Negative cell volumes in Diamond Wing Case

[sean-engelstad]

Issue Type

• [ ] Bug
• [ ] Enhancement

Description

Plan is to run full aerothermoelastic forward analysis with funtofem and ESP/CAPS for diamond wing case. Flow will be laminar meaning viscous forces will be present. However, negative cell volumes appear within the first few iterations due to the fluid cells in the boundary layer being thin triangles. Need to work on improving either the geometry/stiffness settings or the fluid mesh aka pointwise settings for laminar case.

Example

examples/diamondWing/

[sean-engelstad]

We have been messing with the elasticity_gmres settings in the fun3d.nml file for fun3d analysis. The documentation of fun3d suggests that changing some of the mesh elasticity settings can prevent negative cell volumes. However, we were still getting negative cell volumes no matter what in the laminar, diamond airfoil case. We suspected that maybe the sharp corners were playing a role, as nacaWing case does not have any sharp corners, and not a sharpTE. Therefore, I went in and manually chamfered each of the diamond airfoil corners including LE, TE.

After chamfering the diamond wing edges, we did not get the laminar analysis to work yet, but we did get the inviscid to work for the diamond airfoil. Here's the result at mach 2, zero AOA, it just kind of bends backwards ever so slightly.

[sean-engelstad]

Turns out the orientation of xyz axes in my struct and fluid CSM geometries did not match the standard xyz axes from fun3d. Namely I had y vertical, z spanwise, and x along the flow whereas fun3d has x along flow, y spanwise, and z vertical for a wing/aircraft. Therefore, the angle of attack in fun3d caused flow to impinge on the root/tip and cause really weird deformations therefore the fun3d grid convergence would crash and give negative cell volumes. After fixing my orientation, I have rarely gotten negative cell volumes