Hybrid ML-DEM: Rectangular Hopper Particle Discharge Prediction

Saman Kazemi

doi:10.1016/j.cherd.2026.04.059

Hybrid ML-DEM: Rectangular Hopper Particle Discharge Prediction

2026-04-28

0

PRE

AI

S

Saman Kazemi

R

Reza Zarghami *

N

Navid Mostoufi

R

Rahmat Sotudeh‐Gharebagh

K

Kashayar Saleh

DOI：10.1016/j.cherd.2026.04.059

Original

Original request for help

Save

Abstract

En 中文

• Hybrid ML–DEM framework for accelerated particle-scale simulation in hoppers • Continuous CNN prediction with periodic DEM correction • Preserving simulation accuracy with significantly reducing cost • Validated against experimental and conventional DEM results • Acceptable performance across hopper angles and particle loadings • Reduced computational costs by 75% and 30% compared with CPU- and GPU-based DEM

Keywords:

Hybrid ML-DEM

Particle-scale simulation

Computational cost reduction

Continuous CNN prediction

Hopper discharge

Hybrid ML-DEM: Rectangular Hopper Particle Discharge Prediction

Abstract

Journal

Organization