I am interested in H2 flow through a pipeline (transient simulation).
After many failed attemp, I tried to make things small in terms of pipeline length (50 cm), courant number (0.7) and mesh finess (5), however the simulation is quite slow.
Looking at the simulation log, there are 240 time step continuity errors
and 30 PIMPLE: not converged within 4 iterations
.
The residual log of p
seems not alright
The problem is the run stucks at 2% after 30000 minutes, which is really wrong
Link to project
Hey there!
Transient simulations are naturally expensive to run and take lots of time. Are you sure you need a transient simulation? Please consider why do you need the transient response and research ways to achieve the results with a steady-state solution.
If you stick to the transient simulation, one technique that could help to speed-up convergence is to ramp up your inlet boundary condition value, from zero to the final value during a period of time, then allow for your stabilization time.
Another aspect is the simulated time. Typically the end time is picked to allow a number of flow passes through the domain (for example, 2 or 4). So please, estimate your mean flow speed to compute the time it takes a particle to travel from inlet to outlet. Then use a multiple of this value to define your end time to simulate, because I am sure it will be much less than the 5 seconds you currently have.
1 Like