Field-Failure Predictions Based on Failure-Time Data with Dynamic Covariate Information

• Activity Number: 497
• Type: Topic Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): William Q. Meeker*+ and Yili Hong
• Companies: Iowa State University and Virginia Tech
• Keywords: Accelerated failure time model ; Lifetime data ; Usage history ; Cumulative exposure model ; Dynamic data ; Warranty returns

Abstract

Today, there are more and more products installed with automatic data-collecting devices such as smart chips and sensors as well as cellular and network communications capabilities that track how and under which environments the product is being used. While there is tremendous amount of such dynamic data being collected, there is little research on using such data to provide more accurate reliability information for products and systems. Motivated by a consulting problem, this paper focuses on using failure-time data with dynamic covariate information to make warranty and other field-failure predictions. The dynamic covariate information is incorporated into a parametric failure-time model through a cumulative exposure model. A prediction procedure that accounts for unit-to-unit and temporal variability in the use rate is developed to predict field-failure returns at a future time. We also define a metric to quantify the improvements obtained using dynamic information in the prediction accuracy. Simulation studies were conducted to study the effect of different sources of covariate process variability on predictions.