Von Neumann Stability Analysis
Contents
Von Neumann Stability Analysis#
The von Neumann stability analysis method is simple to apply but it cannot handle boundary conditions.
It consists of replacing the spatial variation by a single Fourier component. The method is sufficient for linear equations with constant coefficients.
We shall illustrate the Von Neumann stability method with the FTCS scheme.
Stability of the FTCS scheme#
The linear advection equation#
The FTCS scheme for the linear advection equation is given by:
To apply the Von Neumann stability analysis method we assume a solution of the form:
Substituing in (38), we get
Eliminating the common factor \(e^{ikm\Delta x}\) and defining the amplification factor as \(|B^{n+1}/B^n|\), we can write:
For the scheme to be stable, we require that the amplification factor be \(\leq 1\):
which is impossible to fulfill, because
Therefore, the FTCS scheme applied to the linear advection equation is always unstable, i.e., it is unconditionally unstable.
The diffusion equation#
The one-dimensional diffusion equation is
where \(D\) is the diffusivity.
The FTCS scheme applied to (40) is:
Substituting a solution like (39) in (41), we have
which, after some manipulation, allows to obtain the following expression for the amplification factor:
The Von Neumann stability condition is:
For \(\tau > 0\), the worst case occurs when \(\sin^2 \frac{k \Delta x}{2}=1\), which leads to
Therefore, the FTCS scheme apllied to the linear diffusion equation (40) is conditionally stable, with the stability condition