Optimal tuning of PID controllers for dead-time systems using Stud genetic algorithms

Abstract

The tuning of controllers is a field of intrinsic research in the control and automation industry. Proportional-integral-derivative (PID) controllers and their variant forms have been much studied for automatic controller tuning owing to their simplicity and performance. The challenge in autotuning controllers is increased in systems that involve delay. Use of metaheuristics and soft computing methods are increasingly dominant in the autotuning of PID controllers. This paper deals with the design of a PID controller and Smith’s predictor for processes with delay. The algorithm employs a Stud genetic algorithm that is search intensified using derivative free search methods, namely simplex derivative pattern search and implicit filtering. The resulting algorithms are named as DFSGStudGA1 and DFSGStudGA2 respectively. The effectiveness of the discussed schemes are compared against some existing methods of controller tuning for robustness in terms of vector gain margin (VGM) and tolerance to plant uncertainty.