Integrating reliability improvement modeling into practice-challenges and pitfalls

There is a need to develop methods and models that can be used to estimate reliability not only to demonstrate compliance with requirements but also to monitor reliability growth through the product life. A modeling framework is presented that aims to estimate the reliability of a variant design throughout the product lifecycle to support reliability enhancement decision-making. A Bayesian approach to modeling is adopted where expert judgement data is combined with in-service reliability data. This paper reports one practical implementation within an ongoing electronic design project. A formal process of evaluation is embedded within the modeling process to capture the views of relevant engineers, managers and analysts during the implementation. Through this evaluation we share our experiences in the successes and challenges of modeling, and offer guidance on pitfalls to avoid. The major challenge has been managing the change in reliability culture within the design team required to make this process work. Other problems relate to the acquisition of relevant, good quality data in a timely and cost-effective manner. The major successes relate to getting engineers together to think about reliability issues and share experiences. The modeling process is novel to the engineers, but has encouraged them to review the impact of their decisions on reliability and draw more useful information from the reliability indicators provided by the model.