Open Access

PROCEEDINGS

Neural Network-Based Bubble Interface Prediction

Junhua Gong¹, Yujie Chen^2,*, Bo Yu², Bin Chen¹

1 State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an, 710049, China
2 School of Mechanical Engineering, Beijing Institute of Petrochemical Technology, Beijing, 102617, China

* Corresponding Author: Yujie Chen. Email:

The International Conference on Computational & Experimental Engineering and Sciences 2024, 29(4), 1-2. https://doi.org/10.32604/icces.2024.012531

Abstract

Traditional interface reconstruction methods often rely on numerical approaches, which can be inefficient when dealing with large bubbles, requiring extensive computational resources. To address this issue, we propose a novel model based on convolutional neural networks aimed at rapidly and accurately predicting the equations governing circular bubbles. This model takes the volume fraction of the main-phase fluid surrounding each computational grid cell as input variables and is capable of precisely forecasting the coordinates and radii of bubbles. To further enhance model performance, we employ the Optuna hyperparameter optimizer to fine-tune the model's parameters. Upon training completion, the optimized model achieves prediction errors of less than 1%, demonstrating exceptional accuracy.

---

Keywords

---