A Novel Fault Detection and Location Method for PV Arrays Based on Frequency Analysis

Abstract

This paper presents a fault detection and location method for photovoltaic (PV) arrays. One novelty of the method is the use of the different amplitude changes of the alternating current frequency component of the PV array output voltage. Another novelty is the technique used to generate the PV array output voltage. In this technique, the entire PV array is operated under dark conditions with red pulsed light only applied on one module of the PV array. Under these conditions, the presence of an open circuit fault and the variability in module impedances generate different increases in the total dynamic impedances of the PV array. As a consequence of these increases, the spectral component of the alternating current array output voltage also undergoes different increases, which facilitates fault detection. The discrete Fourier transform and statistical techniques are used to feature extraction and classification indication. The proposed method takes into account reliability and detection time and does not require the use of expensive or specialized equipment or the disconnection of or any modification to the PV array, which reduces the cost of the system. The earlier mentioned is achieved with the use of a pulsed light, which inputs a forward bias voltage into the array. The results of both the simulation and the experiment are presented in order to demonstrate the 100% effectiveness of the proposed method under different temperatures and PV array sizes.