CA2118885C - Process control system - Google Patents
Process control system Download PDFInfo
- Publication number
- CA2118885C CA2118885C CA002118885A CA2118885A CA2118885C CA 2118885 C CA2118885 C CA 2118885C CA 002118885 A CA002118885 A CA 002118885A CA 2118885 A CA2118885 A CA 2118885A CA 2118885 C CA2118885 C CA 2118885C
- Authority
- CA
- Canada
- Prior art keywords
- economic
- obtaining
- variables
- level
- independent
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01J—CHEMICAL OR PHYSICAL PROCESSES, e.g. CATALYSIS OR COLLOID CHEMISTRY; THEIR RELEVANT APPARATUS
- B01J19/00—Chemical, physical or physico-chemical processes in general; Their relevant apparatus
- B01J19/0006—Controlling or regulating processes
- B01J19/0033—Optimalisation processes, i.e. processes with adaptive control systems
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
- G05B13/027—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion using neural networks only
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y10—TECHNICAL SUBJECTS COVERED BY FORMER USPC
- Y10S—TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y10S706/00—Data processing: artificial intelligence
- Y10S706/902—Application using ai with detail of the ai system
- Y10S706/903—Control
- Y10S706/906—Process plant
Abstract
A process control system uses a performance model, in conjunction with an economic model, to meet either performance or economic objectives. The system operates either automatically, on-line, in "real time" or manually off-line, in "real time" to optimize the process to meet a selected objective. The performance model uses neural networks to determine the interrelationships of each of a set of independent variables (controllable inputs to the process) to each other and their relationship to a set of dependent variables (outputs of the process). The performance model is used to optimize the process to meet performance objectives. The economic model is used in conjunction with the performance model to meet economic objectives.
Description
z~t~$es PROCESS CONTROL SYSTEM
Be it known that we, Conrad R. Teran and Delbert Grotewold, citizens of the United States of America, residing in the City of Puerto,La Cruz, Country of Venezuela, and City of The Woodlands, County of Montgomery, State of Texas, respectively, have made a certain new and useful invention in a process control system of which the following is a specification.
SPECIFICATION
BACKGROUND OF THE INVENTION
This invention relates generally to process control systems and more particularly to a system which uses a performance model in real time as well as an economic model to control processes.
. , 211888 At the present time, the operations of industrial processes are analyzed using regression analysis, which applies curve fitting to historical data. In regression analysis, only a single dependent variable can be studied with relation to a set of independent variables. In this type of analysis, the interaction between the independent variables is not taken into account.
Neural networks, which determine the interaction between independent variables, are also used to analyze the operation of existing systems. As is known by those familiar with the art, neural networks solve problems by attempting to simulate the operation of the human brain. The brain has approximately 100 billion cells of more than 100 types, and many of the cells have more than one thousand interconnections to other cells. It can easily outperform any of the current approaches with respect to recognizing patterns and can quickly determine reasonable solutions to complex problems.
Neural network models use groups of "neurons" to emulate brain cells. The neurons are arranged in structural units or layers. In feed forward networks, each of the neurons of a layer gets an input from a preceding layer and sends an output to a succeeding one. Each neuron in a layer is connected to every other neuron of the succeeding layer. Another type of neural network has feedback from layers to preceding layers.
As is the case with the human brain, input, hidden and output layers of neurons are used. The hidden layer can be a single layer of neurons or can comprise multiple layers of neurons..
Be it known that we, Conrad R. Teran and Delbert Grotewold, citizens of the United States of America, residing in the City of Puerto,La Cruz, Country of Venezuela, and City of The Woodlands, County of Montgomery, State of Texas, respectively, have made a certain new and useful invention in a process control system of which the following is a specification.
SPECIFICATION
BACKGROUND OF THE INVENTION
This invention relates generally to process control systems and more particularly to a system which uses a performance model in real time as well as an economic model to control processes.
. , 211888 At the present time, the operations of industrial processes are analyzed using regression analysis, which applies curve fitting to historical data. In regression analysis, only a single dependent variable can be studied with relation to a set of independent variables. In this type of analysis, the interaction between the independent variables is not taken into account.
Neural networks, which determine the interaction between independent variables, are also used to analyze the operation of existing systems. As is known by those familiar with the art, neural networks solve problems by attempting to simulate the operation of the human brain. The brain has approximately 100 billion cells of more than 100 types, and many of the cells have more than one thousand interconnections to other cells. It can easily outperform any of the current approaches with respect to recognizing patterns and can quickly determine reasonable solutions to complex problems.
Neural network models use groups of "neurons" to emulate brain cells. The neurons are arranged in structural units or layers. In feed forward networks, each of the neurons of a layer gets an input from a preceding layer and sends an output to a succeeding one. Each neuron in a layer is connected to every other neuron of the succeeding layer. Another type of neural network has feedback from layers to preceding layers.
As is the case with the human brain, input, hidden and output layers of neurons are used. The hidden layer can be a single layer of neurons or can comprise multiple layers of neurons..
2~.:~~~~a Each of the neurons of the input layer receives a single input (independent variable). A weighting function is associated with, and determines the level of, each of the inputs to an input neuron.
Each neuron in the output layer represents a dependent variable whose value changes due to changes in the values of the independent variables and the interactions between the independent variables. The output of each neuron is modified by a transfer function. Various types of transfer functions, such as linear, linear threshold (which sets a maximum and minimum level to the function), step, Sigmoid (S-shaped), hyperbolic tangent or Gaussian functions can be used.
Historical data representing the independent input variables of a system are provided to the neural network model and the dependent variables (outputs) of the system are determined and compared to the actual output of the system at the time. A
learning rule is inserted into the neural network which enables the neural network to reset the weighting values at the inputs of the network each time historical data is entered so that it may more accurately predict the outputs of the system. Some specific examples of the use of neural networks include improving the accuracy of the analysis of sonar signals, the diagnosis of jet engine problems, and the determination of space craft attitude.
However, models which can predict plural output variables of a process based upon changes in plural input variables which interact with each other, have not been used previous to this invention, to control industrial or chemical processes in "real time" or "on-2.~.~t~~~
line". The term "real time", as used herein, is a process cantrol system which either automatically changes the process, or changes the process within a short time period after the process control system indicates that a change is necessary, while the process is in progress. The term "on-line", as used herein, refers to a system which automatically receives information from, and automatically inserts changes into, the process, while the process is in progress.
In addition to the need for process control systems which operate to control processes on-line and/or in real time, there is also a need for process control systems which employ economic models in conjunction with performance models and which enable the optimization of the system to meet economic goals in addition to other types of performance goals.
OBJECTS OF THE INVENTION
Accordingly, it is a general object of the instant invention to provide a process control system which overcomes the shortcomings of, and improves upon, existing process control systems.
It is a further object of the instant invention 'to provide a process control system which optimizes the process for a system with a plurality of interactive independent (controllable) variables and a plurality of dependent (output) variables.
It is still a further object of the instant invention to provide a process control system which uses an economic model in conjunction with a process model to optimize the process.
Each neuron in the output layer represents a dependent variable whose value changes due to changes in the values of the independent variables and the interactions between the independent variables. The output of each neuron is modified by a transfer function. Various types of transfer functions, such as linear, linear threshold (which sets a maximum and minimum level to the function), step, Sigmoid (S-shaped), hyperbolic tangent or Gaussian functions can be used.
Historical data representing the independent input variables of a system are provided to the neural network model and the dependent variables (outputs) of the system are determined and compared to the actual output of the system at the time. A
learning rule is inserted into the neural network which enables the neural network to reset the weighting values at the inputs of the network each time historical data is entered so that it may more accurately predict the outputs of the system. Some specific examples of the use of neural networks include improving the accuracy of the analysis of sonar signals, the diagnosis of jet engine problems, and the determination of space craft attitude.
However, models which can predict plural output variables of a process based upon changes in plural input variables which interact with each other, have not been used previous to this invention, to control industrial or chemical processes in "real time" or "on-2.~.~t~~~
line". The term "real time", as used herein, is a process cantrol system which either automatically changes the process, or changes the process within a short time period after the process control system indicates that a change is necessary, while the process is in progress. The term "on-line", as used herein, refers to a system which automatically receives information from, and automatically inserts changes into, the process, while the process is in progress.
In addition to the need for process control systems which operate to control processes on-line and/or in real time, there is also a need for process control systems which employ economic models in conjunction with performance models and which enable the optimization of the system to meet economic goals in addition to other types of performance goals.
OBJECTS OF THE INVENTION
Accordingly, it is a general object of the instant invention to provide a process control system which overcomes the shortcomings of, and improves upon, existing process control systems.
It is a further object of the instant invention 'to provide a process control system which optimizes the process for a system with a plurality of interactive independent (controllable) variables and a plurality of dependent (output) variables.
It is still a further object of the instant invention to provide a process control system which uses an economic model in conjunction with a process model to optimize the process.
It is still yet a further object of the instant invention to provide a process control system which optimizes a process in real time while the process is being performed.
It is another object of the instant invention to provide a process control system which optimizes a process for a selected one of the independent variables.
It is still another object of the instant invention to provide a process control system which optimizes the process in relation to a selected one of the dependent variables.
It is still yet another object of the instant invention to provide a process control system which automatically changes the level of a selected independent variable to meet a selected one of a plurality of operating or economic objectives.
It is an additional object of the instant invention to provide a process control system which periodically optimizes the level of a selected independent variable to meet a selected one of a plurality of operating or economic objectives.
SUMMARY OF THE INVENTION
These and other objects of the instant invention are achieved by a process control system having a process model and an economic model. The system accepts information from an operating process and calculates values of dependent (output) variables versus a selected one of the independent variables. Objectives or goals are inserted into the process control system which relate to optimizing the system either with respect to the levels, quantity or quality of the independent variables of the process or with 2:~~.~~5~~';
respect to economic objectives or criteria. If economic objectives are selected, the economic model is used in conjunction with the process model in order to optimize the level of the selected independent variable to meet the goal.
Historical data is used to develop the process model so that it may determine the interrelationship between the output variables and the interactive independent input variables. The process model comprises a neural network which learns from the application of historical data to predict the levels of dependent variables with changes in the controllable or independent variables.
