The objective of this project is to develop a new reliability prediction methodology for space systems aiming to overcome the inherent limitations of prediction practices currently based on outdated or limited handbooks (e.g. MIL-HDBK-217).

The new methodology shall turn reliability predictions into more realistic estimations of the inherent reliability of the system design. This will improve the effectiveness of such estimations to support the analysis of alternative design solutions and verification approaches, as well as the evaluation of the conservativeness of design margins with respect to cost savings opportunities.

The new reliability prediction methodology shall contribute to the sustainability and affordability of future space systems, including within highly competitive markets.

Reliability goals are defined for most ESA projects and reliability predictions are performed at all design levels. The most widely used methodology by industry for reliability prediction is MIL-HDBK-217. This has become out-dated and has limitations due to the obsolescence of its part type failure rate prediction models and the database incompleteness. Moreover, current prediction models are mostly addressing random failures while a significant amount of in-orbit anomalies are not due to random failures and thus are not covered in current methods (e.g. design and/or manufacturing related failures).

Inadequacies of current practices may have adverse impacts on ESA projects, such as: (i) the substantiation of quantitative reliability requirements by the customers, and the relative apportionments of the total system reliability budget from system to lower levels

are questionable; (ii) prediction results provide an inadequate basis to assess the achievement of reliability requirements and are misleading even for design trade-offs; (iii) significant differences exist in the reliability predictions from different manufacturers, with

increased uncertainties for customers to compare their reliability statements and to evaluate their products.

In projects where feedback from industry has highlighted the excellent reliability performances of their systems and products, sometimes exceeding the lifetime expectations, there is a perception of over-design and consequently reduced cost effectiveness during the development process, especially when predictions are proven to be largely conservative with respect to the actual in-orbit performances.

In order to remove the current shortcomings and contribute to the sustainability and affordability of future space systems, including within highly competitive markets, a new reliability prediction methodology satisfying the specific goals and constraints of space systems and products is needed.