Research On DC Series Arc Fault Detection Method And Detector Design In Photovoltaic System

With the constant development of photovoltaic power generation industry in China,photovoltaic power generation brings a series of problems that endanger people’s personal property safety,while bringing some convenience to people’s life,including frequent fire accidents in photovoltaic power generation system.One of primary causes for frequent fire accidents ought to attribute to the DC series arc fault with the high frequency.In order to maintain personal property safety and safeguard safe and stable operation of photovoltaic power stations,this paper is focused on the DC series arc fault of the photovoltaic power generation system to study the detection method and conduct preliminary scheme design for detectors.The paper analyzes the arc classification and causes of DC series arc fault in the photovoltaic power generation system.Based on the actual photovoltaic modules,the DC series arc fault test platform of the photovoltaic power generation system is built,and then DC current signals of the normal state and the series arc fault state under different conditions are collected.The time domain and frequency domain characteristics of normal DC current signal and DC current signal in series arc fault are compared,and the limitations of identifying series arc fault from the perspective of time domain characteristics are identified.In addition,the method of detecting series arc fault based on frequency domain characteristics is proposed.On this basis,the frequency domain characteristics of series arc DC current signal are extracted and quantified.In view of the problem of noise interference in the original DC current signal,the paper constructs a method of noise reduction based on wavelet soft threshold.After the original signal is denoised and the FFT is processed,the characteristic frequency band of DC series arc fault is determined by means of logarithmic average frequency band.A method based on double-tree complex wavelet transform(DT-CWT)is presented to decompose and reconstruct the fault characteristic frequency band and extract the reconstruction coefficient of fault characteristic frequency band.Combined with the entropy complexity algorithm,the complexity of the reconstruction coefficient of faultcharacteristic frequency band is quantified.The DC series arc fault criterion is made according to the complexity of the reconstruction coefficient of fault characteristic frequency.Moreover,the stability and reliability of the fault criterion under external disturbance are analyzed and verified.Through the research of detection method,the paper has completed the preliminary design of DC series arc fault detection device.The device can be embedded in the photovoltaic power station DC combiner box to realize the real-time on-line fault detection feature of DC series arc.In terms of the hardware design part,the main control module,analog input module,trip module,alarm module,fault information storage module,network communication module and other functional modules are designed.Regarding the software design part,the main program flows,such as analog data acquisition flow,fault criterion algorithm flow,and remote communication flow are completed.