Risk Assessment Based on Limited Field Data
 
William Q. Meeker
Department of Statistics
Iowa State University
Ames, IA 50011
wqmeeker@iastate.edu

Abstract

Many consumer products are designed and manufactured so that the probability of failure during the technological life of the product is small. Most product units in the field retire before they fail. Even though the number of failures from such products is small, there is still the need to model and predict field failures in applications that involve safety and risk assessment.   Challenges in modeling and predictions of failures arise because the retirement times are often unknown, and there are delays in field failure reporting.  Motivated by an application to assess the risk of failure, we develop a statistical procedure to predict the field failures of products, which considers the impact of product retirements and reporting delays. Based on the developed method, we provide the point prediction for cumulative number of reported failures in a future time and the corresponding prediction interval to quantify uncertainty. We also conduct sensitivity analysis to show the effects of different assumptions on failure-time and retirement distributions and the values of their parameters.

This is joint work with Zhibing Xu and Yili Hong

2015 QPRC