Designing and Evaluating Distributed Power Processing Systems for Reliability and Maintainability

A method for measuring the reliability and maintainability of distributed power-processing systems is presented. The method utilizes a set of figures of merit which faithfully reflect both the service availability of the system and the impact on service, due to any possible system failures. The procedure used to calculate these figures of merit along with their concomitant cost, is explained. The entire method has been incorporated in a new computer program, which automatically evaluates the cost, reliability, and maintainability performance-of proposed distributed power-processing designs. With the benefit of this program, enough experience has been obtained to suggest a number of design effects and design rules for various architectural features of distributed systems.