Incorporating Product Retirement in Field Performance Reliability Analysis
Abstract: Ã£ 2010 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must first be obtained from the IEEE.
Statistical time-to-failure analysis is a very powerful and versatile analytical tool available to reliability engineers and statisticians for understanding and communicating the failure risk and reliability of a component, device, or system. The typical approach to characterizing time to failure involves fitting a parametric distribution, such as a Weibull probability function, using time series data on sales and records of failure incidents since the launch of a product. Surviving units are treated as having right-censored failure times. If all sold products never retire from service or the retirement rate is so low that it can be ignored, the quantity of surviving units at each age point can be simply derived from the corresponding monthly sales quantity by subtracting the number of failed units sold in that month. However, modern consumer electronic products retire at much faster rates than traditional products such as coffee grinders or audio receivers. For example, most cell phone and consumer printer users upgrade to newer generations of devices even though their older devices work just fine. Properly accounting for the retirement rate in estimating the size of the at-risk population has become one of the most important steps in the field performance reliability analysis of consumer products. Neglecting to account for the shortened age of early retired units can lead to inaccurate characterization of the time-to-failure distribution and an inaccurate estimate of future failures. Retirement rates can be estimated from available information, or analyses can be performed using different assumed rates to examine the sensitivity of inferences. We present several case studies to demonstrate the practical value of accounting for the product retirement of surviving units. Results vary significantly from those obtained without addressing product retirement, thus underscoring the importance of this consideration in practical applications.
Keywords: RAMS 2010 Proceedings - Prediction