Model-free short-term fluid dynamics estimator with a deep 3D-convolutional neural network

作者：

Highlights：

• Novel application of deep learning techniques to fluid dynamics and complex flows.

• Forecasting of the velocity-field of synthetic jet in transitional regime.

• Using 3D convolutional neural network plus intermediate dimensionality reduction.

• Complementary tool to low-rank approximation techniques.

摘要

•Novel application of deep learning techniques to fluid dynamics and complex flows.•Forecasting of the velocity-field of synthetic jet in transitional regime.•Using 3D convolutional neural network plus intermediate dimensionality reduction.•Complementary tool to low-rank approximation techniques.

论文关键词：Computational fluid dynamics,Prediction,Deep learning,Convolutional neural network

论文评审过程：Received 9 April 2020, Revised 12 January 2021, Accepted 17 March 2021, Available online 20 March 2021, Version of Record 8 April 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114924