TY - JOUR
T1 - Deep learning methods for predicting fluid forces in dense particle suspensions
AU - Ashwin, Neil Raj
AU - Cao, Ze
AU - Muralidhar, Nikhil
AU - Tafti, Danesh
AU - Karpatne, Anuj
N1 - Publisher Copyright:
© 2022 Elsevier B.V.
PY - 2022/3
Y1 - 2022/3
N2 - Two deep learning methods, Multi-Layer Perceptron (MLP) network and Convolution Neural Network (CNN) are evaluated to predict drag forces in dense suspensions of ellipsoidal particles using data from Particle Resolved Simulations (PRS). The MLP is trained on the mean flow Reynolds number, solid fraction of the suspension, the aspect ratio of the particle, and orientation to flow direction. The CNN is given an additional 3D spatial map of the particle of interest and its immediate neighborhood via a distance function. The prediction capability of the trained networks is tested at different levels of complexity: on an unseen particle arrangement (Level 1), to all arrangements of an unseen numerical experiment (Level 2), and finally to all experiments of an unseen Reynolds number, solid fraction or aspect ratio (Level 3). The CNN is shown to perform better than the MLP for all testing levels except when testing on an unseen aspect ratio.
AB - Two deep learning methods, Multi-Layer Perceptron (MLP) network and Convolution Neural Network (CNN) are evaluated to predict drag forces in dense suspensions of ellipsoidal particles using data from Particle Resolved Simulations (PRS). The MLP is trained on the mean flow Reynolds number, solid fraction of the suspension, the aspect ratio of the particle, and orientation to flow direction. The CNN is given an additional 3D spatial map of the particle of interest and its immediate neighborhood via a distance function. The prediction capability of the trained networks is tested at different levels of complexity: on an unseen particle arrangement (Level 1), to all arrangements of an unseen numerical experiment (Level 2), and finally to all experiments of an unseen Reynolds number, solid fraction or aspect ratio (Level 3). The CNN is shown to perform better than the MLP for all testing levels except when testing on an unseen aspect ratio.
KW - Convolution neural network (CNN)
KW - Drag force
KW - Ellipsoidal particle suspension
KW - Machine learning
KW - Multi-layer perceptron (MLP)
UR - http://www.scopus.com/inward/record.url?scp=85127495632&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85127495632&partnerID=8YFLogxK
U2 - 10.1016/j.powtec.2022.117303
DO - 10.1016/j.powtec.2022.117303
M3 - Article
AN - SCOPUS:85127495632
SN - 0032-5910
VL - 401
JO - Powder Technology
JF - Powder Technology
M1 - 117303
ER -