A Framework for Warranty Prediction During Product Development

This paper describes the challenges faced by companies to manage warranty performance during product development. Understanding and reducing warranty cost often focuses exclusively on the analysis of product failures. However, warranty costs can also be incurred by events such as misaligned customer expectations that do not involve a product failure, per se. Many experts agree that effective management of system reliability and reliability validation during product development is a key to achieve superior time to market and life cycle quality. The paper first surveys the challenges faced by various organizations ranging from consumer electronics to aircraft engines to experimental high-energy physics accelerators. From the survey emerge some key and common issues that these companies face: identification of failure events; reliability modeling and prediction; prototyping and validation testing. The paper then reviews the current state of the art to identify areas for improvements as well as needed integrations in order to develop a comprehensive framework that will be useful to product developers to manage and predict warranty performance during product development. This framework extends and integrates three areas: 1) extend scenario-based FMEA to include the diagnosis and repair of failure events as part of the scenario; 2) use of Bayesian methods to integrate field data, product development data and engineering judgments; 3) generate costs models that allow tradeoff studies between product design, service model design and warranty policies. The paper concludes by presenting a future research agenda.