Abstract: Reliability prediction is an important phase in product design and development. Comprehensive and precise reliability prediction can be used to evaluate the reliability level of products and provide information and guidance for design. This article summarizes the technical methods currently employed for electronic equipment reliability prediction and holds that the present reliability prediction methods based on physics of failure(PoF)fail to take into consideration the functional composition of products. It,then proposes an electronic equipment reliability prediction method(EERPM)which is based on PoF analysis but also takes circuit function into consideration. This method first establishes the key parts for the performance of a circuit by means of the two methods of sensitivity simulation and principal component analysis(SSPCA), and then decides the probable life distribution of each part by PoF method or statistical experiential data. System life distribution can be obtained by means of mixed distribution; consequently, system reliability data will be obtained. The article takes an aviation power supply circuit as an example to illustrate the method.

Key words: reliability prediction, functional composition analysis, physics of failure, key part, mixed distribution