Gain-adaptive process control
First Claim
Patent Images
1. A method of controlling a process, said method comprising:
- generating a process control signal in response to a comparison of a controller variable input signal and a controller setpoint signal;
applying said process control signal to said process to effect control thereof;
generating, in response to said process control signal, a process prediction signal representative of the predicted response of a measured process variable to said process control signal;
applying a delay factor, modeled after the response delay of said process, to said process prediction signal to produce a delayed prediction signal;
generating a process measurement signal representative of the value of said measured process variable;
generating in response to said process measurement signal and said delayed prediction signal, a correction signal responsive to the ratio of said measured process variable to the delayed process response prediction; and
generating said controller variable input signal in response to the product of said process prediction signal and said correction signal.
1 Assignment
0 Petitions
Accused Products
Abstract
In a process control system, a control signal to the process is provided to a process model having a predictive element and a delay element adapted to reproduce as closely as possible the action of the measured process variable to the control signal. The variable input signal to the process controller, upon which generation of the process control signal is based, is generated by multiplying the signal from the predictive portion of the process model by a correction factor generated in response to the ratio of the actual process measurement signal to the delayed model output signal.
-
Citations
18 Claims
-
1. A method of controlling a process, said method comprising:
-
generating a process control signal in response to a comparison of a controller variable input signal and a controller setpoint signal; applying said process control signal to said process to effect control thereof; generating, in response to said process control signal, a process prediction signal representative of the predicted response of a measured process variable to said process control signal; applying a delay factor, modeled after the response delay of said process, to said process prediction signal to produce a delayed prediction signal; generating a process measurement signal representative of the value of said measured process variable; generating in response to said process measurement signal and said delayed prediction signal, a correction signal responsive to the ratio of said measured process variable to the delayed process response prediction; and generating said controller variable input signal in response to the product of said process prediction signal and said correction signal. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10)
-
7. A method in accordance with claim 4 wherein applying said estimation factor comprises applying to said interim ratio signal a lead-compensation factor having a time constant substantially the same as the time constant of said delay factor.
- 8. A method in accordance with claim 7 wherein said delay factor is of the general form
- space="preserve" listing-type="equation">e.sup.-T.sbsp.md.sup.S
and wherein said estimation factor is of the general form
space="preserve" listing-type="equation">1 + T.sub.dl Swhere e is the natural or Napierian logarithm base, 2.71828 . . . , and where Tmd and Tdl are substantially equal S domain time constants.
-
-
9. A method in accordance with claim 8 wherein the S domain transfer function of said process can be suitably approximated by the general form ##EQU3## where e is the natural or Napierian logarithm base, 2.71828 . . . , Kp is a constant, Td is the process S domain lag time constant, and Tp1 and Tp2 are process S domain gain time constants.
- 10. A method in accordance with claim 9 wherein said process control signal is of the general form
- space="preserve" listing-type="equation">K.sub.1 (SP - V) + K.sub.2 ∫
(SP - V)dt
where SP is said setpoint signal, V is said variable input signal, t is time, and K1 and K2 are constants. - space="preserve" listing-type="equation">K.sub.1 (SP - V) + K.sub.2 ∫
-
11. Apparatus for controlling a process, said apparatus comprising:
-
process controller means for generating a process control signal in response to a comparison of a controller variable input signal and a controller setpoint signal; means for providing a process measurement signal representative of the value of a preselected process variable; process model means comprising model predictor means for simulating the process gain relationship between said process control signal and said measured process variable signal and delivering a process prediction signal responsive to said process control signal and comprising model delay means for applying a model delay factor to said process prediction signal to produce a delayed prediction signal; means for generating, in response to said process measurement signal and said delayed prediction signal, a correction signal responsive to the ratio of said process measurement signal to said delayed prediction signal; means for multiplying said process prediction signal by said correction signal to generate said controller variable input signal. - View Dependent Claims (12, 13, 14, 15, 16, 17, 18)
-
13. Apparatus in accordance with claim 11 wherein said means for generating said correction signal comprises means for dividing said process measurement signal by said delayed prediction signal.
-
14. Apparatus in accordance with claim 11 wherein said means for generating said correction signal comprises:
-
means for dividing said process measurement signal by said delayed prediction signal to produce an interim ratio signal; and estimator means for applying a lead estimation factor to said interim ratio signal to generate said correction signal.
-
-
15. Apparatus in accordance with claim 14 additionally comprising limiter means for maintaining said correction signal within a preselected signal range.
- 16. Apparatus in accordance with claim 14 wherein said delay factor is of the general form
- space="preserve" listing-type="equation">e.sup.-T.sbsp.md.sup.S
and wherein said estimation factor is of the general form
space="preserve" listing-type="equation">1 + T.sub.dl Swhere e is the natural or Napierian logarithm base, 2.71828 . . . , and where Tmd and Tdl are S domain time constants.
-
-
17. Apparatus in accordance with claim 14 wherein the lead time constant of the estimation factor applied by said estimator means is substantially the same as the lag time constant of the delay factor applied by said model delay means.
- 18. Apparatus in accordance with claim 17 wherein said delay factor is of the general formula
- space="preserve" listing-type="equation">e.sup.-T.sbsp.md.sup.S
and wherein said estimation factor is of the general form
space="preserve" listing-type="equation">1 + T.sub.dl Swhere e is the natural or Napierian logarithm base, 2.71828 . . . , and where Tmd and Tdl are substantially equal S domain time constants.
Specification