Various Tuning and Optimization Techniques Employed in PID Controller: A Review

Abstract

PID controllers are implemented in more than 90% of the control system applications. In this review paper, various tuning methods have been presented and comparison of established algorithmic tuning methods has been discussed on system response. There have been many approaches used in the past for tuning and obtaining optimized gain factors such as Ziegler–Nichols method, genetic algorithm (GA), particle swarm optimization (PSO) method, and artificial neural network (ANN). The primary goal of this paper is to establish a proper understanding about different tuning and optimization methods and their effect on process efficiency and stability. The secondary goal is to provide a pathway for future development of a tuning algorithm for a high-temperature research grade furnace controller, based on machine learning (ML). This leads to higher controller efficiency over a predefined finite set of ramp–hold cycles, ensuring lesser rise and settling time, reduced or no overshoot, minimized mean squared error, and maximum stability. Critical manufacturing processes like investment casting, metal injection molding, and other thermal cycling processes like physical vapor deposition/chemical vapor deposition, e-waste processing, which require precise control of temperature are expected to be benefited by ML-integrated PID parameter auto-tuning and control.