Usage space analysis for reliability testing
Article first published online: 22 OCT 2010
Copyright © 2010 John Wiley & Sons, Ltd.
Quality and Reliability Engineering International
Special Issue: Business and Industrial Statistics: Developments and Industrial Practices in Quality and Reliability
Volume 26, Issue 8, pages 877–885, December 2010
How to Cite
Haselgruber, N., Mautner, K. and Thiele, J. (2010), Usage space analysis for reliability testing. Qual. Reliab. Engng. Int., 26: 877–885. doi: 10.1002/qre.1155
- Issue published online: 29 DEC 2010
- Article first published online: 22 OCT 2010
- principal component analysis;
- design of experiments;
- multivariate regression;
- accelerated testing;
- customer usage space
During the development process of a complex technical product, one widely used and important technique is accelerated testing where the applied stress on a component is chosen to exceed the reference stress, i.e. the stress encountered in field operation, in order to reduce the time to failure. For that, the reference stress has to be known. Since a complex technical product may fail regarding numerous failure modes, stress in general is highly dimensional rather than scalar. In addition, customers use their products individually, i.e. field operation should be described by a distribution rather than by one scalar stress value. In this paper, a way to span the customer usage space is shown. It allows the identification of worst case reference stress profiles in significantly reduced dimensions with minimal loss of information. The application example shows that even for a complex product like a combustion engine, stress information can be compressed significantly. With low measurement effort it turned out that only three reference stress cycles were sufficient to cover a broad range of customer stress variety. Copyright © 2010 John Wiley & Sons, Ltd.