betaFeedback

WorksEngineering

Efficient Adjoint Petrov-Galerkin Reduced Order Models for fluid flows governed by the incompressible Navier-Stokes equations

Abstract

This research paper investigates the Adjoint Petrov-Galerkin (APG) method for reduced order models (ROM) and fluid dynamics governed by the incompressible Navier-Stokes equations. The Adjoint Petrov-Galerkin ROM, derived using the Mori-Zwanzig formalism, demonstrates superior accuracy and stability compared to standard Galerkin ROMs. However, challenges arise due to the time invariance of the test basis vectors, resulting in high computational requirements. To address this, we introduce a new efficient Adjoint Petrov-Galerkin (eAPG) ROM formulation, extending its application to the incompressible Navier-Stokes equations by exploiting the polynomial structure inherent in these equations. The offline and online phases partition eliminates the need for repeated test basis vector evaluations. This improves computational efficiency in comparison to the general Adjoint Petrov-Galerkin ROM formulation. A novel approach to augmenting the memory length, a critical factor influencing the stability and accuracy of the APG-ROM, is introduced, employing a data-driven optimization. Numerical results for the 3D turbulent flow around a circular cylinder demonstrate the efficacy of the proposed approach. Error measures and computational cost evaluations, considering metrics such as floating point operations and simulation time, provide a comprehensive analysis.

Introduction

In modern engineering, the important role played by high-fidelity numerical simulations, and especially those addressing fluid flow dynamics governed by the Navier-Stokes equations, cannot be overrated. However, the excessive computational costs associated with full order models (FOM) present a significant hurdle for many applications. Recognizing the impracticality of extensive computational requirements, there has been a notable shift towards the development and use of reduced order models (ROM). The primary objective lies in balancing computational efficiency with accuracy, a trade-off that is central to the implementation of ROMs. An integral aspect of reduced order modeling involves defining reduced spaces in which computational cost and accuracy can be systematically controlled. Among the numerous techniques used for reduced order modeling, projection-based ROMs using the Proper Orthogonal Decomposition (POD) and Galerkin Projection (GP) are popular, particularly in the realm of fluid dynamics, although not exclusively limited to it.

POD is instrumental in constructing a set of orthonormal basis vectors from information extracted from high-fidelity simulations. These basis vectors form the foundation for projection-based ROMs. The order reduction is achieved through the truncation of a substantial number of seemingly insignificant basis vectors, a process designed to retain only a select few basis vectors. We call the basis vectors that are and are not retained the resolved coarse-scale and unresolved fine-scale basis vectors, respectively. The truncation ensures that the subsequent Galerkin Projection yields a Galerkin reduced order model (G-ROM) with significantly fewer degrees of freedom than the FOM. Certainly, the GP stands out as one of the most straightforward approaches in reduced order modeling. Notably, it simplifies the modeling process by employing the same set of basis vectors for both approximation and projection. This significantly mitigates the computational complexity associated with the numerical simulation, making it more applicable to practical applications that are subject to real-world constraints such as real-time processing or limited computational resources. Examples include contaminant airflow [1], magnetic resonance imaging [2], the flow around circular and oscillating cylinders and grooved channels [3, 4, 5], the flow inside of centrifugal and positive replacement pumps [6, 7], diffusion and heat conduction problems in drying processes [8, 9], and thermal flows in integrated circuits [10], to name only a few.

However, the apparent simplicity of G-ROMs can come at a cost. Such models often struggle with stability and accuracy issues, particularly when applied to highly turbulent flows (see, e.g., [11, 12]). The lack of stability and accuracy in simple G-ROMs can be attributed to a key factor - the drastic truncation of basis vectors (see, e.g., [13, 14, 15]). This truncation gives rise to the ROM closure problem that can be interpreted as a model error, related to the challenges encountered in simulations of turbulent flows where closure modeling is required, as seen in Large Eddy Simulations (see, e.g., [16]). The analogy with turbulent flows is evident, as both scenarios involve operating in a coarse, under-resolved framework where closure modeling becomes crucial. In the case of ROMs, the indirect connections arise from the nonlinearity of dynamic systems, where the truncated modes in fact interact with the retained modes. A drastic modal truncation for computational efficiency unintentionally suppresses these interactions (see, e.g., [11, 12]).

