Browsing by Subject "Discontinuous Galerkin"

Now showing 1 - 5 of 5

A hybridized discontinuous Galerkin method for nonlinear dispersive water waves
(2017-05) Samii, Ali; Dawson, Clinton N.; Demkowicz, Leszek F; Gamba, Irene M; Hodges, Ben R; Landis, Chad M; Raja, Laxminarayan L
Simulation of water waves near the coast is an important problem in different branches of engineering and mathematics. For mathematical models to be valid in this region, they should include nonlinear and dispersive properties of the corresponding waves. Here, we study the numerical solution to three equations for modeling coastal water waves using the hybridized discontinuous Galerkin method (HDG). HDG is known to be a more efficient and in certain cases a more accurate alternative to some other discontinuous Galerkin methods, such as local DG. The first equation that we solve here is the Korteweg-de Vries equation. Similar to common HDG implementations, we first express the approximate variables and numerical fluxes in each element in terms of the approximate traces of the scalar variable, and its first derivative. These traces are assumed to be single-valued on each face. We next impose the conservation of the numerical fluxes via two sets of equations on the element boundaries. We solve this equation by Newton-Raphson method. We prove the stability of the proposed method for a proper choice of stabilization parameters. Through numerical examples, we observe that for a mesh with kth order elements, the computed variable and its first and second derivatives show optimal convergence at order k + 1 in both linear and nonlinear cases, which improves upon previously employed techniques. Next, we consider solving the fully nonlinear irrotational Green-Naghdi equation. This equation is often used to simulate water waves close to the shore, where there are significant dispersive and nonlinear effects involved. To solve this equation, we use an operator splitting method to decompose the problem into a dispersive part and a hyperbolic part. The dispersive part involves an implicit step, which has regularizing effects on the solution of the problem. On the other hand, for the hyperbolic sub-problem, we use an explicit hybridized DG method. Unlike the more common implicit version of the HDG, here we start by solving the flux conservation condition for the numerical traces. Afterwards, we use these traces in the original PDEs to obtain the internal unknowns. This process involves Newton iterations at each time step for computing the numerical traces. Next, we couple this solver with the dispersive solver to obtain the solution to the Green-Naghdi equation. We then solve a set of numerical examples to verify and validate the employed technique. In the first example we show the convergence properties of the numerical method. Next, we compare our results with a set of experimental data for nonlinear water waves in different situations. We observe close to optimal convergence rates and a good agreement between our numerical results and the experimental data.
A DPG method for convection-diffusion problems
(2013-08) Chan, Jesse L.; Demkowicz, Leszek; Moser, Robert deLancey
Over the last three decades, CFD simulations have become commonplace as a tool in the engineering and design of high-speed aircraft. Experiments are often complemented by computational simulations, and CFD technologies have proved very useful in both the reduction of aircraft development cycles, and in the simulation of conditions difficult to reproduce experimentally. Great advances have been made in the field since its introduction, especially in areas of meshing, computer architecture, and solution strategies. Despite this, there still exist many computational limitations in existing CFD methods; in particular, reliable higher order and hp-adaptive methods for the Navier-Stokes equations that govern viscous compressible flow. Solutions to the equations of viscous flow can display shocks and boundary layers, which are characterized by localized regions of rapid change and high gradients. The use of adaptive meshes is crucial in such settings -- good resolution for such problems under uniform meshes is computationally prohibitive and impractical for most physical regimes of interest. However, the construction of "good" meshes is a difficult task, usually requiring a-priori knowledge of the form of the solution. An alternative to such is the construction of automatically adaptive schemes; such methods begin with a coarse mesh and refine based on the minimization of error. However, this task is difficult, as the convergence of numerical methods for problems in CFD is notoriously sensitive to mesh quality. Additionally, the use of adaptivity becomes more difficult in the context of higher order and hp methods. Many of the above issues are tied to the notion of robustness, which we define loosely for CFD applications as the degradation of the quality of numerical solutions on a coarse mesh with respect to the Reynolds number, or nondimensional viscosity. For typical physical conditions of interest for the compressible Navier-Stokes equations, the