the measurement of the utilization of the furnace based on the actual coking of the furnace compared to the desired coking of the furnace indicates, in this example, that the furnace is being under utilized and that greater optimization of the plant 10 can occur if the furnace is used more.
Referring now to
Other business decisions can come into play in terms of changing the utilization index of the furnace. For example, at day 6, even though the furnace may be running at or around the fiducial line in terms of the coking parameter, a business person may decide that it is necessary to get five more days use out of the furnace and thereby increase the life of the furnace to 20 days before it is cleaned. This may be done because the cleaning equipment will not be available until 20 days anyway. In this case, the fiducial line between day 6 and day 20 may be redrawn and the control routine or optimizer may then attempt to follow the new fiducial line. However, if the entire fiducial line is redrawn, what was previously being defined as 100% utilization or exact utilization at day 6 for a 15 day fiducial line may now be over utilization for a 20 day fiducial line. In this case, the asset utilization expert 50 may provide a new rate of coking as an optimization or constraint to an optimizer or to a process control routine to attempt to get the utilization index back to 100%. Similarly, if the furnace is to be shut down early such as at the end of 12 days, the furnace may now be under utilized based on a 12 day fiducial line and an optimizer or a process control routine may use a new coking rate in an attempt to get the utilization index back to 100%.
As will be noted, a business decision to run the furnace on a particular fiducial line in conjunction with the use of models for the furnace which measure or estimate the coking amount of the furnace at any particular time can be used to indicate whether a particular entity, such as the furnace, is being over utilized or under utilized during actual process runs. Furthermore, changes to the control of the process can be made to use the furnace more or less, based on the utilization index.
The utilization index for the furnace example given above can be calculated or expressed in any desired manner, such as one based on the difference between or a ratio of the actual coking value and the desired coking value, the new allowable coking rate, the difference between or a ratio of the desired coking rate as defined by the fiducial line and the new allowable coking rate, or any other measure of utilization. Of course, while one manner of determining a utilization index for a furnace has been described herein, there are many other manners of determining or defining utilization indices for furnaces, as well as for other devices, units, loops, etc. within a process plant. In particular, utilization indices for different entities can be measured in any desired manner and in different manners for different types of entities. In one example, the utilization index may be expressed as a percent, where 100% means that the entity is being used the correct or desired amount, a value over 100% means that the entity is being over utilized and a value under 100% means that the entity is being under utilized. Of course, other ways of measuring and expressing utilization could also be used within the context of different kinds of devices.
One important aspect of the system of
Preferably, but not necessarily, the GUI may be implemented using a familiar graphical windows-based structure and appearance, in which a plurality of interlinked graphical views or pages include one or more pull-down menus that enable a user to navigate through the pages in a desired manner to view and/or retrieve a particular type of information. The features and/or capabilities of the asset utilization expert 50 described above may be represented, accessed, invoked, etc. through one or more corresponding pages, views or displays of the GUI. Furthermore, the various displays making up the GUI may be interlinked in a logical manner to facilitate a user's quick and intuitive navigation through the displays to retrieve a particular type of information or to access and/or invoke a particular capability of the asset utilization expert 50.
Generally speaking, the GUI described herein provides intuitive graphical depictions or displays of process control areas, units, loops, devices, etc. Each of these graphical displays may include numerical status and performance indices (some or all of which may be generated by the index generator routine 51 described above) that are associated with a particular view being displayed by the GUI. For example, a display depicting a process control area may provide a set of indices reflecting the status and performance of that area (i.e., a particular portion of the process control system at a particular level of the equipment hierarchy). On the other hand, a display depicting a loop may provide a set of status and performance indices associated with that particular loop. In any event, a user may use the indices shown within any view, page or display to quickly assess whether a problem exists within any of the devices, loops, etc. depicted within that display.
Additionally, the GUI described herein may automatically, or may in response to a request by a user, provide maintenance information to the user. The maintenance information may be provided by any portion of the asset utilization expert 50. Similarly, the GUI may display alarm information, process control information, etc., which may also be provided by the asset utilization expert 50. Still further, the GUI may provide messages to the user in connection with a problem that has occurred or which may be about to occur within the plant 10. These messages may include graphical and/or textual information that describes the problem, suggests possible changes to the system which may be implemented to alleviate a current problem or which may be implemented to avoid a potential problem, describes courses of action that may be pursued to correct or to avoid a problem, etc.
