Robust steady-state target calculation for model predictive control

• US 6,381,505 B1
• Filed: 03/11/1999
• Issued: 04/30/2002
• Est. Priority Date: 09/28/1998
• Status: Expired due to Term

• Current Assignee: Aspentech Corporation
• Original Assignee: Aspen Technology (Telefonaktiebolaget LM Ericsson)
• Inventors: Hawkins, Robert B., Badgwell, Thomas A., Kassmann, Dean E.
• Primary Examiner(s): Grant, William
• Assistant Examiner(s): CABRERA, ZOILA E
• Application Number:
  US09/266,709
• Time in Patent Office: 1,146 Days
• Field of Search: 700/28, 700/29, 700/30, 700/34, 700/40-44, 700/52, 703/2
• US Class Current: 700/44
• CPC Class Codes: G05B 13/048 using a predictor

Abstract

A method and apparatus for steady-state target calculation that explicitly accounts for model uncertainty is disclosed. In accordance with one aspect of the invention, when model uncertainty is incorporated, the linear program associated with the steady-state target calculation can be recast as a highly structured nonlinear program. In accordance with another aspect of the invention, primal-dual interior point methods can be applied to take advantage of the resulting special structure. For a system having characteristic gain parameters G having a known uncertainty description, the present invention provides a method and apparatus for selecting steady-state targets for said system-manipulated variables such that all system-controlled variables will remain feasible at steady-state for all possible values of the parameters G within the known uncertainty description. A nominal estimate {tilde over (G)} of the system parameters G is made, and in accordance with another aspect of the invention, the steady-state targets are selected such that when {tilde over (G)}=G, the system is driven to an operational steady-state in which the objective function is extremized.

Claims

1. A method of controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, G represents system gain parameters, and where J is a function of G, said method comprising the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description and where {tilde over (J)} is said objective function J applied to {tilde over (G)};

(b) computing steady-state targets for said system-manipulated input variables u such that all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description.

2. A method in accordance with claim 1, wherein said step (b) of computing steady-state targets comprises computing said steady-state targets for said system-manipulated input variables u in accordance with said objective function {tilde over (J)}.

3. A method in accordance with claim 1, wherein said step (b) of computing steady-state targets comprises computing said steady-state targets for said system-manipulated input variables u such that said objective function {tilde over (J)} is extremized.

4. A method in accordance with claim 3, wherein as {tilde over (G)} approaches G, said computed steady-state targets approach those steady-state targets that extremize J, subject to said known uncertainty description.

5. A method in accordance with claim 4, wherein if {tilde over (G)}=G, then said computed steady-state targets extremize J, subject to said known uncertainty description.

6. A method in accordance with claim 1, wherein said step (b) of computing steady-state targets for system-manipulated input variables comprises extremizing said objective function J.

7. A method in accordance with claim 1, wherein said step (b) of computing steady-state targets for system-manipulated input variables comprises extremizing said objective function J cast as a second-order cone equation.

8. A method in accordance with claim 1, wherein said step (b) of computing steady-state targets for system-manipulated input variables comprises extremizing said objective function Jcast as an enumerated linear equation.

9. A method of controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters, said method comprising the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description;

(b) extremizing said objective function J by solving the nonlinear equation $\underset{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u,\mathrm{\epsilon <br><br>}}{\min }<br><br>\text{\hspace{1em}<br><br>}<br><br>c^T<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+d^T<br><br>\tilde{G}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+{\mathrm{<br><br>}}^T<br><br>\mathrm{\varepsilon <br><br>}$

 

subject to

10. A method of controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G is a matrix representing system gain parameters, said method comprising the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description;

(b) ascribing minimum and maximum bounds to elements of G;

(c) extremizing said objective function J by solving the linear equation $\underset{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u,\mathrm{\epsilon <br><br>}}{\min }<br><br>\text{\hspace{1em}<br><br>}<br><br>c^T<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+d^T<br><br>\tilde{G}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+{\mathrm{<br><br>}}^T<br><br>\mathrm{\varepsilon <br><br>}$

 

subject to

11. A method of controlling a system in accordance with an objective function J, wherein the operation of said system is describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters, said method comprising the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description;

(b) extremizing said objective function J by solving the nonlinear equation $\underset{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u,\mathrm{\epsilon <br><br>}}{\min }<br><br>\text{\hspace{1em}<br><br>}<br><br>c^T<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+d^T<br><br>\tilde{G}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+{\mathrm{<br><br>}}^T<br><br>\mathrm{\varepsilon <br><br>}$

 

subject to

12. An apparatus for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters having a known uncertainty description, and where J is a function of G, said apparatus comprising;

(a) an estimator for deriving a nominal estimate {tilde over (G)} of said system gain parameters G, where {tilde over (G)} has a known uncertainty description and where {tilde over (J)} is said objective function J applied to {tilde over (G)};

(b) a computing circuit, responsive to at least one past value of said system-controlled output variables and of said system-manipulated input variables to compute steady-state targets for said system-manipulated input variables u such that all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description.