To address the closure problem in ROMs, various approaches have been proposed, including mixing-length ROM closure (see, e.g., [17, 14]), Smagorinsky ROM closure (see, e.g., [18]), spatial filtering (see, e.g., [19, 20, 21, 22]), stochastic closure (see, e.g., [23, 24]), data-driven least-squares regression (see, e.g., [25, 26, 27, 28]) and non-intrusive neural network regression (see, e.g., [29, 30, 31, 32]), and multiscale approaches (see, e.g., [33, 34, 35, 36, 37]). Each method offers a unique perspective and set of techniques to mitigate the challenges posed by the closure problem, aiming to strike a balance between computational efficiency and the accurate representation of complex system dynamics.

A particularly noteworthy approach is the Petrov-Galerkin Projection which involves the application of two distinct sets of basis vectors. Although the literature on Galerkin-based ROMs is extensive, comparatively fewer ROMs have been developed using Petrov-Galerkin projection, including the least squares Petrov-Galerkin ROM [38, 39], streamline upwind Petrov-Galerkin ROM [40, 15, 41], Petrov-Galerkin ROM with nonlinear optimization based on manifolds [42], and the Adjoint Petrov-Galerkin ROM [43].

Parish et al. [43] deals with the derivation of a Petrov-Galerkin Projection approach using the Mori-Zwanzig formalism. The authors construct a ROM exclusively with the resolved coarse-scale basis vectors and introduce a memory term to model the influence of truncated unresolved fine scales. This memory term depends on the temporal history of the resolved coarse-scale variables. Parish et al. establish a ROM known as the Adjoint Petrov-Galerkin ROM (APG-ROM), which is based on a new set of test basis vectors. It has been demonstrated that the accuracy and stability of the APG-ROM considerably surpass those achieved with a standard G-ROM [43]. The main drawback of the APG-ROM is the time invariance of the introduced test basis vectors, which implies that frequent additional calculations in the spatial domain must be carried out. The dimension of the ROM is several orders of magnitude smaller than that of the FOM. Consequently, while the ROM is designed to reduce computational effort, the need for frequent calculations in the spatial domain undermines this benefit, particularly when dealing with a high number of discretized spatial locations.

This paper aims to improve the APG-ROM formulation and extends its applicability to the incompressible Navier-Stokes equations. This is achieved by partitioning the computations into distinct offline and online phases. This division is a common practice for standard G-ROMs of fluid flows. By leveraging the polynomial nature of the Navier-Stokes equations, the computationally demanding operations of the APG-ROM in the spatial dimension can be executed during the offline phase. This eliminates the need for the repeated evaluation of test basis vectors in each time step of the numerical integration in the online phase. We call our approach efficient APG-ROM (eAPG-ROM). While larger than the G-ROM, the eAPG-ROM maintains computational efficiency superior to the general APG-ROM formulation. Additionally, we introduce a novel, straightforward, yet efficient approach aimed at augmenting the number of degrees of freedom in the evaluation of the memory length, a critical factor influencing the stability and accuracy of the APG-ROM. This will involve a data-driven methodology to determine this augmented memory length.

We will present numerical results for the flow around a circular cylinder. The simulation will be conducted in 3D while considering turbulent flow. We will derive error measures for a comparative analysis of the numerical results. Furthermore, an evaluation of numerical computational costs will be conducted, considering metrics such as floating point operations (flops) and simulation time. This comprehensive analysis aims to provide an understanding of the trade-offs between computational efficiency and accuracy achieved by the eAPG-ROM.

The structure of the paper is as follows. Section 2 outlines the standard G-ROM framework applied to fluid flows governed by the incompressible Navier-Stokes equations. Section 3 describes the general APG-ROM, followed by the introduction of the eAPG-ROM framework in Section 4. Additionally, we introduce the concept of an augmented memory length along with an optimization algorithm. Section 5 introduces error measures used to evaluate the quality of the ROMs. Section 6 gives details on the implementation and provides a comparison of computational costs. Section 7 presents the numerical experiment, showcasing the application and performance of the eAPG-ROM. Section 8 summarizes the results with a brief conclusion and outlook.

