Performance And Reliability Evaluation Of Photovoltaic Power Generation System

As an internationally recognized ideal alternative energy, solar energy is a kind of clean energy which is abundant and widely distributed. In recent years, the cost of solar cell components and other components of the power plant has been reduced, and the technology of photovoltaic power generation has become more perfect, the governments have begun to develop and utilize solar energy, and establish photovoltaic stations. Research on the performance and reliability of photovoltaic power plant is of great significance for the maximum utilization of solar energy resources and optimizing the operation of photovoltaic plant.Firstly, this paper introduces the principle and classification of photovoltaic generation systems, and establishes the fine efficiency model of the PV system mainly considering the influence of the temperature rise and the cleaning frequency of the battery on the efficiency of photovoltaic system.Secondly, this paper proposes a set of index system including light resources and photovoltaic power plant performance, and makes a comprehensive evaluation of the photovoltaic plant based on the actual operation data, by the methods of principal component analysis and entropy weight method, then carries on the effective combination of the results of these two methods by the optimal fusion algorithm based on the maximum likelihood estimation.Next, the fault tree of PV system is established on the basis of the electrical architecture diagram of the PV system and some assumptions, considering the relationship between the various parts of the PV system. With exponential distribution as a component’s fault distribution and to the logic structure of the existing fault tree, using the probability function, calculates the minimum cut set and gets the reliability and failure rate of both the whole system and the components in different time period.Finally, the fault tree is transformed into a Bayesian network, the occurrence probability of the top event is calculated by bucket elimination, and the diagnosis reasoning is done to recognize the weak link of the system, and the causal inference is done to estimate the probability of the occurrence of each element and the intermediate node in the system. The effectiveness and superiority of the Bayesian network method in the reliability evaluation of PV system are described.