Main Article Content
Partial discharge analyzer (PDA) has the unique potential for evaluating degradation and possibly failure prediction of XLPE cables as well. The authors have demonstrated its capability towards failure prediction on XLPE cables. In this study, five sets of fifty number of identical XLPE cables were taken up for the investigation. These tests were carried out on both damaged and undamaged conditions. The data acquired during their loading cycles is analyzed using Artificial Neural Network (ANN) which is further developed to predict the failure performance of power cable. Using this approach one could appreciate that impending failure is significant even at 50 to 60 % of maximum expected operating voltage (MEOV) with a reasonable error margin. In fact, till date there is no lucid method spelt out in the open literature for the failure prediction of XLPE Cables. Moreover, this methodology can also be applied to predict in real time the failure of similar XLPE cables made of material systems. Also, in order to verify the results, Gaussian Mixture Models (GMM) is used for data analysis. This work can also be further extended to other equipment like transformers, switch gears and rotating machines.