Intrusive and data-driven reduced order modelling of the rotating thermal shallow water equation
作者:
Highlights:
• Intrusive and data-driven reduced-order solutions are equal up to numerical errors.
• Incorporating the knowledge of parameter dependency, reduced-order solutions are predicted for new parameter values without interpolation.
• The overall system behaviour is captured accurately in the training and prediction phases.
• Preservation of polynomial invariants ensure the long-term stability of the reduced solution.
• Operator inference with re-projection is tailored to learn physics-informed low-order models.
摘要
•Intrusive and data-driven reduced-order solutions are equal up to numerical errors.•Incorporating the knowledge of parameter dependency, reduced-order solutions are predicted for new parameter values without interpolation.•The overall system behaviour is captured accurately in the training and prediction phases.•Preservation of polynomial invariants ensure the long-term stability of the reduced solution.•Operator inference with re-projection is tailored to learn physics-informed low-order models.
论文关键词:Model order reduction,Finite differences,Hamiltonian systems,Fluids,Least-squares
论文评审过程:Received 2 August 2021, Revised 30 November 2021, Accepted 3 January 2022, Available online 10 January 2022, Version of Record 10 January 2022.
论文官网地址:https://doi.org/10.1016/j.amc.2022.126924