LES simulation - Not able to bring down M value of every cell below 0.1

shettysds · February 5, 2018, 15:08

"We tried doing a LES simulation on Tundish (3.18 m*0.8 m*0.78 m). We are refining the cell based on ratio of sgs_tke and total_tke (M = (k_sgs)/(k_sgs + k_res)).

Converting the RANS simulation to a LES simulation we used a field function to refine the mesh at important areas "(${M} > 0.1) ? max(${cellSize}/40, 0.000002) :0" to find the cells with high values of M and splitting them by 40 times i.e. making the cells finer, which eventually resulted in billion cells. Even though cells were as small as 1.6e-5m, there were always some of these cells having M value closer to 1 or 1. when we back traced, we found out that these cells have resolved TKE(k_res) value =0. which basically means M = 1 = (k_sgs)/(k_sgs + k_res) . So no matter how much finer we make the mesh, these cells will always have M value of 1.

Our question would be; Is it fine to have some cells with value of M=1? if yes, then what percentage of these cells are ideal?"

i am getting k_res = 0 in these cells. thus making TKE = K_sgs. which would basically mean that these cells are badly resolved to have M =1. if we had some values for K_res instead of zero, we would have understood it is less than K_sgs so it has to be made finer. with k_res coming as zero at these cells, we are not able to understand why we are not getting any values for K_res at these cells. and if we can ignore these cells.

me3840 · February 7, 2018, 20:39

According to Pope it's acceptable to have the total resolved energy be 80% of the total energy, or 0.2 by your metric. Whether or not you want to follow that is your choice.

However your case is suspicious. Are you using the hybrid upwind discretization? I'm not sure how it's implemented to compute the SGS KE, but the code may lock certain poor cells using the upwind scheme with the SGS, perhaps. Are those areas poor quality?

February 5, 2018, 15:08	LES simulation - Not able to bring down M value of every cell below 0.1	#1
shettysds New Member shetty Join Date: Feb 2018 Posts: 1 Rep Power: 0	"We tried doing a LES simulation on Tundish (3.18 m0.8 m0.78 m). We are refining the cell based on ratio of sgs_tke and total_tke (M = (k_sgs)/(k_sgs + k_res)). Converting the RANS simulation to a LES simulation we used a field function to refine the mesh at important areas "(${M} > 0.1) ? max(${cellSize}/40, 0.000002) :0" to find the cells with high values of M and splitting them by 40 times i.e. making the cells finer, which eventually resulted in billion cells. Even though cells were as small as 1.6e-5m, there were always some of these cells having M value closer to 1 or 1. when we back traced, we found out that these cells have resolved TKE(k_res) value =0. which basically means M = 1 = (k_sgs)/(k_sgs + k_res) . So no matter how much finer we make the mesh, these cells will always have M value of 1. Our question would be; Is it fine to have some cells with value of M=1? if yes, then what percentage of these cells are ideal?" i am getting k_res = 0 in these cells. thus making TKE = K_sgs. which would basically mean that these cells are badly resolved to have M =1. if we had some values for K_res instead of zero, we would have understood it is less than K_sgs so it has to be made finer. with k_res coming as zero at these cells, we are not able to understand why we are not getting any values for K_res at these cells. and if we can ignore these cells.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Other] refineWallLayer Error	Yuby	OpenFOAM Meshing & Mesh Conversion	2	November 11, 2021 12:04
Neighboring cells in tetrahedral mesh	vishwesh	OpenFOAM Programming & Development	9	November 10, 2017 08:06
How to use "translation" in solidBodyMotionFunction in OpenFOAM	rupesh_w	OpenFOAM Running, Solving & CFD	5	August 16, 2016 05:27
Fluent RPvar list for molar-weight	pilou	FLUENT	6	February 21, 2011 04:08
boundary condition scalarField in parallel	nlc	OpenFOAM	4	November 5, 2010 20:26

February 7, 2018, 20:39		#2
me3840 Senior Member Join Date: Nov 2010 Location: USA Posts: 1,232 Rep Power: 25	According to Pope it's acceptable to have the total resolved energy be 80% of the total energy, or 0.2 by your metric. Whether or not you want to follow that is your choice. However your case is suspicious. Are you using the hybrid upwind discretization? I'm not sure how it's implemented to compute the SGS KE, but the code may lock certain poor cells using the upwind scheme with the SGS, perhaps. Are those areas poor quality?