The early design reliability prediction method

Abstract

This research contributes to the ongoing effort toward predicting and improving reliability in early engineering design, specifically during conceptual design. Increasing the rigor of predicting reliability during early design allows the designer to improve the decision-making process. Current reliability prediction methods cannot accurately capture uncertainty during conceptual design. The uncertainty of reliability predictions in early design can be significant, and, as a result, must be utilized in decision-making. To address these limitations, this paper presents the early design reliability prediction method (EDRPM) to calculate function and component failure rate distributions, which capture the parameter uncertainty. If required, these distributions can be converted into reliability values. A new technique, Frequency Weighting, is developed and applied to a Hierarchical Bayesian model to account for the number of times that a component has historically solved a function. This addition is necessary to accurately model the structure of the function-to-component relationship for capturing uncertainty. A case study using the advanced diagnostics and prognostic testbed electrical power system is presented to show the usefulness of EDRPM in facilitating decision-making during conceptual design.