DESCRIPTION OF THE DRAWING
Other objects and many of the intended advantages of this invention will be readily appreciated when the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawing wherein:
Fig. 1 is a block diagram of the process control system constructed in accordance with this invention for controlling the introduction of a chemical into a petroleum reaction (alkylation) system;
Figs. 2A and 2B together comprise a flow chart of the operation of the process control system of Fig. 1;
Fig. 3 is a graph of the levels of various dependent variables of the petroleum reaction process, plotted against the chemical injection rate as established by the system of Fig. 1;
2:~~888 Fig. 4A is a graph of the total product yield (alkylate yield) in thousands of dollars per day plotted against the chemical injection rate of the process;
Fig. 4B is a graph of the incremental acid ( catalyst ) reduction in thousands of dollars per day plotted against the chemical injection rate of the process;
Fig. 4C is a graph of the incremental RON (Research Octane Number) in terms of dollars per day plotted against the chemical injection rate of the process;
Fig. 4D is a graph of the cost of the chemicals in dollars per day plotted against the chemical injection rate of the process;
Fig. 4E is a graph of the total incremental value of the process output in thousands of dollars per day plotted against the chemical injection rate of the process;
Fig. 4F is a graph of the chemical ROI (Return On Investment) in percentages plotted against the chemical injection rate of the process;
Fig. 5 ~ shows in diagrammatic form an exemplary neural network model for an alkylation process;
Figs. 6A and 6H show in tabular form the historical data relating to dependent and independent variables taken for an operating process; and Figs. 6C and 6D tabulate the results obtained using the process control system model and the statistical accuracy of the model.
. 2:1~~:~~
_-.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
The preferred embodiment, which is described in detail below, is for a chemical process control system which optimizes the on-line performance of the reaction or distillation portion of a petroleum refinery which produces gasoline. The independent variable which is optimized to meet various objectives is, in the example given, the flow rate of a chemical additive which is added to the process to improve the efficiency of the process. It should be noted, however, that the process control system described herein is equally applicable to other types of industrial or chemical processes and is capable of determining the optimal flow rate, quantity, quality or other parameter of any independent variable to meet objectives.
The system uses a process model and an economic model, which is utilized when economic goals are to be met. A single independent variable is selected and the system determines the levels of the dependent variables over a pre-selected range of levels of the independent selected variable, as will be explained in detail below. In effect, the levels of the output or dependent variables are "plotted" and stored over the range of the levels of selected variables (See Figs. 3 and 4A-4F). For non-economic goals, the system determines the plot maximum or minimum to calculate the optimum level of the selected independent variable to meet the goal. To meet economic goals, economic values are 2.1.~~~~3 i..serted into the system to modify the plots prior to calculating the optimum level of the selected independent variable to meet the goals.
Referring now in greater detail to the various figures of the drawing wherein like references and characters refer to like parts, a block diagram of the process control system 2 of the instant invention is shown in Fig. 1. The input section A
comprises the process 4. The quantitative levels of the system variables of process 4 are transmitted to process control center 6.
The process control center 6 provides information to personnel at the refinery as to the level or value i.e. quantity, quality, or flow rates, of the variables of the system. The system variables' levels of the process 4 may be changed via controls in the process control center 6. A feedback loop is used between the process control center 6 and the process, to zero-in the level of the system variable which has been adjusted to the desired level from the process control center 6. The levels of the variables of process 4 can, either be automatic and electronic in nature, which can be processed directly by the process control system 2. If they are analogue in nature, they operate indicators such as gauges or meters, which are received in the process control center 6.
If the levels of the system variables are in electronic form, i.e. digital form, the process control center 6 transmits the electronic signals to the Data Acquisition and Processing Section B. The Data Acquisition and Processing Section B has a data acquisition processor 8 which processes and stores the incoming e~::ctronic signals representing the system variables. In the case wherein the levels of the system variables are available only in analogue form, such as on meters, gauges and the like, the information is recorded at the process control center 6 and manually input to the data acquisition processor 8 via a keyboard or other input equipment attached thereto.
The Process Control Section C of the system comprises a process control computer 10. The process control computer 10.
utilizes software which includes a process model 12 and an economic optimizer 14. On command via a trigger signal from the process control computer 10, the data acquisition processor 8 transmits the data, i.e., the levels of the system variables in digital, electronic form, to the process control computer 10. In addition, .
via a keyboard or other suitable entry device, economic values and objectives are entered into the process control computer 10.
The output of the process control computer 10 is an optimized chemical control signal which represents the optimal flow rate of the chemical which should be inserted into the process to meet the objective. The objective and economic basis are also entered and stored into the process control computer 10. If an automatic system is available to control a pump which feeds a chemical additive into the process, a chemical pump controller 16 is used. Equipment such as the Pacesetter, which is available from Hetz Laboratories, Inc., and which is described in United States 2~.1~8~5 F__ent No. 4,897,797, can be used to control the chemical pump 18.
The chemical pump controller 16 feeds back the actual injection rate of the chemical pump to the process control computer 10.
The chemical pump controller 16 transmits a control signal to the chemical pump 18 which controls the action of the pump to change the chemical rate of flow into the process. A
verification signal is also received by the chemical pump controller 16 from the process, which verifies that the desired flow rate has been achieved.
The process model 12 in this embodiment is a neural network model. A suitable neural network model program such as the Professional Two Plus program produced by Neural Ware, Pittsburgh, Pennsylvania, or equivalent may be used and customized for this control system. Historical data must, therefore, be entered into the system to enable the.neural network model to learn, as previously described. An historical data bank is developed by recording the values or levels of the system variables at predetermined intervals over a predetermined period of time.
This historical data is input to the data acquisition processor 8 via its manual input, and then used by the process model to enable the process model 12 to learn to better predict selected outputs of the system with changes in the interacting variables of the system over time. The accuracy of the process model 12 in predicting outputs of the system (the levels of the dependent variables) from 2.~~8885 a _arge number of sets of values of independent variables has been determined using actual operating processes, as will be described in detail later.
Figs. 2A and 2B show a flow diagram of the system.
Continuous inputs from the process, comprising information as to the levels of the variables and the chemical additive rate (the selected independent variable) are transmitted from Block A to the variable data base Block B which is part of the data acquisition processor 8. The starting of the optimization routine, step Sl, can be performed either automatically, at preset timed intervals, while the process is in operation, or it can be started by depressing a manual start button. When the routine is started, the value of the chemical variable is set to a minimum in step S2. The minimum value of the chemical variable and the values of the other variables from the variable data base clock B are sent to the process control computer 10 which then uses the process model 12 to calculate the performance of the process, i.e., the levels of the dependent variables, at current conditions, as shown in step S3.
The calculated performance i.e. predicted levels of the dependent variables and chemical rate is then stored in step S4. The chemical rate is then incremented to its next predetermined value in step S5. Step S6 is a decision step which determines whether the chemical rate that has been incremented is at its maximum allowable rate. If the answer is negative, the process performance is recalculated at the newly incremented chemical rate and step S3 - step S5 are repeated.
~. 211~~~~
For each increment in level of the chemical rate, the levels of the dependent variables are determined, in effect providing a graph or plot, as shown in Fig. 3 ~whicii- will be described later.
When the iterative process is completed, i.e., when the chemical rate has been incremented in steps until it has reached its maximum rate, the output step S6 is in the affirmative.
Decision step S7 determines whether the_objective entered into the system requires economic calculations.
Prior to the start of the optimization routine, if economic goals are to be met, economic data relating to process performance values and chemical unit costs as shown in block D is entered into, and stored in, the process control computer 10. In -' addition, the desired objective for the chemical additive is entered into the system providing an input for decision step S7.
Before proceeding further with a discussion of the flow diagram of Fig. 2 (consisting of Figs. 2A and 2B), the stored calculated performance and chemical rate should be-~discussed.
Fig. 3 shows the results of the calculated performance and chemical rate, taken by the process control system for-wan actual operating process, in graph form.
The chemical injection rate is the ratio of chemical additive to flow into the process to the flow of catalyst sulfuric acid into the process in percent. As can be seen in Fig. 3, the chemical injection rate varies from 0 to 0.6% of the catalyst flow rate.
2~.18~3'~
In this case four dependent variables have been chosen for the process. One of these is the Acid consumption in barrels per day. The Acid cansumption in this case is the amount of sulfuric acid used, which is the catalyst for the reaction operation. As can be seen in the chart as the chemical injection rate is incremented from zero to 0.6, the acid consumption rate, i.e., the consumption of the catalyst, varies from a high of 560 barrels per day for a zero chemical injection rate, to a low of 400 barrels per day for a chemical injection rate of 0.6. A second dependent variable is the Research Octane Number (RON). As can be seen in Fig. 3, the RON varies between 96:1 at a zero chemical injection rate to 95.3 at a 0.6 chemical injection rate, peaking at approximately 0.3 chemical injection rate, at a RON of 96.3.
A third dependent variable is entitled the Alky Make in barrels per day. The Alky Make is the output of the commercial alkylation unit and represents the amount of gasoline that the process is producing. The Alky Make varies from a low of 13,050 to a high of 13,750 barrels per day as the chemical injection rate is increased.
Referring again to Fig. 2 (consisting of Figs. 2A
and 2B), certain objectives do not require economic analysis. For example, if the objectives were to maximize RON, the system would only have to determine, from the calculated performance information, the chemical-injection rate which maximizes RON, in this case a 0,3 chemical injection rate. Similarly, if the objective was to maximize production or Alky Make, or to minimize Acid consumption, the system would only have 1... determine the chemical injection rate which maximizes Alky Make or which minimizes Acid consumption (in this case a chemical injection rate of 0.6). Thus, economic analyses are not required and the path from the negative side of step S7, to step S8, which locates the maxima or minima based on a desired objective, is taken as shown in Fig. 2A.
Figs. 4A-4F show the results of optimization to reach various economic objectives. For example, Fig. 4A shows the value of the Alky Yield in thousands of dollars per day as the chemical injection rate is varied from zero to 0.6. In this case, the dollar value of each barrel of the barrels of gasoline at various octane numbers are entered into the system to generate the graph of Fig. 4A. Thus, the value of the incremental or increased yield of the process in thousands of dollars per day rises to approximately fourteen thousand dollars for chemical injection rates of 0 to 0.6.
Similarly, in Fig. 4B the dollar value of the incremental reduction in Acid. or catalyst requirements peaks at approximately five thousand dollars per day at a chemical injection rate of 0.6.
