This paper proposes a new reliability prediction process which consists of four steps; 1) field data utilization step, 2) similar product data utilization step, 3) degradation test data utilization step, and 4) reliability prediction handbook utilization step. While, conventional reliability prediction process only describes a rough sequence of activities, the proposed process specifies the data priorities to improve the reliability prediction accuracy. For the effective support of the proposed reliability prediction process, we develop a software architecture and implement the software system. The software system consists of four major modules (function module, mapping module, calculation module, and report module), and applied to the reliability prediction of a DC-DC converter for demonstration.
