2. Part 2 - Methods

This part of the handbook, Part 2, is dedicated to presenting the different methods available to perform reliability predictions (RP).

It starts with reminding the main concept, such as the statistical approach, the physics of failures (PoF) approach and the combined approach, in Section 2.4, before fully developing the combined approach most used for space applications, the Bayesian approach in Section 2.5. Finally, Section 2.6 addresses the uncertainties of the models and how to deal with them.

The different reliability prediction inputs considered are: in-orbit return (IOR), manufacturer data, test data, handbook data, and failure mechanism analysis. With these inputs, reliability prediction models can be derived using statistical methods, PoF approaches or a combination of both, see Fig. 2.1.

Fig. 2.1 Relation between different types of reliability prediction methods and the available inputs

This part contains the following paragraphs:

• 2.1. Scope
• 2.2. Normative references
• 2.3. Terms, definitions and abbreviated terms
• 2.4. Model development basic concepts
• 2.5. Bayesian methods for reliability prediction modelling
• 2.6. Uncertainties in the prediction results
• 2.7. Annex A Probability distributions
• 2.8. Annex B Prediction uncertainties for Part III models
• 2.9. Bibliographic references