Can Failure Modes and Effects Analysis Assure a Reliable Product?

Abstract

FMEA us dully regarded as a valuable tool for improvement of product reliability. In many industries, specifically in automotive, FMEA is a mandatory deliverable of product development process documentation. The three attributes, Severity, Probability of Occurrence and Detection, in automotive methodology, or Severity and Probability of Occurrence in the International, IEC methodology, need to be determined to evaluate the risk and mitigate those failure modes that pose a high risk to the product functionality or to the User. While Severity of a potential failure mode is intuitively easy to estimate, Occurrence and Detection need inputs from other reliability methods (reliability prediction or test) to be correctly determined. A very complex FMEA or FMEA with applied Physics of Failure for estimation of probability of occurrence of individual failure modes requires considerable resources for accurate estimates of risk. FMEA also depends on information from other reliability activities such as reliability prediction with failure rate acceleration for stresses, inputs from fault tree or event tree analyses. A thorough FMEA will show failure modes mitigation, risk reduction, and relative reliability improvements of a product, resulting from improvements but will not track overall product reliability improvement since it does not model functionality of the product. When a product also has reliability requirements, performance of just FMEA without support from other reliability methods, will not guarantee or show that the required reliability is achieved. It is the well organized and managed overall reliability program with reliability and engineering tests and analyses (including FMEA) that can assure that the delivered product will meet Customer expectations.