[1Win4] Poster session 1

[1Win4-80] Dynamic analysis of factory and environmental factors affecting properties of rubber materials based on causal inference

〇Huaying Zhang¹, Ren Togo¹, Keisuke Maeda¹, Takahiro Ogawa¹, Miki Haseyama¹ (1.Hokkaido University)

Keywords:causal inference, mateiral development, property estimation

Estimating the physical properties of materials beforehand is crucial for advancing rubber material development. Previous work has studied the influence of external factors, i.e., process factors (mixing methods, etc.) and environmental factors (temperature and humidity, etc.), on the physical properties of rubber materials for accurate estimation. By controlling the adjustable factors that are considered influential, sustainable production is expected. However, previous work did not consider the effect of time series factors such as seasonal change. As a result, the detailed relationship between external factors and physical properties remains unrevealed. In this paper, we analyze the influence of external factors on the physical properties of rubber materials by dynamically updating the analyzing model using the data stream. Specifically, we use a causal discovery method to model the causal relationships between external factors and physical properties and then update the causality model with the data stream in time series. The experimental results reveal temporal changes in these causal relationships. By observing these causal relationships dynamically, a more controllable and sustainable production of high-quality rubber materials is expected.