13. An apparatus in accordance with claim 12, wherein said computing circuit computes said steady-state targets in accordance with said objective function {tilde over (J)}.

14. An apparatus in accordance with claim 12, wherein said computing circuit comprises computational circuitry for computing said steady-state targets for said system-manipulated input variables u such that said objective function objective function {tilde over (J)} is extremized.

15. An apparatus in accordance with claim 14, wherein as {tilde over (G)} approaches G, said computed steady-state targets approach those steady-state targets that extremize J, subject to said known uncertainty description.

16. An apparatus in accordance with claim 15, wherein if {tilde over (G)}=G, then said computed steady-state targets extremize J, subject to said known uncertainty description.

17. An apparatus in accordance with claim 12, wherein said computing circuit comprises computational circuitry for extremizing an objective function cast as a second-order cone equation.

18. An apparatus in accordance with claim 12, wherein said computing circuit comprises computational circuitry for extremizing an objective function cast as an enumerated linear equation.

19. A computer program product for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, G represents system gain parameters, and where J is a function of G, the computer program embodied on a computer readable medium readable by the apparatus of claim 12 and comprising instructions that, when executed, cause the apparatus of claim 12 to perform the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description and where {tilde over (J)} is said objective function J applied to {tilde over (G)};

(b) computing steady-state targets for said system-manipulated input variables u such that all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description.

20. An apparatus for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters having a known uncertainty description, and where J is a function of G, said apparatus comprising;

(a) estimator means for deriving a nominal estimate {tilde over (G)} of said system gain parameters G, where {tilde over (J)} is said objective function applied to {tilde over (G)};

(b) computing means, responsive to at least one past value of said system-controlled output variables and of said system-manipulated input variables to compute steady-state targets for said system-manipulated input variables u such that all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description.

21. An apparatus in accordance with claim 20, wherein said computing means comprises means for computing steady-state targets in accordance with said objective function.

22. An apparatus in accordance with claim 20, wherein said computing means comprises means for computing steady-state targets for said system-manipulated input variables u such that said objective function {tilde over (J)} is extremized.

23. An apparatus in accordance with claim 22, wherein as {tilde over (G)} approaches G, said computed steady-state targets approach those steady-state targets that extremize J, subject to said known uncertainty description.

24. An apparatus in accordance with claim 23, wherein if {tilde over (G)}=G, then said computed steady-state targets extremize J, subject to said known uncertainty description.

25. An apparatus in accordance with claim 20, wherein said computing means comprises means for extremizing said objective function J.

26. An apparatus in accordance with claim 20, wherein said computing means comprises means for extremizing said objective function J cast as a second-order cone equation.

27. An apparatus in accordance with claim 20, wherein said computing means comprises means for extremizing said objective function J cast as an enumerated linear equation.

28. An apparatus for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters having a known uncertainty description, said apparatus comprising;

(a) an estimator for deriving a nominal estimate {tilde over (G)} of said system gain parameters G;

(b) computation circuitry responsive to at least one past value of said system-controlled output variables and of said system-manipulated input variables, and responsive to said nominal estimate {tilde over (G)}, for extremizing said objective function J by solving the equation $\underset{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u,\mathrm{\epsilon <br><br>}}{\min }<br><br>\text{\hspace{1em}<br><br>}<br><br>c^T<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+d^T<br><br>\tilde{G}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+{\mathrm{<br><br>}}^T<br><br>\mathrm{\varepsilon <br><br>}$

 

subject to

29. An apparatus controlling a system in accordance with an objective function J, wherein the operation of said system is describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters having a known uncertainty description, said method comprising the steps of;

(a) estimator circuitry for deriving a nominal estimate {tilde over (G)} of said system gain parameters G;

