Unraveling Oil-Microbubble Interactions via a Ternary Phase-Field Model and Machine Learning for Enhanced Surface
Mohammadreza Shirzadi1, Yuki Yuhara1, Tomonori Fukasawa1
1Chemical Engineering Program, Graduate School of Advanced Science and Engineering, Hiroshima University, 1-4-1, Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8527, Japan.
None:
This study examines the collision dynamics of a rising microbubble (MB) with an oil droplet adhered to a vertical wall for oily surface cleaning. A high-resolution numerical model based on a ternary phase-field model is developed to analyze MB-oil droplet interactions in shear flow. A computational fluid dynamics database with 182 samples is created by varying key physical properties, including the droplet contact angle on the wall surface, interfacial tensions, bubble diameter, and droplet diameter. For each database sample, the MB-oil interaction is classified as "detached" or "not detached". Results reveal that the traditional spreading coefficient method fails to correctly classify the MB-oil interaction due to nonlinear parameter interactions, while a support vector machine (SVM)-based classification significantly improves accuracy. A hybrid model integrating the spreading coefficient and SVM achieves 100% accuracy across various capillary numbers. Additionally, a simplified correlation model is proposed for practical engineering applications.