Also, as shown in Fig. 4C the incremental value of the Research Octane Number ( RON j in dollars per day peaks at six hundred dollars at a chemical injection rate of 0.3.
Fig. 4D shows that the chemical cost in dollars per day rises from 0 to approximately 2,000 as the chemical injection rate increases from 0 to 0.6.
2.~:I~~~
In addition to these economic calculations, more complex economic calculations can be made by the system. For example, as shown in Fig. 4E, the incremental value of the system which is the value of the product produced, less the costs involved in producing the product, is shown in Fig. 4E for various chemical injection rates. As can be seen in the figure, the total value curve begins to flatten out at approximately sixteen thousand dollars per day peaking at a chemical injection rate of 0.47.
Fig. 4F shows the variation in chemical ROI in percent, i.e., the total return minus the chemical cost, divided by the chemical cost. This plot indicates that for the actual process being controlled, a chemical injection rate of 0.35 will maximize the chemical ROI.
Figs. 4A-4F show the actual performance results obtained at a process of a particular alkylation or reaction unit in commercial use. Although six specific economic objectives have been chosen and illustrated in this case, it~will be appreciated that the process control system can optimize the process for a large number of other economic objectives.
Referring again to the flow diagram of Figs. 2A and 2B
the flow of information for combining economic calculations with performance calculations will be described. If the output of step S7 is affirmative, the economic model calculates the process performance economics for each chemical increment using information from the calculated performance base of step S4 and the performance value and chemical unit cost from the stored data base in the . . ~ ' 2 ~. .~. 8 r~; f~
p~ocess control computer, in step S9. In step S10 economics for each chemical increment are summed. The minimum or maximum economics based on the desired objective is determined in step S11.
In step S12 the chemical rate is adjusted to meet the desired objective.
In systems wherein there is automatic control for the chemical pump an output signal is sent to the pump to automatically adjust the chemical rate to the optimal condition for the desired objective, as shown in step S13. If the system does not have an automatic means to reset the chemical rate, a readout or display indicating the optimal chemical rate can be provided to an operator, who then manually adjusts the controls necessary to obtain the desired chemical rate.
The final step S14 resets the system to start the process again. This may be done automatically with a timer or manually at pre-determined times.
An exemplary neural network model for an alkylation unit is shown in Fig. 5. As can be seen in the figure, the model comprises six input neurons which represent independent variables, it has a second layer of six hidden neurons and an output layer of four neurons or dependent variables. In this case, the independent variables of the process are Olefin (the short chain hydrocarbon flow rate); the IC4 (the Isobutane flow rate); the Rx TEMP (the temperature of the reactor of the alkylation unit); the Contaminants which can include methanol, water or sulphur compounds; the Fresh Acid concentration (the strength of the acid - ~ ~ 2~1~J~~~
catalyst at the input to the process); and the Spent Acid concentration (the strength of the catalyst at the output of the process). The four dependent variables of the network are the Alky RON (the Research Octane Number of the gasoline output of the process); Alky MON (the Motor Octane Number of the gasoline output of the process); Alky Make (the output or yield of the process);
and Acid Used (the quantity of the acid used by the process).
As previously described, the independent variables are the variables which are controllable and the dependent variables represent various outputs of the system which can be controlled by varying the levels of the input variables. The number of hidden neurons and the number of layers of hidden neurons are chosen by the designer of the model. If too few hidden neurons are used, the model will not be able to predict outputs accurately with respect to changes in the independent variables. If too many hidden neurons are used the model becomes too sensitive to slight changes and cannot accurately determine general trends . An initial rule of thumb is to use as many hidden neurons as the average of the number of input neurons and output neurons or to use as many hidden neurons as the number of input neurons.
Since each process is to some extent unique, the process control model must learn from historical data taken on the specific process to be controlled. Thus, referring again to Fig. 1, the manual input to the data acquisition processor 8 (Block B), which can be a keyboard or other input device, is used to load historical data into the data acquisition processor. The data representing 2~.~.~3~~rj ou-line values of independent and dependent variables recorded from the process, are presented to the process model. The process model, which in this case is a neural network, predicts the dependent variable values based upon the independent variable values. The neural network then determines the difference between its predicted dependent variable values and the actual dependent variable values and adjusts the weighting assigned to the independent variables in accordance with a learning rule inserted into the model. As more and more historical data is presented to the neural network model, it learns and becomes more accurate in its predictions. This is similar to a human brain, which based upon experience, can more accurately determine future events.
A set of data on an actual operating process using the process model of the present invention is shown in Figs. 6A-6D.
Figs. 6A and 6B show the independent variables of the model, (25 variables); and the dependent variables, (3 variables). In this case the dependent variables are Yield, Acid consumption and Research Octane Number (RON). The limits for each of the variables, i.e., the maxima and minima values as determined by historical data, are given in the table in Figs. 6A and 6B. A
dictionary of the variables is given in Fig. 6B.
Any one of the independent variables may be used as the control variable for optimizing the system with relation to the output or dependent variables. In addition, as discussed previously, the process may be optimized to meet economic goals.
y ~ 2:~~~3~~
--, Figs. 6C and 6D tabulate the results of the testing wherein actual measurements of dependent variables were compared to predicted values of the dependent variables based upon a set of independent variable values from the process. The statistics were compiled from a set of data using individual points of the neural network model, and for a set of data using five point subgroups.
For each of the dependent variables, as shown in the table at the top of Fig. 6C, 293 measurements of the 25 independent variable levels were made. The 293 measurements were also cambined into 59 subgroups for a five point subgroup model. For each of the dependent variables, the standard error of the model SE, the correlation coefficient R2, and the standard deviation SD, are given for the individual point model and the five point subgroup model.
The prediction limits for each of the variables, for the individual points model and for the five point subgroup model are given in the second and third charts of Figs. 6C and 6D. These include the lower control limit (LCL), the average at minus 2 SD
(Standard Deviations), the average at minus 1 SD, the average, the averag8 at plus 1 SD, the average at plus 2 SD, and the upper control limit, are given.
The chart of Fig. 6D gives the validation statistics for the five point subgroup model. As can be seen in the chart, the model prediction for the average Yield is .15 percent less than the actual prediction, the average Acid consumption predicted by the model is 0.17 percent higher than the actual average Acid 2~1~8~5 consumption, and the average Research Octane Number for the model is 0.1 percent above the actual average Research Octane Number.
Also shown in Fig. 6D are the definitions of the statistical parameters used to evaluate the data.
In addition to the controlling of processes in the petroleum refinery field, as illustrated in this embodiment, the process control system of the instant invention is also applicable for the control of other types of processes. For example, in boiler processes, a deposit of solids which are dissolved in the water, tends to build-up on the sides of the boiler. The build-up is called fauling, and anti-foulants are often used to slow down the build-up of deposits. As deposits build-up on the inner walls of the boilers, the system becomes less efficient and more fuel and a higher temperature are required to raise the temperature of the water to the desired level. After a period of time, the boiler must be shut down and the deposits removed from the walls. This process is time consuming and results in increased costs and reduction in capacity of the system.
Additives to prevent fauling build-up (anti-foulants), by slowing down and minimizing the build-up, allow the system to operate more efficiently and for longer periods of time before shutdown.
Application of the instant invention to a boiler system could use for example the amount of fuel, the type of fuel, the measurement of the amount of dissolved solvents in the water, the rate of flow of the fuel, and the anti-foulant flow rate as 211888) independent variables. Objectives for optimization of the system could be maximizing the time between shutdowns, minimizing fuel costs, maximizing return on investment with regard to the chemical additive, maximizing process output or minimizing maintenance, labor anti costs.
The application of this process control system to such a boiler process could entail the use of the flow rate of the anti-foulant chemical additive as the controllable independent variable in a fashion similar to the use of the additive control rate for optimization of the alkylation unit process given previously.
A process control system has been described which is applicable, generally, to industrial commercial processes. It comprises both a process model, such as a neural network model, and an economic model. This enables the process~to be optimized for obtaining objectives relating to maximizing or minimizing dependent output variables or optimization of the system to meet economic criteria such as overall production value, return on investment, minimization of costs, and the like.
Without further elaboration, the foregoing will so fully illustrate my invention that others may, by applying current or future knowledge, readily adapt the same for use under the various conditions of service.
It is another object of the instant invention to provide a process control system which optimizes a process for a selected one of the independent variables.
It is still another object of the instant invention to provide a process control system which optimizes the process in relation to a selected one of the dependent variables.
It is still yet another object of the instant invention to provide a process control system which automatically changes the level of a selected independent variable to meet a selected one of a plurality of operating or economic objectives.
It is an additional object of the instant invention to provide a process control system which periodically optimizes the level of a selected independent variable to meet a selected one of a plurality of operating or economic objectives.
SUMMARY OF THE INVENTION
These and other objects of the instant invention are achieved by a process control system having a process model and an economic model. The system accepts information from an operating process and calculates values of dependent (output) variables versus a selected one of the independent variables. Objectives or goals are inserted into the process control system which relate to optimizing the system either with respect to the levels, quantity or quality of the independent variables of the process or with 2:~~.~~5~~';
respect to economic objectives or criteria. If economic objectives are selected, the economic model is used in conjunction with the process model in order to optimize the level of the selected independent variable to meet the goal.
Historical data is used to develop the process model so that it may determine the interrelationship between the output variables and the interactive independent input variables. The process model comprises a neural network which learns from the application of historical data to predict the levels of dependent variables with changes in the controllable or independent variables.
DESCRIPTION OF THE DRAWING
Other objects and many of the intended advantages of this invention will be readily appreciated when the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawing wherein:
Fig. 1 is a block diagram of the process control system constructed in accordance with this invention for controlling the introduction of a chemical into a petroleum reaction (alkylation) system;
Figs. 2A and 2B together comprise a flow chart of the operation of the process control system of Fig. 1;
Fig. 3 is a graph of the levels of various dependent variables of the petroleum reaction process, plotted against the chemical injection rate as established by the system of Fig. 1;
2:~~888 Fig. 4A is a graph of the total product yield (alkylate yield) in thousands of dollars per day plotted against the chemical injection rate of the process;
Fig. 4B is a graph of the incremental acid ( catalyst ) reduction in thousands of dollars per day plotted against the chemical injection rate of the process;
Fig. 4C is a graph of the incremental RON (Research Octane Number) in terms of dollars per day plotted against the chemical injection rate of the process;
Fig. 4D is a graph of the cost of the chemicals in dollars per day plotted against the chemical injection rate of the process;
Fig. 4E is a graph of the total incremental value of the process output in thousands of dollars per day plotted against the chemical injection rate of the process;
Fig. 4F is a graph of the chemical ROI (Return On Investment) in percentages plotted against the chemical injection rate of the process;
Fig. 5 ~ shows in diagrammatic form an exemplary neural network model for an alkylation process;
Figs. 6A and 6H show in tabular form the historical data relating to dependent and independent variables taken for an operating process; and Figs. 6C and 6D tabulate the results obtained using the process control system model and the statistical accuracy of the model.
