Performance Ratio and Fault Characterization Methods for Photovoltaic Systems

Abstract: This project focuses on investigating performance metrics and fault detectionmethods in photovoltaic systems. The primary goal is to find feasible solutionsproviding high accuracy and reliability in fault detection. The solution shouldbe simple yet robust enough for commercial applicability. For this reason, onlythe minimum essential measurements are utilized for the detection algorithm.These include the output voltage and current of the installation as well asirradiation and temperature from a dedicated sensor. Various existing works inscientific literature are investigated and a viable method is selected for furtheranalysis.The final approach involves a time-series analysis of normalized parameters,providing an overview of the daily performance of a given photovoltaic string.Although the solution is capable of differentiating between three designatedfault types: global shading, line-to-line faults and partial shading, it was foundthat there are numerous limitations that must be overcome before commercialviability, most of which are not directly technical difficulties. The reportends with a conclusive list of shortcomings that need additional thought forfuture works in this area. Overall, the thesis demonstrates the difficulty offault detection for photovoltaic systems when applied outside of controlledlaboratory conditions.