A predictive hybrid reduced order model based on proper orthogonal decomposition combined with deep learning architectures

作者：

Highlights：

• New hybrid reduced order model (ROM) based on physical principles.

• ROM combining POD modes with deep learning architectures.

• ROM applied to accelerate numerical solvers in fluid dynamics.

• Purely data-driven ROM that can be extended to other fields.

摘要

•New hybrid reduced order model (ROM) based on physical principles.•ROM combining POD modes with deep learning architectures.•ROM applied to accelerate numerical solvers in fluid dynamics.•Purely data-driven ROM that can be extended to other fields.

论文关键词：Reduced order models,Deep learning architectures,POD,Modal decompositions,Neural networks,Fluid dynamics

论文评审过程：Received 16 June 2021, Revised 11 September 2021, Accepted 11 September 2021, Available online 25 September 2021, Version of Record 2 October 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115910