Effect of pre-stressing of the test objects on the discharge magnitudes in high density polyethylene

The partial discharges in cavities in solid dielectrics are the major source of insulation failure. In this work, the results of an experimental study of the effects of pre-stressing the high density polyethylene are reported. Artificial cavities of known dimensions were created in test samples, Discharge magnitudes are measured while increasing the voltage in steps to a predetermined value and the discharges are measured with decreasing the voltage and the values are different. The phenomenon was irreversible. The discharge magnitude characteristic which was specific for a given cavity is then used in training the artificial neural network. With the completion of the neural network learning process, the network is used for estimating the cavity size in a given high density polyethylene test specimen