Data processing: generic control systems or specific application – Generic control system – apparatus or process – Optimization or adaptive control
Reexamination Certificate
1998-02-20
2001-02-06
Grant, William (Department: 2786)
Data processing: generic control systems or specific application
Generic control system, apparatus or process
Optimization or adaptive control
C700S053000, C700S072000, C162S253000, C162SDIG004
Reexamination Certificate
active
06185468
ABSTRACT:
INTRODUCTION
The present invention is directed to a decoupling controller for use with a process having two input variables and two output variables and, more specifically, to a paper-making process where the input variables are dry stock flow and steam pressure to a dryer section and the output variables are basis weight and moisture.
BACKGROUND
In the paper-making process, the process itself has long deadtimes relative to the process time constant. This makes control difficult. There are typically two unique deadtimes; one is for the time required for a change in basis weight when the input variable, stock flow, is changed and the other is the time for variation in steam to affect the final moisture carried by the paper sheet. A further difficulty in the paper-making process is the cross-coupling affect; that is, each input variable affects both output variables. Hence, a decoupling controller is desired to regulate the outputs independently; for example, the operator would like to change the setpoint of the basis weight controller without changing the value of the moisture.
Two rather crude techniques have been utilized for decoupling. In a first, a setpoint is changed only once every five minutes, for example, for changes in stock flow and once every minute for changes in steam. This is a much longer period of time than the generation of output data by a sensor which scans across the width of the paper, for example, every 20 seconds.
A second proposed decoupling technique is to provide absolute decoupling constants between changes in stock flow and steam pressure. These might be termed “compensating changes”. However, these are mere guesses and do not compensate for grade changes or speed changes and do not take into account that the coupling effect may be nonlinear. One other problem with a controller for a paper-making machine discussed above, is the fact that the measurements of outputs occur either at long intervals or can occur asynchronously due to sheet breaks or standardization. In any case, scan measurements (which may take up to 120 seconds) only occur every 20 seconds at best.
OBJECT AND SUMMARY OF INVENTION
A general object of the present invention is to provide an improved decoupling controller for use with a process having two input variables and two output variables.
In accordance with the above object, there is provided a decoupling controller for use with a process having two input variables U
1
, U
2
and two output variables X
1
, X
2
where in the process each input variable affects both output variables (that is they are coupled), such process having desired setpoints S
1
, S
2
for the output variables. Such decoupling controller comprising two pairs of linked internal model controllers, each internal model controller (IMC) including a proportional, integral, derivative (P.I.D.) velocity unit C
11
, C
21
, C
12
, and C
22
for respectively receiving from a first pair of difference junctions total process error, et
1
, et
2
, in a feedback loop for the process and producing said input variables U
1
, U
2
, which are control inputs to the process itself, such P.I.D. units taking into account loop, proportional, integral and derivative gains of the feedback loop for both direct and cross-coupling.
Four first order transfer function units K
11
, K
12
, K
21
, and K
22
receive as inputs U
1
, U
2
, the K
11
, K
22
units providing predicted values of X
1
, X
2
, the K
21
, K
12
units providing predicted outputs of X
1
, X
2
due to cross-coupling.
Means feed back to a pair of second summing junctions the outputs of K
11
, K
12
and K
22
, K
21
respectively.
Means couple the outputs of the second summing junctions, which are total predicted values of X
1
and X
2
taking into account cross-coupling, to a pair of third summing junctions, which also receive modeling error signals representing the difference between the actual X
1
and X
2
values and estimated values Y
1
and Y
2
, from a pair of fifth junctions.
Means feed the summed output of the third pair of summing junctions to the first pair of difference junctions, which have as the other difference input the setpoints S
1
, S
2
to provide the total process error inputs et
1
and et
2
to C
11
, C
21
and C
12
, C
22
;
Means take the deadtime of the process into account (that is the lag time between the change of input variables and output variables), including four deadtime units, D
11
, D
21
, and D
12
, D
22
, having their inputs respectively connected to the outputs of K
11
, K
21
, K
12
, and K
22
, including a pair of fourth summing junctions having as outputs the current estimated values Y
1
, Y
2
of the X
1
, X
2
output variables, where one of the pair of fourth summing junctions, sums the outputs of D
11
, D
12
and the other of the pair of summing junctions, sums the outputs of D
22
, D
21
.
Means couple the outputs of the fourth pair of summing junctions, to the fifth pair of difference junctions to take the difference between the actual outputs X
1
, X
2
and the estimated values Y
1
, Y
2
, such differences being the modeling error signals.
Means feed back the modeling error signals to the third pair of summing junctions.
REFERENCES:
patent: 3610899 (1971-10-01), Dahlin
patent: 3619360 (1971-11-01), Persik
patent: 3676295 (1972-07-01), Rice
patent: 3847730 (1974-11-01), Doering
patent: 3852578 (1974-12-01), Rice
patent: 4054780 (1977-10-01), Bartley
patent: 4098641 (1978-07-01), Casey et al.
patent: 5121332 (1992-06-01), Balakrishnan et al.
patent: 5400247 (1995-03-01), He
patent: 5400258 (1995-03-01), He
patent: 5568378 (1996-10-01), Wojsznis
patent: 5791160 (1998-08-01), Mandler et al.
patent: 5853543 (1998-12-01), Hu et al.
patent: 5944955 (1999-08-01), Bossen et al.
Wells, Charles H. “Robust Multivariable IMC Control for Paper Machines.” Tappi Journal, Jul. 1999: 140-144.
Zhang, Weidong and Xiaoming Xu. “Simple Predictor for Processes with Time Delay.” Proceedings of the American Control Conference. San Diego, California. Jun. 1999: 822-826.
Dahlin, E. B. et al. “Designing and Tuning Digital Controllers: Part 1.” Instruments and Control Systems. Part 1, Jun. 1968:77-83. Part 2, Jul. 1968:87-91.
Mahieddine, F. et al. “Decoupling Multivariable Self-Tuning Controller for Varying Time Delays.” IEE Proceedings. Part D, Control Theory and Applications v. 134 (Sep. '89) pp. 209-14.
Heaven, M. et al. “Applications of Model-Based Tuning and Analysis Tools to Paper Machine Control.” Pulp and Paper Canada, vol. 98, No. 7, Jul. 1997, pp. 54-48.
Morari, M. and Zafirious, E., “Robust Process Control.” Englewood Cliffs, 1989, pp. 393-428.
Bahill, A. Terry. “A Simple Adaptive Smith-Predictor for Controlling Time-Delay Systems.” Control Systems Magazine. vol. 3. May, 1983, 16-22.
Dumont, Guy A. “System Identification and Adaptive Control in Papermaking.” Fundamentals of Papermaking: Transactions of the Ninth Fundamental Research Symposium Head at Campridge: Sep. 1989. Mechanical Engineering Publication Limited, London.
Flehr Hohbach Test Albritton & Herbert LLP
Gain Edward F.
Grant William
Impact Systems Inc.
Wright Jerry G.
LandOfFree
Decoupling controller for use with a process having two... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Decoupling controller for use with a process having two..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Decoupling controller for use with a process having two... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2560781