Electrical insulation plays a very important role in motor operations; insulation failure may damage the whole equipment. Partial discharge (PD) is an early indication of an electrical insulation problem; PD fault detection/diagnostics, therefore, becomes a critical and economical tool to avoid in-service failure or property loss. This thesis develops an easy-to-install, low cost PD health monitoring system for motors. The system employs an ANN with multiple sensors inputs for PD diagnostic task. The proposed algorithms and circuits are implemented and tested in the laboratory environment. Simulation results show that the system is sensitive and accurate. |