Reliability Demonstration Within the Framework of Event-Based Endurance Testing considering prior knowledge

Abstract

In order to achieve maximum test coverage and predict reliability in the field, individual test vehicles in development are not subjected to identical load collectives during endurance tests. Instead, the distribution of the individual samples over the routes and their composition comprise individual, vehicle-specific damage parameters and procurement factors. Since some components are not damaged due to the mileage covered in the test, but rather by certain events, a statistically precise reliability statement cannot be formulated solely through a mileage-oriented test. For this reason, endurance testing should focus less on pure mileage accumulation and more on the damaging events themselves. In the context of statistical reliability analysis, it must be considered that the failure behavior of non-mechanical components can deviate from a Weibull distribution. In light of this, the methodological options must allow for the service-life distribution, which best describes the failure behavior in testing and in the field, to be integrated into the reliability demonstration. Existing methods do not offer an option for considering the performance of several event-based lifetime tests with individual vehicle damage in a reliability distribution. Furthermore, prior knowledge from predecessor series or similar series based on the same platform must be taken into account during the validation process, especially during vehicle development. Today’s vehicle manufacturers use a kind of modular box for development. As a result, a wide variety of components can be found in any series, spread across the entire product range. In addition, some series share an identical platform and thus have many common components, enabling the testing effort for a project release for reliability demonstration as part of a vehicle’s development process to be significantly reduced by taking prior knowledge into account when verifying reliability.