GRAPHICAL TECHNIQUE FOR MODELING INTEGRATING (Non–SELF-REGULATING) PROCESSES WITHOUT STEADY-STATE PROCESS DATA

Abstract

Model-fitting techniques for controller tuning that require the process to be initially at steady state cannot generally be used with integrating (non–self-regulating) processes. To address this issue, a graphical model-fitting technique is detailed and demonstrated for determination of first order plus dead time integrating model parameters from integrating process response plots. The resulting model parameters can be used directly in a range of tuning correlations designed specifically for integrating processes. The advantage of this technique is that it requires only two periods of constant manipulated and disturbance variables sustained just long enough for the process variable to respond and establish a clear slope. This is an important benefit because integrating processes generally cannot be maintained at an initial steady state as required when using techniques published for self-regulating processes. The result is an industry-friendly method. The method is demonstrated for level control in a pumped tank, a classical challenge in industrial practice. Both a simulation and a bench-scale experimental system are used in the demonstration studies.

Keywords:

• Controller tuning
• Graphical model fitting
• Integrating processes
• Non–
• self-regulating processes

Acknowledgment

The authors wish to acknowledge John McIlwain for his insight into applying graphical model-fitting techniques to integrating processes.