Learning Forum

[alpyldz]

Hello, I am trying to conduct a mesh independence study in a simple jet in crossflow analysis, but I have been unable to achieve convergence. My case setup is as shown in the visual below. All my equations have converged up to 10^-6, and there is no change in my control points throughout iterations. I am looking at the minimum temperature and velocity values on the facet created along a line at a certain distance from the jet flow. I have created multiple different meshes by dividing the global sizing values on Ansys Fluent's watertight meshing by 1.2 or 1.5 to reduce them. I have tried meshes ranging from 100 thousand cells to 20 million cells, but I have not been able to capture a specific trend. The discovered values, from coarse mesh to fine mesh along the line, are the lowest temperatures as follows: 460, 452, 424, 433, 402, 370, 380, etc. Even with 20 million cells, the values are still changing. I selected the fluid as air with its properties set as time-dependent. I also tried setting them all as constant, but there was no change. I have tried with all turbulence models and wall functions, and similarly, I have tried all solution schemes, but there is a consistent inconsistency, and the results vary as the mesh changes. I am awaiting your help. Thank you.

[Rob]

Please can you post some images? Given the relative speed your poor convergence could be a transient effect in the flow. 

I've just cleared up the duplicate threads, it's Forum gremlins so don't worry about it. 

[alpyldz]

 

Hello Bob, thank you for cleaning up the duplicate posts. I couldn’t figure out how to delete them. The poor converged case is completely absent, meaning all my cases are converging nicely, as I mentioned. The temperature and velocity values at my control points remain stable. However, when I change the mesh, my results change inconsistently. I understand that it’s normal for results to change when the mesh is altered, but as I refine the mesh, the values don’t show a consistent decrease or increase. For example, my temperature values might be 430, 420, 418, and suddenly, as I refine it further, I see values like 402, and with more refinement, even lower values like 375K.

 

[Rob]

Please can you post some results? If you plot with node values off it'll help show how well the mesh refinement looks. Minimum can also be a misleading measure: in theory the minimum velocity on that line could be zero as it picks up the wall velocity. If you overlay the xy plots from the line how do they compare?