PV series arc fault recognition under different working conditions with joint detection method

In recent years, increasing fire accidents in the photovoltaic (PV) system by PV series arc fault cause huge economic losses and pose great threats to its operating safety. However, increasing PV voltage and maximum power point track (MPPT) from the inverter during arc ignition make PV series arc fault complex to be detected. This paper aims at providing a joint detection method to arc fault circuit interrupters (AFCI) serving for smart micro grid. In this paper, two methods to bring PV series arc fault into the PV system have been recorded by intensified charge-coupled device (ICCD). By introducing PV series arc fault into the system in due time, normal and fault electric signals have been acquired under different imitated working conditions. Statistic method from time domain and short time Fourier transformation (STFT) from time-frequency domain are chosen to diagnose this kind of fault. A detection variable from each method has been proposed to recognize it accurately. A relatively satisfying joint algorithm based on two proposed detection variables has been put forward to prevent unwanted nuisance trips from system transient process. To fit constantly varying electric signals in PV system, this detection algorithm also adopts dynamic threshold value.