Puckett's method

Puckett's method is a high-resolution finite volume method for the numerical solution of hyperbolic partial differential equations, particularly those governing compressible flow. Developed by Elbridge Gerry Puckett in the late 20th century, it is designed to accurately capture discontinuities such as shock waves and contact discontinuities while maintaining high-order accuracy in smooth regions of the flow. The method is recognized for its application within the framework of operator splitting and its use of sophisticated slope limiter techniques to prevent numerical oscillations.

Overview

Puckett's method emerged from ongoing research into high-resolution schemes during the 1980s and 1990s, a period marked by significant advances in computational fluid dynamics for simulating complex phenomena like supersonic and hypersonic flows. It builds upon the foundational work of Sergei K. Godunov and the subsequent development of total variation diminishing schemes by Amiram Harten and others. The method is specifically tailored for systems of conservation laws, such as the Euler equations, and is often implemented within multi-dimensional codes using a dimensionally split approach. Its development was influenced by contemporaneous work on the piecewise parabolic method and other Godunov-type schemes.

Mathematical formulation

The core of Puckett's method involves discretizing the computational domain into control volumes and solving an integral form of the governing conservation laws. For a system like the Navier-Stokes equations, the method typically employs a Riemann solver, such as the HLLC Riemann solver, to compute fluxes at cell interfaces. A key component is its reconstruction step, where piecewise polynomial profiles, often linear or parabolic, are constructed within each cell using data from neighboring cells like in the MUSCL scheme. This reconstruction utilizes a flux limiter or slope limiter, such as the monotonized central limiter, to enforce the TVD property and suppress Gibbs phenomenon. The time integration frequently relies on Runge-Kutta methods or a method of lines approach to achieve higher-order temporal accuracy.

Applications in computational fluid dynamics

Puckett's method has been extensively applied to challenging problems in aerospace engineering and astrophysics. It is used in simulating shock tube problems, such as the classic Sod shock tube, and more complex configurations like shock–boundary layer interaction in supersonic inlets. The method is implemented in major research codes, including those at institutions like Lawrence Livermore National Laboratory and NASA, for studying re-entry vehicle heating and supernova remnants. Its ability to handle strong discontinuities makes it suitable for modeling detonation waves in combustion and blast wave propagation, as well as in multiphase flow simulations involving sharp material interfaces.

Comparison with other numerical methods

Compared to first-order methods like the Lax–Friedrichs method, Puckett's method provides significantly sharper resolution of discontinuities with reduced numerical diffusion. Against other high-order methods such as WENO or Discontinuous Galerkin schemes, it often offers a favorable balance of computational cost and accuracy for many practical engineering simulations. While spectral methods excel in smooth flows, Puckett's method is fundamentally superior for flows containing shocks. Its use of explicit limiters differentiates it from artificial viscosity approaches, providing a more physically consistent means of shock capturing. However, like other Godunov-type schemes, it can be more computationally intensive per time step than simpler finite difference methods.

Implementation considerations

Efficient implementation of Puckett's method requires careful attention to data structures for handling adaptive mesh refinement, often used in codes like FLASH or AMReX. Parallel computing via Message Passing Interface is crucial for scaling to large problems on supercomputers such as those at the Texas Advanced Computing Center. Stability is governed by the Courant–Friedrichs–Lewy condition, necessitating robust time step control algorithms. The choice of Riemann solver and limiter function can significantly impact results for specific problems like Rayleigh–Taylor instability or Kelvin–Helmholtz instability. Furthermore, extending the method to unstructured grids, as used in commercial software like ANSYS Fluent, presents additional challenges in gradient reconstruction and flux computation.

Category:Computational fluid dynamics Category:Numerical analysis Category:Finite volume methods