Artificial neural network based series active power filter for self excited induction generator

Abstract

This paper presented ANN (Artificial Neural Network) based SAPF (series active power filter) for voltage control and power quality improvement of constant speed prime mover driven self excited induction generator. The constant speed prime mover such as biogas, biomass, gas-turbine, and diesel turbine driven standalone generating system has problem of poor voltage regulation and power quality with linear and nonlinear load conditions. The SAPF is inject series voltage between source and load to maintain constant source and load voltage with power quality improvement of source and load current. The SAPF is modeled using Insulated Gate Bipolar Junction Transistor based Voltage Source Converter with self supported DC bus. A 7.5 kW, 415 V, and 50 Hz asynchronous machine with SAPF and loads are designed, modeled, and simulated in MATLAB environment. The performance of ANN based SAPF is also compared with a Proportional Integral controller based SAPF.