1.1.8.3. The Riemann Problem¶
The Riemann Problem is a family of problems, (which includes the 1D Euler Equations).
There are two components for the Riemann problem:
A conservation law (e.g. the 1D Euler Equations)
Piecewise constant initial data with a single jump discontinuity
e.g. for Euler Equations:
For \(x \lt x_0\) \(\quad \mathbf{U}(x, t_0) = \mathbf{U}_L\)
For \(x \gt x_0\) \(\quad \mathbf{U}(x, t_0) = \mathbf{U}_R\)
\(\mathbf{U}_L\) and \(\mathbf{U}_R\) are constant vectors
Take \(x_0 = 0\) and \(t_0=0\)
1.1.8.3.1. Usefulness of the Riemann Problem¶
It has an exact analytical solution for the Euler Equations (and any scalar conservation laws, or any linear system of equations)
1.1.8.3.2. Features of the Riemann Problem¶
Solution is self-similar - solution stretches in space and time, but doesn’t change shape.
\(\mathbf{U}(x, t_1)\) and \(\mathbf{U}(x, t_2)\) are “similar”
In other words, it really depends on a single variable \(x / t\)
Or the solution is constant along a characteristic line \(x = ct\) (c = constant) on x-t plane
Generally, self-similarity means only one independent variable like: \(x / t\) or \(t / \sqrt{x}\), similar to the Blasius solution from boundary layer theory.
There is an analytical solution \(\Rightarrow\) Riemann problem useful to test numerical schemes
Riemann problem appears as part of numerical formulation of several CFD methods, e.g. “wave capturing” methods \(\Rightarrow\) “Riemann solvers” are at the heart of the method - may need to solve these multiple times - major cost of the solution