Using autocorrelation to detect model mismatch in a process controller

Patent metadata

Field	Value
Publication number	US-9904257-B2
Application number	US-201213465764-A
Country	US
Kind code	B2
Filing date	May 7, 2012
Priority date	Jan 31, 2008
Publication date	Feb 27, 2018
Grant date	Feb 27, 2018

Abstract

A process controller adaptation and tuning technique uses a closed loop adaptation cycle that performs an autocorrelation analysis on the prediction error or the control error of a process control system to determine if significant process model mismatch exists or to determine an increase or a decrease in process model mismatch over time. The adaptation and tuning technique may perform a controller tuning cycle when the determined model mismatch raises above a predetermined level.

First claim

The invention claimed is:

1. A method of detecting process model mismatch between a controller model used by a predictive process controller and a process plant, the method comprising:
determining, via a computing device, an error signal associated with the control of the process plant;
performing, via a computing device, an autocorrelation analysis on the error signal to obtain an autocorrelation output value;
analyzing, via a computing device, the autocorrelation output value to detect process model mismatch between the controller model and the process plant, the analyzing including comparing the autocorrelation output value to a predetermined threshold; and
triggering a retuning of the predictive process controller, without regenerating the controller model, during on-line controller operation when the comparison indicates that the autocorrelation output value exceeds the predetermined threshold.

2. The method of claim 1 , wherein determining the error signal associated with the control of the process plant includes determining a control error signal associated with the predictive process controller as a difference between a measured value of a process variable and a set point for the process variable.

3. The method of claim 2 , wherein performing the autocorrelation analysis on the error signal includes performing the autocorrelation analysis on control error data collected during a time period when the process variable is in a steady state condition.

4. The method of claim 1 , wherein determining the error signal associated with the control of the process plant includes determining a prediction error signal associated with the predictive process controller as a difference between a predicted value of a process variable and a measured value of the process variable.

5. The method of claim 4 , wherein performing the autocorrelation analysis on the error signal includes performing the autocorrelation analysis on prediction error data collected during a time period when the process variable is being controlled in response to a disturbance or upset.

6. The method of claim 5 , wherein determining the error signal associated with the control of the process plant includes obtaining a first control error signal for a first time period as a difference between a measured value of a process variable and a set point for the process variable during the first time period and obtaining a second control error signal for a second time period later than the first time period as a difference between a measured value of the process variable and a set point for the process variable during the second time period, and wherein performing an autocorrelation analysis on the error signal includes performing an autocorrelation analysis on the first control error signal and performing an autocorrelation analysis on the second control error signal, and wherein analyzing the autocorrelation analysis to detect process model mismatch includes comparing the results of the autocorrelation analyses on the first and second control error signals to determine a change in process model mismatch between the first and second time periods.

7. The method of claim 6 , wherein obtaining the first control error signal and obtaining the second control error signal includes obtaining the first and second control error signals during first and second time periods in which the process variable is in a steady state condition.

8. A method of detecting process model mismatch for predictive process controllers, the method comprising:
determining, via a computing device, an error signal associated with a controller model used by a predictive process controller in a process plant;
performing, via a computing device, an autocorrelation analysis on the error signal to obtain an autocorrelation output having a value indicative of a degree of process model mismatch between the controller model and the process plant; and
retuning the predictive process controller, without regenerating the controller model, during on-line controller operation by generating one or more tuning parameter values based on the value of the autocorrelation output.

Assignees

• Thiele Dirk
• Fisher Rosemount Systems Inc

Inventors

• Thiele Dirk

Classifications

• G05B13/048

  using a predictor · CPC title

• G05B13/042 Primary

  in which a parameter or coefficient is automatically adjusted to optimise the performance · CPC title

• G05B17/02

  electric · CPC title

• G05B19/41885

  characterised by modeling, simulation of the manufacturing system · CPC title

• G05B13/04 Primary

  involving the use of models or simulators · CPC title