Prof. Dr. Matthias Ehrhardt
|
Lecture Winter Term 2018/19:
Numerical Methods for Hyperbolic Problems
(Advanced Topic)
Outline
15.10.2018
Part I: Mathematical Theory
§ 0. Introduction, Examples
0.1 Hyperbolic conservation laws in 1D
0.2 The derivation of conservation laws
- Example 1: 1D gas flow in a tube
§ 1. Scalar conservation laws
1.1 Examples
- Example 1: Linear Advection Equation
- Example 2: Advection equation with variable coefficients
- Example 3: (viscous) Burgers equation
1.2 shock development, weak solutions
- Example 1: Shock development in Burgers equation
16.10.2018
- Definition 1: weak solution
1.3 The Riemann problem - shock waves - rarefaction waves
- Riemann Problem
- Rankine-Hugoniot jump condition
- Example 1: Burgers equation
22.10.2018
- Manipulation with conservation laws
1.4 The entropy condition
- Entropy condition (1st version)
- Entropy functions
- Entropy condition (2nd version)
1.5 Summary
1.6 Further examples of scalar equations
- Example 1: Traffic Flow Model
23.10.2018
- Example 2: Buckley-Leverett equation (two-phase flow)
- Example 3: Kynch equation (sedimentation processes)
§ 2. Linear Hyperbolic Systems
- Definition 1: hyperbolic
- Definition 2: strictly hyperbolic
2.1 Examplee
2.2 Characteristic variables
- Domain of dependence
- p-characteristics
- Simple waves
- Example 1: wave equation
29.10.2018
2.3 Linearization of nonlinear systems
- Example 1: 1D Euler equations of gas dynamics, sound waves
- Example 2: isothermal Euler equations
2.4 The Riemann problem
- Riemann problem
- Rankine-Hugoniot jump condition for systems
2.5 The Hugoniot locus in the phase space
- Definition 1: Phase space, Hugoniot locus
- Example 1:
30.10.2018
§ 3. Nonlinear hyperbolic systems
3.1 Hugoniot locus, shocks
- Riemannp roblem
- Definition 1: Hugoniot locus
- Hugoniot curves
- Theorem 1:
- Example 1: isothermal Euler equations
- Theorem 2: (local) unique solvability of the Riemann problem
05.11.2018
3.2 The entropy condition of Lax
- entropy condition of Lax
- pure nonlinear field
- Example 1: isothermal Euler equations
- Linear degeneration
- contact discontinuity
- modified entropy condition
06.11.2018
3.3 Rarefaction waves, Riemann invariants
- Example 1: isothermal Euler equations
- Calculation of rarefaction waves (1st method)
- General solution to the Riemann problem
- Example 1: isothermal Euler equations
- Theorem 1: (local) unique solvability of the entropy-fulfilling Riemann problem
- Calculation of contact discontinuities and rarefaction waves using Riemann invariants (2nd method)
- Definition 1: p-Riemann invariants
- Example 1: isothermal Euler equations
12.11.2018
3.4 The Riemann problem for the full Euler equations
- isentropic Euler equations
- Riemann invariants
- Solution of the Riemann problem
Part II: Numerical Methods
§ 4. Numerical methods for linear equations
- Grid (points), approximation, hyperbolic mesh ratio
- Finite difference method
- Example 1: explicit Euler method
- Example 2: implicit Euler method
- Example 3: Lax-Friedrichs method
- Example 4: Leapfrog method
- Example 5: Lax-Wendroff method
- General explicit one-step methods Un+1=Hk(Un)
13.11.2018
4.1 Error, Consistency
- Error (function)
- Definition 1: Convergence
- Example 1: explicit Euler method
- Definition 2: local discretization error, consistency (order)
- Example 2: Lax-Friedrichs method
4.2 Stability
- Recursion formula for the error
- Definition 1: Stability
- Theorem 1: Lax-Richtmyer's equivalence theorem
- Example 1: Lax-Friedrichs method
- Definition 2: (numerical) domain of dependence
- CFL (Courant-Friedrichs-Levy) condition
- Example 2: one-sided method
- Example 3: 3-point scheme
- Example 1: Lax-Friedrichs Procedure (continued)
19.11.2018
- Upwind methods
- Example 2: one-sided method
- Example 4: Upwind method
4.3 Calculation of discontinuous solutions
- Example 1: ut+ux=0, methods 1./2. Order, numerical results
- Modified equations (for 1st order methods)
- Example 2: Lax-Friedrichs method
- artificial diffusion
- Example 3: Upwind method
- Error for discontinuous data
- Modified equations (for 2nd order methods)
- Example 4: Lax-Wendroff method
- Example 5: Beam Warming scheme
- Dispersion relation
- Phase velocity, dispersion
- group velocity
20.11.2018
§ 5. Conservative methods for nonlinear equations
5.1 The Lax-Wendroff Theorem
- Example 1: Example 1: Upwind method for ut+u ux=0, numerical Results
- Definition 1: conservative, numerical flow, consistent
- Interpretation
- Example 2: (conservative) upwind method for ut+(u2/2)x=0
- Example 3:Lax-Friedrichs method for ut+f(u)x=0
- Example 4: Lax-Wendroff generalizations for ut+f(u)x=0
- Example 5: Two-step method
- Richtmyer-two step Lax-Wendroff method
- MacCormack method
26.11.2018
- Definition 2: Total variation
- Definition 3: Convergence
- Theorem 1: Theorem of Lax-Wendroff
- Entropy condition
- Example 6: Riemann problem for ut+(u2/2)x=0, generalized upwind