. 2:1~~:~~
_-.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
The preferred embodiment, which is described in detail below, is for a chemical process control system which optimizes the on-line performance of the reaction or distillation portion of a petroleum refinery which produces gasoline. The independent variable which is optimized to meet various objectives is, in the example given, the flow rate of a chemical additive which is added to the process to improve the efficiency of the process. It should be noted, however, that the process control system described herein is equally applicable to other types of industrial or chemical processes and is capable of determining the optimal flow rate, quantity, quality or other parameter of any independent variable to meet objectives.
The system uses a process model and an economic model, which is utilized when economic goals are to be met. A single independent variable is selected and the system determines the levels of the dependent variables over a pre-selected range of levels of the independent selected variable, as will be explained in detail below. In effect, the levels of the output or dependent variables are "plotted" and stored over the range of the levels of selected variables (See Figs. 3 and 4A-4F). For non-economic goals, the system determines the plot maximum or minimum to calculate the optimum level of the selected independent variable to meet the goal. To meet economic goals, economic values are 2.1.~~~~3 i..serted into the system to modify the plots prior to calculating the optimum level of the selected independent variable to meet the goals.
Referring now in greater detail to the various figures of the drawing wherein like references and characters refer to like parts, a block diagram of the process control system 2 of the instant invention is shown in Fig. 1. The input section A
comprises the process 4. The quantitative levels of the system variables of process 4 are transmitted to process control center 6.
The process control center 6 provides information to personnel at the refinery as to the level or value i.e. quantity, quality, or flow rates, of the variables of the system. The system variables' levels of the process 4 may be changed via controls in the process control center 6. A feedback loop is used between the process control center 6 and the process, to zero-in the level of the system variable which has been adjusted to the desired level from the process control center 6. The levels of the variables of process 4 can, either be automatic and electronic in nature, which can be processed directly by the process control system 2. If they are analogue in nature, they operate indicators such as gauges or meters, which are received in the process control center 6.
If the levels of the system variables are in electronic form, i.e. digital form, the process control center 6 transmits the electronic signals to the Data Acquisition and Processing Section B. The Data Acquisition and Processing Section B has a data acquisition processor 8 which processes and stores the incoming e~::ctronic signals representing the system variables. In the case wherein the levels of the system variables are available only in analogue form, such as on meters, gauges and the like, the information is recorded at the process control center 6 and manually input to the data acquisition processor 8 via a keyboard or other input equipment attached thereto.
The Process Control Section C of the system comprises a process control computer 10. The process control computer 10.
utilizes software which includes a process model 12 and an economic optimizer 14. On command via a trigger signal from the process control computer 10, the data acquisition processor 8 transmits the data, i.e., the levels of the system variables in digital, electronic form, to the process control computer 10. In addition, .
via a keyboard or other suitable entry device, economic values and objectives are entered into the process control computer 10.
The output of the process control computer 10 is an optimized chemical control signal which represents the optimal flow rate of the chemical which should be inserted into the process to meet the objective. The objective and economic basis are also entered and stored into the process control computer 10. If an automatic system is available to control a pump which feeds a chemical additive into the process, a chemical pump controller 16 is used. Equipment such as the Pacesetter, which is available from Hetz Laboratories, Inc., and which is described in United States 2~.1~8~5 F__ent No. 4,897,797, can be used to control the chemical pump 18.
The chemical pump controller 16 feeds back the actual injection rate of the chemical pump to the process control computer 10.
The chemical pump controller 16 transmits a control signal to the chemical pump 18 which controls the action of the pump to change the chemical rate of flow into the process. A
verification signal is also received by the chemical pump controller 16 from the process, which verifies that the desired flow rate has been achieved.
The process model 12 in this embodiment is a neural network model. A suitable neural network model program such as the Professional Two Plus program produced by Neural Ware, Pittsburgh, Pennsylvania, or equivalent may be used and customized for this control system. Historical data must, therefore, be entered into the system to enable the.neural network model to learn, as previously described. An historical data bank is developed by recording the values or levels of the system variables at predetermined intervals over a predetermined period of time.
This historical data is input to the data acquisition processor 8 via its manual input, and then used by the process model to enable the process model 12 to learn to better predict selected outputs of the system with changes in the interacting variables of the system over time. The accuracy of the process model 12 in predicting outputs of the system (the levels of the dependent variables) from 2.~~8885 a _arge number of sets of values of independent variables has been determined using actual operating processes, as will be described in detail later.
Figs. 2A and 2B show a flow diagram of the system.
Continuous inputs from the process, comprising information as to the levels of the variables and the chemical additive rate (the selected independent variable) are transmitted from Block A to the variable data base Block B which is part of the data acquisition processor 8. The starting of the optimization routine, step Sl, can be performed either automatically, at preset timed intervals, while the process is in operation, or it can be started by depressing a manual start button. When the routine is started, the value of the chemical variable is set to a minimum in step S2. The minimum value of the chemical variable and the values of the other variables from the variable data base clock B are sent to the process control computer 10 which then uses the process model 12 to calculate the performance of the process, i.e., the levels of the dependent variables, at current conditions, as shown in step S3.
The calculated performance i.e. predicted levels of the dependent variables and chemical rate is then stored in step S4. The chemical rate is then incremented to its next predetermined value in step S5. Step S6 is a decision step which determines whether the chemical rate that has been incremented is at its maximum allowable rate. If the answer is negative, the process performance is recalculated at the newly incremented chemical rate and step S3 - step S5 are repeated.
~. 211~~~~
For each increment in level of the chemical rate, the levels of the dependent variables are determined, in effect providing a graph or plot, as shown in Fig. 3 ~whicii- will be described later.
When the iterative process is completed, i.e., when the chemical rate has been incremented in steps until it has reached its maximum rate, the output step S6 is in the affirmative.
Decision step S7 determines whether the_objective entered into the system requires economic calculations.
Prior to the start of the optimization routine, if economic goals are to be met, economic data relating to process performance values and chemical unit costs as shown in block D is entered into, and stored in, the process control computer 10. In -' addition, the desired objective for the chemical additive is entered into the system providing an input for decision step S7.
Before proceeding further with a discussion of the flow diagram of Fig. 2 (consisting of Figs. 2A and 2B), the stored calculated performance and chemical rate should be-~discussed.
Fig. 3 shows the results of the calculated performance and chemical rate, taken by the process control system for-wan actual operating process, in graph form.
The chemical injection rate is the ratio of chemical additive to flow into the process to the flow of catalyst sulfuric acid into the process in percent. As can be seen in Fig. 3, the chemical injection rate varies from 0 to 0.6% of the catalyst flow rate.
2~.18~3'~
In this case four dependent variables have been chosen for the process. One of these is the Acid consumption in barrels per day. The Acid cansumption in this case is the amount of sulfuric acid used, which is the catalyst for the reaction operation. As can be seen in the chart as the chemical injection rate is incremented from zero to 0.6, the acid consumption rate, i.e., the consumption of the catalyst, varies from a high of 560 barrels per day for a zero chemical injection rate, to a low of 400 barrels per day for a chemical injection rate of 0.6. A second dependent variable is the Research Octane Number (RON). As can be seen in Fig. 3, the RON varies between 96:1 at a zero chemical injection rate to 95.3 at a 0.6 chemical injection rate, peaking at approximately 0.3 chemical injection rate, at a RON of 96.3.
A third dependent variable is entitled the Alky Make in barrels per day. The Alky Make is the output of the commercial alkylation unit and represents the amount of gasoline that the process is producing. The Alky Make varies from a low of 13,050 to a high of 13,750 barrels per day as the chemical injection rate is increased.
Referring again to Fig. 2 (consisting of Figs. 2A
and 2B), certain objectives do not require economic analysis. For example, if the objectives were to maximize RON, the system would only have to determine, from the calculated performance information, the chemical-injection rate which maximizes RON, in this case a 0,3 chemical injection rate. Similarly, if the objective was to maximize production or Alky Make, or to minimize Acid consumption, the system would only have 1... determine the chemical injection rate which maximizes Alky Make or which minimizes Acid consumption (in this case a chemical injection rate of 0.6). Thus, economic analyses are not required and the path from the negative side of step S7, to step S8, which locates the maxima or minima based on a desired objective, is taken as shown in Fig. 2A.
Figs. 4A-4F show the results of optimization to reach various economic objectives. For example, Fig. 4A shows the value of the Alky Yield in thousands of dollars per day as the chemical injection rate is varied from zero to 0.6. In this case, the dollar value of each barrel of the barrels of gasoline at various octane numbers are entered into the system to generate the graph of Fig. 4A. Thus, the value of the incremental or increased yield of the process in thousands of dollars per day rises to approximately fourteen thousand dollars for chemical injection rates of 0 to 0.6.
Similarly, in Fig. 4B the dollar value of the incremental reduction in Acid. or catalyst requirements peaks at approximately five thousand dollars per day at a chemical injection rate of 0.6.
Also, as shown in Fig. 4C the incremental value of the Research Octane Number ( RON j in dollars per day peaks at six hundred dollars at a chemical injection rate of 0.3.
Fig. 4D shows that the chemical cost in dollars per day rises from 0 to approximately 2,000 as the chemical injection rate increases from 0 to 0.6.
2.~:I~~~
In addition to these economic calculations, more complex economic calculations can be made by the system. For example, as shown in Fig. 4E, the incremental value of the system which is the value of the product produced, less the costs involved in producing the product, is shown in Fig. 4E for various chemical injection rates. As can be seen in the figure, the total value curve begins to flatten out at approximately sixteen thousand dollars per day peaking at a chemical injection rate of 0.47.
Fig. 4F shows the variation in chemical ROI in percent, i.e., the total return minus the chemical cost, divided by the chemical cost. This plot indicates that for the actual process being controlled, a chemical injection rate of 0.35 will maximize the chemical ROI.
