Fault diagnosis method of photovoltaic array

Abstract

The invention discloses a fault diagnosis method of a photovoltaic array. According to the method, fault diagnosis of the photovoltaic array is carried out on the basis of a multi-stage cluster support vector machine (SVM) algorithm. The method comprises the steps as follows: (1) firstly, hierarchical cluster is carried out on various class centers of a training sample set; with similarity measurement of cluster analysis as a mean distance, the mean distance of front two classes in each stage is supposed to be the minimal; the other similarity measurement is in a similar way; and two samples with the minimal means are merged into one class; (2) secondly, the sample obtained from the step (1) is an SVW input and output sample; data samples are trained to obtain various stages of SVM modules, which comprise an SVM1, an SVM2 and an SVM3; and (3) finally, a data test sample is input into the trained SVM model; and hierarchical classification diagnosis is carried out to further determine the fault type of the system. The implementation process of the fault diagnosis method is simple, clear and easy to realize; fault diagnosis can be timely carried out; and operation maintenance of a photovoltaic system is facilitated.

Description

technical field [0001] The invention relates to the field of photovoltaic power generation, in particular to a photovoltaic array fault diagnosis method. Background technique [0002] As the global energy problem becomes increasingly prominent, people gradually turn their attention to renewable energy. As a clean energy source, solar energy has a large development space. In the past few years, due to the decline in the price of photovoltaic materials and the policy support of various governments, the photovoltaic industry has developed rapidly. It is estimated that the annual output of my country's photovoltaic industry will reach 135 million m in 2015 and 2020. 2 and 273 million m 2 . At present, the life span of photovoltaic modules is about 25 years, which requires that photovoltaic power plants can generate electricity stably, continuously and efficiently during these 25 years. However, photovoltaic power plants are often installed in relatively harsh environments, r...

AI Technical Summary

Problems solved by technology

This type of method has poor real-time performance, and the accuracy and efficiency of fault diagnosis and positioning depend on the level of the thermal imager or camera, which is not conducive to online fault diagnosis, positioning and alarm, etc.

In the electrical method, the commonly used ground capacitance detection method and time domain reflection method need to detect the photovoltaic system and then locate the fault when the system stops working, which has a great impact on the real-time operation of the system.

Images