Importantly, the GUI display shown in
Before discussing the GUI and the manner in which asset information, process control information, maintenance information, diagnostic information or any other type of information is displayed to a user thereby, a brief discussion of the manner in which the performance and status indices are generated is provided below. Also, it should be recognized that while a performance index, a health index, a variability index and a utilization index are described in detail herein in connection with the various displays of the GUI, additional and/or different indices may be generated by the asset utilization expert 50 and displayed via the GUI without departing from the scope of the invention.
In general, each of the indices generated by the index generator routine 51 and displayed via the GUI may be calculated for individual devices, for logical and/or physical groupings of devices, for logical processes (e.g., control loops), for logical groupings of devices such as units and areas, etc. In other words, the indices may, in principal, be calculated at each level of the equipment and logical hierarchy of a process control system or, more generally, an asset utilization system, which may include one or more process control systems. However, the meaning of a particular index may depend on the context (i.e., whether the index corresponds to a logical or a physical grouping of devices and/or parameters) in which the index is generated and displayed and may depend on the level of the hierarchy at which it is displayed. For example, at the lowest level of the equipment hierarchy, indices correspond to physical devices such as valves, temperature sensors, actuators, etc. Thus, each device may have a unique set of indices that may be generated within the device or for the device based on information stored within the device at the time the device is manufactured. Accordingly, each device may generate and provide its indices to higher levels of the hierarchy and to the asset utilization expert 50 as needed.
Similarly, units or loops, each of which is composed of one or more devices or function blocks may each have a unique set of indices. However, the index values for each unit or loop may be generated by mathematically combining the index values for the individual devices or function blocks used within the unit or loop. Thus, if a unit or loop is composed of a pressure transmitter, a valve and a pump (or function blocks associated with the operation of these devices), the index values for the unit or loop may be based on various mathematical combinations of the index values generated for or by each of those devices or function blocks making up the unit or the loop. Likewise, because sub-unit and unit levels of the hierarchy are composed of one or more loops which, in turn, are composed of devices, the index values for each sub-unit and unit may be generated by mathematically combining loop or device index values. Still further, area indices may be determined as combinations of the units, loops, devices, etc. within the area.
As discussed in greater detail below, the mathematical combination of device index values to form index values for loop, sub-unit, unit and area levels of the hierarchy may use weighted summations or averages, or any other suitable mathematical combination. Of course, the calculation of one or more of the performance, health, variability and utilization indices may not be appropriate, required or useful for every level of the logical and equipment hierarchies.
Further,
The HI for a loop, on the other hand, is preferably, but not necessarily, a mathematical combination (such as, for example, a weighted summation or average) of the HI values for the individual devices or functions blocks that make up the loop. Likewise, the HI values for the sub unit and unit levels may also be a mathematical combination of the underlying HI values for loops and sub units. Thus, ultimately, the HI values hierarchy for levels above the device level are based on one or more HI values for devices that have been formed into composite values.
As is also shown in
In a similar manner, the HI for the unit 500 may be calculated as a weighted average of the HI values for all of the devices (and/or loops) that make up the unit 500. A table such as that shown in
Still further, the GUI may also provide textual messages within the graphical display shown in
It will also be understood that a user may view successively lower and lower entities within a plant and be provided information about the indices associated with each of these different entities or views. Thus, for example, a user may look at a view of the plant and see a particular set of indices for the plant. The user may then focus on one area, such as by clicking on one of the areas within the plant view, and see the indices associated with that area. Similarly, by clicking on units within the displayed area, the indices for different units may be viewed. Likewise indices for loops, sub units, devices etc. may then be viewed by focusing in on these different entities from a view of an entity in which these entities are located. In this manner, a user can quickly find the cause of a lower than (or higher than) expected index at any point or level of the plant.
