Apparatus and method for adjustable identification of controller feasibility regions to support cascaded model predictive control (MPC)

Patent metadata

Field	Value
Publication number	US-10466684-B2
Application number	US-201715604903-A
Country	US
Kind code	B2
Filing date	May 25, 2017
Priority date	May 25, 2017
Publication date	Nov 5, 2019
Grant date	Nov 5, 2019

Abstract

A method includes receiving, at a master MPC controller from a slave MPC controller, proxy limit values indicating to what extent the slave controller is able to change multiple manipulated variables in multiple directions within a variable space without violating process variable constraints of the slave controller. The variable space includes a first feasibility region defined by the process variable constraints. The method also includes estimating a second feasibility region associated with the slave controller. At least part of the second feasibility region resides within the first feasibility region. The method further includes performing plantwide optimization at the master controller. A solution generated during the plantwide optimization includes a combination of manipulated variable values within the second feasibility region. Estimating the second feasibility region includes identifying edges of the second feasibility region based on the proxy limit values. At least some of the edges meet at locations away from the proxy limit values.

First claim

What is claimed is:

1. A method comprising:
receiving, at a master model predictive control (MPC) controller from a slave MPC controller, proxy limit values indicating to what extent the slave MPC controller is able to change multiple manipulated variables in multiple directions within a variable space without violating process variable constraints of the slave MPC controller, the variable space including a first feasibility region defined by the process variable constraints;
estimating a second feasibility region associated with the slave MPC controller, at least part of the second feasibility region residing within the first feasibility region; and
performing plantwide optimization at the master MPC controller, wherein a solution generated during the plantwide optimization includes a combination of manipulated variable values within the second feasibility region;
wherein estimating the second feasibility region comprises identifying edges of the second feasibility region based on the proxy limit values, at least some of the edges meeting at locations away from the proxy limit values.

2. The method of claim 1 , wherein at least one portion of the second feasibility region resides outside of the first feasibility region.

3. The method of claim 1 , wherein identifying the edges of the second feasibility region comprises:
identifying a user-defined scaling factor;
scaling the proxy limit values based on the user-defined scaling factor; and
identifying edges that pass through the scaled proxy limit values.

4. The method of claim 3 , wherein:
a first value of the user-defined scaling factor causes the second feasibility region to reside completely within the first feasibility region; and
a second value of user-defined scaling factor causes at least one portion of the second feasibility region to reside outside of the first feasibility region.

5. The method of claim 1 , wherein identifying the edges of the second feasibility region comprises identifying edges that pass through the proxy limit values.

6. The method of claim 1 , wherein identifying the edges of the second feasibility region comprises identifying a scalar rectangle, a size of the scalar rectangle based on a user-defined scaling factor.

7. The method of claim 1 , further comprising:
repeating the receiving and estimating steps for multiple slave MPC controllers.

8. An apparatus comprising:
a master model predictive control (MPC) controller comprising:
at least one interface configured to receive, from a slave MPC controller, proxy limit values indicating to what extent the slave MPC controller is able to change multiple manipulated variables in multiple directions within a variable space without violating process variable constraints of the slave MPC controller, the variable space including a first feasibility region defined by the process variable constraints; and
at least one processing device configured to:
estimate a second feasibility region associated with the slave MPC controller, at least part of the second feasibility region residing within the first feasibility region; and
perform plantwide optimization, wherein a solution generated during the plantwide optimization includes a combination of manipulated variable values within the second feasibility region;

wherein, to estimate the second feasibility region, the at least one processing device is configured to identify edges of the second feasibility region based on the proxy limit values, at least some of the edges meeting at locations away from the proxy limit values.

9. The apparatus of claim 8 , wherein at least one portion of the second feasibility region resides outside of the first feasibility region.

10. The apparatus of claim 8 , wherein, to identify the edges of the second feasibility region, the at least one processing device is configured to:
identify a user-defined scaling factor;
scale the proxy limit values based on the user-defined scaling factor; and
identify edges that pass through the scaled proxy limit values.

11. The apparatus of claim 10 , wherein:
a first value of the user-defined scaling factor causes the second feasibility region to reside completely within the first feasibility region; and
a second value of user-defined scaling factor causes at least one portion of the second feasibility region to reside outside of the first feasibility region.

12. The apparatus of claim 8 , wherein, to identify the edges of the second feasibility region, the at least one processing device is configured to identify edges that pass through the proxy limit values.

13. The apparatus of claim 8 , wherein, to identify the edges of the second feasibility region, the at least one processing device is configured to identify a scalar rectangle, a size of the scalar rectangle based on a user-defined scaling factor.

14. The apparatus of claim 8 , wherein the at least one processing device is further configured to repeat the receiving and estimating operations for multiple slave MPC controllers.

15. A non-transitory computer readable medium containing computer readable program code that when executed causes at least one processing device to:
receive, at a master model predictive control (MPC) controller from a slave MPC controller, proxy limit values indicating to what extent the slave MPC controller is able to change multiple manipulated variables in multiple directions within a variable space without violating process variable constraints of the slave MPC controller, the variable space including a first feasibility region defined by the process variable constraints;
estimate a second feasibility region associated with the slave MPC controller, at least part of the second feasibility region residing within the first feasibility region; and
perform plantwide optimization at the master MPC controller, wherein a solution generated during the plantwide optimization includes a combination of manipulated variable values within the second feasibility region;
wherein the computer readable program code that when executed causes the at least one processing device to estimate the second feasibility region comprises:
computer readable program code that when executed causes the at least one processing device to identify edges of the second feasibility region based on the proxy limit values, at least some of the edges meeting at locations away from the proxy limit values.

16. The non-transitory computer readable medium of claim 15 , wherein at least one portion of the second feasibility region resides outside of the first feasibility region.

17. The non-transitory computer readable medium of claim 15 , wherein the computer readable program code that when executed causes the at least one processing device to identify the edges of the second feasibility region comprises:
computer readable program code that when executed causes the at least one processing device to:
identify a user-defined scaling factor;
scale the proxy limit values based on the user-defined scaling factor; and
identify edges that pass through the scaled proxy limit values.

18. The non-transitory computer readable medium of claim 17 , wherein:
a first value of the user-defined scaling factor causes the second feasibility region to reside completely within the first feasibility region; and
a second value of user-defined scaling factor causes at least one portion of the second feasibility region to reside outside of the first feasibility region.

19. The non-transitory computer readable medium of claim 15 , wherein the computer

Assignees

• Honeywell Int Inc

Inventors

• Lu Joseph Z

Classifications

• G05B2219/2231

  Leader-follower · CPC title

• G05B19/41805 Primary

  characterised by assembly · CPC title

• G05B13/048 Primary

  using a predictor · CPC title

• Y02P90/02

  Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS] · CPC title