From reliability-prediction to a reliability-budgetedddd

Abstract

For many years, prediction according to MIL-HDBK-217 was a major activity in most projects. A prediction was required in the contract, but the customer seldom required the assumptions to be stated in the report and compliance with the assumptions to be verified. This leads to the so called "number game". Today it is a "hot issue" if experience data can be used for anything, and whether a component does have a meaningful hazard rate. This paper discusses which advantages a prediction can offer to a manufacturer and a customer. This leads to the conclusion that instead of prediction one should talk about a reliability budget, which will guide the project manager to the part of the design where early analysis and tests should be made. This reliability budget is made using a Weibull probability paper, so that nonconstant hazard rate can be included in the budget. This allows the inclusion of wear-out in the analysis, allowing a comparison of the reliability budget with the reliability target for the system. Reliability prediction has also been criticized for blocking for improved design techniques. The experience of the well known Danish company Bang and Olufsen, using a modified prediction technique to encourage designers to improve their design technique, is described.