[PierceH]

Hello,


I am trying to replicate solving the X-velocity field using a UDS. The initial case study is simple, using an inviscid case study and specifying the source term, S = -dp/dx, and turning off the diffusion term. This leaves the UDS to solve for velocity.


Away from the body this is fine, however, near the inviscid wall there are spikes in data. I'm guessing that this is due to corrections in the momentum equation that cannot be implemented to a UDS. Is there any way that I can work this to calculate a velocity field that will agree well with that of the solution resolved by the main flow solver?


Thanks!

[Martin_Sz]

The spikes in data you're encountering near the inviscid wall in your UDS solution are likely due to the absence of viscous effects in your model. Here are some ways to address this issue and improve the agreement between your UDS solution and the main flow solver:
* Implement a Near-Wall Treatment: While a full viscous term might not be necessary, consider incorporating a near-wall treatment that mimics the effect of the boundary layer. This can help dampen the unphysical spikes near the wall.
* Relaxation Techniques: Apply relaxation techniques to the UDS near the wall. This progressively reduces the influence of the UDS solution as you approach the wall, allowing the main flow solver's results to dominate in that region.
* Alternative UDF Formulations: Explore alternative UDF formulations that account for wall-bounded effects to some extent. These formulations might include empirical correlations or analytical approximations for near-wall behavior.
* Consult CFD Documentation: Refer to ANSYS Fluent's documentation for specific guidance on implementing UDFs for wall-bounded flows. There might be recommended practices or examples that address this challenge.