Multiscale physics‐informed population balance model and neural surrogates for particle size prediction in jet milling

Abstract Particle size reduction is costly to optimize during early‐stage specialty chemical and pharmaceutical development because both material availability and experimental time are limited. Conventional population balance models (PBMs) for jet milling commonly require repeated calibration across operating conditions and materials. In this work, a Physics‐informed PBM (PI‐PBM) was developed by integrating material dependent breakage properties obtained from single‐particle crushing tests with pressure dependent collision descriptors derived from coupled computational fluid dynamics–discrete element method (CFD–DEM simulations. The PI‐PBM reproduced the experimental particle size distributions (PSDs) with accuracy comparable to that of independently calibrated pressure‐specific models, while reducing calibration time by 20%–47% and potentially reducing the material requirement for model calibration. A neural network surrogate, denoted as population balance neural network (PBNN), was subsequently trained using PI‐PBM generated datasets. The trained PBNN reproduced both unimodal and bimodal product PSDs, predicted product PSDs for an unseen feed distribution, and demonstrated partial transferability to an external material without retraining.

Authors

Institutions

Publication Details

Journal
AIChE Journal
Published
2026-11-16
DOI
https://doi.org/10.1002/aic.70744
Primary Topic
Mineral Processing and Grinding
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Multiscale physics‐informed population balance model and neural surrogates for particle size prediction in jet milling

Sajad Khodadadi, Xizhong Chen, Dingeman L.H. van der Haven, Leqi Lin et al.
AIChE Journal
Mineral Processing and Grinding
article

Multiscale physics‐informed population balance model and neural surrogates for particle size prediction in jet milling

Sajad Khodadadi, Xizhong Chen, Dingeman L.H. van der Haven, Leqi Lin, Wenbo Zhu, Lijun Xue, Yang Liu, Xingyu Zhou
article en

Abstract

Abstract Particle size reduction is costly to optimize during early‐stage specialty chemical and pharmaceutical development because both material availability and experimental time are limited. Conventional population balance models (PBMs) for jet milling commonly require repeated calibration across operating conditions and materials. In this work, a Physics‐informed PBM (PI‐PBM) was developed by integrating material dependent breakage properties obtained from single‐particle crushing tests with pressure dependent collision descriptors derived from coupled computational fluid dynamics–discrete element method (CFD–DEM simulations. The PI‐PBM reproduced the experimental particle size distributions (PSDs) with accuracy comparable to that of independently calibrated pressure‐specific models, while reducing calibration time by 20%–47% and potentially reducing the material requirement for model calibration. A neural network surrogate, denoted as population balance neural network (PBNN), was subsequently trained using PI‐PBM generated datasets. The trained PBNN reproduced both unimodal and bimodal product PSDs, predicted product PSDs for an unseen feed distribution, and demonstrated partial transferability to an external material without retraining.

AIChE Journal
Novartis (China) (CN)
Mineral Processing and Grinding
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon