Enhanced fifth order WENO Shock-Capturing Schemes with Deep Learning

In this paper we enhance the well-known fifth order WENO shock-capturing scheme by using deep learning techniques. This fine-tuning of an existing algorithm is implemented by training a rather small neural network to modify the smoothness indicators of the WENO scheme in order to improve the numerical results especially at discontinuities. In our approach no further post-processing is needed to ensure the consistency of the method, which simplifies the method and increases the effect of the neural network. Moreover, the convergence of the resulting scheme can be theoretically proven. We demonstrate our findings with the inviscid Burgers’ equation, the Buckley-Leverett equation and the 1-D Euler equations of gas dynamics. Hereby we investigate the classical Sod problem and the Lax problem and show that our novel method outperforms the classical fifth order WENO schemes in simulations where the numerical solution is too diffusive or tends to overshoot at shocks.

[1]  Nikolaus A. Adams,et al.  A new class of adaptive high-order targeted ENO schemes for hyperbolic conservation laws , 2018, J. Comput. Phys..

[2]  S. Osher,et al.  Efficient implementation of essentially non-oscillatory shock-capturing schemes,II , 1989 .

[3]  P. Roe.  Approximate Riemann Solvers, Parameter Vectors, and Difference Schemes , 1997 .

[4]  Wai-Sun Don,et al.  An improved weighted essentially non-oscillatory scheme for hyperbolic conservation laws , 2008, J. Comput. Phys..

[5]  G. Naga Raju,et al.  A modified fifth-order WENO scheme for hyperbolic conservation laws , 2016, Comput. Math. Appl..

[6]  Chi-Wang Shu,et al.  Efficient Implementation of Weighted ENO Schemes , 1995 .

[7]  Nikolaus A. Adams,et al.  A family of high-order targeted ENO schemes for compressible-fluid simulations , 2016, J. Comput. Phys..

[8]  P. Lax.  Weak solutions of nonlinear hyperbolic equations and their numerical computation , 1954 .

[9]  Natalia Gimelshein,et al.  PyTorch: An Imperative Style, High-Performance Deep Learning Library , 2019, NeurIPS.

[10]  Dimitrios I. Fotiadis,et al.  Artificial neural networks for solving ordinary and partial differential equations , 1997, IEEE Trans. Neural Networks.

[11]  Lin Fu,et al.  A Hybrid Method with TENO Based Discontinuity Indicator for Hyperbolic Conservation Laws , 2019, Communications in Computational Physics.

[12]  Chi-Wang Shu,et al.  High Order Weighted Essentially Nonoscillatory Schemes for Convection Dominated Problems , 2009, SIAM Rev..

[13]  Hong-bo Wang,et al.  Efficient WENOCU4 scheme with three different adaptive switches , 2020, Journal of Zhejiang University-SCIENCE A.

[14]  Alfio Quarteroni,et al.  Advanced numerical approximation of nonlinear hyperbolic equations : lectures given at the 2nd session of the Centro Internazionale Matematico Estivo (C.I.M.E.) held in Cetraro, Italy, June 23-28, 1997 , 1998 .

[15]  Andrea D. Beck,et al.  A Neural Network based Shock Detection and Localization Approach for Discontinuous Galerkin Methods , 2020, J. Comput. Phys..

[16]  Sergio Pirozzoli,et al.  Conservative Hybrid Compact-WENO Schemes for Shock-Turbulence Interaction , 2002 .

[17]  Chi-Wang Shu.  Essentially non-oscillatory and weighted essentially non-oscillatory schemes for hyperbolic conservation laws , 1998 .

[18]  Ignasi Colominas,et al.  A reduced-dissipation WENO scheme with automatic dissipation adjustment , 2021, J. Comput. Phys..

[19]  ShuChi-Wang,et al.  Efficient implementation of essentially non-oscillatory shock-capturing schemes, II , 1989 .

[20]  Stephan Hoyer,et al.  Learning data-driven discretizations for partial differential equations , 2018, Proceedings of the National Academy of Sciences.

[21]  Zhi J. Wang,et al.  Optimized weighted essentially nonoscillatory schemes for linear waves with discontinuity: 381 , 2001 .

[22]  F. ARÀNDIGA,et al.  Analysis of WENO Schemes for Full and Global Accuracy , 2011, SIAM J. Numer. Anal..

[23]  P. Alam.  ‘G’ , 2021, Composites Engineering: An A–Z Guide.

[24]  Yang Liu,et al.  Globally optimal finite-difference schemes based on least squares , 2013 .