Figs. 4A-4F show the actual performance results obtained at a process of a particular alkylation or reaction unit in commercial use. Although six specific economic objectives have been chosen and illustrated in this case, it~will be appreciated that the process control system can optimize the process for a large number of other economic objectives.
Referring again to the flow diagram of Figs. 2A and 2B
the flow of information for combining economic calculations with performance calculations will be described. If the output of step S7 is affirmative, the economic model calculates the process performance economics for each chemical increment using information from the calculated performance base of step S4 and the performance value and chemical unit cost from the stored data base in the . . ~ ' 2 ~. .~. 8 r~; f~
p~ocess control computer, in step S9. In step S10 economics for each chemical increment are summed. The minimum or maximum economics based on the desired objective is determined in step S11.
In step S12 the chemical rate is adjusted to meet the desired objective.
In systems wherein there is automatic control for the chemical pump an output signal is sent to the pump to automatically adjust the chemical rate to the optimal condition for the desired objective, as shown in step S13. If the system does not have an automatic means to reset the chemical rate, a readout or display indicating the optimal chemical rate can be provided to an operator, who then manually adjusts the controls necessary to obtain the desired chemical rate.
The final step S14 resets the system to start the process again. This may be done automatically with a timer or manually at pre-determined times.
An exemplary neural network model for an alkylation unit is shown in Fig. 5. As can be seen in the figure, the model comprises six input neurons which represent independent variables, it has a second layer of six hidden neurons and an output layer of four neurons or dependent variables. In this case, the independent variables of the process are Olefin (the short chain hydrocarbon flow rate); the IC4 (the Isobutane flow rate); the Rx TEMP (the temperature of the reactor of the alkylation unit); the Contaminants which can include methanol, water or sulphur compounds; the Fresh Acid concentration (the strength of the acid - ~ ~ 2~1~J~~~
catalyst at the input to the process); and the Spent Acid concentration (the strength of the catalyst at the output of the process). The four dependent variables of the network are the Alky RON (the Research Octane Number of the gasoline output of the process); Alky MON (the Motor Octane Number of the gasoline output of the process); Alky Make (the output or yield of the process);
and Acid Used (the quantity of the acid used by the process).
As previously described, the independent variables are the variables which are controllable and the dependent variables represent various outputs of the system which can be controlled by varying the levels of the input variables. The number of hidden neurons and the number of layers of hidden neurons are chosen by the designer of the model. If too few hidden neurons are used, the model will not be able to predict outputs accurately with respect to changes in the independent variables. If too many hidden neurons are used the model becomes too sensitive to slight changes and cannot accurately determine general trends . An initial rule of thumb is to use as many hidden neurons as the average of the number of input neurons and output neurons or to use as many hidden neurons as the number of input neurons.
Since each process is to some extent unique, the process control model must learn from historical data taken on the specific process to be controlled. Thus, referring again to Fig. 1, the manual input to the data acquisition processor 8 (Block B), which can be a keyboard or other input device, is used to load historical data into the data acquisition processor. The data representing 2~.~.~3~~rj ou-line values of independent and dependent variables recorded from the process, are presented to the process model. The process model, which in this case is a neural network, predicts the dependent variable values based upon the independent variable values. The neural network then determines the difference between its predicted dependent variable values and the actual dependent variable values and adjusts the weighting assigned to the independent variables in accordance with a learning rule inserted into the model. As more and more historical data is presented to the neural network model, it learns and becomes more accurate in its predictions. This is similar to a human brain, which based upon experience, can more accurately determine future events.
A set of data on an actual operating process using the process model of the present invention is shown in Figs. 6A-6D.
Figs. 6A and 6B show the independent variables of the model, (25 variables); and the dependent variables, (3 variables). In this case the dependent variables are Yield, Acid consumption and Research Octane Number (RON). The limits for each of the variables, i.e., the maxima and minima values as determined by historical data, are given in the table in Figs. 6A and 6B. A
dictionary of the variables is given in Fig. 6B.
Any one of the independent variables may be used as the control variable for optimizing the system with relation to the output or dependent variables. In addition, as discussed previously, the process may be optimized to meet economic goals.
y ~ 2:~~~3~~
--, Figs. 6C and 6D tabulate the results of the testing wherein actual measurements of dependent variables were compared to predicted values of the dependent variables based upon a set of independent variable values from the process. The statistics were compiled from a set of data using individual points of the neural network model, and for a set of data using five point subgroups.
For each of the dependent variables, as shown in the table at the top of Fig. 6C, 293 measurements of the 25 independent variable levels were made. The 293 measurements were also cambined into 59 subgroups for a five point subgroup model. For each of the dependent variables, the standard error of the model SE, the correlation coefficient R2, and the standard deviation SD, are given for the individual point model and the five point subgroup model.
The prediction limits for each of the variables, for the individual points model and for the five point subgroup model are given in the second and third charts of Figs. 6C and 6D. These include the lower control limit (LCL), the average at minus 2 SD
(Standard Deviations), the average at minus 1 SD, the average, the averag8 at plus 1 SD, the average at plus 2 SD, and the upper control limit, are given.
The chart of Fig. 6D gives the validation statistics for the five point subgroup model. As can be seen in the chart, the model prediction for the average Yield is .15 percent less than the actual prediction, the average Acid consumption predicted by the model is 0.17 percent higher than the actual average Acid 2~1~8~5 consumption, and the average Research Octane Number for the model is 0.1 percent above the actual average Research Octane Number.
Also shown in Fig. 6D are the definitions of the statistical parameters used to evaluate the data.
In addition to the controlling of processes in the petroleum refinery field, as illustrated in this embodiment, the process control system of the instant invention is also applicable for the control of other types of processes. For example, in boiler processes, a deposit of solids which are dissolved in the water, tends to build-up on the sides of the boiler. The build-up is called fauling, and anti-foulants are often used to slow down the build-up of deposits. As deposits build-up on the inner walls of the boilers, the system becomes less efficient and more fuel and a higher temperature are required to raise the temperature of the water to the desired level. After a period of time, the boiler must be shut down and the deposits removed from the walls. This process is time consuming and results in increased costs and reduction in capacity of the system.
Additives to prevent fauling build-up (anti-foulants), by slowing down and minimizing the build-up, allow the system to operate more efficiently and for longer periods of time before shutdown.
Application of the instant invention to a boiler system could use for example the amount of fuel, the type of fuel, the measurement of the amount of dissolved solvents in the water, the rate of flow of the fuel, and the anti-foulant flow rate as 211888) independent variables. Objectives for optimization of the system could be maximizing the time between shutdowns, minimizing fuel costs, maximizing return on investment with regard to the chemical additive, maximizing process output or minimizing maintenance, labor anti costs.
The application of this process control system to such a boiler process could entail the use of the flow rate of the anti-foulant chemical additive as the controllable independent variable in a fashion similar to the use of the additive control rate for optimization of the alkylation unit process given previously.
A process control system has been described which is applicable, generally, to industrial commercial processes. It comprises both a process model, such as a neural network model, and an economic model. This enables the process~to be optimized for obtaining objectives relating to maximizing or minimizing dependent output variables or optimization of the system to meet economic criteria such as overall production value, return on investment, minimization of costs, and the like.
Without further elaboration, the foregoing will so fully illustrate my invention that others may, by applying current or future knowledge, readily adapt the same for use under the various conditions of service.
Claims (46)
1. A system for optimizing and controlling a process comprising:
(a) means for obtaining and receiving data, on-line and in real-time, representing the levels of a plurality of independent input variables of said process and the levels of at least one dependent output variable of said process;
(b) means for storing said data;
(c) means for determining the interrelationships among said plurality of independent input variables and the relationships between said plurality of independent variables and said at least one dependent output variable;
(d) means for generating a plot of said level of a selected one of said at least one dependent output variable at incremental levels of a selected one of said plurality of input variables over a pre-selected range of levels of said selected one of said plurality of said independent input variables;
(e) means for providing data representing at least one objective for said system relating to said at least one dependent output variable;
(f) means for obtaining a calculated level of a selected one of said plurality of independent input variables to meet said objective by determining the maximum or minimum level of said plot;
(g) means for moving said selected one of said plurality of independent variables to said calculated level; and (h) means for obtaining a calculated level for each of said plurality of independent variables in turn and for iteratively moving each of said plurality of independent variables to said respective calculated level,
(a) means for obtaining and receiving data, on-line and in real-time, representing the levels of a plurality of independent input variables of said process and the levels of at least one dependent output variable of said process;
(b) means for storing said data;
(c) means for determining the interrelationships among said plurality of independent input variables and the relationships between said plurality of independent variables and said at least one dependent output variable;
(d) means for generating a plot of said level of a selected one of said at least one dependent output variable at incremental levels of a selected one of said plurality of input variables over a pre-selected range of levels of said selected one of said plurality of said independent input variables;
(e) means for providing data representing at least one objective for said system relating to said at least one dependent output variable;
(f) means for obtaining a calculated level of a selected one of said plurality of independent input variables to meet said objective by determining the maximum or minimum level of said plot;
(g) means for moving said selected one of said plurality of independent variables to said calculated level; and (h) means for obtaining a calculated level for each of said plurality of independent variables in turn and for iteratively moving each of said plurality of independent variables to said respective calculated level,
2. The system of claim 1 wherein said means for determining said interrelationships and said relationships comprises a neural network model, which is trained off-line.
3. The system of claim 2 wherein said means for obtaining and receiving data comprises automatic means and said means for inserting said calculated value comprises automatic means.
4. The system of claim 2 wherein said means for obtaining and receiving data comprises automatic means and said means for inserting said calculated value comprises manual means.
5. The system of claim 2 wherein said means for obtaining and receiving data comprises manual means and said means for inserting said calculated value comprises manual means.
6. The system of claim 2 wherein said means for obtaining and receiving data comprises manual means and said means for inserting said calculated value comprises manual means.
7. The system of claim 2 wherein said system further comprises a computer with a memory for storing historical data relating to said process.
8. The system of claim 7 wherein said neural network comprises software in said computer which determines said interrelationships and relationship.
9. The system of claim 2 wherein said system further comprises an economic model.
10. The system of claim 9 wherein said system further provides means for receiving data representing at least one economic objective of said system.
11. The system of claim 10 wherein said economic model further comprises means for obtaining a calculated level of a selected one of said plurality of independent input variables to meet said economic objective.