Still further, the work order generation routine 54 could include a business to business communication function that, based on detected actual or predicted problems within the plant 10, will automatically communicate with a supplier or other business to order parts, supplies, etc. with or without operator or maintenance person intervention. More particularly, the routine 54 can receive notifications of current problems or predicted future problems with devices or other assets based on data provided by or predictions made by the asset utilization expert 50 or any of the data analysis tools, such as the rotating equipment analysis tools. The routine 54 then automatically contacts a supplier via, for example an internet, telephone, wireless or other communication connection and orders the parts, equipment or supplies to be delivered to the plant 10 before the device needs to be replaced. In this manner, the work order generation routine 54 limits the down time or helps to assure that there is little or no down time caused by the need to wait for parts, equipment or supplies to fix the problem when it actually occurs. This fact, then makes the plant 10 more efficient.
Referring now to
Referring now to
As illustrated in
During operation, any of the processes 900-903, may at convenient times collect input and output data associated with the process and provide such data to the remote monitoring facility 910 via one of the servers 900A-903A and the world wide web, internet or other communication network connected to the server 912. Upon receipt of the data from a plant, an appropriate one of the processors 914 accesses that data and executes the appropriate process monitoring and condition monitoring tools for that plant to detect problems within the plant based on the collected data, to provide condition, plant or process monitoring for the plant, or to perform optimization for the plant. Of course, the data collected at the plant and sent to the remote monitoring facility 910 is data previously determined to be necessary to run the desired models 916, optimizers 918 or other data analysis tools 920 and is collected and sent to the facility 910 at a periodic or non-periodic rate that is appropriate for the tools or models being executed. Thus, for optimizers, data may need to be collected and sent at a different rate than for models or for performance, process or asset monitoring tools. Of course, as part of an optimization or performance, condition or process monitoring exercise, any appropriate models or other tools may be executed and the execution of these models or other tools generally follows the principles discussed above with respect to these same tools in the plant 10 of FIG. 1.
In any event, after executing the models, data analysis or optimizer tools, the processor 914 places the results back on the server 912 where these results can be picked up by the appropriate one of the plants 900-903 at any desired time. Alternatively or additionally, these results may be sent directly to the appropriate one of the plants 900-903 by the server 912. The data resulting from the analysis can be any desired performance modeling data, plots or charts including, for example, those described above with respect to the user interface routines or GUI routine 58. The results could also be suggestions by, for example, an optimizer for making changes to the plants, indices for the plants, or any other results capable of being provided by these types of tools.
In one embodiment, the a real time optimizer, such as that described above, may be executed on a real time basis, assuming that the plants 900-903 provide sufficient data in a timely periodically manner so as to enables the proper execution of this optimizer. If desired, the servers 900A-903A may automatically collect and send the appropriate data to enable proper operation of the optimizer. In one embodiment, the plants may include the asset utilization expert 50 described herein or any other expert data collection tools for use in assuring that the proper data is sent to the remote monitoring facility 910 in a timely or periodic manner.
In this manner, the remote monitoring facility 910 can execute the software for asset, performance, condition and process monitoring as well as executing one or more optimizers for different plants. This, in turn, means that the plants 900-903 do not need to include the processing power or applications for these purposes, which may be less expensive for the plants. Of course, the plants may pay on a per usage basis or on some other predetermined fee schedule for use of the remote monitoring facility 910. If desired, the remote monitoring facility 910 may contract to get a portion of the profits and/or losses of the plant based on the use of the tools at the facility 910 and the implementation of the results of those tools.
If desired, each of the plants 900-903 may update the models 916 stored within the remote monitoring facility 910 applicable to those plants by sending new or updated models to the server 912 using any desired communication format such as XML, HTML, etc. Still further, the remote monitoring facility 910 may include generic templates for different process plants, areas, units, devices, loops, etc. which may be downloaded to each of the plants 900-903 via the server 912 and these templates may be altered at the plants 900-903 to reflect the actual operation of those plants. The updated models may then be sent back to the remote monitoring facility 910 as models to be implemented in asset, condition or process monitoring or in the optimizers for the plant. In this manner, changes to the plants 900-903 may be adequately or accurately reflected within the remote monitoring facility 910.
