Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show...
Transcript of Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show...
![Page 1: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/1.jpg)
1Romain Teyssier5th JETSET School
Numerical Solutions for
Hyperbolic Systems of
Conservation Laws:
from Godunov Method to
Adaptive Mesh Refinement
Romain Teyssier
CEA Saclay
![Page 2: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/2.jpg)
2Romain Teyssier5th JETSET School
- Euler equations, MHD, waves, hyperbolic systems of conservationlaws, primitive form, conservative form, integral form
- Advection equation, exact solution, characteristic curve, Riemanninvariant, finite difference scheme, modified equation, Von Neumananalysis, upwind scheme, Courant condition, Second order scheme
- Finite volume scheme, Godunov method, Riemann problem,approximate Riemann solver, Second order scheme, Slope limiters,Characteristic tracing
- Multidimensional scheme, directional splitting, Godunov, Runge-Kutta, CTU, 3D slope limiting
- AMR, patch-based versus cell-based, octree structure, gradedoctree, flux correction, EMF correction, restriction and prolongation,divB conserving interpolation
- Parallel computing with the RAMSES code
Outline
![Page 3: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/3.jpg)
3Romain Teyssier5th JETSET School
Lagrangian form: fluid element of unit mass
- Mass conservation:
- Euler equation:
- First law of thermodynamics:
Chain rule:
from Lagrange derivative to Euler derivative
Volume expansion rate:
The Euler equations
![Page 4: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/4.jpg)
4Romain Teyssier5th JETSET School
Eulerian form: fixed volume element
- Mass conservation
- Momentum conservation
- Energy conservation
The Euler equations
Kinetic energy Internal energy Equation of state
![Page 5: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/5.jpg)
5Romain Teyssier5th JETSET School
The Euler equations
System of conservation laws
- Vector of conservative variables
- Flux function
Integral form
x
t
x2x1
t1
t2
![Page 6: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/6.jpg)
6Romain Teyssier5th JETSET School
The Euler equations
Primitive form of the Euler equations:
- Vector of primitive variables
- Quasi-linear form
![Page 7: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/7.jpg)
7Romain Teyssier5th JETSET School
The Euler equations
Hyperbolic system: J and A have real eigenvalues.
Eigenvectors are travelling waves.
Perturbation of an equilibrium state:
Wave equation:
Eigenvalues:A=
Sound speed: Eigenvectors:
![Page 8: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/8.jpg)
8Romain Teyssier5th JETSET School
Linear scalar partial differential equation:
Initial conditions:
Defining and using the chain rule, we get:
Along the characteristic curve
we have the Riemann invariant
The advection equation
x
u
t=0
t>0
a x t
![Page 9: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/9.jpg)
9Romain Teyssier5th JETSET School
Piecewise constant initial states:
Solution:
The Riemann problem
x
u
uL
uR
![Page 10: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/10.jpg)
10Romain Teyssier5th JETSET School
Finite difference scheme
Finite difference approximation of the advection equation
![Page 11: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/11.jpg)
11Romain Teyssier5th JETSET School
The Modified Equation
Taylor expansion in time up to second order
Taylor expansion in space up to second order
The advection equation becomes the advection-diffusion equation
Negative diffusion coefficient: the scheme is unconditionally unstable
![Page 12: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/12.jpg)
12Romain Teyssier5th JETSET School
The Upwind scheme
a>0: use only upwind values, discard downwind variables
Taylor expansion up to second order:
Upwind scheme is stable if C<1, with
![Page 13: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/13.jpg)
13Romain Teyssier5th JETSET School
Von Neumann analysis
Fourier transform the current solution:
Evaluate the amplification factor of the 2 schemes.
Fromm scheme:
Upwind scheme:
ω>1: the scheme is unconditionally unstable
ω<1 if C<1: the scheme is stable under the Courant condition.
