PEA Feeder Losses Calculation by using Artificial Neural Networks

M. Patamakajonpong and S. Jiriwibhakorn (Thailand)


Energy and Power Systems, Power Flow Analysis


This paper proposes the determination of technical losses in the distribution feeder system of Provincial Electricity Authority (PEA) of Thailand by using artificial neural networks (ANNs). Input features of ANNs compose of voltage, real power, reactive power and power factor. Feeder 6, Rayong 3 substation is studied. Voltage, real power, reactive power and power factor for training and testing ANNs are obtained from the Computer-Based Substation Control System (CSCS). The Power System Simulator/Advanced Distribution Engineering Productivity Tool (PSS/ADEPT) is used for simulating the load flow in the distribution system including calculating technical losses for the training and testing outputs of ANNs. Program PSS/ADEPT has been used as the benchmark for comparing the results of ANNs. MAPE of test data of ANNs is very good (about 0.0204 %). Therefore, the engineers of PEA could use ANNs to predict the feeder losses of the distribution feeders accurately and comfortably.

Important Links:

Go Back