I am modeling a gas-in-gas process, gas exiting from a cirular tube into still air.

I am using a 3D slice with periodic boundary condition and an axis boundary condition.

Could someone tell me what does an axis boundary condition do? Is it symmetry or some kind of reflecting boundary condition?

how are the flow variables computed near the axis type?

Thanks for the clarification

CFDtoy

HI cfd toy i think you are modeling 3-d, generally axis boundary condition is applicable for 2-d geomtry. for 3-d symetry is applicable, since we have planes so it will consider as symetry plane which ever you define. and also if u are modelling flow in which hot fluid flows to comparatively ambient it is not advisible to use symmetry condition ok, this is according to my knowledge bye

Hello Ravindra: Thanks for your reply. I am modeling 3D flow with more than 5 cells in circumferential direction. Now, at the axis, which is actually an edge, I was wondering how the flow variables are calculated.

I know axis is not a boundary condition. But I am not sure about the implementation. Could you detail me on some numerics into axis type boundary implementation?

Thanks

CFDtoy

hi if you are doing 2-d axi-symmetric problem then only you can calculate the flow variables at that axis,if you give axis as one of the boundary condition. but if you are doing computations in 3-dimension then i think you should select that plane of symmetry by using symmetry boundary condition and try to calculate the flow variables at that plane. this is ny opinion. bye

hi cfdtoy i am finding temperature and velocity distribution of engine plume.i have used k-e method for solving it. the k and e residuals are oscillating around 1e-2 and continuity is oscillating around 1e-3 for some eighty thousand iterations. can u give me any idea how to bring the residuals down. whether it means that the solution may not converge at all? waiting for your reply ravindra

Reduce ur under relaxation. Typically, k,e use 0.4 or less. Not higher. What kind of convective schemes are you using? Use Upwind to see if the diffusion effects lead to convergence. If you get convergence with upwind then we can start playing with other things. If in effect you are unable to obtain convergence with upwinding process, it would then imply that even with sufficient numerical diffusion your solution diverges.

Please let me know once you are through these steps. Normally, k and eps models are strongly dependent on the boundary and intiial conditions as well.

Btw, Are you running transient or steady state?

CFDtoy!

hi im using first order upwinding scheme for density,momentum, energy... etc. pressure and velocity coupling i have used the simple algorithm, since im simulating only steady state flow. i will apply and see , what wll happen and come back to u. thanks for reply ravindra