12. A system for optimizing and controlling a process comprising:
(a) means for obtaining and receiving data, on-line and in real-time, representing the levels of a plurality of independent input variables of said process and the level of at least one dependent output variable of said process;
(b) means for storing said data;
(c) means for determining the interrelationships among said plurality of independent input variables and the relationships between said plurality of independent input variables and said at least one dependent output variable;
(d) means for generating a plot of said level of a selected one of said at least one dependent output variable at incremental levels of a selected one of said plurality of input variables over a pre-selected range of levels of said selected one of said plurality of said independent input variables;
(e) means for providing data representing at least one economic objective relating to an economic criteria to said system;
(f) means for generating a plot of the values of said economic criteria over said pre-selected range of levels of said selected one of said plurality of independent input variables;
(g) means for obtaining a calculated level of said selected one of said plurality of independent input variables to meet said economic objective by determining the maximum or minimum level of said plot;
(h) means for moving said selected one of said plurality of independent variables to said calculated level; and (i) means for obtaining a calculated level for each of said plurality of independent variables in turn and for iteratively moving each of said plurality of independent variables to said respective calculated level.
(a) means for obtaining and receiving data, on-line and in real-time, representing the levels of a plurality of independent input variables of said process and the level of at least one dependent output variable of said process;
(b) means for storing said data;
(c) means for determining the interrelationships among said plurality of independent input variables and the relationships between said plurality of independent input variables and said at least one dependent output variable;
(d) means for generating a plot of said level of a selected one of said at least one dependent output variable at incremental levels of a selected one of said plurality of input variables over a pre-selected range of levels of said selected one of said plurality of said independent input variables;
(e) means for providing data representing at least one economic objective relating to an economic criteria to said system;
(f) means for generating a plot of the values of said economic criteria over said pre-selected range of levels of said selected one of said plurality of independent input variables;
(g) means for obtaining a calculated level of said selected one of said plurality of independent input variables to meet said economic objective by determining the maximum or minimum level of said plot;
(h) means for moving said selected one of said plurality of independent variables to said calculated level; and (i) means for obtaining a calculated level for each of said plurality of independent variables in turn and for iteratively moving each of said plurality of independent variables to said respective calculated level.
13. The system of claim 1 wherein said means for determining said interrelationships and said relationships comprises a neural network model, which is trained off-line.
14. The system of claim 2 wherein said means for providing said economic basis and desired objective comprises an economic model.
15. The system of claim 3 wherein said means for obtaining and receiving data comprises automatic means and said means for inserting said calculated value comprises automatic means.
16. The system of claim 3 wherein said means for obtaining and receiving data comprises automatic means and said means for inserting said calculated value comprises manual means.
17. The system of claim 3 wherein said means for obtaining and receiving data comprises manual means and said means for inserting said calculated value comprises manual means.
18. The system of claim 3 wherein said means for obtaining and receiving data comprises manual means and said means for inserting said calculated value comprises automatic means.
19. The system of claim 3 wherein said system further comprises a computer with a memory for storing historical data relating to said process.
20. The system of claim 8 wherein said neural network comprises software in said computer which determines said interrelationships and relationships.
21. The system of claim 9 wherein said at least one economic objective comprises maximizing the total value of the output of said process.
22. The system of claim 9 wherein said at least one economic objective comprises maximizing the return on investment of the output of said process.
23. The system of claim 9 wherein said system controls a chemical process.
24. The system of claim 12 wherein said selected parameter comprises the injection rate of a chemical additive to said chemical process.
25. The system of claim 12 wherein said at least one economic objective comprises minimizing the level of consumption of catalytic material of said process.
26. The system of claim 12 wherein said at least one economic objective comprises maximizing the quality of the output of said process.
27. The system of claim 4 wherein said automatic means for storing said data comprises a data acquisition processor.
28. The system of claim 5 wherein said automatic means for storing said data comprises a data acquisition processor.
29. The system of claim 13 wherein said means for inserting said calculated level comprises a chemical pump controller and a chemical pump.
30. A system for controlling a petroleum reaction process for producing gasoline, said process comprising:
(a) means for obtaining and receiving data, on-line and in real-time, representing the levels of a plurality of independent input variables of said process and the level of at least one dependent output variable of said process;
(b) means for storing said data;
(c) means for determining the interrelationships among said plurality of independent input variables and the relationships between said plurality of independent input variables and said at least one dependent output variable;
(d) means for generating a plot of said level of a selected one of said at least one dependent output variable at incremental levels of a selected one of said plurality of input variables over a pre-selected range of levels of said selected one of said plurality of said independent input variables;
(e) means for providing data representing at least one economic objective relating to an economic criteria to said system;
(f) means for generating a plot of the values of said economic criteria over said pre-selected range of levels of said selected one of said plurality of independent input variables;
(g) means for obtaining a calculated level of said selected one of said plurality of independent input variables to meet said economic objective by determining the maximum or minimum level of said plot;
(h) means for moving said selected one of said plurality of independent variables to said calculated level; and (i) means for obtaining a calculated level for each of said plurality of independent variables in turn and for iteratively moving each of said plurality of independent variables to said respective calculated level.
(a) means for obtaining and receiving data, on-line and in real-time, representing the levels of a plurality of independent input variables of said process and the level of at least one dependent output variable of said process;
(b) means for storing said data;
(c) means for determining the interrelationships among said plurality of independent input variables and the relationships between said plurality of independent input variables and said at least one dependent output variable;
(d) means for generating a plot of said level of a selected one of said at least one dependent output variable at incremental levels of a selected one of said plurality of input variables over a pre-selected range of levels of said selected one of said plurality of said independent input variables;
(e) means for providing data representing at least one economic objective relating to an economic criteria to said system;
(f) means for generating a plot of the values of said economic criteria over said pre-selected range of levels of said selected one of said plurality of independent input variables;
(g) means for obtaining a calculated level of said selected one of said plurality of independent input variables to meet said economic objective by determining the maximum or minimum level of said plot;
(h) means for moving said selected one of said plurality of independent variables to said calculated level; and (i) means for obtaining a calculated level for each of said plurality of independent variables in turn and for iteratively moving each of said plurality of independent variables to said respective calculated level.
31. The system of claim 30 wherein said means for determining said interrelationships and said relationships comprises a neural network model, which is trained off-line.
32. The system of claim 31 wherein said means for providing said economic basis and desired objective comprises an economic model.
33. The system of claim 32 wherein said means for obtaining and receiving data comprises automatic means and said means for inserting said calculated value of said chemical additive comprises automatic means.
34. The system of claim 33 wherein said means for obtaining and receiving data comprises automatic means and said means for inserting said calculated value of said chemical additive comprises manual means.
35. The system of claim 34 wherein said means for obtaining and receiving data comprises manual means and said means for inserting said calculated value of said chemical additive comprises manual means.
36. The system of claim 35 wherein said means for obtaining and receiving data comprises manual means and said means for inserting said calculated value of said chemical additive comprises automatic means.
37. The system of claim 36 wherein said system further comprises a computer with a memory for storing historical data relating to said process.
38. The system of claim 37 wherein said neural network comprises software in said computer which determines said interrelationships and relationships.
39. The system of claim 38 wherein said at least one economic objective comprises maximizing the total value of the output of said gasoline of said process.
40. The system of claim 39 wherein said at least one economic objective comprises maximizing the return on investment of the chemical additive of said process.
41. The system of claim 40 wherein said at least one economic objective comprises minimizing the level of consumption of catalytic material of said process.
42. The system of claim 41 wherein said at least one economic objective comprises maximizing the octane number of said gasoline.
43. The system of claim 33 wherein said automatic means for storing said data comprises a data acquisition processor.
44. The system of claim 34 wherein said automatic means for storing said data comprises a data acquisition processor.
45. The system of claim 33 wherein said automatic means for inserting said calculated level comprises a chemical pump controller and a chemical pump.
