Galerkin v. least-squares Petrov-Galerkin projection in nonlinear model reduction

Published on Feb 1, 2017in Journal of Computational Physics3.553
路 DOI :10.1016/J.JCP.2016.10.033
Kevin Carlberg18
Estimated H-index: 18
(DOE: United States Department of Energy),
Matthew F. Barone21
Estimated H-index: 21
(SNL: Sandia National Laboratories),
Harbir Antil18
Estimated H-index: 18
(GMU: George Mason University)
Sources
Abstract
Least-squares Petrov-Galerkin (LSPG) model-reduction techniques such as the Gauss-Newton with Approximated Tensors (GNAT) method have shown promise, as they have generated stable, accurate solutions for large-scale turbulent, compressible flow problems where standard Galerkin techniques have failed. However, there has been limited comparative analysis of the two approaches. This is due in part to difficulties arising from the fact that Galerkin techniques perform optimal projection associated with residual minimization at the time-continuous level, while LSPG techniques do so at the time-discrete level.This work provides a detailed theoretical and computational comparison of the two techniques for two common classes of time integrators: linear multistep schemes and Runge-Kutta schemes. We present a number of new findings, including conditions under which the LSPG ROM has a time-continuous representation, conditions under which the two techniques are equivalent, and time-discrete error bounds for the two approaches. Perhaps most surprisingly, we demonstrate both theoretically and computationally that decreasing the time step does not necessarily decrease the error for the LSPG ROM; instead, the time step should be 'matched' to the spectral content of the reduced basis. In numerical experiments carried out on a turbulent compressible-flow problem with over one million unknowns, we show that increasing the time step to an intermediate value decreases both the error and the simulation time of the LSPG reduced-order model by an order of magnitude.
Figures & Tables
Download
馃摉 Papers frequently viewed together
2016
References58
Newest
#1Peter Benner (MPG: Max Planck Society)H-Index: 54
#2Serkan Gugercin (VT: Virginia Tech)H-Index: 31
Last. Karen Willcox (MIT: Massachusetts Institute of Technology)H-Index: 48
view all 3 authors...
Numerical simulation of large-scale dynamical systems plays a fundamental role in studying a wide range of complex physical phenomena; however, the inherent large-scale nature of the models often leads to unmanageable demands on computational resources. Model reduction aims to reduce this computational burden by generating reduced models that are faster and cheaper to simulate, yet accurately represent the original large-scale system behavior. Model reduction of linear, nonparametric dynamical s...
Source
#1David Amsallem (Stanford University)H-Index: 20
#2Matthew J. Zahr (Stanford University)H-Index: 14
Last. Kyle M. Washabaugh (Stanford University)H-Index: 3
view all 3 authors...
Projection-based model reduction techniques rely on the definition of a small dimensional subspace in which the solution is approximated. Using local subspaces reduces the dimensionality of each subspace and enables larger speedups. Transitions between local subspaces require special care and updating the reduced bases associated with each subspace increases the accuracy of the reduced-order model. In the present work, local reduced basis updates are considered in the case of hyper-reduction, fo...
Source
We propose a novel model reduction approach for the approximation of non linear hyperbolic equations in the scalar and the system cases. The approach relies on an offline computation of a dictionary of solutions together with an online L^1norm minimization of the residual. It is shown why this is a natural framework for hyperbolic problems and tested on nonlinear problems such as Burgers' equation and the one-dimensional Euler equations involving shocks and discontinuities. Efficient algorith...
#1Kevin Carlberg (SNL: Sandia National Laboratories)H-Index: 18
#2Raymond S. TuminaroH-Index: 32
Last. Paul T. BoggsH-Index: 17
view all 3 authors...
This work proposes a model-reduction methodology that preserves Lagrangian structure and achieves computational efficiency in the presence of high-order nonlinearities and arbitrary parameter dependence. As such, the resulting reduced-order model retains key properties such as energy conservation and symplectic time-evolution maps. We focus on parameterized simple mechanical systems subjected to Rayleigh damping and external forces, and consider an application to nonlinear structural dynamics. T...