(b) computation circuitry, responsive to at least one past value of said system-controlled output variables and of said system-manipulated input variables, and responsive to said nominal estimate {tilde over (G)}, for extremizing said objective function J by solving the equation $\underset{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u,\mathrm{\epsilon <br><br>}}{\min }<br><br>\text{\hspace{1em}<br><br>}<br><br>c^T<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+d^T<br><br>\tilde{G}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+{\mathrm{<br><br>}}^T<br><br>\mathrm{\varepsilon <br><br>}$

 

subject to

30. An apparatus for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=G Δ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters having a known uncertainty description, said apparatus comprising;

(a) estimator means for deriving a nominal estimate {tilde over (G)} of said system gain parameters G;

(b) computing means responsive to at least one past value of said system-controlled output variables and of said system-manipulated input variables, and responsive to said nominal estimate {tilde over (G)}, for extremizing said objective function J by solving the equation $\underset{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u,\mathrm{\epsilon <br><br>}}{\min }<br><br>\text{\hspace{1em}<br><br>}<br><br>c^T<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+d^T<br><br>\tilde{G}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+{\mathrm{<br><br>}}^T<br><br>\mathrm{\varepsilon <br><br>}$

 

subject to

31. An apparatus controlling a system in accordance with an objective function J, wherein the operation of said system is describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters having a known uncertainty description, said apparatus comprising;

(a) estimator means for deriving a nominal estimate {tilde over (G)} of said system gain parameters G;

(b) computing means, responsive to at least one past value of said system-controlled output variables and of said system-manipulated input variables, and responsive to said nominal estimate {tilde over (G)}, for extremizing said objective function J by solving the equation $\underset{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u,\mathrm{\epsilon <br><br>}}{\min }<br><br>\text{\hspace{1em}<br><br>}<br><br>c^T<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+d^T<br><br>\tilde{G}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>u+{\mathrm{<br><br>}}^T<br><br>\mathrm{\varepsilon <br><br>}$

 

subject to

32. A controller for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, G represents system gain parameters, and where J is a function of G, the controller comprising (a) a processor and (b) a program storage device that is readable by said processor and that comprises instructions that, when executed, cause the controller to perform the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description and where {tilde over (J)} is said objective function J applied to {tilde over (G)};

(b) computing steady-state targets for said system-manipulated input variables u such that all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description.

33. A computer program product for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, G represents system gain parameters, and where J is a function of G, the computer program embodied on a computer readable medium and comprising instructions that, when executed, cause a computing apparatus to perform the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description and where {tilde over (J)} is said objective function J applied to {tilde over (G)};

(b) computing steady-state targets for said system-manipulated input variables u such that all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description.

34. A method of controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, G represents system gain parameters, and where J is a function of G, said method comprising the steps of;

(a) modeling said system according to an equation Δ

y={tilde over (G)}Δ

u, where {tilde over (G)} is a nominal estimate of G and has a known uncertainty description and where {tilde over (J)} is said objective function J applied to {tilde over (G)};

(b) computing steady-state targets for said system-manipulated input variables u such that;

(i) all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description; and

(ii) said objective function {tilde over (J)} is extremized;

wherein as {tilde over (G)} approaches G, said computed steady-state targets approach those steady-state targets that extremize J, subject to said known uncertainty description.

35. A method in accordance with claim 34, wherein if {tilde over (G)}=G, then said computed steady-state targets extremize J, subject to said known uncertainty description.

36. An apparatus for controlling a system in accordance with an objective function J, the operation of said system being describable by an equation Δ

y=GΔ

u, where y represents one or more system-controlled output variables, u represents one or more system-manipulated input variables, and G represents system gain parameters having a known uncertainty description, and where J is a function of G, said apparatus comprising;

(a) estimator means for deriving a nominal estimate {tilde over (G)} of said system gain parameters G, where {tilde over (J)} is said objective function applied to {tilde over (G)};

(b) computing means, responsive to at least one past value of said system-controlled output variables and of said system-manipulated input variables to compute steady-state targets for said system-manipulated input variables u such that;

(i) all of said system-controlled output variables will remain feasible at steady-state for all possible values of the system gain parameters within said known uncertainty description; and

(ii) said objective function {tilde over (J)} is extremized;

wherein as {tilde over (G)} approaches G, said computed steady-state targets approach those steady-state targets that extremize J, subject to said known uncertainty description.

37. An apparatus in accordance with claim 36, wherein if {tilde over (G)}=G, then said computed steady-state targets extremize J, subject to said known uncertainty description.