Papers by Author: Adrien Zéanh

Abstract: This paper presents a Finite Elements Modelling (FEM) based methodology dedicated to the evaluation of the lifetime and the reliability of assemblies involving brittle materials under cyclic loading. It focuses on the particular case of metal bonded Aluminium Nitride (AlN) substrates used in power electronic switch modules. The ceramic fracture criterion was formulated according to the weakest link concept, under Weibull's approach. The material's parameters were determined by running three points bending tests. In order to check the relevancy of the proposed methodology, a non linear thermomechanical Finite Elements Model allowed computing the number of thermal cycles before substrate brittle fracture within a test vehicle, which was then compared to experimental results. Once validated, the methodology was applied to two different configurations of a power switch module, designed for harsh environment aeronautic applications. The corresponding external loading profile was considered to compute and monitor the evolution of the maximal principal stresses within the ceramic substrates whole volumes. Their lifetimes and reliabilities was finally assessed and compared to the applications requirements.

Abstract: This work studies the reliability of power electronic component in aeronautical environment to the ageing eect of the thermal cycling. The structure fatigue is sensitive to the process assembly conditions especially of the soldering process. To correclty evaluate the reliability of the power module, the identication of the solder behavior is one of the rst steps. Anand Model is here identied. Experimental test have to be established to evaluate the parameters of the law. A srt study is made to evaluate the indetiability of the law according to the dierent experimantal test. Then, the scatter of the parameters is evaluated in a context of time series. In the end, the scatter of the parameters is included in a nite element model to understand the inuence of this scatter on the evaluation of the number of cycle before failure.