[25]  Jianxian Qiu,et al.  A hybrid WENO method with modified ghost fluid method for compressible two-medium flow problems , 2020, Numerical Mathematics: Theory, Methods and Applications.

[26]  Wai-Sun Don,et al.  Accuracy of the weighted essentially non-oscillatory conservative finite difference schemes , 2013, J. Comput. Phys..

[27]  Tim Colonius,et al.  Enhancement of shock-capturing methods via machine learning , 2020, ArXiv.

[28]  P. Wesseling.  Principles of Computational Fluid Dynamics , 2000 .

[29]  Wang Chi-Shu,et al.  Essentially Non-Oscillatory and Weighted Essentially Non-Oscillatory Schemes for Hyperbolic Conservation Laws , 1997 .

[30]  S. Osher,et al.  Uniformly high order accuracy essentially non-oscillatory schemes III , 1987 .

[31]  M. Crandall,et al.  Monotone difference approximations for scalar conservation laws , 1979 .

[32]  Bin Dong,et al.  Learning to Discretize: Solving 1D Scalar Conservation Laws via Deep Reinforcement Learning , 2019, Communications in Computational Physics.

[33]  R. LeVeque.  Finite Volume Methods for Hyperbolic Problems: Characteristics and Riemann Problems for Linear Hyperbolic Equations , 2002 .

[34]  Deep Ray,et al.  Controlling oscillations in high-order Discontinuous Galerkin schemes using artificial viscosity tuned by neural networks , 2020, J. Comput. Phys..

[35]  C. Tam,et al.  Dispersion-relation-preserving finite difference schemes for computational acoustics , 1993 .

[36]  Rong Wang,et al.  Linear Instability of the Fifth-Order WENO Method , 2007, SIAM J. Numer. Anal..

[37]  Jan S. Hesthaven,et al.  Detecting troubled-cells on two-dimensional unstructured grids using a neural network , 2019, J. Comput. Phys..

[38]  Kun Wang,et al.  A Characteristic-Featured Shock Wave Indicator for Conservation Laws Based on Training an Artificial Neuron , 2020, J. Sci. Comput..

[39]  G. Sod.  A survey of several finite difference methods for systems of nonlinear hyperbolic conservation laws , 1978 .

[40]  Jungho Yoon,et al.  Modified Non-linear Weights for Fifth-Order Weighted Essentially Non-oscillatory Schemes , 2016, J. Sci. Comput..

[41]  Jianxian Qiu,et al.  Hermite WENO schemes and their application as limiters for Runge-Kutta discontinuous Galerkin method II: Two dimensional case , 2005 .

[42]  Kaj Nyström,et al.  A unified deep artificial neural network approach to partial differential equations in complex geometries , 2017, Neurocomputing.

[43]  J. M. Powers,et al.  Mapped weighted essentially non-oscillatory schemes: Achieving optimal order near critical points , 2005 .

[44]  P. Alam.  ‘S’ , 2021, Composites Engineering: An A–Z Guide.

[45]  V. Guinot.  Approximate Riemann Solvers , 2010 .

[46]  Stefano Ermon,et al.  Learning Neural PDE Solvers with Convergence Guarantees , 2019, ICLR.

[47]  Justin A. Sirignano,et al.  DGM: A deep learning algorithm for solving partial differential equations , 2017, J. Comput. Phys..

[48]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[49]  Yeon Ju Lee,et al.  An improved weighted essentially non-oscillatory scheme with a new smoothness indicator , 2013, J. Comput. Phys..

[50]  Jianxian Qiu,et al.  Hermite WENO schemes and their application as limiters for Runge-Kutta discontinuous Galerkin method: one-dimensional case , 2004 .

[51]  S. Osher,et al.  Uniformly high order accurate essentially non-oscillatory schemes, 111 , 1987 .

[52]  Zhoushun Zheng,et al.  Solution of two-dimensional time-fractional Burgers equation with high and low Reynolds numbers , 2017 .

[53]  Wai-Sun Don,et al.  High order weighted essentially non-oscillatory WENO-Z schemes for hyperbolic conservation laws , 2011, J. Comput. Phys..

[54]  A. Harten.  High Resolution Schemes for Hyperbolic Conservation Laws , 2017 .

[55]  D. Pullin,et al.  Hybrid tuned center-difference-WENO method for large eddy simulations in the presence of strong shocks , 2004 .

[56]  CostaBruno,et al.  High order weighted essentially non-oscillatory WENO-Z schemes for hyperbolic conservation laws , 2011 .