[Jon Faried]

Hi all!

I'm running a quick validation simulation on the DARPA Suboff submarine, and have split the general surface area into multiple surfaces (hull, sail, rudders) in order to see the drag applied on each one.

For some bizarre reason, I've been getting negative pressure drag on the rudders (friction drag seems fine), which makes no physical sense. The direction of the drag, as specified in the drag reports, is correct (and is in the same direction as the hull and the sail, regardless), so why might I be getting a force applied in the opposite direction? Any leads would be greatly appreciated.

Thanks in advance!

Jon

[Jon Faried]

Forgot to note something important: it seems like the total drag (from summation of all forces applied on the various surfaces) is giving feasible results, but once the problem is split into surfaces, the pressure drag doesn't make sense. (i.e. the negative force on the rudders is offset by a larger positive force on the hull)

[fluid23]

Are you sure you aren't looking at axial force rather than drag? Drag will point along your velocity vector, axial force points along the chord of the rudder section. You can definitely have axial forces that point upstream (i.e. negative), however, when transformed into the flow axis you will get a negative drag because the normal force is much larger and tilts downstream a little.

In some airfoils you can end up with leading edge suction that is strong enough to dominate the other pressure forces and result in a negative pressure drag value. However, your friction drag will dominate that so the net is still a positive total drag. If your pressure+friction drag makes sense then you are probably OK. It's hard to say for sure without more details of what you are modeling, how you are modeling it and how you are obtaining your forces.

If you end up with negative total drag (press+fric) then you definitely screwed up somewhere.

[Jon Faried]

Hi, fluid23, and thanks for the reply! I'll try to clarify a bit:

1. The report used is a Force report, in the direction of -X, which is the direction of the flow (i.e. no angle of attack or drift). This considered, is it wrong to use the Force report? If so, what alternative would you suggest?

2. The negative pressure is certainly dominating, to the extent that for specific surfaces (e.g. the rudders) the total drag is (very) negative, to the extent that it's apparent that something is certainly wrong.

3. I'll be happy to follow up with any details you think might be useful. Thanks so much for your help!

[fluid23]

1. Force report is the right choice. I would double check your input parts to make sure you are capturing what you planned on. For a sanity check you could always place probes along the surface at a mid-span chord line to obtain pressure and shear distributions. Integrating those to get lift and drag can give you a reality good understanding of what is going on.

2. I thought you said the total drag looked feasible...

3. It might be helpful if you explain your model setup a bit (i.e. steady/transient, laminar/turbulent, mesh and physical models, etc...) I would also be curious to know what your mesh looks like. Did you meet y+ targets? Are you capturing the wake properly?

[Jon Faried]

Thanks again!

1. Which integral did you use for drag and lift? A surface integral on the probes? I checked the pressure and skin friction coefficient distribution on the mid-chord span line and both they seem fine.

2. By 'total drag' I'm referring to the total drag on the hull, sail, and rudders together. The problem here is that when I try to split the drag into individual regions, I'm getting negative pressure forces which strike me as unfeasible.

3. Steady, segregated, turbulent (K-omega SST) flow with y+ values ranging from 1 (on the hull, where I added a few prism layers) to approx. 300 (on a few trimmed cells on the rudders and sail, since adding the prism there was giving me hell). Surface remesher with trimmed cells and automatic surface repair. Attaching a few images for your convenience shortly...

[Jon Faried]

Hmmm, it seems I can't attach photos here. Shame. In any case, the wakes are certainly captured properly

[fluid23]

1. See attached.

2. You could see local negative pressure drag due to suction. I would definitely include friction in those estimates.

3. 300 y+, even locally, can mess up an SST analysis. I would try to bring that down closer to 60 if you can. Modify geometry slightly, use different surface mesh parameters, but try to bring it down.