Methodological framework for implementation of a prediction reliability model of IGBT power modules used in railway applications

The control of critical electronic components reliability is one of the main issues in railway traction applications. Insulated Gate Bipolar Transistors (IGBT) modules are part of these components. They are subjected to high stresses due to severe conditions of use of the train. The increase of requirements in terms of reliability and safety imposes to be able to assess these dependability measures. This paper introduces a methodological framework for predicting reliability of IGBT based on an innovative and structured Bayesian approach.