Spectral<i>/hp</i> element methods:recent developments, applications, and perspectives

Hui Xu, Chris D. Cantwell, Carlos Monteserin, Claes Eskilsson, Allan Peter Engsig‐Karup, Spencer J. Sherwin, Xu, Hui, Cantwell, Chris, Monteserin, Carlos, Eskilsson, Claes, Engsig-Karup, Allan, Sherwin, SpencerVBN 2018

TL;DR

This review paper provides a comprehensive overview of the spectral/hp element method, a powerful numerical technique for solving complex fluid dynamics problems. It highlights the method's recent advancements, diverse applications in areas like transitional flows and ocean engineering, and discusses ongoing challenges and future directions for its broader adoption in science and industry.

What Problem Does This Paper Address?

The paper addresses the increasing demand for high-accuracy and efficient numerical simulations in computational fluid dynamics (CFD). Traditional "low-order" methods, while robust, often struggle to capture the intricate details of unsteady flows in complex geometries with the necessary precision. This limitation hinders accurate predictions in many engineering and scientific applications. The spectral/hp element method offers a solution by combining the geometric flexibility of finite element methods with the superior accuracy and rapid convergence of spectral methods, enabling high-fidelity simulations at a potentially lower computational cost.

Methods

The core of the spectral/hp element method lies in its approximation strategy. It uses high-degree polynomial basis functions within each element of a computational mesh. This approach allows for arbitrary spatial accuracy, provided the solution is sufficiently smooth. The paper details:

  • Spectral/hp Element Formulation: This involves partitioning the domain into elements and representing the solution within each element using a basis of orthogonal polynomials (like Legendre or Chebyshev). The "h" refers to the mesh refinement (element size), and the "p" refers to the polynomial degree.
  • Continuous and Discontinuous Galerkin Formulations: The paper discusses how these methods are implemented, differentiating between enforcing continuity of the solution across element boundaries (Continuous Galerkin) and focusing on flux continuity (Discontinuous Galerkin).
  • Nektar++ Framework: The paper highlights Nektar++, an open-source spectral/hp element framework, as a key enabler for implementing these complex methods, making them more accessible.
  • Dealiasing and Spectral Vanishing Viscosity (SVV): For under-resolved simulations, techniques like dealiasing (to mitigate aliasing errors from numerical quadrature) and SVV (to suppress spurious oscillations by adding controlled dissipation) are discussed as crucial stabilization methods.
  • High-Order Mesh Generation: The challenges and recent advancements in generating high-order curved meshes, particularly for complex geometries and boundary layers, are presented.
  • Combined Spectral/hp-Fourier Discretization: For problems with geometrically homogeneous directions, this approach leverages Fourier methods to decouple modes, enhancing efficiency.

Datasets

Not discussed in the paper. The paper focuses on the methodology and its application to various fluid dynamics problems, often using established benchmark cases or demonstrating capabilities with illustrative examples rather than relying on specific datasets for training or evaluation.

Key Insights and Ideas

  • The Power of "p": Increasing the polynomial order ("p") in spectral/hp methods can lead to exponential convergence for smooth solutions, offering a significant accuracy advantage over traditional "h" refinement alone.
  • Bridging Geometric Flexibility and Spectral Accuracy: The spectral/hp method successfully marries the geometric adaptability of finite element methods with the high-precision capabilities of spectral methods, a combination that is highly desirable for complex engineering problems.
  • Nektar++ as a Game Changer: The development and availability of open-source frameworks like Nektar++ are crucial for democratizing the use of these advanced numerical techniques, lowering the barrier to entry for researchers and engineers.
  • Stabilization is Key for Real-World Flows: For turbulent and marginally resolved flows, techniques like dealiasing and Spectral Vanishing Viscosity are not just refinements but essential tools for achieving stable and meaningful simulations.
  • Future is Hybrid: Combining spectral/hp methods with other techniques like Fourier methods or leveraging GPU acceleration is seen as a critical path forward for tackling the most challenging problems.

Key Results

The paper showcases the effectiveness of the spectral/hp element method across a wide range of applications:

  • Transitional Flows: Accurate capture of flow structures with low numerical dissipation and dispersion, leading to better predictions of turbulent phenomena.
  • Ocean Engineering: Successful application in modeling wave propagation and wave-body interactions, including complex scenarios like tsunami simulations and wave scattering.
  • Turbulent Flows: Demonstrated ability to simulate turbulent channel and pipe flows, and complex phenomena like wingtip vortices with high fidelity.
  • Hydrodynamic Stability: Support for linearised Navier-Stokes equations enables global flow stability analyses.
  • Efficiency Gains: For high polynomial orders, sum-factorization and element-wise operations offer significant performance improvements over traditional global assembly.

Limitations and Future Work

The authors acknowledge several challenges and areas for future development:

  • High-Order Mesh Generation: Robustly generating high-order meshes that conform to complex CAD geometries remains a significant hurdle for widespread industrial adoption.
  • hp-Adaptivity Optimization: Further research is needed to develop optimal algorithms for choosing between "h" and "p" adaptivity to balance computational cost and accuracy.
  • Stabilization Refinement: Techniques like SVV require further calibration and improvement, especially for marginal or under-resolved computations, to avoid artificial numerical pollution.
  • Verification and Validation: Continued efforts are needed to rigorously verify and validate spectral/hp simulations for under-resolved flows, ensuring that high-precision benefits are not lost.
  • Tip Vortex Modeling: Accurate modeling of three-dimensional boundary layers and the low-pressure regions within vortex cores remains challenging.
  • Implicit Large Eddy Simulation (iLES): While promising, the SVV-based iLES strategy requires further refinement and calibration for various high-Reynolds number flows.
  • Fluid-Structure Interaction: While spectral elements offer superior wave propagation, their application in fluid-structure interaction, especially with surface waves, needs further exploration.
  • Cavitating Flows: The implementation of discontinuous spectral/hp methods for multiphase flows with cavitation is a recent development with significant potential.
  • Free Surface Flows: While spectral/hp methods are being used, their application to free surface problems, particularly with h/p-adaptivity, offers substantial opportunities.

Summary updated

Abstract

The spectral/hp element method combines the geometric flexibility of the classical h-type finite element technique with the desirable numerical properties of spectral methods, employing high-degree piecewise polynomial basis functions on coarse finite element-type meshes. The spatial approximation is based upon orthogonal polynomials, such as Legendre or Chebychev polynomials, modified to accommodate a C0 - continuous expansion. Computationally and theoretically, by increasing the polynomial order p, high-precision solutions and fast convergence can be obtained and, in particular, under certain regularity assumptions an exponential reduction in approximation error between numerical and exact solutions can be achieved. This method has now been applied in many simulation studies of both fundamental and practical engineering flows. This paper briefly describes the formulation of the spectral/hp element method and provides an overview of its application to computational fluid dynamics. In particular, it focuses on the use of the spectral/hp element method in transitional flows and ocean engineering. Finally, some of the major challenges to be overcome in order to use the spectral/hp element method in more complex science and engineering applications are discussed.
BibTeX
@inproceedings{xu2018spectral,
  author = {Hui Xu and Chris D. Cantwell and Carlos Monteserin and Claes Eskilsson and Allan Peter Engsig‐Karup and Spencer J. Sherwin and Xu, Hui and Cantwell, Chris and Monteserin, Carlos and Eskilsson, Claes and Engsig-Karup, Allan and Sherwin, Spencer},
  title = {{Spectral\<i\>/hp\</i\> element methods:recent developments, applications, and perspectives}},
  booktitle = {VBN},
  year = {2018},
}
Figure