Source
#1Charbel Farhat (Stanford University)H-Index: 88
#2Philip Avery (Stanford University)H-Index: 17
Last. Julien Cortial (Stanford University)H-Index: 10
view all 4 authors...
Source
#1Harbir Antil (GMU: George Mason University)H-Index: 18
#2Matthias Heinkenschloss (Rice University)H-Index: 31
Last. Danny C. Sorensen (Rice University)H-Index: 55
view all 3 authors...
Projection based methods lead to reduced order models (ROMs) with dramatically reduced numbers of equations and unknowns. However, for nonlinear or parametrically varying problems the cost of evaluating these ROMs still depends on the size of the full order model and therefore is still expensive. The Discrete Empirical Interpolation Method (DEIM) further approximates the nonlinearity in the projection based ROM. The resulting DEIM ROM nonlinearity depends only on a few components of the original...
Source
#1Harbir Antil (GMU: George Mason University)H-Index: 18
#2Scott E. Field (UMD: University of Maryland, College Park)H-Index: 20
Last. Manuel Tiglio (UMD: University of Maryland, College Park)H-Index: 28
view all 5 authors...
We present an algorithm to generate application-specific, global reduced order quadratures (ROQ) for multiple fast evaluations of weighted inner products between parameterized functions. If a reduced basis or any other projection-based model reduction technique is applied, the dimensionality of integrands is reduced dramatically; however, the cost of approximating the integrands by projection still scales as the size of the original problem. In contrast, using discrete empirical interpolation po...
Source
#1Maciej Balajewicz (Stanford University)H-Index: 13
#2Earl H. Dowell (Duke University)H-Index: 78
Last. Bernd R. Noack (University of Poitiers)H-Index: 43
view all 3 authors...
We generalize the POD-based Galerkin method for post-transient flow data by incorporating Navier鈥揝tokes equation constraints. In this method, the derived Galerkin expansion minimizes the residual like POD, but with the power balance equation for the resolved turbulent kinetic energy as an additional optimization constraint. Thus, the projection of the Navier鈥揝tokes equation on to the expansion modes yields a Galerkin system that respects the power balance on the attractor. The resulting dynamica...
Source
This paper puts forth several closure models for the proper orthogonal decomposition (POD) reduced order modeling of fluid flows. These new closure models, together with other standard closure models, are investigated in the numerical simulation of the Burgers equation. This simplified setting represents just the first step in the investigation of the new closure models. It allows a thorough assessment of the performance of the new models, including a parameter sensitivity study. Two challenging...
#1Kevin Carlberg (SNL: Sandia National Laboratories)H-Index: 18
#2Charbel Farhat (Stanford University)H-Index: 88
Last. David Amsallem (Stanford University)H-Index: 20
view all 4 authors...
The Gauss-Newton with approximated tensors (GNAT) method is a nonlinear model-reduction method that operates on fully discretized computational models. It achieves dimension reduction by a Petrov-Galerkin projection associated with residual minimization; it delivers computational efficiency by a hyper-reduction procedure based on the 'gappy POD' technique. Originally presented in Ref. [1], where it was applied to implicit nonlinear structural-dynamics models, this method is further developed her...
Source
Cited By151
Newest
#1Cheng Huang (UM: University of Michigan)H-Index: 26
#2Christopher R. Wentland (UM: University of Michigan)H-Index: 3
Last. Charles L. Merkle (Purdue University)H-Index: 22
view all 4 authors...
Abstract null null A projection-based formulation is presented for non-linear model reduction of problems with extreme scale disparity. The approach allows for the selection of an arbitrary, but complete, set of solution variables while preserving the structure of the governing equations. Least-squares-based minimization is leveraged to guarantee symmetrization and discrete consistency with the full-order model (FOM). Two levels of scaling are used to achieve the conditioning required to effecti...
