A Dual–Stage Aging Process Model for Reliability Analysis of Warranty Claims
Publication Date: 2016-Jan-15
The IP.com Prior Art Database
Page 01 of 3
A Dual-Stage Aging Process Model for Reliability Analysis of Warranty Claims
In root-cause analysis of field reliability issues, warranty cost forecasting and field service action planning, there arises a need to model a population of some identical vehicle components with non-homogenous aging history. The traditional approach is to model the time origin of the aging history of as the vehicle sales date, commensurate with the start of customer usage cycle. However, some components (e.g., subject to corrosion) start aging right after they are produced, i.e., before entering the customer usage cycle. The time interval between the production date and the sales date varies across vehicle population and can be as high as 12 months.
We are modelling a dual-stage aging process consisting of two consecutive time intervals between vehicles: 1) production and sales dates, 2) sales and repair dates, respectively. The underlying hazard functions that correspond to the two stages can considerably differ from each other both in their value and/or the gradient. We further model the two stages as the time-dependent covariates of a reliability (survival) model. This approach considerably improves the estimation and prediction accuracy of the dual- stage aging process, which, in turn, enables an efficient root-cause analysis, warranty cost forecasting and field service action planning. Full mathematical details of the proposed solution are discussed below.
Page 02 of 3
Mathematical details of the proposed solution
Warranty Start Date (WSD)
Production date Warranty Start Date Repair Date
Censoring (cut-off) Date