Symplectic integration of PDEs using Clebsch variables

Robert I McLachlan; Christian Offen; Benjamin K Tapley

doi:10.3934/jcd.2019005

Article Contents

2019, Volume 6, Issue 1: 111-130. Doi: 10.3934/jcd.2019005

This issue Previous Article The dependence of Lyapunov exponents of polynomials on their coefficients Next Article Numerical efficacy study of data assimilation for the 2D magnetohydrodynamic equations

Symplectic integration of PDEs using Clebsch variables

1.
School of Fundamental Sciences, Massey University, Private Bag 11 222, Palmerston North, 4442, New Zealand
2.
Department of Mathematical Sciences, Norwegian University of Science and Technology, Sentralbygg 2, Gløshaugen, Norway

Published: August 2019

This research was supported by the Marsden Fund of the Royal Society Te Apārangi

Abstract / Introduction Full Text(HTML) Figure(8) / Table(1) Related Papers Cited by

Abstract

Many PDEs (Burgers' equation, KdV, Camassa-Holm, Euler's fluid equations, …) can be formulated as infinite-dimensional Lie-Poisson systems. These are Hamiltonian systems on manifolds equipped with Poisson brackets. The Poisson structure is connected to conservation properties and other geometric features of solutions to the PDE and, therefore, of great interest for numerical integration. For the example of Burgers' equations and related PDEs we use Clebsch variables to lift the original system to a collective Hamiltonian system on a symplectic manifold whose structure is related to the original Lie-Poisson structure. On the collective Hamiltonian system a symplectic integrator can be applied. Our numerical examples show excellent conservation properties and indicate that the disadvantage of an increased phase-space dimension can be outweighed by the advantage of symplectic integration.

Keywords:

Mathematics Subject Classification: Primary: 37M15, 65P10; Secondary: 37K05, 35Q31, 37J15, 53D20.

Citation:

FullText(HTML)

Figure 1. Uniform periodic grids on $ S^1 \cong \mathbb{R}/L\mathbb{Z} $, $ L>0 $

Download: Full-size image PowerPoint slide

Figure 2. Order-two convergence for the travelling wave solution of the extended Burgers' equation outlined in section 6.2. The plots correspond to the conventional solution (○) and the collective solution (△) and an order-two reference line (). The error is calculated after 512 timesteps, with $ L = 8 $, $ \Delta t = 2^{-14} $ and $ \Delta x = L/2^{k} $ for $ k = 1,2,3 $ and $ 4 $

Download: Full-size image PowerPoint slide

Figure 3. Inviscid Burgers' equation solutions of the conventional method () and collective method (). The grid parameters are $ n_x = 64 $, $ \Delta x = 0.125 $, $ L = 8 $ and $ \Delta t = 2^{-12} $. A shock forms at about $ t = 0.4 $

Download: Full-size image PowerPoint slide

Figure 4. The errors corresponding to the conventional () and collective () methods for the inviscid Burgers' equation and $ \mathcal{O}(t^2) $ reference lines ()

Download: Full-size image PowerPoint slide

Figure 5. Travelling wave solutions of the perturbed Burgers' equation (top row) and the positive Fourier modes (bottom row) at $ t = 109 $ (left column), $ t = 218 $ (middle column) and $ t = 437 $ (right column). The plots correspond to the conventional method (), collective method () and the exact travelling wave solution (). The grid parameters are $ n_x = 16 $, $ \Delta x = 0.5 $, $ L = 8 $ and $ \Delta t = 2^{-6} $

Download: Full-size image PowerPoint slide

Figure 6. The errors corresponding to the conventional () and collective () methods for the travelling wave experiment. The reference lines () are $ \mathcal{O}(t) $ in figures (a) and (b) and exponential in figure (c)

Download: Full-size image PowerPoint slide

Figure 7. Periodic bump solutions of the extended Burgers' equation (top row) and the positive Fourier modes (bottom row) at $ t = 10 $ (left column), $ t = 100 $ (middle column) and $ t = 1000 $ (right column). The plots correspond to the conventional method () and the collective method (). The grid parameters are $ n_x = 32 $, $ \Delta x = 0.25 $, $ L = 8 $ and $ \Delta t = 2^{-8} $

Download: Full-size image PowerPoint slide

Figure 8. The errors corresponding to the conventional () and collective () methods for the periodic bump example. The reference line () in figure (a) is $ \mathcal{O}(t) $

