Mahalanobis Distance and Projection Pursuit Analysis for Health Assessment of Electronic Systems

This paper presents a Mahalanobis distance and projection pursuit analysis based prognostic and diagnostic approach for early detection of anomalies in electronic products and systems. These have been used to detect deviations in system performance from normal operation, and are efficient at characterizing products with short field histories. A case study is presented to demonstrate that an "abnormal" system can be distinguished from a "normal" system and that a new system can be characterized based on existing baselines from different computer models.