Reliability Estimation Techniques Applied in ALSTOM

Many organizations and industrial sectors have created databases containing the histories of critical events affecting items of equipment to determine unreliability of equipment. The intention is to record and maintain data about the root cause analysis of such events and to determine the main component / equipment or cause of error which triggered the forced event. Failure rate evaluations are the basis for this estimation. For confidence in these estimations large and accurate resources of data are necessary. When the event information is limited or not enough, the application of mathematical concepts is the only way forward. In this paper the ALSTOM approach to such failure rate estimation using a number of methods is presented. These include development of statistical methodologies that combine frequentist and Bayesian ideas as well as using information that is available as standard industry-wide. An overview shall also be provided on the methodology and application of reliability concepts as developed and customised to the requirements of its product development within ALSTOM power business.