Optimisation and rule firing analysis in fuzzy logic based maximum power point tracking

Abstract

This paper presents a technique for implementing population based metaheuristic algorithms during dynamic optimisation of a non–linear system with time varying inputs. The system dynamics due to the presence of multiple inputs and large signal variations are simulated when designing controller parameters. The proposed method is used to implement the particle swarm optimisation (PSO) algorithm and used to optimise fuzzy logic controllers for photovoltaic (PV) array maximum power point tracking. Controller optimisation is carried out using a large signal average model of the dc–dc converter. A rule firing analysis technique for interpretation of fuzzy logic controller rule participation at run–time is formulated. The rule inference parameters used for analysis are firing frequency, firing strength, and contribution to the control effort. The performance of optimised fuzzy logic controllers consisting of 9, 25, and 49 rules is analysed at run–time. Simulation results show that the fuzzy logic controller rules centred on the equilibrium point have the most significant contribution to the control effort. A fuzzy logic controller (FLC) with 9 rules can therefore give a good performance. The robustness of the 9–rule FLC is verified using the Lyapunov stability theory.