New approach of the solution of physical fields of fluid dynamics: Physics-informed long short-term memory network

Zhiwei Li

doi:10.1016/j.ijheatfluidflow.2025.110024

New approach of the solution of physical fields of fluid dynamics: Physics-informed long short-term memory network

2025-08-28

0

PRE

AI

Z

Zhiwei Li

G

Guihua Hu *

DOI：10.1016/j.ijheatfluidflow.2025.110024

Original

Original request for help

Save

Abstract

En 中文

• Combining CFD and deep learning establishes a fast response high fidelity model. • Integrating LSTM with PINN model obtains reliable physical field predictions. • The proposed PI-LSTM outperforms PINN and BI-LSTM in terms of predictive metrics. • The PI-LSTM increases computing efficiency and reduces storage space compared to CFD. • The validity and accuracy of proposed method are proved by three physical fields.

Keywords:

CFD

deep learning

LSTM

PINN

PI-LSTM

New approach of the solution of physical fields of fluid dynamics: Physics-informed long short-term memory network

Abstract

Journal

Organization

Cited Papers

Citing Papers