Nomenclature

SymbolDefinition
CG,N∈RNC^{G,N} \in\mathbb{R}^{N}constant coefficients of the G-ROM in RN\mathbb{R}^{N}
CG∈RrC^{G} \in\mathbb{R}^{r}constant coefficients of the G-ROM
CΠ∈RNC^{\Pi} \in\mathbb{R}^{N}orthonormally projected constant coefficients of the G-ROM
CAPG,N∈RNC^{\mathrm{APG},N} \in\mathbb{R}^{N}constant coefficients of the efficient APG-ROM in RN\mathbb{R}^{N}
CAPG∈RrC^{\mathrm{APG}} \in\mathbb{R}^{r}constant coefficients of the efficient APG-ROM
EREC∈R\mathcal{E}_{\mathrm{REC}} \in\mathbb{R}reconstruction error
EROM∈R\mathcal{E}_{\mathrm{ROM}} \in\mathbb{R}error introduced by the reduced order model
ETRU∈R\mathcal{E}_{\mathrm{TRU}} \in\mathbb{R}truncation error
ETotal∈R\mathcal{E}_{\mathrm{Total}} \in\mathbb{R}total error
flopsAPG∈R\mathit{flops}_{\mathrm{APG}} \in\mathbb{R}flop count for the General APG-ROM
flopsG,Offline∈R\mathit{flops}_{G,\mathrm{Offline}} \in\mathbb{R}flop count for the G-ROM Offline-Phase
flopsG,Online∈R\mathit{flops}_{G,\mathrm{Online}} \in\mathbb{R}flop count for the G-ROM Online-Phase
flopseAPG,Offline∈R\mathit{flops}_{e\mathrm{APG},\mathrm{Offline}} \in\mathbb{R}flop count for the efficient APG-ROM Offline-Phase
flopseAPG,Online∈R\mathit{flops}_{e\mathrm{APG},\mathrm{Online}} \in\mathbb{R}flop count for the efficient APG-ROM Online-Phase
J(⋅)∈RJ(\cdot) \in\mathbb{R}jacobian
KAPG,N∈RN×r3K^{\mathrm{APG},N} \in\mathbb{R}^{N \times r^{3}}cubic coefficients of the efficient APG-ROM in RN\mathbb{R}^{N}
KAPG∈Rr×r3K^{\mathrm{APG}} \in\mathbb{R}^{r \times r^{3}}cubic coefficients of the efficient APG-ROM
LG,N∈RN×rL^{G,N} \in\mathbb{R}^{N \times r}linear coefficients of the G-ROM in RN\mathbb{R}^{N}
LG∈Rr×rL^{G} \in\mathbb{R}^{r \times r}linear coefficients of the G-ROM
LΠ∈RN×rL^{\Pi} \in\mathbb{R}^{N \times r}orthonormally projected linear coefficients of the G-ROM
LAPG,N∈RN×rL^{\mathrm{APG},N} \in\mathbb{R}^{N \times r}linear coefficients of the efficient APG-ROM in RN\mathbb{R}^{N}
LAPG∈Rr×rL^{\mathrm{APG}} \in\mathbb{R}^{r \times r}linear coefficients of the efficient APG-ROM
L\mathcal{L}Liouville operator
M∈RM \in\mathbb{R}number of snapshots
M\mathcal{M}memory term
N=dNgrid∈RN = dN_{\mathrm{grid}} \in\mathbb{R}
Ngrid∈RN_{\mathrm{grid}} \in\mathbb{R}number of spatial locations on the grid
QG,N∈RN×r2Q^{G,N} \in\mathbb{R}^{N \times r^{2}}quadratic coefficients of the G-ROM in RN\mathbb{R}^{N}
QG∈Rr×r2Q^{G} \in\mathbb{R}^{r \times r^{2}}quadratic coefficients of the G-ROM
QΠ∈RN×r2Q^{\Pi} \in\mathbb{R}^{N \times r^{2}}orthonormally projected quadratic coefficients of the G-ROM
QAPG,N∈RN×r2Q^{\mathrm{APG},N} \in\mathbb{R}^{N \times r^{2}}quadratic coefficients of the efficient APG-ROM in RN\mathbb{R}^{N}
QAPG∈Rr×r2Q^{\mathrm{APG}} \in\mathbb{R}^{r \times r^{2}}quadratic coefficients of the efficient APG-ROM
R(⋅)∈RNR(\cdot) \in\mathbb{R}^{N}right-hand side operator
ReRereynolds number
Tp∈RT_{p} \in\mathbb{R}time of the simulation for one flow period
U∗∈RN×MU^{*} \in\mathbb{R}^{N \times M}snapshot matrix of the time-variant velocity contributions