- transonic rarefaction waves
- Approximation of the entropy solution
- Discrete entropy condition
- Numerical entropy flow
- Theorem 2: Limiting solution is the entropy solution
27.11.2018
5.2 The Godunov Scheme
- Example 1: Upwind, Lax-Friedrichs for ut+aux=0
- Example 2: CIR (Courant-Isaacson-Rees) method
- Godunov method
- Calculation of the entropy solution
- Theorem 1: Godunov method satisfies discrete entropy condition
- Corollary 1: Limiting solution of the Godunov method is entropy solution
03.12.2018
5.3 Approximate solution of the Riemann problem
- approximative Riemann solver
- 2 approaches for approximate Godunov method
- Theorem 1: Conservativity and consistency of the approximate Godunov method
- The method of Roe
- Definition 1: Roe matrix
- Advantages and disadvantages of the Roe method
- Construction of a "sonic entropy fix"
- scalar problems
- Example 1:
- Example 2:
- Systems
- Example 3:
- Example 4:
- Construction of a Roe matrix
- Example 5: Roe matrix for isothermal Euler equations
04.12.2018
§ 6. Convergence of scalar methods (nonlinear stability)
6.1 TV-Stability
- concept of convergence
- Definition 1: TV-Stability
- Theorem 1: criterion for TV stability
- Theorem 2: Convergence Theorem
6.2 Stability criteria
- Theorem 3: TVD property of Entropy solutions
- Definition 2: TVD (total variation diminishing) method
- Theorem 4: monotonicity preservation property of entropy solutions
- Definition 3: monotonicity preserving
10.12.2018
- Theorem 5: TVD methods are monotonicity preserving
- Theorem 6: L1-contraction property of entropy solutions
- Definition 4: l1-contracting
- Theorem 7: l1-contracting methods are TVD
- Example 1: left-side Upwind method
- Theorem 8: Monotonicity of Entropy Solutions
- Korollar 1: Minimum-Maximum Property of Entropy Solutions
- Definition 5: monotone
- Example 2: Lax-Friedrichs method
- Theorem 9: monotone schemes are l1-contracting
- Theorem 10: monotone methods have at most Consistency Order 1
- Theorem 11: Limiting solution of consistent, monotone methods is entropy solution
11.12.2018
§ 7. High-order methods (construction of high-resolution TVD methods)
7.1 Artificial viscosity
- Example 1: Lax-Wendroff for ut+aux=0 with Artificial viscosity
- Theorem 1: Linear monotonicity preserving methods have at most consistency order 1
- Corollary 1: Stability concepts from Chapter 6.2 are equivalent for linear methods
- solution-dependent artificial viscosity
7.2 Methods with flux limitation (flux-limiter)
- Example 2: FL: Upwind flux, FH:Lax-Wendroff flux for ut+aux=0
- Theorem 2: TVD methods close to extreme points not 2nd order
- 2nd order resolution
- Theorem 3: TVD criterion for general method with flux limitation
- ``Superbee limiter`` of Roe
- Flux limiter of van Leer
17.12.2018
- Generalization for linear systems
7.3 Methods with slope limitation (slope-limiter)
- Algorithm: reconstruction, evolution, projection
- Theorem 4: TVD criterion for general method with slope limitation
- Example 3: Interpretation of Lax-Wendroff for ut+aux=0
- minmod-slope
- Generalization for linear systems
- Generalization for nonlinear systems
18.12.2018
§ 8. Semi-discrete Methods
- method of lines (MOL)
- Evolution equations for cell averages
- improved approximation of the flux
- Example 1: piecewise linear approximation
- Reconstruction by primitive functions
- ENO (essentially nonoscillatory) method
- Definition 1: ENO-method of order p
07.01.2019
§ 9. Multidimensional problems
- Finite Volume Methods
- Semi-discrete Methods
- Splitting Schemes
- Example 1: scalar Advection equation in 2D
- Splitting of order 1
- Strang-Splitting
- Example 2: Linear conservation law in 2D
- TVD-Methods in 2D
- Theorem 1: TVD-Methoden in 2D have almost consistency order 1
08.01.2019
Part III: Excursion
§ 10. Initial boundary value problems (IBVPs) for conservation laws
10.1 Boundary conditions for linear Systems
- Example 1: linear System in 1D
- Example 2: 1D Euler equations
- sub/supersonic inflow boundary
- Outflow extrapolation
- Example 3: No-flux Boundary conditions in R2
- Example 4: Absorbing Boundary conditions
- Definition 1: strong solution
- Theorem 1: Existence / uniqueness of strong solution of IBVP in 2D
- Definition 2: well-posed IBVP
- Theorem 2: ill-posed IBVP
- Theorem 3: Necessary condition for well-posedness
- Lemma 1: Uniform Kreiss condition
14.01.2019
10.2 Boundary Conditions for Nonlinear Scalar Equations
- Definition 1: BV(Omega)
- Theorem 1: Compactness in L1(Omega)
- Lemma 1: Existence of a trace mapping
- Theorem 2: Convergence of the viscosity solution
- Motivation of the Boundary Conditions
- Definition 2: Weak solution of the IBVP
- Theorem 3: Existence / uniqueness of weak solution of IBVP
15.01.2019
§ 11. A kinetic Approach for hyperbolic Systems
- Example 1: Advection equation
- Example 2: Burgers equation
- Kinetic schemes
21.01.2019
22.01.2019