A Novel Approach to Corona Monitoring

Corona discharge (CD) and partial discharge (PD) indicate early stages of insulation problems in motors. Early detection of CD/PD will enable better coordination and planning of resources such as maintenance personnel, ordering of parts, etc. Most importantly, one can prevent catastrophic failures during normal operations. In this proposal, Intelligent Automation, Incorporated (IAI) and its subcontractor, Prof. W. Lee at University of Texas at Arlington (UTA) propose an integrated and low approach to detect PC/CD. It consists of several major steps: preprocessing to remove background and sensor noise, feature extraction using Principal Component Analysis (PCA) or Power Spectral Density (PSD), and efficient fault classifiers such as Support Vector Machine (SVM), Multilayer Perceptron Neural Net (MLNN) for fault detection and isolation. In addition, anomaly detection, trending, and time-to-failure prediction algorithms will be developed and incorporated into this Phase 2 proposal. The approach uses low frequency signals and hence is much cheaper (less than ten thousand dollars per unit) than existing methods in the markets. In Phase 2 we will develop motor/switchgear test beds, develop and implement various algorithms, and perform extensive tests in both the test beds and the actual motor/switchgears at Arnold Air Force Base.