Modeling of Czochralski silicon single crystal growth process based on improved stack autoencoding

The growth process of Czochralski silicon is a strong nonlinear time-varying process with multiphase and multi-field coupling. Its quality is closely related to modeling, so it is necessary to accurately model it. Among them, how to better learn the characteristics of silicon crystal growth process has become a difficult. To solve this problem, this paper proposes an improved stack autoencoding method to realize the deeper mining of features. According to the experimental data of crystal growth in single crystal furnace, the performance of the improved model is compared with that of the stack autoencoding before the improvement, and the effectiveness of the improved model is verified.