Modelling of various meteorological effects on leakage current level for suspension type of high voltage insulators using HMLP neural network

Electrical power system lines sometimes pass along the coastal regions and transverse through the industrial areas of the Peninsular Malaysia. The phenomenon of salt blown from the sea to the land at the coastal area was causing salt deposition to the transformer bushing which contaminating the bushing surfaces and produced leakage current. Hence, it triggering to insulator flashover and finally the hot power arc will damage the bushing. This paper estimates leakage current level by modeling it as a function of various meteorological parameters using Hybrid Multilayered Perceptron Networks (HMLP) with Modified Recursive Prediction Error (MRPE) learning algorithms. The results are also compared with the regression analysis done previously. Meteorological parameters and leakage current data are based on the real measured data collected at YTL Paka Power Station in Terengganu.