V∈RM×MV \in\mathbb{R}^{M \times M}right-hand side eigenvector matrix of the POD
W∈Rr×rW \in\mathbb{R}^{r \times r}matrix memory length weighting matrix
a∈RNa \in\mathbb{R}^{N}generalized coordinates
a~∈Rr\tilde{a} \in\mathbb{R}^{r}modal coefficients for the coarse-scale
a~POD∈Rr\tilde{a}^{\mathrm{POD}} \in\mathbb{R}^{r}POD modal coefficients for the coarse-scale
a~G∈Rr\tilde{a}^{G} \in\mathbb{R}^{r}G-ROM modal coefficients for the coarse-scale
a~APG∈Rr\tilde{a}^{\mathrm{APG}} \in\mathbb{R}^{r}general APG-ROM modal coefficients for the coarse-scale
a~eAPG∈Rr\tilde{a}^{e\mathrm{APG}} \in\mathbb{R}^{r}eAPG-ROM modal coefficients for the coarse-scale
a~eAPG,τ∈Rr\tilde{a}^{e\mathrm{APG},\tau} \in\mathbb{R}^{r}eAPG-ROM modal coefficients for the coarse-scale with arbitrarily chosen scalar memory length
a~eAPG,τ,opt∈Rr\tilde{a}^{e\mathrm{APG},\tau,\mathrm{opt}} \in\mathbb{R}^{r}eAPG-ROM modal coefficients for the coarse-scale with optimized scalar memory length
a~eAPG,T,opt∈Rr\tilde{a}^{e\mathrm{APG},T,\mathrm{opt}} \in\mathbb{R}^{r}eAPG-ROM modal coefficients for the coarse-scale with optimized matrix memory length
d∈Rd \in\mathbb{R}number of dimensions of the spatial domain
dx, dy, dzdx,\,dy,\,dzcell dimensions in x, y and z directions
p∈RNgridp \in\mathbb{R}^{N_{\mathrm{grid}}}pressure
r∈Rr \in\mathbb{R}number of reduced/coarse-scale base vectors
Δt∈R\Delta t \in\mathbb{R}time step for numerical integration
tm∈Rt_{m} \in\mathbb{R}discrete time
u∈RNu \in\mathbb{R}^{N}velocity
u′∈RNu' \in\mathbb{R}^{N}time-averaged velocity
u∗∈RNu^{*} \in\mathbb{R}^{N}time-variant velocity
u~∈RN\tilde{u} \in\mathbb{R}^{N}velocity recovered with the coarse-scale base vectors
w∈Rw \in\mathbb{R}memory length weight
xn∈Rx_{n} \in\mathbb{R}discrete spatial location
Greek letters
Δ, Δ\Delta,\,\Deltaspatial second differential operator
Πˉ\bar{\Pi}resolved coarse-scale projection operator
Πˉ\bar{\Pi}unresolved fine-scale projection operator
Σ∈RM×M, σ∈R\Sigma\in\mathbb{R}^{M \times M},\,\sigma\in\mathbb{R}singular values
T∈Rr×rT \in\mathbb{R}^{r \times r}matrix memory length
Topt∈Rr×rT^{\mathrm{opt}} \in\mathbb{R}^{r \times r}optimized matrix memory length
Φ∈RN×M, ϕ∈RN, φ∈Rd\Phi\in\mathbb{R}^{N \times M},\,\phi\in\mathbb{R}^{N},\,\varphi\in\mathbb{R}^{d}trial base vectors, POD modes
Φ~∈RN×r, ϕ~∈RN, φ~∈Rd\tilde{\Phi} \in\mathbb{R}^{N \times r},\,\tilde{\phi} \in\mathbb{R}^{N},\,\tilde{\varphi} \in\mathbb{R}^{d}coarse-scale trial base vectors, coarse-scale POD modes
Φˉ∈RN×r, ϕˉ∈RN, φˉ∈Rd\bar{\Phi} \in\mathbb{R}^{N \times r},\,\bar{\phi} \in\mathbb{R}^{N},\,\bar{\varphi} \in\mathbb{R}^{d}fine-scale trial base vectors, fine-scale POD modes
Ψ∈RN×M, ψ∈RN, ψ∈Rd\Psi\in\mathbb{R}^{N \times M},\,\psi\in\mathbb{R}^{N},\,\psi\in\mathbb{R}^{d}test base vectors
ΨAPG∈RN×r\Psi^{\mathrm{APG}} \in\mathbb{R}^{N \times r}adjoint Petrov-Galerkin test base vectors
δik\delta_{ik}Kronecker delta
∇, ∇\nabla,\,\nablaspatial first differential operator

