9 - Reliability estimation considering warranty data, usage rate profile, and qualitative information on design changes

Abstract

Designing warranty policies based on reliability estimates obtained by using only failure/warranty data can be unrealistic and incur a huge cost due to inaccurate estimates. This suggests that to get more accurate reliability estimates, one should be able to utilize all sorts of information available for assessing the reliability. This chapter summarizes a comprehensive framework that facilitates the integration of information and/or data on surviving population as well as qualitative or subjective information on the impact of design changes on product reliability. The framework uses customer usage behavior to derive information on surviving population and treated survival time as censored data. The subjective information on design changes and their impact on reliability improvement are derived using appropriate indices and fuzzy logic. The inclusion of the censored data on surviving population and impact of design changes in reliability improvement provide better reliability estimates and warranty claims prediction. The case examples along with simulation study are provided to demonstrate the applicability and the validity of the proposed framework.