Closed Loop transient response System Identification
Abstract
Closed-loop Transient step response analysis SID technique is known for its simplicity and effectiveness in accommodating practical constraints and operational restrictions of particular system, and thus giving customized models for the particular process. Different algorithms have therefore been developed within the field the closed-loop step-response transient analysis.
The objective of this thesis is to review the most common transient step response SID algorithms, and analyse the algorithm’s performance in model estimation and PID controller auto-tuning.
Nine different transient step-response SID algorithms are reviewed and studied by means of simulating different numerical examples which represent processes with various dynamic characteristics. Among the discussed algorithms, five (YS, JR, JL, CLC and MF) are primarily for FOPDT and the remaining (JS, JSDR, DR and DR1) are for SOPDT. The simulation takes also into account the stochastic framework where the systems are influenced with measurement noise, and weighs up the performance of the algorithms based on accuracy in estimating model parameters, robustness of PID-setting, and strength in withstanding the measurement noise.
The study reveals that JL has the strongest performance among FOPDT algorithms which also provides the most robust PI-controller auto-tuning. For SOPDT algorithms, the study points out the DR1 algorithm as superb and the most robust when not exposed to measurement noise.