Classification of Superimposed Partial Discharge Patterns

Phase resolved partial discharge patterns (PRPD) are routinely used to assess the condition of power transformers. In past, classification systems have been developed in order automate fault identification task. Most those work with assumption that only one source is active. reality, however, multiple PD sources can be active at same time. Hence, PRPD overlap and cannot separated easily, e.g., by visual inspection. Multiple a single represent multi-label problem. We present system based on long short-term memory (LSTM) neural networks resolve this The generally able classify overlapping while being trained class sources. achieves accuracy 99% mean 43% for an imbalanced dataset. This method identify main and, depending data, also second source. works conventional electrical measuring devices. Within detailed discussion presented approach, both its benefits but problems regarding different repetition rates evaluated.