While the asset utilization expert 50 and other process elements have been described as preferably being implemented in software, they may be implemented in hardware, firmware, etc., and may be implemented by any other processor associated with the process control system 10. Thus, the elements described herein may be implemented in a standard multi-purpose CPU or on specifically designed hardware or firmware such as an application-specific integrated circuit (ASIC) or other hard-wired device as desired. When implemented in software, the software routine may be stored in any computer readable memory such as on a magnetic disk, a laser disk, or other storage medium, in a RAM or ROM of a computer or processor, in any database, etc. Likewise, this software may be delivered to a user or a process plant via any known or desired delivery method including, for example, on a computer readable disk or other transportable computer storage mechanism or over a communication channel such as a telephone line, the internet, wireless communication, etc. (which are viewed as being the same as or interchangeable with providing such software via a transportable storage medium). Also, while the expert 50 is described as possibly being a rule-based expert, other types of expert engines could be used as well, including those which use other known data mining techniques.
Thus, while the present invention has been described with reference to specific examples, which are intended to be illustrative only and not to be limiting of the invention, it will be apparent to those of ordinary skill in the art that changes, additions or deletions may be made to the disclosed embodiments without departing from the spirit and scope of the invention.
This patent claims priority from U.S. Provisional Application Ser. No. 60/273,164 which was filed on Mar. 1, 2001.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4527271 | Hallee et al. | Jul 1985 | A |
| 4607325 | Horn | Aug 1986 | A |
| 4657179 | Aggers et al. | Apr 1987 | A |
| 4734873 | Malloy et al. | Mar 1988 | A |
| 4763243 | Barlow et al. | Aug 1988 | A |
| 4764862 | Barlow et al. | Aug 1988 | A |
| 4885694 | Pray et al. | Dec 1989 | A |
| 4907167 | Skeirik | Mar 1990 | A |
| 4910691 | Skeirik | Mar 1990 | A |
| 4944035 | Aagardl et al. | Jul 1990 | A |
| 4956793 | Bonne et al. | Sep 1990 | A |
| 4965742 | Skeirik | Oct 1990 | A |
| 5006992 | Skeirik | Apr 1991 | A |
| 5008810 | Kessel et al. | Apr 1991 | A |
| 5015934 | Holley et al. | May 1991 | A |
| 5018215 | Nasr et al. | May 1991 | A |
| 5043863 | Bristol et al. | Aug 1991 | A |
| 5050095 | Samad | Sep 1991 | A |
| 5070458 | Gilmore et al. | Dec 1991 | A |
| 5121467 | Skeirik | Jun 1992 | A |
| 5134574 | Beaverstock et al. | Jul 1992 | A |
| 5140530 | Guha et al. | Aug 1992 | A |
| 5142612 | Skeirik | Aug 1992 | A |
| 5161013 | Rylander et al. | Nov 1992 | A |
| 5167009 | Skeirik | Nov 1992 | A |
| 5187674 | Bonne | Feb 1993 | A |
| 5193143 | Kaemmerer et al. | Mar 1993 | A |
