Process and Disturbance Identification by Curve Fitting The Closed Loop Response

Abstract

Methods are described to estimate process and disturbance parameters based on curve fitting the transient response of a closed loop control system. The objectives are : 1) to mathc the process transfer function to a first order plus dead time model and 2) to match the disturbance to a first order model unit step response that enters the loop at the end of the process. The curve fitting methods identify : process parameters kp(gain), ¿p, (time constant), and ¿d (dead time); and disturbance parameters kl(gain), and ¿l(time constant). The methods are based on fitting time domain equations for closed loop response. These are derived using both continuous system analysis and sampled data system analysis.