A numerical investigation of data-driven closure techniques for POD-Galerkin reduced order models

Author(s)

Zucatti, Victor ; Wolf, William R.

• Document type: Journal article
• Source: AIAA AVIATION 2022 FORUM; v. N/A, p. 13-pg., 2022-01-01.

Abstract

Data-driven closure techniques are assessed for applications in POD-Galerkin reduced order modeling. Closure is based on optimization with respect to temporal modes of the proper orthogonal decomposition (POD) basis and significantly improve the models' performance for computing unsteady compressible flows. Model reduction is obtained via Galerkin projection of the spatial POD modes on the non-conservative compressible Navier-Stokes equations. Closure is performed by adding linear and nonlinear coefficients to the original ROMs and minimizing the error with respect to the POD temporal modes. The effects of adding energy preserving constraints to this minimization problem is also analysed. In this work, the test problems consist of the canonical low Reynolds compressible flow around a circular cylinder and a moderate Reynolds number subsonic flow over an airfoil. In the latter case, boundary layer instabilities are responsible for tonal noise emission at multiple frequencies due to frequency and amplitude modulations induced by a separation bubble developing on the airfoil suction side. Results show that nonlinear calibration coefficients outperform their linear counterparts for the present simulations. (AU)

• FAPESP's process: 13/08293-7 - CCES - Center for Computational Engineering and Sciences
• Grantee: Munir Salomao Skaf
• Support Opportunities: Research Grants - Research, Innovation and Dissemination Centers - RIDC
• FAPESP's process: 21/06448-0 - High-fidelity numerical simulations applied in unsteady aerodynamics, turbulence and aeroacoustics
• Grantee: William Roberto Wolf
• Support Opportunities: Research Grants - Young Investigators Grants - Phase 2
• FAPESP's process: 18/11410-9 - On the application of principal component analysis for the construction of reduced order models
• Grantee: Victor Zucatti da Silva
• Support Opportunities: Scholarships in Brazil - Master