Fourier Transform and Short-Time Fourier Transform Decomposition for Photovoltaic Arc Fault Detection

Arc faults are one of the leading causes of photovoltaic (PV) system failure resulting in electrical fires. Thus, arc fault detection is crucial for safe operation and is a prerequisite for high penetration of PV. Developing an effective arc fault detector requires a thorough understanding of the electrical characteristics of PV arcs. However, since arcs produce chaotic waveforms, a large sample of voltage and current waveforms are needed. This paper describes a computer-controlled mechatronics testbed to facilitate scientifically repeatable arc generation conditions and waveform recording. The recorded waveforms are analyzed using the Fast Fourier Transform (FFT) and the Short Time Fourier Transform (STFT). The results of using both methods are compared to discuss their effectiveness as tools to detect the onset of PV arc faults as per the UL1699B criteria.