MicroNanoReliability Congress, 2007

Embedded Remaining Life Prognostics and Diagnostics of Electronics


Vincent Rouet1
Arnaud Delye1
Nikhil Vichare2*
Michael G. Pecht2
Bruno Foucher1


1EADS France, Suresnes, 92152 - France
2CALCE EPSC, University of Maryland, College Park, MD, 20742 – USA
*Now with DELL inc., One Dell Way, MS 6628, Round Rock, TX, 78682 – USA

Abstract:

Health and Usage Monitoring Systems (HUMS) have been developed to improve the reliability as well as the functionality and the performance of high cost systems such as helicopters. As electronics is increasingly integrated into systems, the cost of electronic equipment failures is increasing similarly. Thus, HUMS are more and more valuable for diagnostics and prognosis. This trend is supported by aircraft manufacturers and the MoDs as well, which intend to extend the use of prognostics technologies on weapon platforms, vehicles and ammunitions. To assess the feasibility of an onboard Prognostic Health Management (PHM) system, this paper discusses the demonstration of a real time PHM system for electronic interconnect fatigue prediction based on in-situ sensors, data fusion and relevant algorithms processing. The study focuses on thermo-mechanical degradation of electronic assemblies, because a significant percentage of field failures are related to the high temperature and temperature cycling operating environments of electronic equipment.

The prototype presented within this study is an autonomous embedded PHM system which integrates an onboard real time methodology for remaining life prediction based on a physics of failure approach. The results of accelerated failure testing applied on a Printed Wired Assembly (PWA) are compared to the modelling and predictions. They validate the identified prognostic features and show good agreement between the actual degradation status and the life expectancy forecasted for the PWA.

Complete article is available to CALCE PHM Consortium Members.



[Home Page] [Articles Page]
Copyright © 2008 by CALCE and the University of Maryland, All Rights Reserved