Development of Nonlinear Optimization Techniques Based PID Controller for SEPIC Converter

Abstract

This thesis presents an investigation on closed loop stability of SEPIC (Single-Ended Primary Inductor Converter) converter by employing an optimized PID controller where the parameters are tuned by Genetic Algorithm (GA) and Simulated Annealing (SA) algorithm. Genetic Algorithm is a stochastic algorithm inspired by natural evolution and is extensively used as an optimization technique in power electronics in recent years. SA refers to an optimization technique, based on the principles of thermodynamics where the analogy of cooling of metal and freezing into a minimum energy level is utilized. Here, State Space Average method is deployed to model and obtain the transfer function of the converter based system. Hence, GA and SA based PID controller is studied and implemented in the system so that the stability of the converter can be evaluated and compared with the conventional PID controller. Different fitness functions (IAE, ITAE, ISE, ITSE) and performance parameters like percentage of overshoot, rise time, settling time and peak amplitude are taken into account to investigate the stability of the system. The step responses of the closed loop system are obtained through rigorous simulation in MATLAB. It is observed that, GA-PID-2 (IAE) is the most optimized controller among the entire GA based PID controllers as the overshoot is the lowest (3.43%) and peak amplitude is 1.03. Moreover, rise time (0.000102s) and settling time (0.00418s) are also in acceptable limit. However, GA-PID-2 (ISE) shows quick rise (0.000102s) and settling time (0.000676s) but the overshoot of ISE is 11.3% which is much greater than IAE of GA-PID-2. Meanwhile, SA-PID-2 (IAE) is the most optimized controller among all other SA based controllers where the value of overshoot is 2.2% and peak amplitude is 1.01. For Rise time and settling time, the values for SA-PID-2 (IAE) are 0.000665s and 0.000104s respectively which are in acceptable limit. Though SA-PID-3 (IAE) shows quick rise time (0.000118s), the overshoot is high (22.2%). Therefore, SA-PID-2 (IAE) is selected as the most optimized and suitable controller for SEPIC converter. For changed value of inductor, the simulation is done for GA-PID-2 (for IAE) and SA-PID-2 (for IAE), it is observed that less overshoot is obtained than conventional PID controller (overshoot 17.3%). For GA-PID-2, the overshoot is 3.36% and for SA-PID-2 the value is 9.12%. So, it is evident that, nonlinear optimization technique based PID controller is providing more optimized result in case of any change in the value of inductor of the circuit than conventional PID controller. However, GA is more suitable option than SA in terms of designing the controller for less overshoot. Therefore, GA-PID-2 (IAE) is the most optimized controller for SEPIC converter in this investigative study.