|
[Sponsors] |
December 3, 2018, 16:47 |
Convergence issues despite refinement cycles
|
#1 |
New Member
Join Date: Sep 2018
Posts: 5
Rep Power: 8 |
Hello everyone,
I am currently designing a front wing that I want to optimize through CFD. I have done research about my problem before posting. Here are the parameters of my simulation : - Steady state - Constant density (air at 25°C) - Segregated flow - SST K-Omega turbulence - All y+ wall treatment (all wall Y+ values on the wing are below 1) - 6.5 million polyhedral cells with many surface & volumetric controls around the body and in the wake zone. The things I have noticed about this simulation are the high amount of residuals even with some mesh refinement cycles and the non-convergence of the physical values I want to measure. First, I have created a threshold for each residual (X Y Z Momentum, Specific Dissipation Rate and Turbulent Kinetic Energy) to show them on a geometry scene. I used the "Query" command to have the min and max value of each residual and took the 20% extreme values to define the thresholds. Then I created a volumetric control around the concerned cells and set the cell size to a smaller value. I have repeated this cycle 5 times and found out that the RMS values showed on the Residuals plots decreased a little bit at each refinement cycle, reaching a non-normalized value of about 1E-04 (except for the Sdr). However the auto normalized values are about 1E-01. I think that there is not much more to do according to the enclosed screenshot named "Residuals visualizing". The Momentum residuals are in the far wake but the top 20% of the TKE are still just behind the body. The physical values I want to measure (Drag and Lift) do not seem converged as they oscillate randomly. I have plotted pressure contour around the body and it seems rather logical. I have checked the mesh with a complete diagnostic and used thresholds to show Skewness angle (>85°), Cell quality (<1E-05) and Volume change (<1E-02) but no cell is concerned and the mesh is topologically valid. I do not think I can consider this simulation converged. What else do you think I could do to make it converge ? Thank you in advance. Please find some screenshots below. Last edited by karchouille; December 4, 2018 at 04:11. |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
decomposePar problem: Cell 0contains face labels out of range | vaina74 | OpenFOAM Pre-Processing | 37 | July 20, 2020 06:38 |
Influence of the boundary layer refinement on convergence | MrNavierStokes | FLUENT | 0 | June 18, 2015 06:54 |
Tutorial on Fluidized Bed has convergence issues | luca.delbene | STAR-CCM+ | 0 | December 2, 2014 07:05 |
killed "snappyHexMesh" | parkh32 | OpenFOAM Pre-Processing | 2 | April 8, 2012 18:12 |
Convergence Issues 2D Airfoil | AeroEngiNerd | STAR-CCM+ | 1 | February 28, 2012 10:34 |