System availability analysis for a multi-megawatt photovoltaic power plant

In the design of large solar installations, accurate prediction of subsystem availability is essential to understanding energy harvest and, consequently, financial rates of return for the energy producer or user. However, making accurate reliability predictions is difficult for a number of reasons, including variations in environmental conditions, uncertainty about subsystem and component reliability, and the unavailability of proper analytical methods tailored for the task. This paper presents an analytical method for time-dependent modeling of subsystem availability in any geographic location. It can be applied to subsystems such as solar panels, tracking devices, and power inverters. The results for each subsystem can then be combined in a system-level reliability and performance analysis of the complete photovoltaic power plant. Although absolute reliability predictions come with high uncertainty for relatively new types of equipment such as those found in the solar power industry, this methodology can be useful to help assess a PV plant's financial viability, obtain more realistic estimates of expected downtime, plan preventative maintenance schedules, and budget for spares.