46. The system of claim 35 wherein said automatic means for inserting said calculated level comprises a chemical pump controller and a chemical pump.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US5473893A | 1993-04-29 | 1993-04-29 | |
US08/054,738 | 1993-04-29 |
Publications (2)
Publication Number | Publication Date |
---|---|
CA2118885A1 CA2118885A1 (en) | 1994-10-30 |
CA2118885C true CA2118885C (en) | 2005-05-24 |
Family
ID=21993186
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CA002118885A Expired - Fee Related CA2118885C (en) | 1993-04-29 | 1994-03-11 | Process control system |
Country Status (2)
Country | Link |
---|---|
US (1) | US5521814A (en) |
CA (1) | CA2118885C (en) |
Families Citing this family (120)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6456982B1 (en) * | 1993-07-01 | 2002-09-24 | Dragana N. Pilipovic | Computer system for generating projected data and an application supporting a financial transaction |
US7222079B1 (en) * | 1994-06-23 | 2007-05-22 | Ingenix, Inc. | Method and system for generating statistically-based medical provider utilization profiles |
US5557514A (en) | 1994-06-23 | 1996-09-17 | Medicode, Inc. | Method and system for generating statistically-based medical provider utilization profiles |
US5696696A (en) * | 1994-10-11 | 1997-12-09 | Betzdearborn, Inc. | Apparatus and method for automatically achieving and maintaining congruent control in an industrial boiler |
US5923571A (en) * | 1994-10-11 | 1999-07-13 | Betzdearborn, Inc. | Apparatus and method for automatic congruent control of multiple boilers sharing a common feedwater line and chemical feed point |
US5704011A (en) * | 1994-11-01 | 1997-12-30 | The Foxboro Company | Method and apparatus for providing multivariable nonlinear control |
WO1996020438A1 (en) * | 1994-12-28 | 1996-07-04 | Omron Corporation | Method and system for inference using hierarchy model, and method and system for control |
US5649064A (en) * | 1995-05-19 | 1997-07-15 | The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | System and method for modeling the flow performance features of an object |
WO1997007471A1 (en) * | 1995-08-11 | 1997-02-27 | Hausman Robert E | Apparatus and method for modeling linear and quadratic programs |
WO1997048063A1 (en) * | 1996-06-07 | 1997-12-18 | Peter Schimitzek | Computer system for company management |
US5850339A (en) * | 1996-10-31 | 1998-12-15 | Giles; Philip M. | Analysis of data in cause and effect relationships |
US5943244A (en) * | 1997-02-17 | 1999-08-24 | I2 Technologies, Inc. | System for optimizing a network plan and method of operation |
US5764513A (en) * | 1997-02-27 | 1998-06-09 | Nec Corporation | Check system for production process procedure |
US5999918A (en) * | 1997-04-02 | 1999-12-07 | Rational Investors, Inc. | Interactive color confidence indicators for statistical data |
DE19715503A1 (en) * | 1997-04-14 | 1998-10-15 | Siemens Ag | Integrated computer and communication system for the plant area |
CA2289473A1 (en) * | 1997-05-21 | 1998-11-26 | Khimetrics, Inc. | Method for controlled optimization of enterprise planning models |
US6988076B2 (en) * | 1997-05-21 | 2006-01-17 | Khimetrics, Inc. | Strategic planning and optimization system |
US6153149A (en) * | 1997-08-06 | 2000-11-28 | The Trustees Of Princeton University | Adaptive feedback control flow reactor |
US6151585A (en) * | 1998-04-24 | 2000-11-21 | Microsoft Corporation | Methods and apparatus for determining or inferring influential rumormongers from resource usage data |
US7603286B2 (en) * | 1998-05-21 | 2009-10-13 | Sap Ag | Demand-model based price image calculation method and computer program therefor |
CA2336785A1 (en) | 1998-07-02 | 2000-01-13 | Kepner-Tregoe, Inc. | Method and apparatus for problem solving, decision making, storing, analyzing, retrieving enterprisewide knowledge and conclusive data |
US6249768B1 (en) * | 1998-10-29 | 2001-06-19 | International Business Machines Corporation | Strategic capability networks |
US7572169B1 (en) | 1998-11-06 | 2009-08-11 | Beaver Creek Concepts Inc | Advanced finishing control |
US7575501B1 (en) | 1999-04-01 | 2009-08-18 | Beaver Creek Concepts Inc | Advanced workpiece finishing |
US7037172B1 (en) | 1999-04-01 | 2006-05-02 | Beaver Creek Concepts Inc | Advanced wafer planarizing |
US7131890B1 (en) | 1998-11-06 | 2006-11-07 | Beaver Creek Concepts, Inc. | In situ finishing control |
US20130189801A1 (en) * | 1998-11-06 | 2013-07-25 | Semcon Tech, Llc | Advanced finishing control |
US7878882B2 (en) * | 1999-04-01 | 2011-02-01 | Charles J. Molnar | Advanced workpiece finishing |
US6986698B1 (en) | 1999-04-01 | 2006-01-17 | Beaver Creek Concepts Inc | Wafer refining |
US8353738B2 (en) * | 1998-11-06 | 2013-01-15 | Semcon Tech, Llc | Advanced finishing control |
US7008300B1 (en) | 2000-10-10 | 2006-03-07 | Beaver Creek Concepts Inc | Advanced wafer refining |
US6739947B1 (en) | 1998-11-06 | 2004-05-25 | Beaver Creek Concepts Inc | In situ friction detector method and apparatus |
US7220164B1 (en) | 2003-12-08 | 2007-05-22 | Beaver Creek Concepts Inc | Advanced finishing control |
US6068012A (en) * | 1998-12-29 | 2000-05-30 | Ashland, Inc. | Performance-based control system |
US8044793B2 (en) | 2001-03-01 | 2011-10-25 | Fisher-Rosemount Systems, Inc. | Integrated device alerts in a process control system |
US7562135B2 (en) * | 2000-05-23 | 2009-07-14 | Fisher-Rosemount Systems, Inc. | Enhanced fieldbus device alerts in a process control system |
US7206646B2 (en) * | 1999-02-22 | 2007-04-17 | Fisher-Rosemount Systems, Inc. | Method and apparatus for performing a function in a plant using process performance monitoring with process equipment monitoring and control |
US6975219B2 (en) * | 2001-03-01 | 2005-12-13 | Fisher-Rosemount Systems, Inc. | Enhanced hart device alerts in a process control system |
US7346404B2 (en) * | 2001-03-01 | 2008-03-18 | Fisher-Rosemount Systems, Inc. | Data sharing in a process plant |
DE19910910A1 (en) * | 1999-03-11 | 2000-09-28 | Voith Sulzer Papiertech Patent | Process for carrying out an optimized fiber or paper production process |
US6768928B1 (en) * | 1999-05-20 | 2004-07-27 | Olympus Optical Co., Ltd. | Mechanism component design support system |
AU5496100A (en) * | 1999-06-18 | 2001-01-09 | Thomson Financial Inc. | System, method and computer readable medium containing instructions for evaluating and disseminating investor performance information |
US7016872B1 (en) * | 1999-06-18 | 2006-03-21 | Thomson Financial Inc. | System, method and computer readable medium containing instructions for evaluating and disseminating investor performance information |
US6708155B1 (en) * | 1999-07-07 | 2004-03-16 | American Management Systems, Inc. | Decision management system with automated strategy optimization |
US6952679B1 (en) | 1999-07-26 | 2005-10-04 | Texas Shopper Network, Inc. | Method and system for evaluating quality services |
US6243615B1 (en) * | 1999-09-09 | 2001-06-05 | Aegis Analytical Corporation | System for analyzing and improving pharmaceutical and other capital-intensive manufacturing processes |
AU8001800A (en) * | 1999-10-06 | 2001-05-10 | Accenture Llp | Method and estimator for production scheduling |
WO2001027835A1 (en) * | 1999-10-08 | 2001-04-19 | Autonomous Effects, Inc. | System and method for efficiently solving complex systems of time-varying and time-invariant variables |
US7058197B1 (en) | 1999-11-04 | 2006-06-06 | Board Of Trustees Of The University Of Illinois | Multi-variable model for identifying crop response zones in a field |
US6775647B1 (en) | 2000-03-02 | 2004-08-10 | American Technology & Services, Inc. | Method and system for estimating manufacturing costs |
US6446721B2 (en) | 2000-04-07 | 2002-09-10 | Chevron U.S.A. Inc. | System and method for scheduling cyclic steaming of wells |
US6681153B1 (en) * | 2000-08-02 | 2004-01-20 | General Electric Company | Automated, interactive data collection, data transmittal and data assessment method |
US6931365B1 (en) * | 2000-12-18 | 2005-08-16 | Purdue Research Foundation | Industry simulation environment |
JP4678943B2 (en) * | 2000-12-22 | 2011-04-27 | 富士通株式会社 | New product initial input amount prediction method, new product initial input amount prediction device, and recording medium |
US6735541B2 (en) * | 2001-02-16 | 2004-05-11 | Exxonmobil Research And Engineering Company | Process unit monitoring program |
US8073967B2 (en) | 2002-04-15 | 2011-12-06 | Fisher-Rosemount Systems, Inc. | Web services-based communications for use with process control systems |
US7720727B2 (en) * | 2001-03-01 | 2010-05-18 | Fisher-Rosemount Systems, Inc. | Economic calculations in process control system |
US8214501B1 (en) | 2001-03-02 | 2012-07-03 | At&T Intellectual Property I, L.P. | Methods and systems for electronic data exchange utilizing centralized management technology |
US7716332B1 (en) | 2001-06-20 | 2010-05-11 | At&T Intellectual Property I, L.P. | System and method for server-based predictive caching of back-end system data |
US6549864B1 (en) | 2001-08-13 | 2003-04-15 | General Electric Company | Multivariate statistical process analysis systems and methods for the production of melt polycarbonate |
US20030046130A1 (en) * | 2001-08-24 | 2003-03-06 | Golightly Robert S. | System and method for real-time enterprise optimization |
US6810291B2 (en) * | 2001-09-14 | 2004-10-26 | Ibex Process Technology, Inc. | Scalable, hierarchical control for complex processes |
CH701481B1 (en) * | 2001-09-25 | 2011-01-31 | Roland Pulfer | Process management. |
US6915172B2 (en) | 2001-11-21 | 2005-07-05 | General Electric | Method, system and storage medium for enhancing process control |
US7797139B2 (en) | 2001-12-07 | 2010-09-14 | Chevron U.S.A. Inc. | Optimized cycle length system and method for improving performance of oil wells |
US7110924B2 (en) * | 2002-05-15 | 2006-09-19 | Caterpillar Inc. | Method for controlling the performance of a target system |
US7801976B2 (en) * | 2002-05-28 | 2010-09-21 | At&T Intellectual Property I, L.P. | Service-oriented architecture systems and methods |
US7376472B2 (en) * | 2002-09-11 | 2008-05-20 | Fisher-Rosemount Systems, Inc. | Integrated model predictive control and optimization within a process control system |
DE10393080T5 (en) * | 2002-10-08 | 2005-09-29 | Invensys Systems, Inc., Foxboro | service portal |
US7685085B2 (en) | 2003-11-10 | 2010-03-23 | James Ralph Heidenreich | System and method to facilitate user thinking about an arbitrary problem with output and interfaces to external systems, components and resources |
US7730009B1 (en) | 2002-11-11 | 2010-06-01 | Zxibix, Inc. | System and methods for archetype enabled research and search |
US7949617B1 (en) | 2002-11-11 | 2011-05-24 | Linda Shawn Higgins | System and methods for facilitating user thinking and learning utilizing enhanced interactive constructs |
US7720780B1 (en) | 2003-11-10 | 2010-05-18 | Zxibix, Inc. | System and method for facilitating collaboration and related multiple user thinking and cooperation regarding an arbitrary problem |