Reynolds number dictates the scale of shock and boundary layer phenomena, and can be extremely high -- on the order of 10⁷ in a unit domain. For an under-resolved mesh, the Galerkin finite element method develops large oscillations which prevent convergence and pollute the solution. The issue of robustness for finite element methods was addressed early on by Brooks and Hughes in the SUPG method, which introduced the idea of residual-based stabilization to combat such oscillations. Residual-based stabilizations can alternatively be viewed as modifying the standard finite element test space, and consequently the norm in which the finite element method converges. Demkowicz and Gopalakrishnan generalized this idea in 2009 by introducing the Discontinous Petrov-Galerkin (DPG) method with optimal test functions, where test functions are determined such that they minimize the discrete linear residual in a dual space. Under the ultra-weak variational formulation, these test functions can be computed locally to yield a symmetric, positive-definite system. The main theoretical thrust of this research is to develop a DPG method that is provably robust for singular perturbation problems in CFD, but does not suffer from discretization error in the approximation of test functions. Such a method is developed for the prototypical singular perturbation problem of convection-diffusion, where it is demonstrated that the method does not suffer from error in the approximation of test functions, and that the L² error is robustly bounded by the energy error in which DPG is optimal -- in other words, as the energy error decreases, the L² error of the solution is guaranteed to decrease as well. The method is then extended to the linearized Navier-Stokes equations, and applied to the solution of the nonlinear compressible Navier-Stokes equations. The numerical work in this dissertation has focused on the development of a 2D compressible flow code under the Camellia library, developed and maintained by Nathan Roberts at ICES. In particular, we have developed a framework allowing for rapid implementation of problems and the easy application of higher order and hp-adaptive schemes based on a natural error representation function that stems from the DPG residual. Finally, the DPG method is applied to several convection diffusion problems which mimic difficult problems in compressible flow simulations, including problems exhibiting both boundary layers and singularities in stresses. A viscous Burgers' equation is solved as an extension of DPG to nonlinear problems, and the effectiveness of DPG as a numerical method for compressible flow is assessed with the application of DPG to two benchmark problems in supersonic flow. In particular, DPG is used to solve the Carter flat plate problem and the Holden compression corner problem over a range of Mach numbers and laminar Reynolds numbers using automatically adaptive schemes, beginning with very under-resolved/coarse initial meshes.
The enriched Galerkin method for linear elasticity and phase field fracture propagation
(2015-12) Mital, Prashant; Wheeler, Mary F. (Mary Fanett); Wick, Thomas
This thesis focuses on the application of the discontinuous Galerkin (DG) and enriched Galerkin (EG) methods to the problems of linear elasticity and phase field fracture propagation. The use of traditional and popular continuous Galerkin method (CG) for linear elasticity has posed some challenges. For example, nonphysical stress oscillations often occur in CG solutions for linearly elastic, nearly incompressible materials. Furthermore, CG solutions produce discontinuous stresses at the finite element boundaries which need to be post-processed. Based on the success of the DG methods in solving these challenges, we attempt resolution of the same problems with the yet untested EG method. For phase field fracture propagation, the CG method has been ubiquitously used in the literature. Since the phase field displacement solution is essentially discontinuous across the crack, we hypothesize that the discontinuous DG and EG methods could offer some advantages when applied to the fracture problem. We then perform a comparative analysis of CG, DG and EG applied to the phase field equations to determine if this is indeed the case. We begin by applying a family of DG and EG methods, including Nonsymmetric Interior Penalty Galerkin (NIPG), Symmetric Interior Penalty Galerkin (SIPG), and Incomplete Interior Penalty Galerkin (IIPG) to 2D linear elasticity problems. It is shown that the EG methods are simple and robust for dealing with linear elasticity. They are also shown to converge at the same rates as the corresponding DG methods. A detailed comparison of the performance of NIPG, IIPG, and SIPG is also made. We then propose a novel monolithic scheme with an augmented-Lagrangian method for phase field fracture propagation. We apply CG, DG and EG methods to the scheme and establish convergence in space and time through numerical studies. It is shown that the Newton method used for solving the system of nonlinear equations converges faster for DG and EG than it does for CG.