| 5197114 | Skeirik | Mar 1993 | A |
| 5212765 | Skeirik | May 1993 | A |
| 5224203 | Skeirik | Jun 1993 | A |
| 5282261 | Skeirik | Jan 1994 | A |
| 5287505 | Calvert et al. | Feb 1994 | A |
| 5291190 | Scarola et al. | Mar 1994 | A |
| 5301101 | MacArthur et al. | Apr 1994 | A |
| 5311447 | Bonne | May 1994 | A |
| 5311562 | Palusamy et al. | May 1994 | A |
| 5333298 | Bland et al. | Jul 1994 | A |
| 5347449 | Meyer et al. | Sep 1994 | A |
| 5351184 | Lu et al. | Sep 1994 | A |
| 5353207 | Keeler et al. | Oct 1994 | A |
| 5369599 | Sadjadi et al. | Nov 1994 | A |
| 5373452 | Guha | Dec 1994 | A |
| 5384698 | Jelinek | Jan 1995 | A |
| 5390326 | Shah | Feb 1995 | A |
| 5396415 | Konar et al. | Mar 1995 | A |
| 5398303 | Tanaka | Mar 1995 | A |
| 5408406 | Mathur et al. | Apr 1995 | A |
| 5428679 | French | Jun 1995 | A |
| 5442544 | Jelinek | Aug 1995 | A |
| 5486920 | Killpatrick et al. | Jan 1996 | A |
| 5486996 | Samad et al. | Jan 1996 | A |
| 5488697 | Kaemmerer et al. | Jan 1996 | A |
| 5515266 | Meyer | May 1996 | A |
| 5533093 | Horton et al. | Jul 1996 | A |
| 5537310 | Tanake et al. | Jul 1996 | A |
| 5541833 | Bristol et al. | Jul 1996 | A |
| 5546301 | Agrawal et al. | Aug 1996 | A |
| 5559690 | Keeler et al. | Sep 1996 | A |
| 5561599 | Lu | Oct 1996 | A |
| 5566065 | Hansen et al. | Oct 1996 | A |
| 5570282 | Hansen et al. | Oct 1996 | A |
| 5572420 | Lu | Nov 1996 | A |
| 5574638 | Lu | Nov 1996 | A |
| 5596704 | Geddes et al. | Jan 1997 | A |
| 5608845 | Ohtsuka et al. | Mar 1997 | A |
| 5640491 | Bhat et al. | Jun 1997 | A |
| 5640493 | Skeirik | Jun 1997 | A |
| 5666297 | Britt et al. | Sep 1997 | A |
| 5680409 | Qin et al. | Oct 1997 | A |
| 5687090 | Chen et al. | Nov 1997 | A |
| 5687224 | Alley et al. | Nov 1997 | A |
| 5692158 | Degeneff et al. | Nov 1997 | A |
| 5704011 | Hansen et al. | Dec 1997 | A |
| 5715158 | Chen | Feb 1998 | A |
| 5729661 | Keeler et al. | Mar 1998 | A |
| 5740324 | Mathur et al. | Apr 1998 | A |
| 5742513 | Bouhenguel et al. | Apr 1998 | A |
| 5761518 | Boehling et al. | Jun 1998 | A |
| 5764891 | Warrior | Jun 1998 | A |
| 5777872 | He | Jul 1998 | A |
| 5781432 | Keeler et al. | Jul 1998 | A |
| 5790898 | Kishima et al. | Aug 1998 | A |
| 5796609 | Tao et al. | Aug 1998 | A |
| 5798939 | Ochoa et al. | Aug 1998 | A |
| 5809437 | Breed | Sep 1998 | A |
| 5809490 | Guiver et al. | Sep 1998 | A |
| 5819050 | Boehling et al. | Oct 1998 | A |
| 5819232 | Shipman | Oct 1998 | A |
| 5825645 | Konar et al. | Oct 1998 | A |
| 5826249 | Skeirik | Oct 1998 | A |
| 5842189 | Keeler et al. | Nov 1998 | A |
| 5847952 | Samad | Dec 1998 | A |
| 5859773 | Keeler et al. | Jan 1999 | A |
| 5859964 | Wang et al. | Jan 1999 | A |
| 5873009 | Yamashita et al. | Feb 1999 | A |
| 5877954 | Klimasauskas et al. | Mar 1999 | A |
| 5892679 | He | Apr 1999 | A |
| 5892939 | Call et al. | Apr 1999 | A |
| 5893906 | Daffin et al. | Apr 1999 | A |
| 5898869 | Anderson | Apr 1999 | A |
| 5901058 | Steinman et al. | May 1999 | A |
| 5905989 | Biggs | May 1999 | A |
