Haidar Samet, Dariush Keihan Asl, Teymoor Ghanbari, Alaa Hamza Omran
الملخص الانجليزي
Since photovoltaic (PV) arrays are outdoor components and may experience harsh environmental conditions, early fault detection in such systems is vital. High capacity PV farms are generally established in a wide area and so having an efficient and cost effective monitoring system for them is a challenge. This paper is aimed to find the optimal number and scheme of the required measurement units to have a desirable monitoring system. To this end, a neural network (NN) based fault detection method for a typical PV system is proposed. The NN is trained using the gathered voltage and current signals in different normal and fault conditions. All the possible schemes for distribution of the measurement units in terms of scheme and number are considered in the training in the same condition. Accuracy of the NN is assessed in these scenarios as a criterion for selection of the optimal scheme for the measurement units.
تاريخ النشر
10/06/2018
الناشر
2018 IEEE International Conference on Environment and Electrical Engineering and 2018 IEEE Industrial and Commercial Power Systems Europe (EEEIC/I&CPS Europe)
رقم المجلد
رقم العدد
الكلمات المفتاحية
Artificial neural networks , Measurement units , Sensors , Fault detection , Training , Current measurement
الاء حمزة عمران | Alaa Hamza Omran | 6620