Artificial Neural Network Based Design of Circuit Parameters of Quasi Switched Boost DC-AC Inverter

Abstract

The impedance source inverters are greatly preferred for photovoltaic applications as they can boost the voltage. There are numerous numbers of impedance and quasi-impedance network topologies available. This paper proposes the design of a High Gain Switched Boost Quasi-Impedance Source Inverter (HG-SBqZSI) for photovoltaic applications. There exists a set of parameters that affect the performance of the inverter. Hence, a proper design of the inverter is vital. Some of the parameters that influence the performance of the inverter are shoot through duty ratio, capacitor voltage ripple, inductor ripple, modulation index etc. This paper discusses the design and optimisation of HG-SBqZSI using Neural Networks (NN). The neural network-based prediction is carried out using NN toolbox in MATLAB/SIMULINK. The shoot through Duty ratio (D) is predicted using NN and on proper selection of shoot- through duty ratio value, the values of impedance elements and the gain of the inverter is optimised depending on the application.