Equipment Health Factor prediction for complex semiconductor manufacturing facility
Résumé
This work takes place within the IMPROVE European project aimed at increasing the availability of manufacturing equipment and to avoid rejection of the products in the semiconductor field. The thermal furnaces are one of the important production equipment. They are composed of two processing reactors for the gas deposition on silicon wafers at low pressure and high temperature. Due to the occurrence and the severity of registered drifts, this equipment requires special attention. In this context, we propose a probabilistic model to predict failures based on Bayesian belief networks. The sequential data are strongly present on the extracted databases and their modeling is important. For their simplicity and flexibility, the dynamic Bayesian networks are used for this. They allow predict the future failures according to their causes and in a dynamic way.