[aerok_25]

So i have given an inlet condition to the end of the cylinder that is inside the domain. The outlet is given a pressure outlet condition, right left and top wall were given the symmeter boundary condition. The issue is with the upstream wall, i have tried pressure outlet and pressure inlet with both hybrid and standard initialization but im getting bad residuals for 2 of the 3 models that im using ( SA,KE and Komega ), im literally stuck and im not sure on what to do

[Martin_Sz]

Here are some insights that might help you resolve the bad residuals you're encountering with the upstream wall:
Understanding the Problem:
* Pressure Outlet vs. Pressure Inlet: A pressure outlet is typically used at the flow exit where the pressure is known. A pressure inlet is used at the flow entrance where the mass flow rate or velocity is known. Since you have a cylinder inside the domain, using a pressure outlet at the upstream wall might not be suitable.
Possible Solutions:
* Symmetry Boundary Condition:
* If the flow upstream of the cylinder is symmetrical, using a symmetry boundary condition on the upstream wall is likely the most appropriate approach. This assumes the flow on one side of the wall mirrors the flow on the other side.
* Velocity Inlet (if known):
* If you know the velocity profile at the inlet of the domain (upstream of the cylinder), you can set a velocity inlet boundary condition on the upstream wall. This provides more specific information about the incoming flow.
Additional Considerations:
* Hybrid Initialization vs. Standard Initialization: While initialization can affect convergence, it's usually not the primary cause of persistently bad residuals. Focus on refining the boundary conditions first.
* Model Selection: The k-ε (KE) and k-ω (KW) models are Reynolds-Averaged Navier-Stokes (RANS) models suitable for steady-state simulations. However, for complex flows with separation or large pressure gradients, consider using a more advanced model like Shear Stress Transport (SST) variants.
Recommendations:
* Double-Check Boundary Conditions: Ensure that the boundary conditions you've applied are consistent with the actual flow physics. A pressure outlet at the upstream wall might disrupt the flow entering the domain.
* Review Mesh Quality: A poor-quality mesh can lead to convergence issues. Check for excessively skewed or elongated elements, especially near the cylinder and upstream wall.
* Monitor Residuals: Pay attention to which continuity and velocity component residuals are highest. This can provide clues about where the convergence problems are originating.
* Simplify and Test: If you're still struggling, try simplifying your model by removing the cylinder temporarily. Simulate the flow in the empty domain with the proposed upstream wall boundary condition to isolate any issues unrelated to the cylinder's presence.
By carefully considering these points and refining your boundary conditions, mesh quality, and potentially the turbulence model, you should be able to achieve better convergence and resolve the bad residuals you're encountering.

[aerok_25]

Thank you for your detailed reply! I want the upstream wall to represent atomoshpeheric conditions. Additionally I am noticing converging residuals for the SA and Komega models but not for the Kepsilon models ( Diverges after 70 iterations ). The mesh has around 500K elements with a average orthogonality of 0.96 using polyhexcore meshes.