High-order (hybridized) discontinuous Galerkin method for geophysical flows
(2019-08) Kang, Shinhoo, 1980-; Bui-Thanh, Tan; Bisetti, Fabrizio; Willcox, Karen E; Demkowicz, Leszek F; Ghattas, Omar; Arbogast, Todd
As computational research has grown, simulation has become a standard tool in many fields of academic and industrial areas. For example, computational fluid dynamics (CFD) tools in aerospace and research facilities are widely used to evaluate the aerodynamic performance of aircraft or wings. Weather forecasts are highly dependent on numerical weather prediction (NWP) model. However, it is still difficult to simulate the complex physical phenomena of a wide range of length and time scales with modern computational resources. In this study, we develop a robust, efficient and high-order accurate numerical methods and techniques to tackle the challenges. First, we use high-order spatial discretization using (hybridized) discontinuous Galerkin (DG) methods. The DG method combines the advantages of finite volume and finite element methods. As such, it is well-suited to problems with large gradients including shocks and with complex geometries, and large-scale simulations. However, DG typically has many degrees-of-freedoms. To mitigate the expense, we use hybridized DG (HDG) method that introduces new “trace unknowns” on the mesh skeleton (mortar interfaces) to eliminate the local “volume unknowns” with static condensation procedure and reduces globally coupled system when implicit time-stepping is required. Also, since the information between the elements is exchanged through the mesh skeleton, the mortar interfaces can be used as a glue to couple multi-phase regions, e.g., solid and fluid regions, or non-matching grids, e.g., a rotating mesh and a stationary mesh. That is the HDG method provides an efficient and flexible coupling environment compared to standard DG methods. Second, we develop an HDG-DG IMEX scheme for an efficient time integrating scheme. The idea is to divide the governing equations into stiff and nonstiff parts, implicitly treat the former with HDG methods, and explicitly treat the latter with DG methods. The HDG-DG IMEX scheme facilitates high-order temporal and spatial solutions, avoiding too small a time step. Numerical results show that the HDG-DG IMEX scheme is comparable to an explicit Runge-Kutta DG scheme in terms of accuracy while allowing for much larger timestep sizes. We also numerically observe that IMEX HDG-DG scheme can be used as a tool to suppress the high-frequency modes such as acoustic waves or fast gravity waves in atmospheric or ocean models. In short, IMEX HDG-DG methods are attractive for applications in which a fast and stable solution is important while permitting inaccurate processing of the fast modes. Third, we also develop an EXPONENTIAL DG scheme for an efficient time integrators. Similar to the IMEX method, the governing equations are separated into linear and nonlinear parts, then the two parts are spatially discretized with DG methods. Next, we analytically integrate the linear term and approximate the nonlinear term with respect to time. This method accurately handles the fast wave modes in the linear operator. To efficiently evaluate a matrix exponential, we employ the cutting-edge adaptive Krylov subspace method. Finally, we develop a sliding-mesh interface by combining nonconforming treatment and the arbitrary Lagrangian-Eulerian (ALE) scheme for simulating rotating flows, which are important to estimate the characteristics of a rotating wind turbine or understanding vortical structures shown in atmospheric or astronomical phenomena. To integrate the rotating motion of the domain, we use the ALE formulation to map the governing equation to the stationary reference domain and introduce mortar interfaces between the stationary mesh and the rotating mesh. The mortar structure on the sliding interface changes dynamically as the mesh rotates
Task-based parallelism for hurricane storm surge modeling
(2020-07-30) Bremer, Maximilian Heimo Moritz; Dawson, Clinton N.; Biros, George; Gamba, Irene; Heimbach, Patrick; Pingali, Keshav
Hurricanes are incredibly devastating events, constituting seven of the ten most costly U.S. natural disasters since 1980. The development of real-time forecasting models that accurately capture a storm's dynamics play an essential role in informing local officials' emergency management decisions. ADCIRC is one such model that is operationally active in the National Oceanic and Atmospheric Administration's Hurricane Surge On-Demand Forecast System. However, ADCIRC faces several limitations. It struggles solving highly advective flows and is not locally mass conservative. These aspects limit applicable flow regimes and can cause unphysical behavior. One proposed alternative which addresses these limitations is the discontinuous Galerkin (DG) finite element method. However, the DG method's high computational cost makes it unsuitable for real-time forecasting and has limited adoption among coastal engineers. Simultaneously, efforts to build an exascale machine and the resulting power constrained computing architectures have led to massive increases in the concurrency applications are expected to manage. These architectural shifts have in turn caused some groups to turn away from the traditional flat MPI or MPI+OpenMP programming models to more functional task-based programming models, designed specifically to be performant on these next generation architectures. The aim of this thesis is to utilize these new task-based programming models to accelerate DG simulations for coastal applications. We explore two strategies for accelerating the DG method for storm surge simulation. The first strategy addresses load imbalance caused by coastal flooding. During the simulation of hurricane storm surge, cells are classified as either wet or dry. Dry cells can trivially update, while wet cells require full evaluation of the physics. As the storm makes landfall and causes flooding, this generates a load imbalance. We present two load balancing strategies---an asynchronous diffusion-based approach and semi-static approach---to optimize compute resource utilization. These load balancing strategies are analyzed using a discrete-event simulation that models the task-based storm surge simulation. We find speed-ups of up to 56% over the currently used mesh partitioning and up to 97% of the theoretical speed-up. The second strategy focuses on a first order adaptive local timestepping scheme for nonlinear conservation laws. For problems such as hurricane storm surge, the global CFL timestepping constraint is overly stringent for the majority of cells. We present a timestepping scheme that allows cells to stably advance based on local stability constraints. Since allowable timestep sizes depend on the state of the solution, care must be taken not to incur causality errors. The algorithm is accompanied with a proof of formal correctness that ensures that with a sufficiently small minimum timestep, the solution exhibits desired characteristics such as a maximum principle and total variation stability. The algorithm is parallelized using a speculative discrete event simulator. Performance results show that the implementation recovers 59%-77% of the optimal speed-up.