(methods)= # Part 2 - Methods This part of the handbook, Part 2, is dedicated to presenting the different methods available to perform reliability predictions ({term}`RP`). It starts with reminding the main concept, such as the statistical approach, the physics of failures (PoF) approach and the combined approach, in {numref}`method_5`, before fully developing the combined approach most used for space applications, the Bayesian approach in {numref}`method_6`. Finally, {numref}`method_7` 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 {numref}`Figure_method_0`. ```{figure} pictures/method_figure1.png --- width: 600px name: Figure_method_0 --- Relation between different types of reliability prediction methods and the available inputs ``` This part contains the following paragraphs: ```{tableofcontents} ```