[11p-N107-1~16] 23.1 Joint Session N "Informatics"

△ [11p-N107-11] Estimation of melt state dynamics in floating zone method using Gaussian mixture model

〇Ryo Omae¹, Shogo Sumitani¹, Yusuke Tosa¹, Shunta Harada² (1.Anamorphosis Networks, 2.Nagoya Univ.)

Keywords:Gaussian mixture model, Crystal growth, System identification

In some material manufacturing processes, a person checks the time-varying state of the process and controls it accordingly, and it is often difficult to automate such processes. If the dynamics of the process (how the state changes in response to the input) can be estimated, the optimal operation trajectory can be estimated by reinforcement learning, etc., and automatic operation will become possible, similar to robot walking control. However, in general, it is difficult to collect a large amount of operation data, and it is necessary to estimate the dynamics from a small amount of data. In this study, we estimated the dynamics of crystal growth in the floating zone (FZ) method using a Gaussian mixture model, noting that operational data in manufacturing processes often show similar time-series changes.