A. Zare - Department of Electrical Engineering- Marvdasht Branch, Islamic Azad University, Marvdasht, Iran.
M. Simab - Department of Electrical Engineering- Marvdasht Branch, Islamic Azad University, Marvdasht, Iran.
M. Nafar - Department of Electrical Engineering- Marvdasht Branch, Islamic Azad University, Marvdasht, Iran.

Due to the growing demand in the electricity sector and the shift to the operation of renewable sources, the use of solar arrays has been at the forefront of consumers' interests. In the meantime, since the production capacity of each solar cell is limited, in order to increase the production capacity of photovoltaic (PV) arrays, several cells are arranged in parallel or in series to form a panel in order to obtain the expected power. Short circuit (SC) and open circuit (OC) faults in the solar PV systems are the main factors that reduce the amount of solar power generation, which has different types. Partial shadow, cable rot, un-achieved maximum power point tracking (MPPT) and ground faults are some of these malfunctions that should be detected and located as soon as possible. Therefore, effective fault detection strategy is very essential to maintain the proper performance of PV systems to minimize network interruptions. The detection method must also be able to detect, locate and differentiate between SC and OC modules in irradiated PV arrays and non-uniform temperature distributions. In this paper, based on artificial intelligence (AI) and neural networks (NN), neutrons can be utilized, as they have been trained in machine learning process, to detect various types of faults in PV networks. The proposed technique is faster than other artificial neural networks (ANN) methods, since it uses an additional hidden layer that can also increase processing accuracy. The output results prove the superiority of this claim.

Photovoltaic arrays, Fault detection, Machine learning, Neural network