[alex_03]

Hi,

I’m running simulations of flow past a partially submerged vertical cylinder using ANSYS Fluent.

The setup I have is:

• Multiphase model: VOF (implicit, with implicit body force)
• Turbulence model: SST k-ω SBES with curvature correction
• Geometry: fixed-end, surface-piercing circular cylinder

One of the main quantities I’m tracking is total drag on the cylinder.

I’ve been testing momentum discretisation schemes, and I’ve run into an issue I can't seem to figure out how to fix.

Bounded Central Differencing (BCD)

• After ~1 s, drag looks as though it is going to reach a steady value.
• Then it transitions into a low-frequency, damped sinusoidal oscillation – amplitude decreases over time, trending toward a mean.
• On finer meshes (Stage 3), the oscillations are much less pronounced and the drag looks closer to “settled.”
• These values are actually in the ballpark of what I expect from literature (~20N).

Second Order Upwind (SOU)

• Drag force is much steadier in time (no big oscillations).
• But the magnitude is consistently too low, and mesh refinement makes it worse – finer meshes push the drag farther from expected values, and the gap between refinement stages actually increases instead of shrinking (opposite of convergence).

I have included screenshots of drag vs time plots for the three refinement stages of both momentum discretisation schemes to help visualise what I am trying to explain above.

Right now it feels like I’m stuck choosing between better drag values but with that damped sinusoidal behaviour present for the coarse refinement stages (BCD) or stable time histories but wrong mean values that diverge with refinement (SOU).

I was hoping someone would know how I can preferably get rid of the oscillations in the BCD simulations or at least damp them out more quickly, without having to fall back on SOU and its under-prediction issue.

I'd really appreciate any advice.