Arc fault detection method for photovoltaic system

Abstract

The invention discloses a photovoltaic system fault arc detection method of mixing quadratic form time frequency distribution characteristics and self-adapting product function analysis. A local mean is used to decompose and analyze a current signal, and a first-order product function is selected to be used as a characteristic value. Quadratic form time frequency distribution is used to analyze the current signal, and a plurality of frequency components in a frequency range used for indicating generating of a fault arc effectively are selected to be used as another group of characteristic values. A real-time range method is used to process a plurality of groups of characteristic values to acquire a real-time range value, which is input in an extreme learning machine, and whether the fault arc exists is determined by the output value of the extreme learning machine, and finally, a cut-off signal is transmitted according to a set trigger standard. The selected characteristic values are used to discover the fault arc accurately, and essential differences before and after the occurrence of the fault arc are distinctly distinguished, and the extreme learning machine is used to determine the fault arc in various load working conditions quickly and accurately, and a fault arc branch is cut off. Compared with a conventional direct current fault arc detection technology, the photovoltaic system fault arc detection method has advantages of wide application range and strong expandability.

Description

technical field [0001] The invention belongs to the technical field of photovoltaic electrical fault detection, and specifically relates to a method of obtaining a product function of a certain order current signal through local mean value decomposition, obtaining multiple frequency components of the current signal through quadratic time-frequency distribution, and processing multiple frequency components by the real-time range method. The corresponding real-time extreme difference values ​​are obtained from each eigenvalue, which is input to the extreme learning machine for arc fault identification, thereby realizing accurate detection of arc faults in photovoltaic systems under multi-load conditions. Background technique [0002] With the increasingly severe problem of energy shortage and the concept of sustainable development, photovoltaic power generation, a new type of environmentally friendly renewable energy, has been widely used in residential electricity and industri...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem that the existing photovoltaic system arc fault detection method has a narrow scope of application, can only be applied to a single photovoltaic system load, and may cause photovoltaic system fault arc rejection due to direct transplantation, and provides a hybrid secondary Arc Fault Detection Method for Photovoltaic System Based on Type Time-Frequency Distribution Characteristics and Adaptive Product Function Analysis

Images