ONLINE INTEGRATION OF MODEL-BASED OPTIMIZATION AND MODEL-LESS CONTROL

• US 20160170393A1
• Filed: 02/23/2016
• Published: 06/16/2016
• Est. Priority Date: 11/05/2012
• Status: Active Grant

• Current Assignee: Rockwell Automation Technologies Incorporated (Allen-Bradley Co., Inc.)
• Original Assignee: Rockwell Automation Technologies Incorporated (Allen-Bradley Co., Inc.)
• Inventors: Sayyarrodsari, Bijan, Liano, Kadir, Smith, Alexander B.
• Granted Patent: US 10,620,604 B2
• Time in Patent Office: Days
• CPC Class Codes: G05B 13/042 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: G05B 19/045 using logic state machines, consisting only of a memory or a programmable logic device containing the logic for the controlled machine and in which the state of its outputs is dependent on the state of its inputs or part of its own output states, e.g. binary decision controllers, finite state controllers
• CPC Class Codes: G05B 2219/25273 Fuzzy logic combined with delay element

Abstract

In certain embodiments, a control system includes a model-less controller configured to control operation of a plant or process. The control system also includes a model-based controller that includes a model of the plant or process being controlled by the model-less controller. The model-based controller is configured to modify parameters of the model-less controller.

Claims

1. A control system comprising:

a model-less controller configured to determine a manipulated variable of an industrial plant or process based at least in part on a tuning parameter of the model-less controller without use of any models that model operation of the industrial plant or process; and

an automation controller communicatively coupled to the model-less controller, wherein the automation controller is configured to;

determine a model configured to model operation of the industrial plant or process; and

execute an explicit optimization procedure to determine a trajectory of the tuning parameter over a prediction horizon based at least in part on the model.

2. The control system of claim 1, wherein:

the model-less controller comprises a fuzzy logic controller; and

the tuning parameter comprises a mean value configured to indicate shape of a fuzzy membership function, wherein the fuzzy logic controller is configured to use the fuzzy membership function to determine the manipulated variable.

3. The control system of claim 1, wherein the model-less controller comprises an expert system.

4. The control system of claim 1, wherein the model-less controller is configured to determine the manipulated variable based at least in part on the tuning parameter multiplied by a difference between an output of the industrial plant or process and a desired output.

5. The control system of claim 1, wherein the explicit optimization procedure comprises an explicit search strategy that explicitly uses the model.

6. An industrial system comprising:

a model-based controller configured to;

determine a model configured to model operation of the industrial system; and

perform an explicit optimization procedure during the operation of the industrial system based at least in part on the model to determine a trajectory of a tuning parameter over a prediction horizon; and

a model-less controller communicatively coupled to the model-based controller, wherein the model-less controller is configured to;

determine a manipulated variable of the industrial system based at least in part on the trajectory of the tuning parameter without using the model; and

instruct the industrial system to implement the manipulated variable during a single time step in the prediction horizon to control operation of the industrial system.

7. The industrial system of claim 6, wherein:

the model-less controller comprises a fuzzy logic controller; and

the tuning parameter comprises a mean value configured to indicate shape of a fuzzy membership function, wherein the fuzzy logic controller is configured to use the fuzzy membership function to determine the manipulated variable.

8. The industrial system of claim 6, wherein the model-less controller comprises an expert system.

9. The industrial system of claim 6, wherein the industrial system comprises a control device comprising a multi-core processor, wherein:

a first core of the multi-core processor is configured to implement the model-less controlled; and

a second core of the multi-core processor is configured to implement the model-based controller.

10. The industrial system of claim 6, wherein the model-based controller is configured to determine the trajectory of the tuning parameter subject constraints on value of the tuning parameter, constraints on rate of change of the tuning parameter, or both.

11. The industrial system of claim 6, wherein the model-less controller is configured to determine the manipulated variable based at least in part on value of the tuning parameter associated with the single time step.

12. The industrial system of claim 6, wherein value of the tuning parameter changes each execution cycle of the model-less controller.

13. The industrial system of claim 6, wherein the prediction horizon comprises a plurality of time steps.

14. The industrial system of claim 6, wherein the industrial system comprises an industrial plant or an industrial process.

15. A tangible, non-transitory, computer-readable medium configured to store instructions executable by one or more processors communicatively coupled to an industrial system, wherein the instructions comprise instructions to:

determine, using the one or more processors, a first trajectory of a tuning parameter over a first prediction horizon by performing a first explicit optimization procedure based at least in part on a model that describes operation of the industrial system;

determine, using the one or more processors, a first value of a manipulated variable based at least in part on the first trajectory of the tuning parameter without explicitly using the model; and

instruct, using the one or more processors, the industrial system to implement the first value of the manipulated variable during a first time step in the first prediction horizon to control operation of the industrial system.

16. The tangible, non-transitory, computer-readable medium of claim 15, comprising instructions to:

determine, using the one or more processors, a second trajectory of the tuning parameter over a second prediction horizon by performing a second explicit optimization procedure based at least in part on the model, wherein the second prediction horizon is different from the first prediction horizon and the second trajectory of the tuning parameter is different from the first trajectory of the tuning parameter;

determine, using the one or more processors, a second value of the manipulated variable based at least in part on the second trajectory of the tuning parameter; and

instruct, using the one or more processors, the industrial system to implement the second value during a second time step in the second prediction horizon to control operation of the industrial system.

17. The tangible, non-transitory, computer-readable medium of claim 16, wherein the instructions to determine the second value of the manipulated variable comprise instructions to:

determine, using the one or more processors, a third trajectory of the tuning parameter over the second prediction horizon by time shifting the first trajectory of the tuning parameter to the second prediction horizon;

determine, using the one or more processors, a first cost associated with using the second trajectory of the tuning parameter to determine the second value of the manipulated variable based at least in part on an objective function;

determine, using the one or more processors, a second cost associated with using the third trajectory of the tuning parameter to determine the second value of the manipulated variable based at least in part on the objective function; and

determine, using the one or more processors, the second value of the manipulated variable using the second trajectory of the tuning parameter when the first cost is less than the second cost and using the third trajectory of the tuning parameter when the second cost is less than the first cost.

18. The tangible, non-transitory, computer-readable medium of claim 15, comprising instructions to parameterize, using the one or more processors, a model-less controller into a plurality of tuning parameters, wherein the plurality of tuning parameters comprises the tuning parameter and the model-less controller is configured to determine the first value of the manipulated variable based at least in part on the first trajectory of the tuning parameter without explicitly accessing the model.

19. The tangible, non-transitory, computer-readable medium of claim 18, wherein:

the model-less controller comprises a fuzzy logic controller; and

the tuning parameter comprises a mean value configured to indicate shape of a fuzzy membership function, wherein the fuzzy logic controller is configured to use the fuzzy membership function to determine the manipulated variable.

20. The tangible, non-transitory, computer-readable medium of claim 18, wherein the model-less controller comprises an expert system.