![Page 14: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/14.jpg)
14Romain Teyssier5th JETSET School
The advection-diffusion equation
Finite difference approximation of the advection equation:
Central differencing unstable:
Upwind differencing is stable:
Smearing of initialdiscontinuity:
“numerical diffusion”
Thickness increases
as
![Page 15: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/15.jpg)
15Romain Teyssier5th JETSET School
Finite volume scheme
Finite volume approximation of the advection equation:
Use integral form of the conservation law:
Exact evolution of volume averaged quantities:
Time averaged flux function:
![Page 16: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/16.jpg)
16Romain Teyssier5th JETSET School
Sergei Konstantinovich Godunov
![Page 17: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/17.jpg)
17Romain Teyssier5th JETSET School
Godunov scheme for the advection equation
The time averaged flux function:
is computed using the solution of the Riemann problem defined at cell interfaces with piecewise constant initial data.
x
ui
ui+1
For all t>0:
The Godunov scheme for the advection equation is identical tothe upwind finite difference scheme.
![Page 18: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/18.jpg)
18Romain Teyssier5th JETSET School
The time average flux function iscomputed using the self-similarsolution of the inter-cell Riemannproblem:
Godunov scheme for hyperbolic systemsThe system of conservation laws
is discretized using the followingintegral form:
This defines the Godunov flux:
Advection: 1 wave, Euler: 3 waves, MHD: 7 waves
Piecewise constantinitial data
![Page 19: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/19.jpg)
19Romain Teyssier5th JETSET School
Riemann solvers
Exact Riemann solution is costly: involves Raphson-Newtoniterations and complex non-linear functions.
Approximate Riemann solvers are more useful.
Two broad classes:
- Linear solvers
- HLL solvers
![Page 20: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/20.jpg)
20Romain Teyssier5th JETSET School
Linear Riemann solvers
Define a reference state as the arithmetic average or the Roe average
Evaluate the Jacobian matrix at this reference state.
Compute eignevalues and (left and right) eigenvectors
The flux function is given by the linear Riemann solution.
where
A simple example, the upwind Riemann solver:
![Page 21: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/21.jpg)
21Romain Teyssier5th JETSET School
Approximate the true Riemann fan by 2 waves and 1 intermediate state:
HLL Riemann solver
x
t
UL UR
U*
Compute U* using the integral form between SLt and SRt
Compute F* using the integral form between SLt and 0.
![Page 22: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/22.jpg)
22Romain Teyssier5th JETSET School
Other HLL-type Riemann solvers
Lax-Friedrich Riemann solver:
HLLC Riemann solver: add a third wave for the contact (entropy) wave.
x
t
UL UR
U*L
SL SRS*
U*R
See Toro (1997) for details.
![Page 23: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/23.jpg)
23Romain Teyssier5th JETSET School
Higher Order Godunov schemes
Bram Van Leer
Godunov method is stable but very diffusive. It wasabandoned for two decades, until…
![Page 24: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/24.jpg)
24Romain Teyssier5th JETSET School
Second Order Godunov scheme
x
ui
ui+1Piecewise linearapproximation of the solution:
The linear profile introduces a length scale: theRiemann solution is not self-similar anymore:
The flux function is approximated using a predictor-corrector scheme:
The corrected Riemann solver has now predicted states as initial data:
![Page 25: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/25.jpg)
25Romain Teyssier5th JETSET School
Predictor Step for the advection equation
x
ui
ui+1
The predicted states are computedusing a Taylor expansion in spaceand time:
Second order predicted states are the new initial conditions forthe Riemann solver:
The corrected flux function is the upwind predicted state:
![Page 26: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/26.jpg)
26Romain Teyssier5th JETSET School
Modified equation for the second order scheme
Taylor expansion in space and time up to third order:
We obtain a dispersive term and the scheme is stable for C<1.
![Page 27: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/27.jpg)
27Romain Teyssier5th JETSET School
Summary: the MUSCL scheme for systems
Compute second order predicted states using a Taylor expansion:
Update conservative variables using corrected Godunov fluxes
![Page 28: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/28.jpg)
28Romain Teyssier5th JETSET School
Monotonicity preserving schemes
We use the central finite difference approximation for the slope:
In this case, the solution is oscillatory, and therefore non physical.
Second order linear scheme.
firstorder
secondorder
Oscillations are due to the non monotonicity of the numerical scheme.
A scheme is monotonicity preserving if:
- No new local extrema are created in the solution
- Local minimum (maximum) non decreasing (increasing) function of time.
Godunov theorem: only first order linear schemes are monotonicity preserving !
![Page 29: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/29.jpg)
29Romain Teyssier5th JETSET School
Slope limiters
Harten introduced the Total Variation of the numerical solution:
Harten’s theorem: a Total Variation Diminishing (TVD) scheme ismonotonicity preserving.