ν∈R\nu\in\mathbb{R} kinematic viscosity ρ(⋅)∈R\rho(\cdot) \in\mathbb{R} spectral radius τ∈R\tau\in\mathbb{R} memory length τopt∈R\tau^{\mathrm{opt}} \in\mathbb{R} optimized memory length φ(⋅)\varphi(\cdot) resolved, coarse modal coefficient operator ω1∈R,ω2∈R\omega_1 \in\mathbb{R}, \omega_2 \in\mathbb{R} Flop count for spatial differential operations

Abbreviations

APG-ROM adjoint Petrov-Galerkin reduced order model eAPG-ROM efficient adjoint Petrov-Galerkin reduced order model Flops floating point operations FOM full order model G-ROM Galerkin reduced order model GP Galerkin projection POD proper orthogonal decomposition ROM reduced order model SST shear-stress transport SVD singular value decomposition URANS unsteady Reynolds-Averaged Navier-Stokes

Notation

Matrices are denoted by bold uppercase letters, vectors by bold lowercase letters, and scalars by lowercase letters. Functions represented by bold letters followed by parentheses indicate matrix or vector functions. Bold spatial differential operators Δ\Delta and ∇\nabla apply the Δ\Delta and ∇\nabla operators to every dd-dimensional Euclidean space vector. Consider, for example, a column vector g=(g(x1)T ⋯ g(xNgrid)T)T∈RN\mathbf{g} = \left(\mathbf{g}(x_1)^T \ \cdots\ \mathbf{g}(x_{N_{\mathrm{grid}}})^T\right)^T \in\mathbb{R}^N, with g(xn)∈Rd\mathbf{g}(x_n) \in\mathbb{R}^d, n=1,…,Ngridn = 1,\ldots,N_{\mathrm{grid}}, where NgridN_{\mathrm{grid}} is the total number of spatial points on a discrete spatial grid, d=3d = 3 throughout the paper and N=dNgridN = dN_{\mathrm{grid}}. Then ∇g=(∇g(x1)T ⋯ ∇g(xNgrid)T)T\nabla\mathbf{g} = \left(\nabla\mathbf{g}(x_1)^T \ \cdots\ \nabla\mathbf{g}(x_{N_{\mathrm{grid}}})^T\right)^T and Δg=(Δg(x1)T ⋯ Δg(xNgrid)T)T\Delta\mathbf{g} = \left(\Delta\mathbf{g}(x_1)^T \ \cdots\ \Delta\mathbf{g}(x_{N_{\mathrm{grid}}})^T\right)^T.

POD-Galerkin Reduced Order Model for fluid flows

Our objective is to implement a projection-based reduced order model for fluid flows governed by the incompressible Navier-Stokes equations,

∂∂tu(t)=−(u(t)⋅∇)u(t)+νΔu(t)−∇p(t),∇⋅u(t)=0,u(0)=u0,p(0)=p0,(1)\begin{aligned} \frac{\partial}{\partial t}\mathbf{u}(t) &= -(\mathbf{u}(t)\cdot\nabla)\mathbf{u}(t) + \nu\Delta\mathbf{u}(t) - \nabla p(t),\\ \nabla\cdot\mathbf{u}(t) &= 0,\\ \mathbf{u}(0) &= \mathbf{u}_0,\quad p(0) = p_0, \tag*{(1)} \end{aligned}

