Evolutionary optimization methodology for resonant converter design considering photovoltaic systems

Abstract

Series Resonant Converter (SRC) design parameters are calculated by evaluating circuit characteristics of resonant tank circuit, transformer and rectifier circuit that constructs SRC. Evaluation of these characteristics helps to investigate various effects of parameters on the overall system performance and optimizes converter parts to be more efficient and compatable with each other. Transformer is the main isolation part of SRC and optimal design of this part leads to reduce transformer losses as well as cost and circuit size. In order to provide more accurate design, 5-parameter and 9-parameter transformer models are suggested and compared to minimize transformer no-load loss (NLL) and load loss (LL). Moreover, reduction of transformer cost and size are also studied as another objective function by providing better circuit performance. Optimization is carried out utilizing three evolutionary algorithms (EAs): Genetic Algorithm, Differential Evolutionary Algorithm and Nondominated Sorting Genetic Algorithm for transformer loss minimization according the suggested two models. Flux density is calculated using Finite Element Method (FEM) after transformer design parameters are optimized. Total owning cost is used as a measurement tool by taking into account of no-load and load loss values. SRC is designed for photovoltaic (PV) system which is a renewable energy system. First, PV array is constructed for 6 modules in series connection taking into account of weather conditions. Array characteristics are collected according to selected region (Yildiz Technichal University – Dept. of Electronics and Communication Building Roof) and according to non shaded or shaded configuration functions. Annual generated power and output voltage variation are gathered from this design to provide better performance when designing a converter topology. Azimuth angle, tilt angle, distances between two panels and tracking types are the main optimization parameters for photovoltaic data collection process. Helioscope and Design Builder program are used to simulate the real photovoltaic design module.