[edwinrajeev]

I am running a water wave simulation with a VOF method (waves2Foam) and trying to compute wave disspiation and forces on a submerged structure. My orbital velocities vary in time from about *+1 m/s to ~0 and then -1 m/s and back over the wave period. As a consequence, my y+ varies from 15 to 2000 over the period. I am using RANS with k-Omega-SST and log law of the wall. I know y+ has to be smaller than 100-300 in order to be in the region where log-law is valid. But I feel like when velocities are small and y+ is 2000, the error in computing the flow field and its impact on shear/pressure should be small, right? * What do you suggest?

[LuckyTran]

I don't follow how velocities can be "low" and y+ be in the high 2000's. If velocity is low, y+ will be low.

Still you should be careful because high y+ doesn't mean arbitrarily high. y+ of 2000 is more outside the boundary layer than inside it, which means you have no resolution of the boundary layer at all and you might as well be running a simulation with slip walls and no need to talk about y+ in the first place.

[Lupocci]

, and since increases with , is large for large velocities. However, since the motion is unsteady, it is a wave, so in the same spot he first has very large velocities and large y+, then small velocities and small y+, and so on. So the suggestion would be to make sure to make sure that for large velocities the max y+ is below 100-300. You will have very small Y+ for small velocities and you will be in the viscous sublayer, where the log law is not valid. Since Edwin is using a lag law of the wall, he needs to make sure to use an approach that works through the whole viscous/buffer/log law region. Is the Spalding approach maybe the best one in Openfoam to deal with this type of unsteady flow? I am not sure since I do not run unsteady flows. Any other suggestions?

[FMDenaro]

Originally Posted by edwinrajeev

I am running a water wave simulation with a VOF method (waves2Foam) and trying to compute wave disspiation and forces on a submerged structure. My orbital velocities vary in time from about *+1 m/s to ~0 and then -1 m/s and back over the wave period. As a consequence, my y+ varies from 15 to 2000 over the period. I am using RANS with k-Omega-SST and log law of the wall. I know y+ has to be smaller than 100-300 in order to be in the region where log-law is valid. But I feel like when velocities are small and y+ is 2000, the error in computing the flow field and its impact on shear/pressure should be small, right? * What do you suggest?

First of all, if you want an accurate evaluation of the wall stress along your body, I am sorry to say that you have only one way: resolving the BL in any phase of the wave cycle, that is ensuring you have 3-4 nodes with y+<1. Differently, your wall stress is wrong.

Using wall-modelled BCs can be useful if your goal is different from computing the drag.
What is more is that your flow problem has a strongly variable local Reynolds number and some transitional regime are plausible.
A strategy could be an adaptive grid, adding points when the velocity increases, however you should be able to resolve the BL when the velocity reaches the max magnitude.

[Lupocci]

Filippo, thank you for your insight. I always read that having y+=1 on the first computational point next to the wall should be enough to resolve the viscous layer. Isn't 3-4 points with y+<=1 overkill? Any reference with a comparison between results with 1 point at y+=1 and 3-4 points with y+<=1, please? Also, does have anyone experience with the Spalding law in Openfoam?

[FMDenaro]

https://www.researchgate.net/publica...ddy_simulation