Machine-Learning Capillary Force Model

<1.5% error

A Catboost-trained capillary force model predicts inter-particle and particle-wall liquid bridge forces with lower errors than existing simulation models.

Hanwen, Cui · Zhao, Honghua · 70991989 · Xingli, Zhang · Waleed, Awadalseed · Hongxiang, Tang

Construction and Building Materials 2024

Specifications

球相对误差低于

{'zh': '1.5', 'en': ''} %

壁相对误差低于

{'zh': '2', 'en': ''} %

Advantages

Sphere-sphere database contains 163,440,000 data points

Sphere-wall database contains 2043,000 data points