References

  1. [1]T. John, M. Guay, N. Hariharan, and S. Naranayan, “POD-based observer for estimation in Navier-Stokes flow,” Computers and Chemical Engineering, vol. 34, no. 6, pp. 965 – 975, 2010.DOI
  2. [2]M. Seoane, P. D. Ledger, A. J. Gil, S. Zlotnik, and M. Mallett, “A combined reduced order-full order methodology for the solution of 3D magneto-mechanical problems with application to magnetic resonance imaging scanners,” International Journal for Numerical Methods in Engineering, vol. 121, no. 16, pp. 3529–3559, 2020.DOI
  3. [3]M. Bergmann and L. Cordier, “Optimal control of the cylinder wake in the laminar regime by trust-region methods and POD reduced-order models,” Journal of Computational Physics, vol. 8, pp. 7813–7840, 2008.
  4. [4]A. Deane, I. G. Kevrekidis, G. Karniadakis, and S. A. Orszag, “Low-dimensional models for complex geometry flows: Application to grooved channels and circular cylinders,” Physics of Fluids, vol. 3, pp. 2337–2354, 1991.DOI
  5. [5]E. Liberge and A. Hamdouni, “Reduced order modelling method via proper orthogonal decomposition (POD) for flow around an oscillating cylinder,” Journal of Fluids and Structures, vol. 26, no. 2, pp. 292 – 311, 2010.DOI
  6. [6]T. Gunder, A. Sehlinger, R. Skoda, and M. Mönnigmann, “Sensor placement for reduced-order model-based observers in hydraulic fluid machinery,” IFAC-PapersOnLine, vol. 51, no. 13, pp. 414–419, 2018.DOI
  7. [7]K. D. Sommer, L. Reineking, Y. P. Ravichandran, R. Skoda, and M. Mönnigmann, “Estimating flow fields with reduced order models,” Heliyon, vol. 9, no. 11, p. e20930, 2023.DOI
  8. [8]M. O. Berner, F. Sudbrock, V. Scherer, and M. Mönnigmann, “POD and Galerkin-based reduction of a wood chip drying model,” IFAC-PapersOnLine, vol. 50, no. 1, pp. 6619–6623, 2017.DOI
  9. [9]M. O. Berner, V. Scherer, and M. Mönnigmann, “An observer for partially obstructed wood particles in industrial drying processes,” Computers & Chemical Engineering, vol. 141, p. 107013, 2020.arxiv.org/abs/1911.04881
  10. [10]D. S. Meyer, B. T. Helenbrook, and M.-C. Cheng, “Proper Orthogonal Decomposition-based reduced basis element thermal modeling of integrated circuits,” International Journal for Numerical Methods in Engineering, vol. 112, no. 5, pp. 479–500, 2017.DOI
  11. [11]I. Akhtar, Z. Wang, J. Borggaard, and T. Iliescu, “A new closure strategy for proper orthogonal decomposition reduced-order models,” Journal of Computational and Nonlinear Dynamics, vol. 7, no. 3, p. 034503, 2012.DOI
  12. [12]J. Baiges, R. Codina, and S. Idelsohn, “Reduced-order subscales for POD models,” Computer Methods in Applied Mechanics and Engineering, vol. 291, pp. 173–196, 2015.DOI
  13. [13]S. E. Ahmed, S. Pawar, O. San, A. Rasheed, T. Iliescu, and B. R. Noack, “On closures for reduced order models—a spectrum of first-principle to machine-learned avenues,” Physics of Fluids, vol. 33, no. 9, p. 091301, 2021.DOI
  14. [14]M. Couplet, P. Sagaut, and C. Basdevant, “Intermodal energy transfers in a proper orthogonal decomposition–Galerkin representation of a turbulent separated flow,” Journal of Fluid Mechanics, vol. 491, p. 275–284, 2003.DOI
  15. [15]M. Bergmann, C.-H. Bruneau, and A. Iollo, “Enablers for robust POD models,” Journal of Computational Physics, vol. 228, no. 2, pp. 516–538, 2009.DOI
  16. [16]P. Sagaut, Large Eddy Simulation for Incompressible Flows. Berlin: Scientific Computation, Springer-Verlag, 3 ed., 2006.
  17. [17]N. Aubry, P. Holmes, J. L. Lumley, and E. Stone, “The dynamics of coherent structures in the wall region of a turbulent boundary layer,” Journal of Fluid Mechanics, vol. 192, p. 115–173, 1988.DOI
  18. [18]S. Ullmann and J. Lang, “A POD-Galerkin reduced model with updated coefficients for Smagorinsky LES,” in V European conference on computational fluid dynamics, ECCOMAS CFD, p. 2010, 2010.
  19. [19]X. Xie, D. Wells, Z. Wang, and T. Iliescu, “Approximate deconvolution reduced order modeling,” Computer Methods in Applied Mechanics and Engineering, vol. 313, pp. 512–534, 2017.arxiv.org/abs/1510.02726
  20. [20]X. Xie, M. Mohebujjaman, L. G. Rebholz, and T. Iliescu, “Data-driven filtered reduced order modeling of fluid flows,” SIAM Journal on Scientific Computing, vol. 40, no. 3, pp. B834–B857, 2018.arxiv.org/abs/1709.04362
