A53. Non-stationary convergence test
It was demonstrated by Claessens et al. (??) that
in case of a rapidly changing wind velocity the results of Swan
are very sensitive to numerical choices such as time step and number
of iterations per time step.
In their test the wind velocity changed from 8 m/s to 18 m/s in one hour.
They simulated a front passage.
In this series of tests we simplify even further. We assume a homogeneous
situation, represented by only one computational point. Claessens et al.
do not give the depth in their tests, we use a (uniform) depth of 10 m.
The graphs in file A53.plt Compare a number of numerical settings.
8 cases are considered:
- Δt = 20 minutes, 1 iteration per time step
- Δt = 20 minutes, 2 iterations per time step
- Δt = 20 minutes, 4 iterations per time step
- Δt = 10 minutes, 2 iterations per time step
- Δt = 5 minutes, 2 iterations per time step
- Δt = 2 minutes, 2 iterations per time step
- Δt = 1 minute, 2 iterations per time step
- Δt = 10 minutes, 2 iterations per time step, relaxed limiter (=0.2)
- Claessens, E.J., J.P. de Waal, H.C. van Twuiver and M. Bottema (??)
- The application of a coupled wind, waterlevel and wave model
in warning system against flooding,
Directorate-General of Public Works and Water Management, Institute for Inland
Water Management and Waste Water Treatment (RIZA), Lelystad, the Netherlands