Estimation of Photovoltaic System Reliability and Performance Metrics

Abstract

A framework to integrate reliability and performance analysis of grid-tied photovoltaic (PV) systems is formulated using Markov reward models (MRM). The framework allows the computation of performance metrics such as capacity and energy yield, and reliability metrics such as availability. The paper also provides an analytical method to compute the sensitivity of performance metrics to MRM-parameter variations. The approach to sensitivity analysis is demonstrated to be particularly useful to formulate optimal operational policies, e.g., repair strategies, as the impact of variations in model parameters on system performance can be rapidly evaluated. Case studies demonstrate several applications of the proposed framework, including analysis of residential and large utility-level installations, and emerging distributed inverter architectures.