  21. [21]D. Wells, Z. Wang, X. Xie, and T. Iliescu, “An evolve-then-filter regularized reduced order model for convection-dominated flows,” International Journal for Numerical Methods in Fluids, vol. 84, no. 10, pp. 598–615, 2017.DOI
  22. [22]F. Sabetghadam and A. Jafarpour, “a Regularization of the POD-Galerkin dynamical systems of the Kuramoto–Sivashinsky equation,” Applied Mathematics and Computation, vol. 218, no. 10, pp. 6012–6026, 2012.DOI
  23. [23]T. P. Sapsis and P. F. Lermusiaux, “Dynamically orthogonal field equations for continuous stochastic dynamical systems,” Physica D: Nonlinear Phenomena, vol. 238, no. 23, pp. 2347–2360, 2009.DOI
  24. [24]V. Resseguier, E. Mémin, and B. Chapron, “Stochastic fluid dynamic model and dimensional reduction,” in Ninth International Symposium on Turbulence and Shear Flow Phenomena, pp. 649–654, 2015.DOI
  25. [25]M. Couplet, C. Basdevant, and P. Sagaut, “Calibrated reduced-order POD-Galerkin system for fluid flow modelling,” Journal of Computational Physics, vol. 207, pp. 192–220, 2005.DOI
  26. [26]M. Buffoni, S. Camarri, A. Iollo, and M. V. Salvetti, “Low-dimensional modelling of a confined three-dimensional wake flow,” Journal of Fluid Mechanics, vol. 569, p. 141–150, 2006.
  27. [27]E. C. Laurent Perret and J. Delville, “Polynomial identification of POD based low-order dynamical system,” Journal of Turbulence, vol. 7, p. N17, 2006.DOI
  28. [28]V. Zuccatti and W. Wolf, “Data-driven closure of projection-based reduced order models for unsteady compressible flows,” Computer Methods in Applied Mechanics and Engineering, vol. 386, p. 114120, 2021.DOI
  29. [29]O. San and T. Iliescu, “A stabilized proper orthogonal decomposition reduced-order model for large scale quasi-geostrophic ocean circulation,” Advances in Computational Mathematics, vol. 41, pp. 1289–1319, 2015.
  30. [30]S. Pawar, S. M. Rahman, H. Vaddir­eddy, O. San, A. Rasheed, and P. Vedula, “A deep learning enabler for nonintrusive reduced order modeling of fluid flows,” Physics of Fluids, vol. 31, no. 8, p. 085101, 2019.DOI
  31. [31]R. Jendersie, A. Mjalled, X. Lu, L. Reineking, A. Kharaghani, M. Mönnigmann, and C. Lessig, “NeuroPNM: Model Reduction of Pore Network Models using Neural Networks,” Particuology, vol. 86, pp. 239–251, 2023.DOI
  32. [32]A. Mjalled, E. Torres, and M. Mönnigmann, “Reduced-order modeling framework using two-level neural networks,” PAMM, vol. 23, no. 2, p. e202300061, 2023.DOI
  33. [33]J. Borggaard, A. Duggleby, A. Hay, T. Iliescu, and Z. Wang, “Reduced-order modeling of turbulent flows,” in Proceedings of MTNS, vol. 2008, 2008.
  34. [34]Z. Wang, I. Akhtar, J. Borggaard, and T. Iliescu, “Proper orthogonal decomposition closure models for turbulent flows: A numerical comparison,” Computer Methods in Applied Mechanics and Engineering, vol. 237-240, pp. 10–26, 2012.DOI
  35. [35]G. Stabile, F. Ballarin, G. Zuccarino, and G. Rozza, “A reduced order variational multiscale approach for turbulent flows,” Advances in Computational Mathematics, vol. 45, pp. 2349–2368, 2019.DOI
  36. [36]R. Reyes and R. Codina, “Projection-based reduced order models for flow problems: A variational multiscale approach,” Computer Methods in Applied Mechanics and Engineering, vol. 363, p. 112844, 2020.DOI
  37. [37]C. Mou, B. Koc, O. San, L. G. Rehbolz, and T. Iliescu, “Data-driven variational multiscale reduced order models,” Computer Methods in Applied Mechanics and Engineering, vol. 373, p. 113470, 2021.DOI
  38. [38]K. Carlberg, C. Bou-Mosleh, and C. Farhat, “Efficient non-linear model reduction via a least-squares Petrov–Galerkin projection and compressive tensor approximations,” International Journal for Numerical Methods in Engineering, vol. 86, no. 2, pp. 155–181, 2011.DOI
  39. [39]K. Carlberg, M. Barone, and H. Antil, “Galerkin v. least-squares Petrov–Galerkin projection in nonlinear model reduction,” Journal of Computational Physics, vol. 330, pp. 693–734, 2017.arxiv.org/abs/1504.03749
  40. [40]B. Kragel, Streamline Diffusion POD Models in Optimization. PhD thesis, Universität Trier, 2005.DOI
