Neuro-evolutionary approaches to power system harmonics estimation

Abstract

This paper focuses on exploiting two computational intelligence techniques such as artificial neural network and evolutionary computation techniques in estimation of harmonics in power system. Accurate estimation of harmonics in distorted power system current/voltage signal is essential to effectively design filters for eliminating harmonics. No standard design is available for handling of local minima and training of NN but Evolutionary Computation (EC) techniques are capable of resolving local minima. Neural Network and Evolutionary Computing (Bacterial Foraging Optimization (BFO)) are combined to achieve accurate estimation of different harmonics components of a distorted power system signal. First estimation of unknown parameters are carried out using BFO, then optimized output of BFO are taken as initial values of the unknown parameters for Adaline. Amplitude and phase of fundamental and harmonics components are determined from final updated values of unknown parameters using Adaline. This Adaline based Bacterial Foraging Optimization (Adaline-BFO) hybrid estimation algorithm addresses the problems of slow convergence and reduction of time generation of off-springs happening in Genetic Algorithm (GA), and to avoid local minima in Particle Swarm Optimization (PSO). The proposed Adaline-BFO algorithm has been applied for estimation of harmonics of the voltage obtained across the inverter terminals of a prototype Photovoltaic (PV) system. From the obtained results, it is confirmed that the proposed Adaline-BFO algorithm provides superior estimation performance in terms of improvement in % error in estimation, processing time in computation and performance in presence of inter and sub-harmonic components when compared with the Discrete Fourier Transform (DFT), Kalman Filter (KF) and BFO algorithms.