1 - Reliability-Based Design Optimization

Abstract

In order to increase and maintain their businesses, mechatronics manufacturers develop innovative products and reduce product development costs. Economic constraints motivate them to reduce the duration of the testing phase and the number of prototypes and to develop simulations. Introducing innovations enables them to meet customers’ expectations and stand out from their competitors. However the risk that their products might not function correctly in operational conditions must be evaluated with precision as defects during the warranty period have a negative impact on the bottom line. To reduce these industrial and financial risks and respect the levels of performance and reliability required, reliability actions need to be part of design. This chapter describes a reliability-based design optimization approach which is adapted to mechatronics systems. The first step involves establishing predictive reliability of the electronics system in operational conditions based on the FIDES reliability manual. This recent manual based on probability laws is regularly updated according to field returns. The objectives of the following steps are to identify the potentially faulty elements in the life profile conditions and to establish stress distributions leading to failures. In order to understand failure mechanisms, the effects of mechanical, thermal and electromagnetic loads on several prototypes are characterized experimentally. After this, tests are used to provoke failures. Consecutive failure analysis helps to develop multiphysics failure models. These models are optimized and validated by comparing model responses to thermal or vibratory solicitations and those obtained experimentally. Developing meta models capable of treating the variability of life profile loads and of fabrication enables reliability predictions. Design is then optimized by adjusting the architecture elements which are critical for reliability.