Machine Learning Based Power Loss Modeling and Fault Diagnostics for Power Transformer and AC Drive System

Degradation of component insulation, indicated by increased power loss, causes various faults in power conversion systems. However, direct power loss measurement is challenging. In this paper, a novel machine learning based power loss modeling method is presented and adopted for inter-turn short circuit diagnostics in power transformer. Dynamic behaviors of the power loss in the system are modeled using machine learning, and the decision boundary is determined based on system operation conditions. Proposed method is validated through simulation and experimental test.