Hi everyone,

I am trying to simulate the case about flow around obstruction, which need the fully-developed flow when approaching the obstruction, however, for the inlet boundary condition of specified velocity, only the uniform distribution is allowed, so how to set up the Log-law distribution at the inlet boundary. Thanks a lot.

mice

There are at least two ways I could do that: 1) The easy way -> Given your hydrodynamic conditions, calculate the length that it is required to get fully developed conditions (given as a function of the Hydraulic Radius or something like that), and put an additional block in F3D, so make sure you have Fully developed turbulent conditions where you have your obstacle. You may need to check and compare against an analytical solution to see if your mesh resolution is working.

2) The not so easy way -> In the utility folder (~/FLOW3D/source/utility) there is a file bc.F, you can add a code there that depending on smooth or rough wall, you can calculate a parabolic profile and then put that as a boundary condition, but you have to check that you may need also boundary conditions for TKE and DTKE, right? Besides, you may still need to add a portion of straight channel upstream of your obstacle, since your implementation won't be perfect as desired due to lateral walls, etc, etc. However the length until you have fully developed conditions will be smaller for sure. We had a similar problem, where we had velocity measurements upstream and we add some lines into bc.F to input those measurements as boundary conditions.

Good luck

Jorge

Referencia: Muller, D., Abad, J. D., Garcia, C., Gatner, J. W. and Garcia, M. H. (2007) "Errors in Acoustic Doppler Velocity measurements caused by flow disturbance". Journal of Hydraulic Engineering, 133 (12): 1411-1420

Hi Jorge,

Thanks for your detailed reply. For your suggestion of methods 1), what is the bottom boundary condition?, if the bottom boundary condition is set up to WALL, there is no roughness option provided by WALL type, the velocity distribution closed to the wall or bed is uniform not the Log-law distribution; if adding an obstacle over the bed, the roughness of that obstacle can be set up, however, because of the existing obstacle the inlet boundary is blocked partially, it seems like that the channel length needed to reach the fully-developed condition depends on the height of that obstacle ( or the degree blocking the inlet boundary). For methods 2),could you paste or send your code here or to me,I think that it is very useful for me to read your code; could you describe the portion of straight channel that you mentioned. By the way, I am simulating the scour around cylinder, have you done the similar cases?

Thanks again

Regards,

micetorice