Source
#1Michael Sleeman (U of T: University of Toronto)H-Index: 1
#2Masayuki Yano (U of T: University of Toronto)H-Index: 14
Abstract null null We present a projection-based model reduction formulation for parametrized time-dependent nonlinear partial differential equations (PDEs). Our approach builds on the following ingredients: reduced bases (RB), which provide rapidly convergent approximations of the parameter-temporal solution manifold; reduced quadrature (RQ) rules, which provide hyperreduction of the nonlinear residual; and the dual-weighted residual (DWR) method, which provides an error representation formula ...
Source
#1Yukiko S. Shimizu (SNL: Sandia National Laboratories)
#2Eric J. Parish (SNL: Sandia National Laboratories)H-Index: 9
Abstract null null Projection-based reduced-order models (PROMs) restrict the full-order model (FOM) to a low-dimensional subspace. Space鈥搕ime PROMs in particular restrict the FOM to a temporally space鈥搕ime trial subspace, and compute approximate solutions through a residual orthogonalization or minimization process. One such technique of interest is the space鈥搕ime least-squares Petrov鈥揋alerkin method (ST-LSPG), which reduces both the spatial and temporal dimensions. However, ST-LSPG is computat...
Source
#1A. Schein (TUM: Technische Universit盲t M眉nchen)H-Index: 1
#2Michael W. Gee (TUM: Technische Universit盲t M眉nchen)H-Index: 23
This work proposes a framework for projection-based model order reduction (MOR) of computational models aiming at a mechanical analysis of abdominal aortic aneurysms (AAAs). The underlying full-order model (FOM) is patient-specific, stationary and nonlinear. The quantities of interest are the von Mises stress and the von Mises strain field in the AAA wall, which result from loading the structure to the level of diastolic blood pressure at a fixed, imaged geometry (prestressing stage) and subsequ...
Source
Machine learning is rapidly becoming a core technology for scientific computing, with numerous opportunities to advance the field of computational fluid dynamics. This paper highlights some of the areas of highest potential impact, including to accelerate direct numerical simulations, to improve turbulence closure modelling, and to develop enhanced reduced-order models. In each of these areas, it is possible to improve machine learning capabilities by incorporating physics into the process, and ...
#1Chong Ke (WD: Western Digital)H-Index: 1
#2Xingyong Song (A&M: Texas A&M University)H-Index: 10
Source
#1Chi Hoang (SNL: Sandia National Laboratories)H-Index: 1
#2Youngsoo Choi (LLNL: Lawrence Livermore National Laboratory)H-Index: 12
Last. Kevin Carlberg (UW: University of Washington)H-Index: 18
view all 3 authors...
While reduced-order models (ROMs) have demonstrated success in many applications across computational science, challenges remain when applied both to extreme-scale models due to the prohibitive cost of generating requisite training data, and to decomposable systems due to many-query problems often requiring repeated reconfigurations of system components. Therefore, we propose the domain-decomposition least-squares Petrov-Galerkin (DD-LSPG) model-reduction method applicable to parameterized syste...
Source
#1Elnaz RezaianH-Index: 3
#2Cheng HuangH-Index: 26
Last. Karthik Duraisamy (UM: University of Michigan)H-Index: 25
view all 3 authors...
Balanced truncation (BT) is a model reduction method that utilizes a coordinate transformation to retain eigen-directions that are highly observable and reachable. To address realizability and scalability of BT applied to highly stiff and lightly-damped systems, a non-intrusive data-driven method is developed for balancing discrete-time systems via the eigensystem realization algorithm (ERA). The advantage of ERA for balancing transformation makes full-state outputs tractable. Further, ERA enabl...
Source
#1Wenqian ChenH-Index: 1
#2Qian WangH-Index: 7
Last. Chuhua ZhangH-Index: 1
view all 4 authors...
Source
This website uses cookies.
We use cookies to improve your online experience. By continuing to use our website we assume you agree to the placement of these cookies.
To learn more, you can find in our Privacy Policy.