| 5907701 | Hanson | May 1999 | A |
| 5909370 | Lynch | Jun 1999 | A |
| 5909541 | Sampson et al. | Jun 1999 | A |
| 5909586 | Anderson | Jun 1999 | A |
| 5917840 | Cheney et al. | Jun 1999 | A |
| 5918233 | La Chance et al. | Jun 1999 | A |
| 5924086 | Mathur et al. | Jul 1999 | A |
| 5940290 | Dixon | Aug 1999 | A |
| 5948101 | David et al. | Sep 1999 | A |
| 5949417 | Calder | Sep 1999 | A |
| 5960214 | Sharpe, Jr. et al. | Sep 1999 | A |
| 5960441 | Bland et al. | Sep 1999 | A |
| 5984502 | Calder | Nov 1999 | A |
| 5988847 | McLaughlin et al. | Nov 1999 | A |
| 6006199 | Berlin et al. | Dec 1999 | A |
| 6008985 | Lake et al. | Dec 1999 | A |
| 6017143 | Eryurek et al. | Jan 2000 | A |
| 6033257 | Lake et al. | Mar 2000 | A |
| 6041263 | Boston et al. | Mar 2000 | A |
| 6047221 | Piche et al. | Apr 2000 | A |
| 6055483 | Lu | Apr 2000 | A |
| 6067505 | Bonoyer et al. | May 2000 | A |
| 6076124 | Korowitz et al. | Jun 2000 | A |
| 6078843 | Shavit | Jun 2000 | A |
| 6093211 | Hamielec et al. | Jul 2000 | A |
| 6106785 | Havlena et al. | Aug 2000 | A |
| 6108616 | Borchers et al. | Aug 2000 | A |
| 6110214 | Klimasaukas | Aug 2000 | A |
| 6122555 | Lu | Sep 2000 | A |
| 6128279 | O'Neil et al. | Oct 2000 | A |
| 6144945 | Garg et al. | Nov 2000 | A |
| 6144952 | Keeler et al. | Nov 2000 | A |
| 6169980 | Keeler et al. | Jan 2001 | B1 |
| 6246325 | Chittipeddi | Jun 2001 | B1 |
| 6259956 | Myers et al. | Jul 2001 | B1 |
| 6298377 | Hartikainen et al. | Oct 2001 | B1 |
| 6298454 | Schleiss et al. | Oct 2001 | B1 |
| 6317701 | Pyotsia et al. | Nov 2001 | B1 |
| 6321348 | Kobata | Nov 2001 | B1 |
| 6397114 | Eryurek et al. | May 2002 | B1 |
| 6411678 | Tomlinson et al. | Jun 2002 | B1 |
| 6421571 | Spriggs et al. | Jul 2002 | B1 |
| 6445963 | Blevins et al. | Sep 2002 | B1 |
| 6460070 | Turek et al. | Oct 2002 | B1 |
| 6557118 | Schleiss et al. | Apr 2003 | B2 |
| 6584432 | Holzinger et al. | Jun 2003 | B1 |
| 6587879 | Reynolds | Jul 2003 | B1 |
| 6614882 | Beamon et al. | Sep 2003 | B1 |
| 6633782 | Schleiss et al. | Oct 2003 | B1 |
| 6643608 | Hershey et al. | Nov 2003 | B1 |
| 6654801 | Mann et al. | Nov 2003 | B2 |
| 6721609 | Wojsznis et al. | Apr 2004 | B1 |
| 6738811 | Liang | May 2004 | B1 |
| 6751525 | Crisp, III | Jun 2004 | B1 |
| 6774786 | Havekost et al. | Aug 2004 | B1 |
| 6809292 | Spear et al. | Oct 2004 | B2 |
| 20020077711 | Nixon et al. | Jun 2002 | A1 |
| 20020082753 | Guskov et al. | Jun 2002 | A1 |
| 20020099638 | Coffman et al. | Jul 2002 | A1 |
| 20020133254 | Elwood et al. | Sep 2002 | A1 |
| Number | Date | Country |
|---|---|---|
| 0 467 257 | Jan 1992 | EP |
| 0 509 817 | Oct 1992 | EP |
| 0 626 697 | Nov 1994 | EP |
| 0 965 897 | Dec 1999 | EP |
| 2001-318966 | Nov 2001 | JP |
| 2002-049799 | Feb 2002 | JP |
| 2002-092205 | Mar 2002 | JP |
| WO 9116481 | Oct 1991 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 20020169514 A1 | Nov 2002 | US |
| Number | Date | Country | |
|---|---|---|---|
| 60273164 | Mar 2001 | US |