Design non-linear TVD second order scheme using slope limiters:
where the slope limiter is a non-linear function satisfying:
![Page 30: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/30.jpg)
30Romain Teyssier5th JETSET School
No local extrema
We define 3 local slopes: left, right and central slopes
and
x
ui
ui+1
ui-1
New maximum !
For all slope limiters:
![Page 31: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/31.jpg)
31Romain Teyssier5th JETSET School
The minmod slope
x
ui
ui+1
ui-1
Linear reconstruction is monotone at time tn
Slope limiting is never truly second order !
![Page 32: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/32.jpg)
32Romain Teyssier5th JETSET School
The moncen slope
x
ui
ui+1
ui-1
Extreme values don’t overshoot initial average states.
![Page 33: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/33.jpg)
33Romain Teyssier5th JETSET School
The superbee slope
Predicted states don’t overshoot initial average states.
TVD constraint is preserved by the Riemann solver.
The Courant factor now enters the slope definition.
![Page 34: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/34.jpg)
34Romain Teyssier5th JETSET School
The ultrabee slope
Use the final state to compute the slope limiter.
Upwind Total Variation constraint.
Strict Total Variation preserving limiter.
![Page 35: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/35.jpg)
35Romain Teyssier5th JETSET School
Summary: slope limiters
first order
minmod moncen
superbee ultrabee
![Page 36: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/36.jpg)
36Romain Teyssier5th JETSET School
Non linear systems: characteristics tracing.
x
ui
ui+1 Non-linear Riemann problems: waves speedsdepend on the input states.
TVD schemes are not necessary monotone.
Modify the predictor step according to the localRiemann solution: Piecewise Linear Method(PLM) and Piecewise Parabolic Method (PPM).
If
else
If
else
![Page 37: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/37.jpg)
37Romain Teyssier5th JETSET School
Multidimensional Godunov schemes
2D Euler equations in integral (conservative) form
Flux functions are now time and space average.
2D Riemann problems interact along cell edges:
Even at first order, self-similarity does not apply to theflux functions anymore.
Predictor-corrector schemes ?
![Page 38: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/38.jpg)
38Romain Teyssier5th JETSET School
Perform 1D Godunov scheme along each direction in sequence.
X step:
Y step:
Change direction at the next step using the same time step.Compute Δt, X step, Y step, t=t+Δt Y step, X step t=t+Δt
Courant factor per direction:
Courant condition:
Cost: 2 Riemann solves per time step.Second order based on corresponding 1D higher order method.
Directional (Strang) splitting
![Page 39: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/39.jpg)
39Romain Teyssier5th JETSET School
Unsplit schemes
Godunov schemeNo predictor step.Flux functions computed using 1DRiemann problem at time tn in eachnormal direction.2 Riemann solves per step.Courant condition:
Runge-Kutta schemePredictor step using the Godunovscheme and Δt/2.Flux functions computed using 1DRiemann problem at time tn+1/2 ineach normal direction.4 Riemann solves per step.Courant condition:
Corner Transport UpwindPredictor step in transverse directiononly using the 1D Godunov scheme.Flux functions computed using 1DRiemann problem at time tn+1/2 ineach normal direction.4 Riemann solves per step.Courant condition:
Second order schemes: multidimensional Taylorexpansions and multidimensional slope limiters.
![Page 40: Numerical Solutions for Hyperbolic Systems of Conservation ... · Spectral methods can show exponential convergence. More flexible approaches: use ultra-high-order shock-capturing](https://reader035.fdocuments.net/reader035/viewer/2022071403/60f54364f179d975d23d1245/html5/thumbnails/40.jpg)
40Romain Teyssier5th JETSET School
Beyond second order Godunov schemes ?
Smooth regions of the flowMore efficient to go to higher order.Spectral methods can show exponential convergence.More flexible approaches: use ultra-high-order shock-capturing schemes: WENO, discontinuous Galerkin anddiscontinuous element methods
Discontinuity in the flowMore efficient to refine the mesh, since higher order schemesdrop to first order.Adaptive Mesh Refinement is the most appealing approach.
What about the future ?Combine the 2 approaches.Usually referred to as “h-p adaptivity”.