US8660972B1 (en) | 2002-11-11 | 2014-02-25 | Zxibix, Inc. | System and method to provide a customized problem solving environment for the development of user thinking about an arbitrary problem |
AU2003295430A1 (en) | 2002-11-11 | 2004-06-03 | Zxibix, Inc. | System and method of facilitating and evaluating user thinking about an arbitrary problem |
US10395173B1 (en) | 2002-11-11 | 2019-08-27 | Zxibix, Inc. | System and methods for exemplary problem solving, thinking and learning using an exemplary archetype process and enhanced hybrid forms |
US7016742B2 (en) * | 2002-11-27 | 2006-03-21 | Bahelle Memorial Institute | Decision support for operations and maintenance (DSOM) system |
US7092893B2 (en) * | 2003-01-28 | 2006-08-15 | Praxair Technology, Inc. | Control of liquid production of air separation plant network |
CH698890B1 (en) * | 2003-03-19 | 2009-11-30 | Roland Pulfer | Modeling a complex system. |
CH703081B1 (en) * | 2003-03-19 | 2011-11-15 | Roland Pulfer | Analysis of a model of a complex system. |
CH703073B1 (en) * | 2003-03-19 | 2011-11-15 | Roland Pulfer | Comparing models a complex system. |
US7711588B2 (en) * | 2003-04-03 | 2010-05-04 | Sap Ag | Method and computer program for field spectrum optimization |
US20040215697A1 (en) * | 2003-04-11 | 2004-10-28 | Vasileios Vergopoulos | Systems and methods for managing unit operating data |
US7030747B2 (en) * | 2004-02-26 | 2006-04-18 | Fisher-Rosemount Systems, Inc. | Method and system for integrated alarms in a process control system |
US8340981B1 (en) | 2004-03-02 | 2012-12-25 | Cave Consulting Group, Inc. | Method, system, and computer program product for physician efficiency measurement and patient health risk stratification utilizing variable windows for episode creation |
US7580852B2 (en) * | 2004-04-15 | 2009-08-25 | Sap Ag | System and method for modeling non-stationary time series using a non-parametric demand profile |
US7603304B2 (en) * | 2005-03-08 | 2009-10-13 | International Business Machines Corporation | Domain specific return on investment model system and method of use |
US20060229921A1 (en) * | 2005-04-08 | 2006-10-12 | Mr. Patrick Colbeck | Business Control System |
US8005647B2 (en) | 2005-04-08 | 2011-08-23 | Rosemount, Inc. | Method and apparatus for monitoring and performing corrective measures in a process plant using monitoring data with corrective measures data |
US9201420B2 (en) | 2005-04-08 | 2015-12-01 | Rosemount, Inc. | Method and apparatus for performing a function in a process plant using monitoring data with criticality evaluation data |
US7451004B2 (en) * | 2005-09-30 | 2008-11-11 | Fisher-Rosemount Systems, Inc. | On-line adaptive model predictive control in a process control system |
US20070202603A1 (en) * | 2006-02-27 | 2007-08-30 | Steven Wayne Counts | Apparatus and method for sampling and correcting fluids |
US7979840B2 (en) * | 2006-10-31 | 2011-07-12 | International Business Machines Corporation | Method and apparatus for service-oriented architecture process decomposition and service modeling |
US7991499B2 (en) * | 2006-12-27 | 2011-08-02 | Molnar Charles J | Advanced finishing control |
US7881825B2 (en) * | 2007-03-28 | 2011-02-01 | Praxair Technology, Inc. | Production control utilizing real time optimization |
US8489240B2 (en) * | 2007-08-03 | 2013-07-16 | General Electric Company | Control system for industrial water system and method for its use |
US8301676B2 (en) | 2007-08-23 | 2012-10-30 | Fisher-Rosemount Systems, Inc. | Field device with capability of calculating digital filter coefficients |
US20090055156A1 (en) * | 2007-08-24 | 2009-02-26 | Rockwell Automation Technologies, Inc. | Using commercial computing package models to generate motor control code |
US7702401B2 (en) | 2007-09-05 | 2010-04-20 | Fisher-Rosemount Systems, Inc. | System for preserving and displaying process control data associated with an abnormal situation |
US8055479B2 (en) | 2007-10-10 | 2011-11-08 | Fisher-Rosemount Systems, Inc. | Simplified algorithm for abnormal situation prevention in load following applications including plugged line diagnostics in a dynamic process |
US8357286B1 (en) | 2007-10-29 | 2013-01-22 | Semcon Tech, Llc | Versatile workpiece refining |
US8145334B2 (en) * | 2008-07-10 | 2012-03-27 | Palo Alto Research Center Incorporated | Methods and systems for active diagnosis through logic-based planning |
US8219437B2 (en) * | 2008-07-10 | 2012-07-10 | Palo Alto Research Center Incorporated | Methods and systems for constructing production plans |
US8266092B2 (en) | 2008-07-10 | 2012-09-11 | Palo Alto Research Center Incorporated | Methods and systems for target value path identification |
US8165705B2 (en) * | 2008-07-10 | 2012-04-24 | Palo Alto Research Center Incorporated | Methods and systems for continuously estimating persistent and intermittent failure probabilities for production resources |
US20100163469A1 (en) * | 2008-12-26 | 2010-07-01 | Zhaoyang Wan | Control system for monitoring localized corrosion in an industrial water system |
US8359110B2 (en) | 2009-03-23 | 2013-01-22 | Kuhn Lukas D | Methods and systems for fault diagnosis in observation rich systems |
US8505881B2 (en) | 2009-10-12 | 2013-08-13 | Enviromix, Llc | Mixing systems and methods of mixing |
US8323498B2 (en) * | 2010-10-04 | 2012-12-04 | Enviromix, Llc | Systems and methods for automated control of mixing and aeration in treatment processes |
US9927788B2 (en) | 2011-05-19 | 2018-03-27 | Fisher-Rosemount Systems, Inc. | Software lockout coordination between a process control system and an asset management system |
US11599892B1 (en) | 2011-11-14 | 2023-03-07 | Economic Alchemy Inc. | Methods and systems to extract signals from large and imperfect datasets |
US9031892B2 (en) | 2012-04-19 | 2015-05-12 | Invensys Systems, Inc. | Real time safety management system and method |
US20140095268A1 (en) * | 2012-09-28 | 2014-04-03 | Avaya Inc. | System and method of improving contact center supervisor decision making |
TR201515610A2 (en) * | 2015-12-08 | 2017-06-21 | Koc Ueniversitesi | A control system for modeling and optimizing a diesel hydroprocessing unit |
US10930535B2 (en) | 2016-12-02 | 2021-02-23 | Applied Materials, Inc. | RFID part authentication and tracking of processing components |
US10370599B2 (en) | 2017-01-06 | 2019-08-06 | Saudi Arabian Oil Company | Methods and systems for optimizing demulsifier and wash water injection rates in gas oil separation plants |
US11603326B2 (en) | 2017-09-29 | 2023-03-14 | EnviroMix, Inc. | Systems and methods for treatment processes |
US11449578B2 (en) * | 2019-09-27 | 2022-09-20 | Botty Todorov DIMANOV | Method for inspecting a neural network |
US11956513B2 (en) * | 2020-12-14 | 2024-04-09 | Exfo Inc. | Automated calibration of QoE assessment between content delivery network (CDN) entities |
US20220405817A1 (en) * | 2021-06-16 | 2022-12-22 | Pod Foods Co. | Virtualized wholesaling |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US3725653A (en) * | 1968-04-11 | 1973-04-03 | Gulf Research Development Co | Apparatus for controlling chemical processes |
US3648035A (en) * | 1969-06-02 | 1972-03-07 | Industrial Nucleonics Corp | System and method for optimizing processor or equipment profit |
US4858147A (en) * | 1987-06-15 | 1989-08-15 | Unisys Corporation | Special purpose neurocomputer system for solving optimization problems |
IE71168B1 (en) * | 1988-09-30 | 1997-01-29 | Torpey Patrick | A method and an apparatus for extracting liquid from a sludge |
JP2821189B2 (en) * | 1989-09-01 | 1998-11-05 | 株式会社日立製作所 | Learning type decision support system |
GB8927567D0 (en) * | 1989-12-06 | 1990-02-07 | Gen Electric Co Plc | Detection of chemicals |
US5111531A (en) * | 1990-01-08 | 1992-05-05 | Automation Technology, Inc. | Process control using neural network |
US5159660A (en) * | 1990-08-09 | 1992-10-27 | Western Thunder | Universal process control using artificial neural networks |
US5268834A (en) * | 1991-06-24 | 1993-12-07 | Massachusetts Institute Of Technology | Stable adaptive neural network controller |
US5282131A (en) * | 1992-01-21 | 1994-01-25 | Brown And Root Industrial Services, Inc. | Control system for controlling a pulp washing system using a neural network controller |
-
1994
- 1994-03-11 CA CA002118885A patent/CA2118885C/en not_active Expired - Fee Related
- 1994-10-18 US US08/324,863 patent/US5521814A/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
US5521814A (en) | 1996-05-28 |
CA2118885A1 (en) | 1994-10-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CA2118885C (en) | Process control system | |
Thompson et al. | Modeling chemical processes using prior knowledge and neural networks | |
US7058617B1 (en) | Method and apparatus for training a system model with gain constraints | |
CN102841538B (en) | For controlling the method and system of batch process | |
KR940005029B1 (en) | Supporting method and system for process control | |
CN104932488A (en) | Model predictive control performance evaluation and diagnosis method | |
KR102440372B1 (en) | Providing method, apparatus and computer-readable medium of managing influent environmental information of sewage treatment facilities based on big data and artificial intelligence | |
Lu et al. | Fuzzy modeling and expert optimization control for industrial processes | |
CN104361153A (en) | Method for predicting coking amount of heavy oil catalytic cracking settler | |
Duarte et al. | Combined mechanistic and empirical modelling | |
Ghosh et al. | Application of data-driven modeling approaches to industrial hydroprocessing units | |
Rippin | Control of batch processes | |
JPH03134706A (en) | Knowledge acquiring method for supporting operation of sewage-treatment plant | |
Chen et al. | Neural network-based predictive control for multivariable processes | |
Santander et al. | Deep learning economic model predictive control for refinery operation: A fluid catalytic cracker-fractionator case study | |
Gutierrez et al. | Integration of Design and Control using an MPC-based superstructure for control structure selection. | |
MacKinnon et al. | Robust closed-loop dynamic real-time optimization | |
CN101339408A (en) | Sodium aluminate solution continuous carbonation decomposition course intelligence control system | |
Xiong et al. | Optimal control of fed-batch processes based on multiple neural networks | |
Cahyadi et al. | Salt Fields Productivity Forecasting Based On Sunlight Duration, Wind Speed and Temperature Data | |
Yu et al. | Predictive control based on neural networks of the chemical process | |
CN115795315B (en) | Abnormality sensing method and system for chemical liquid supply equipment | |
Zhang et al. | Nonlinear model predictive control based on multiple local linear models | |
US20240077860A1 (en) | Method for controlling a plant of separation and treatment industrial processes without chemical reaction | |
MacKinnon et al. | Robust Multi-Scenario Dynamic Real-Time Optimization with Embedded Closed-Loop Model Predictive Control |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
EEER | Examination request | ||
MKLA | Lapsed |