A Prediction Based Design-for-Reliability Tool

摘要

In today's short product lifecycles, it is especially important to apply design for reliability (DFR) practices as early as possible. Since decisions about which reliability features should be included in a design occur in the context of a system design trade-off process, there then needs to be some way to help design engineers decide which DFR actions have sufficient benefit to outweigh other concerns. We present an approach to giving design engineers feedback to their decisions from a reliability perspective by using the reliability prediction process. Specific DFR practices are linked to parameters in a reliability prediction model, and a sensitivity analysis allows the design engineer understand the relationship between their choices and the predicted failure rate of the system when it is fielded. We describe an initial implementation of this approach using a standard reliability prediction model, the Telcordia (Bellcore) parts stress prediction model, and some standard rules of thumb from the Reliability Analysis Center's Reliability Toolkit. Finally we discuss plans for future evolution and improvement in this approach.