Download: Full-size image PowerPoint slide

Table 1. Overview of the setting

Continuous system	Spatially discretised system
Collective Hamiltonian system on an infinite-dimensional symplectic vector space in Clebsch variables $ q_t = \frac{\delta \bar H}{\delta p}, \quad p_t = -\frac{\delta \bar H}{\delta q}. $ Exact solutions preserve the symplectic structure, the Hamiltonian $ \bar H=H\circ J $, all quantities related to the Casimirs of the original PDE and the fibres of the Clebsch map $ J(q,p)=u $.	Canonical Hamiltonian ODEs in $ 2N $ variables $ \hat q_t = \nabla_{\hat p} \hat {\bar H}, \quad \hat p_t = - \nabla_{\hat q} \hat {\bar H}. $ The exact flow preserves the symplectic structure and the Hamiltonian $ \hat {\bar H} $. Time-integration with the midpoint rule is symplectic.
Original PDE, interpreted as a Lie-Poisson equation $ u_t = \mathrm{ad}^\ast_{\frac {\delta H}{\delta u}}u. $ Exact solutions preserve the Poisson structure, the Hamiltonian $ H $ and all Casimirs.	Non-Hamiltonian ODEs in $ N $ variables $ \hat u_t = K(\hat u) \nabla_{\hat u} \hat H, \qquad K^T=-K. $ Exact solutions conserve $ \hat H $. Time-integration with the midpoint rule is not symplectic.

| Show Table

DownLoad: CSV

Related Papers

Cited by

References

[1]	L. Brugnano, M. Calvo, J. Montijano and L. Rández, Energy-preserving methods for Poisson systems, Journal of Computational and Applied Mathematics, 236 (2012), 3890–3904, 40 years of numerical analysis: "Is the discrete world an approximation of the continuous one or is it the other way around". doi: 10.1016/j.cam.2012.02.033.
[2]	A. Chern, F. Knöppel, U. Pinkall, P. Schröder and S. Weiẞmann, Schrödinger's smoke, ACM Transactions on Graphics (TOG), 35 (2016), 77. doi: 10.1145/2897824.2925868.
[3]	D. Cohen and E. Hairer, Linear energy-preserving integrators for Poisson systems, BIT Numerical Mathematics, 51 (2011), 91-101. doi: 10.1007/s10543-011-0310-z.
[4]	M. Dahlby, B. Owren and T. Yaguchi, Preserving multiple first integrals by discrete gradients, Journal of Physics A: Mathematical and Theoretical, 44 (2011), 305205, 14pp. doi: 10.1088/1751-8113/44/30/305205.
[5]	D. M. de Diego, Lie-Poisson integrators, preprint, arXiv: 1803.01427, URL https://arXiv.org/abs/1803.01427.
[6]	E. Hairer, C. Lubich and G. Wanner, Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations, Second edition. Springer Series in Computational Mathematics, 31. Springer-Verlag, Berlin, 2006. doi: 10.1007/3-540-30666-8.
[7]	B. Khesin and R. Wendt, Infinite-dimensional lie groups: Their geometry, orbits, and dynamical systems, The Geometry of Infinite-Dimensional Groups, 2009, 47–153. doi: 10.1007/978-3-540-77263-7_2.
[8]	A. Kriegl and P. W. Michor, The Convenient Setting of Global Analysis, vol. 53, AMS, 1997. doi: 10.1090/surv/053.
[9]	E. Kuznetsov and A. Mikhailov, On the topological meaning of canonical Clebsch variables, Physics Letters A, 77 (1980), 37-38. doi: 10.1016/0375-9601(80)90627-1.
[10]	J. Marsden and A. Weinstein, Coadjoint orbits, vortices, and Clebsch variables for incompressible fluids, Physica D: Nonlinear Phenomena, 7 (1980), 305-323. doi: 10.1016/0167-2789(83)90134-3.
[11]	J. E. Marsden and R. Abraham, Foundations of Mechanics, 2nd edition, Addison-Wesley Publishing Co., Redwood City, CA., 1978, URL http://resolver.caltech.edu/CaltechBOOK:1987.001.
[12]	J. E. Marsden, S. Pekarsky and S. Shkoller, Discrete Euler-Poincaré and Lie-Poisson equations, Nonlinearity, 12 (1999), 1647-1662. doi: 10.1088/0951-7715/12/6/314.
[13]	J. E. Marsden and T. S. Ratiu, Introduction to Mechanics and Symmetry: A Basic Exposition of Classical Mechanical Systems, Springer New York, New York, NY, 1999. doi: 10.1007/978-0-387-21792-5.
[14]	R. I. McLachlan, Spatial discretization of partial differential equations with integrals, IMA Journal of Numerical Analysis, 23 (2003), 645-664. doi: 10.1093/imanum/23.4.645.
[15]	R. I. McLachlan, K. Modin and O. Verdier, Collective symplectic integrators, Nonlinearity, 27 (2014), 1525-1542. doi: 10.1088/0951-7715/27/6/1525.
[16]	I. Vaisman, Symplectic realizations of poisson manifolds, Lectures on the Geometry of Poisson Manifolds, 1994,115–133. doi: 10.1007/978-3-0348-8495-2_9.
[17]	C. Vizman, Geodesic equations on diffeomorphism groups, SIGMA Symmetry Integrability Geom. Methods Appl., 4 (2008), Paper 030, 22 pp. doi: 10.3842/SIGMA.2008.030.