Method of designing model predictive control for cross directional flat sheet manufacturing processes to guarantee spatial robustness and to prevent actuator picketing

Patent metadata

Field	Value
Publication number	US-10358771-B2
Application number	US-201615273709-A
Country	US
Kind code	B2
Filing date	Sep 23, 2016
Priority date	Sep 23, 2016
Publication date	Jul 23, 2019
Grant date	Jul 23, 2019

Abstract

Automated parameter tuning techniques for cross-directional model predictive control for paper-making under user-specified parametric uncertainties to reduce variability of the actuator and measurement profiles in the spatial domain is proposed. Decoupling properties of the spatial and temporal frequency components permit separate controller design and parameter tuning. CD-MPC design that explicitly accounts for parametric model uncertainty while finding MPC cost function weighing matrices that prevent actuator picketing and guarantee robust stability of the spatial CD profile. Picketing refers to periodic variation patterns in the actuator array. The inventive technique includes: (i) determining the worst case cutoff frequency of all process models, given parametric uncertainty, (ii) designing a weighing matrix to penalize high frequency actuator variability based on the process model and worst case cutoff frequency, and (iii) finding a multiplier for the spatial frequency weighted actuator variability term in the MPC cost function that assures robust spatial stability.

First claim

What is claimed is:

1. A system which forms a material in a spatially-distributed multivariable-array cross-directional process wherein the system comprises:
at least one set of actuator arrays each distributed adjacent the material in the cross direction (CD), wherein each set of actuator arrays is controllable to vary the properties of the material;
means for measuring and acquiring properties data about the properties of the material; and
a multivariable model predictive controller (MPC) for providing CD control to the cross-directional process, wherein the MPC, in response to signals that are indicative of the properties of the properties data, provides signals to the at least one set of actuator arrays to vary properties of the material, wherein the MPC includes means for spatially tuning the MPC which comprises:
(a) means for inputting nominal spatial model parameters and corresponding parametric uncertainty specifications;
(b) means for calculating worst-case cut-off frequency v w for all possible process models given the parametric uncertainty specification wherein the worst-case cut-off frequency comprises the smallest cut-off frequency of all possible process models given the parametric uncertainty specifications;
(c) means for designing a weighting matrix S b to penalize high frequency actuator variability based on the process model and worst-case cut-off frequency;
(d) means for developing a robust spatial condition based on the parametric uncertainty specifications;
(e) means for adjusting a multiplier of S b in a MPC cost function to satisfy the robust stability condition wherein S b is a weighting matrix in penalty term of a MPC cost function which controls actuator bending and picketing behavior; and
(f) means for outputting the weighting matrix S b and its multiplier.

2. The system of claim 1 wherein the S b is a spatial filter whose spatial frequency response is a mirror of an actuator frequency response.

3. The system of claim 1 wherein the parametric uncertainty comprises trust ranges around the nominal spatial model parameters which characterize possible mismatch between an identified model and an actual process.

4. The system of claim 1 wherein high frequency actuator variability comprises actuator bending and picketing behavior that includes frequency components beyond the worst-case cut-off frequency.

5. The system of claim 1 wherein the nominal spatial model parameters comprise parameters of a mathematical model of the process obtained from a bump test and model identification procedure.

6. The system of claim 1 wherein the MPC includes means for spatially tuning the MPC with respect to one of the actuator arrays.

7. In a process control system having a multivariable model predictive controller (MPC) for providing control to a spatially-distributed multiple-array, sheetmaking cross-directional (CD) process having at least one manipulated actuator array and at least one controlled measurement array, a method of providing control of the multiple-array process that comprises the steps of:
(a) tuning the MPC by the steps of:
(i) inputting nominal spatial model parameters and corresponding parametric uncertainty specifications;
(ii) calculating worst-case cut-off frequency v w for all possible process models given the parametric uncertainty specification;
(iii) designing a weighting matrix S b to penalize high frequency actuator variability based on the process model and worst-case cut-off frequency wherein the worst-case cut-off frequency comprises the smallest cut-off frequency of all possible process models given the parametric uncertainty specifications;
(iv) developing a robust spatial condition based on the parametric uncertainty specifications;
(v) adjusting a multiplier of S b in a MPC cost function to satisfy the robust stability condition wherein S b is a weighting matrix in penalty term of a MPC cost function which controls actuator bending and picketing behavior; and
(vi) outputting the weighting matrix S b and its multiplier;

(b) inputting the tuning parameters into the MPC; and
(c) controlling the multiple-array CD process with the MPC.

8. The system of claim 7 wherein the S b is a spatial filter whose spatial frequency response is a mirror of an actuator frequency response.

9. The system of claim 7 wherein the parametric uncertainty comprises trust ranges around the nominal spatial model parameters which characterize possible mismatch between an identified model and an actual process.

10. The system of claim 7 wherein high frequency actuator variability comprises actuator bending and picketing behavior that includes frequency components beyond the worst-case cut-off frequency.

11. The system of claim 7 wherein the nominal spatial model parameters comprise parameters of a mathematical model of the process obtained from a bump test and model identification procedure.

12. The system of claim 7 wherein the spatially-distributed sheetmaking process is a paper-making process.

13. A non-transitory computer readable medium embodying a computer program for automatically tuning a model predictive controller (MPC) employed to control a cross-directional process having a manipulated actuator array comprising a plurality of actuators and at least one controlled measurement array wherein the program comprises readable program code for:
(a) inputting nominal spatial model parameters and corresponding parametric uncertainty specifications;
(b) calculating worst-case cut-off frequency v w for all possible process models given the parametric uncertainty specifications wherein the worst-case cut-off frequency comprises the smallest cut-off frequency of all possible process models given the parametric uncertainty specifications;
(c) designing a weighting matrix Sb to penalize high frequency actuator variability based on the process model and worst-case cut-off frequency;
(d) developing a robust spatial condition based on the parametric uncertainty specifications;
(e) adjusting a multiplier of St in a MPC cost function to satisfy the robust stability condition wherein S b is a weighting matrix in penalty term of a MPC cost function which controls actuator bending and picketing behavior;
(f) outputting the weighting matrix S b and its multiplier;
(g) inputting tuning parameters into the MPC; and
(h) controlling the cross-directional process with the MPC.

14. The computer readable medium claim 13 wherein the S b is a spatial filter whose spatial frequency response is a mirror of an actuator frequency response.

15. The computer readable medium of claim 13 wherein the parametric uncertainty comprises trust ranges around the nominal spatial model parameters which characterize possible mismatch between an identified model and an actual process.

16. The system of claim 1 wherein the at least one set of actuator arrays comprise a single array comprising of a plurality of manipulated actuators that are arranged in the CD and wherein the means for measuring and acquiring properties data about the properties of the material comprises a corresponding single controlled measurement array.

17. The system of claim 7 wherein the at least one manipulated actuator array comprises a single array comprising of a plurality of manipulated actuators that are arranged in the CD and wherein the at least one controlled measurement array comprises a corresponding single controlled measurement array.

18. The computer readable medium of claim 13 wherein the manipulated actuator array comprises a single array comprisi

Assignees

• Honeywell Ltd

Inventors

• He Ning
• Liu Xiaotao
• Forbes Michael
• Backstrom Johan
• Chen Tongwen

Classifications

• G05B13/048 Primary

  using a predictor · CPC title

• G06F17/16

  Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization (matrix transposition G06F7/78)} · CPC title

• D21G9/0054 Primary

  details of algorithms or programs · CPC title