Method and a system for tuning multivariable PID controller

Patent metadata

Field	Value
Publication number	US-10082772-B2
Application number	US-201414261109-A
Country	US
Kind code	B2
Filing date	Apr 24, 2014
Priority date	Oct 24, 2011
Publication date	Sep 25, 2018
Grant date	Sep 25, 2018

Abstract

A method and system are disclosed for tuning a Multivariable Proportional-Integral-Derivative (PID) controller in a process plant. The plant can have at least one PID controller for controlling the processes of the plant. The method can include formulating Model Predictive Control (MPC) for the process model of the plant; obtaining a desired (e.g., optimal control) solution for at least one designated (e.g., critical) region in relation to the MPC; determining PID tuning parameters corresponding to the control solution; and tuning the PID controller with the PID tuning parameters.

First claim

I claim:

1. A method for tuning an adaptive Multivariable Proportional-Integral-Derivative (PID) controller within an operating range of the Multivariable PID controller, in a plant using a Model Predictive Control (MPC), the operating range comprising a plurality of critical regions, the method comprising:
formulating the MPC for a process model of the plant, wherein a solution of the MPC accounts for constraints on the plant;
obtaining an optimal control solution of the MPC in one or more critical regions from the plurality of critical regions using a look-up table comprising a plurality of rows, wherein each row of the look-up table is associated with a critical region from the plurality of critical regions, wherein each critical region from the plurality of critical regions is associated with a corresponding affine control law in the look-up table, the corresponding affine control law having the form U=M·θ+m, where U is an output, θis an input, and M and m are constant matrices, and wherein the optimal control solution accounts for the constraints on the plant and the optimal control solution is for providing the affine control law corresponding to each of the one or more critical regions;
determining one or more equivalent PID tuning parameters for the one or more critical regions corresponding to the optimal control solution of the MPC in the one or more critical regions based on the one or more corresponding affine control laws from the plurality of affine control laws, wherein determining the one or more equivalent PID tuning parameters comprises one of (i) merging the one or more critical regions and (ii) clustering the one or more critical regions by merging the one or more equivalent PID tuning parameters corresponding to the one or more critical regions; and
tuning the adaptive Multivariable PID controller with the one or more equivalent PID tuning parameters to control the plant in the one or more critical regions.

2. The method as claimed in claim 1 , comprising:
formulating the MPC and obtaining the optimal control solution offline.

3. The method as claimed in claim 1 , comprising:
obtaining the optimal control solution for the one or more critical regions purports to an affine function of one or more of states, inputs, or error in a critical region of operation.

4. The method as claimed in claim 1 , wherein determining the one or more equivalent PID tuning parameters comprises:
reducing the number of sets of PID parameters.

5. The method as claimed in claim 4 , wherein reducing the number of sets of PID tuning parameters comprises:
merging a plurality of critical regions of operation, and determining PID tuning parameters corresponding to the plurality of critical regions of operations thereof.

6. The method as claimed in claim 4 , wherein reducing the number of sets of PID tuning parameters comprises:
grouping two or more sets of PID tuning parameters corresponding to different critical regions, and merging the different critical regions.

7. The method as claimed in claim 1 , wherein obtaining the optimal control solution, in relation to the MPC comprises:
accounting for one or more constraints of the plant.

8. The method as claimed in claim 1 , wherein the PID tuning parameters account for the predictive capability of the MPC in relation to the process model, by which an output error is minimized in a predictive manner.

9. The method as claimed in claim 1 , wherein the PID tuning parameters and their corresponding critical regions purport to the adaptive Multivariable PID controller with adaptive gain.

10. The method as claimed in claim 1 , comprising:
implementing the adaptive Multivariable PID controller online in a closed loop in the plant.

Assignees

• Abb Research Ltd

Inventors

• Gupta Arun

Classifications

• G05B13/04 Primary

  involving the use of models or simulators · CPC title

• G05B13/048 Primary

  using a predictor · CPC title