Health Assessment of Electronic Products using Mahalanobis Distance and Projection Pursuit Analysis

With increasing complexity in electronic systems there is a need for system level anomaly detection and fault isolation. Anomaly detection based on vector similarity to a training set is used in this paper through two approaches, one the preserves the original information, Mahalanobis Distance (MD), and the other that compresses the data into its principal components, Projection Pursuit Analysis. These methods have been used to detect deviations in system performance from normal operation and for critical parameter isolation in multivariate environments. The study evaluates the detection capability of each approach on a set of test data with known faults against a baseline set of data representative of such “healthy" systems.

Authors:

• Sachin Kumar
• Vasilis Sotiris
• Michael Pecht

Keywords:

• Mahalanobis distance
• Principle components
• Projection pursuit
• Health assessment
• Anomaly.

References:

[1] N. Vichare, P. Rodgers; V. Eveloy; and M. Pecht; "Environment and Usage
Monitoring of Electronic Products for Health Assessment and Product
Design," International Journal of Quality Technology and Quantitative
Management. vol. 2, no. 4, 2007, pp. 235-250.
[2] J. Gu; N. Vichare; T. Tracy; and M. Pecht; "Prognostics Implementation
Methods for Electronics," 53rd Annual Reliability and Maintainability
Symposium (RAMS), Florida, 2007.
[3] G. Zhang, C. Kwan; R. Xu; N. Vichare; and M. Pecht; "An Enhanced
Prognostic Model for Intermittent Failures in Digital Electronics," IEEE
Aerospace Conference, Big Sky, MT, March 2007.
[4] N. Vichare; and M. Pecht; "Enabling Electronic Prognostics Using Thermal
Data," Proceedings of the 12th International Workshop on Thermal
Investigation of ICs and Products, Nice, C├┤te d'Azur, France, 27-29
September 2006.
[5] N. Vichare, P. Rodgers; and M. Pecht; "Methods for Binning and Density
Estimation of Load Parameters for Prognostics and Health Management,
International Journal of Performability Engineering, vol. 2, no. 2, April
2006.
[6] A. Fraser; N. Hengartner; K. Vixie; and B. Wohlberg; "Incorporating
Invariants in Mahalanobis Distance based Classifiers: Application to Face
Recognition," in International Joint Conference on Neural Networks
(IJCNN), (Portland, OR, USA), Jul 2003.
[7] J. E. Jackson; and G. S. Mudholkar; "Control Procedures for Residuals
Associated With Principal Component Analysis," Technometrics, vol. 21,
no. 3, 1979.
[8] J. Liu, K. Lim; R. Srinivasan; and X. Doan; "On-Line Process Monitoring
and Fault Isolation Using PCA," Proceedings of the 2005 IEEE
International Symposium on, Mediterranean Conference on Control and
Automation, 2005, pp. 658 - 66.
[9] G. Taguchi, and R. Jugulum; The Mahalanobis-Taguchi Strategy: A
Pattern Technology System, Wiley, 2002.
[10] G. Taguchi, S. Chowdhury; and Y. Wu; The Mahalanobis-Taguchi System,
New York: McGraw-Hill. 2001.
[11] E. B. Martin, A. J. Morris; and J. Zhang; "Process Performance Monitoring
Using Multivariate Statistical Process Control," IEEE Proceeding of
Control Theory Application, vol. 143, no.2, March 1996.
[12] H. Chen, G. Jiang; C. Ungureanu; and K. Yoshihira; "Failure Detection and
Localization in Component Based Products by Online Tracking," KDD,
2005.
[13] H. Wang, Z. Song; and P. Li; "Fault Detection Behavior and Performance
Analysis of Principal Component Analysis Based Process Monitoring
Methods," American Chemical Society, vol. 41, 2002, pp. 2455 - 2464.
[14] H. Yue, and S. J. Qin; "Reconstruction-Based Fault Identification Using a
Combined Index," American Chemical Society, vol. 40, 2001, pp. 4403-
4414.