  41. [41]S. Giere, T. Iliescu, V. John, and D. Wells, “Supg reduced order models for convection-dominated convection–diffusion–reaction equations,” Computer Methods in Applied Mechanics and Engineering, vol. 289, pp. 454–474, 2015.
  42. [42]L. Reineking, K. Sommer, Y. P. Ravichandran, R. Skoda, and M. Mönnigmann, “Long-term stable reduced models for hydraulic systems governed by reynolds averaged navier-stokes equations,” IFAC-PapersOnLine, vol. 55, no. 7, pp. 254–259, 2022.DOI
  43. [43]E. J. Parish, C. R. Wentland, and K. Duraisamy, “The Adjoint Petrov-Galerkin method for non-linear model reduction,” Computer Methods in Applied Mechanics and Engineering, vol. 365, p. 112991, 2020.arxiv.org/abs/1810.03455
  44. [44]T. J. R. Hughes, The Finite Element Method: Linear Static and Dynamic Finite Element Analysis. Dover Publications, 2000.
  45. [45]L. Sirovich, “Turbulence and the dynamics of coherent structures. Part I: Coherent structures,” Quarterly of applied Mathematics, vol. 45, no. 3, pp. 561–571, 1987.DOI
  46. [46]Z. Bai, J. Demmel, J. Dongarra, A. Ruhe, and H. van der Vorst, Templates for the solutions of algebraic eigenvalue problems a practical guide. Philadelphia: Society for Industrial and Applied Mathematics, 2000.
  47. [47]A. J. Chorin and O. H. Hald, Stochastic Tools in Mathematics and Science, vol. 2. Springer, 2009.
  48. [48]R. Zwanzig, “Nonlinear generalized Langevin equations,” Journal of Statistical Physics, vol. 9, no. 3, pp. 215–220, 1973.DOI
  49. [49]H. Mori, “Transport, Collective Motion, and Brownian Motion*),” Progress of Theoretical Physics, vol. 33, no. 3, pp. 423–455, 1965.DOI
  50. [50]E. J. Parish and K. Duraisamy, “Non-Markovian closure models for large eddy simulations using the Mori-Zwanzig formalism,” Phys. Rev. Fluids, vol. 2, p. 014604, 2017.DOI
  51. [51]G. H. Golub and C. F. van Loan, Matrix Computations., ch. 2.4. The Johns Hopkins University Press, fourth ed., 2013.
  52. [52]J. Dormand and P. Prince, “A family of embedded Runge-Kutta formulae,” Journal of Computational and Applied Mathematics, vol. 6, no. 1, pp. 19–26, 1980.DOI
  53. [53]C. Addison, J. Allwright, N. Binsted, N. Bishop, B. Carpenter, P. Dalloz, D. Gee, V. Getov, T. Hey, R. Hockney, M. Lemke, J. Merlin, M. Pinches, C. Scott, and I. Wolton, “The Genesis distributed-memory benchmarks. Part 1: Methodology and general relativity benchmark with results for the SUPRENUM computer,” Concurrency: Practice and Experience, vol. 5, no. 1, pp. 1–22, 1993.DOI
  54. [54]C. W. Ueberhuber, Numerical Computation 1. Heidelberg: Springer-Verlag Berlin, 1997.DOI
  55. [55]F. R. Menter, “Two-equation eddy-viscosity turbulence models for engineering applications,” AIAA Journal, vol. 32, no. 8, pp. 1598–1605, 1994.DOI
  56. [56]N. Casimir, Z. Xiangyuan, G. Ludwig, and R. Skoda, “Assessment of statistical eddy-viscosity turbulence models for unsteady flow at part and overload operation of centrifugal pumps,” in Proceedings of 13th European Conference on Turbomachinery Fluid Dynamics and Thermodynamics (ETC’13), Lausanne, Switzerland, p. 13, 2019.DOI
  57. [57]N. Casimir, X. Zhu, M. Hundshagen, G. Ludwig, and R. Skoda, “Numerical study of rotor-stator interaction of a centrifugal pump at part load with special emphasis on unsteady blade load,” Journal of Fluids Engineering, vol. 142, no. 8, p. 30, 2020.DOI
  58. [58]H. G. Weller, G. Tabor, H. Jasak, and C. Fureby, “A tensorial approach to computational continuum mechanics using object-oriented techniques,” Computers in Physics, vol. 12, no. 6, pp. 620–631, 1998.
  59. [59]S. V. Patankar and D. B. Spalding, “A calculation procedure for heat, mass and momentum transfer in three-dimensional parabolic flows,” International Journal of Heat and Mass Transfer, vol. 15, no. 3, pp. 1787–1806, 1972.
  60. [60]R. Issa, “Solution of the implicitly discretised fluid flow equations by operator-splitting,” Journal of Computational Physics, vol. 62, no. 1, pp. 40–65, 1986.DOI

Paper details

vlnc:W_4pwb5vms
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Engineering

Not recorded for this paper

View statement
Looking for related papers…
Contents