I am performing a CFD study to analyze the effect of different swirl numbers on the shock structure in a CD nozzle.I have been trying to simulate the shock for very long time.I have obtained satisfactory results for over-expanded flow in the nozzles with and without the swirl, however when i try to simulate the shock the solution never converges(flat line with oscillating residuals). I have tried a lot of stuff from changing the turbulence model to changing the under relaxation factors but to no avail.My mesh quality(created using ICEM) is pretty good and has close to a million nodes.I am certain the mesh is not a problem.My BCs are ,inlet pres=1atm and outlet pressure is 65kPa.Any guidance will be of great help.Thank you in advance!
Thank you for the reply sir.However my problem is solution convergence.I have the critical pressure ratios at which the shocks will be obsereved.But when i try to simulate the same my residuals keep oscillating and do not converge to the desired values.
this might mean that you have unsteady flow. shock boundary layer interaction is an unsteady phenomena and the shock may be moving back and forth. make sure you are resolving the boundary layer well with 30 layers of inflation. start from about 0.1 micron away from the wall.
i am sure you are using the density based solver. transition-sst might be the best for steady calculations.
Yes i am using density based solver.I had been using k-omega SST model.I will try giving inflation with transition SST model and let you know of my results.Thank you for the advice!
Try to reduce the time step and using a truly time dependent solution.. i guess you are solving the problem in a steady state fashion without time marchin.
Hi Julio,
Sir i tried transient simulations and it seems to have worked.I have verified it with an existing paper and the results are close.The shock is also well developed.Although when i tried the same simulation after i troducing swirl at the inlet,the plot which i obtain has two lines for each quantity.(mach number plots,pressure plots and so on).Does this have something to do with the swirl phenomenon?
Quaai, called just julio, drop the ”Sir”. Can you please elaborate more on your explanation ? Providing us with plots and more details? I was not able to follow your thoughts.
Yes julio.Here are the plots.The red plots are normal axial flow and the white plots are flow with swirl(swirl number 0.4)
I see, this does not have anything to do because of the model, I guess is a problem in your postprocessing data. I guess your are plotting these variable along the center line, so I expect to see only one line representing the primitive variables and Mach number. the second line has to be an error in your postprocessing. that is easy to see, since a point cannot have two different pressure values, as well as the other values for the primitive variables.
Check your postprocessing routines.
Do i need to run the sims again?Or just figure out from the post processing?
I did not plot these in the post processor.I just plotted it in the solution component.
The very last one makes a lot of sense, although you have a small oscillation behind the shock. Back to your previous question, you don’t need to run the case again.
Hi julio.I was trying to make these plots in excel as well.Suprisingly i am getting two sets of points for my primitive variables.And the graph again looks funny.Why is fluent giving me two sets of values for each variable.I am still confused.
Hi Quasai, I insist on the fact that you may have a problem in your post processing. I think you are plotting different primitive variables at the same time. Why don’t you take a look carefully at your raw data? It just doesn‘t make sense for me this since one point can only have one single value. Unless your are plotting different things.
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators