Hi Foad,

what conditions are you imposing at the outlet of the domain? I am thinking that a variation in the length of the chamber should not affect significantly the flow upstream, unless the outlet gets too close to the swirling flow and alters the flow. Can you provide a sketch of the test article?

Additional question: is the simulation converged with low residuals?

Best regards,

Lorenzo

Hi Foad,

I dont know exactly about your experiment's geometry, but I think it can happen. Definitely, something is beyond the physics of this problem and it might be related to pressure drop or sth like that. For instance, at a constant pressure difference, it woul be possible to increase flow rate from a discharged flow by installing sth to adjust the pressure drop... Lets say diffuser.... However, if I knew your gerometry, we would be able discuss more about it.

btw, for the numerical somulation, you had better set the inlet pressure and based on your problem, mesh quality should be enough to get rid of any numerical losses.

good luck,

Hi Lorenzo, 

Thanks for your feedback,

I clearly understand your points. In start ccm+ I`m stuck with mass flow and velocity inlet boundary conditions and as far as outlet goes, I`m using pressure outlet with 0.0 gauge pressure. In terms of convergence, all are stable and below e-5, expert the dissipation rate which is stable but of e-3 order. the origin of dissipation rate was found to be at the discharge section (please find the figure I have attached) and in particular near the wall and prism layers where the velocity reaches to 50 m/s. However, the smooth transition was considered to the core mesh and also y+ was less than 1.0.

and then, is the swirl chamber distance, which is actually very narrow and our study shows that same as what you mentioned, it will definitely alters the swirling flow field by increasing the regions of negative axial velocity and moving them upstream close to the exit of tangential entries.

considering all these, currently I am using the experimentally measured discharge mass flow as an inlet boundary condition for simulation so that I can make my result to look reasonable. This in turn led to underestimation of spray angle since the surface average pressure from the simulation at inlet has accordingly decreased and does not fit the experiment anymore.

Best Regards,

Hi Hamid, 

Thank you for your answer, 

I found your answer interesting and useful. Please find the schematic I have attached. Unfortunately I am not able to use pressure inlet as boundary condition since it is not provided by star CCM+. Grids are tailored precisely so sadly I have to cross this one out.

Best Regards.

it seems to be complex geometry I think. You need a high quality mesh with appropriate schemes to solve the flow for this case. In Star-ccm I dont know the options but I think there should be an option for your problem. However, I can direct you to a CFD gourp where you can find out more. 

good luck,