Learning Forum

[Ashen Perera]

Hi, 

I am doing a CFD of a shell and tube heat exchnager ( simple model with 7 tubes(. I am using a standard wall function with a K Epsilon standard turbulence model. The target y plus is 30 so I use a Yh of 2.2 mm with 2 layers. 

1. The average y plus is around 14 at the tube surface after CFD post, I can increase it but the inflation mesh coarsens. I did increase but there isnt a signifcant increase ( increases to 19 around). Solutions are converging to reasonable results ( plot is shown below). 
2. I am able to easily refine the wall mesh to get Y plus values around 5 , then is it reasonable to use the standard wall function in this case. As in literature i see such cases. Or should I opt for a enhanced wall treatment or scalable.

TIA, 

Ashen 

[Rob]

Have a look at the courses in Learning. They should explain what y+ is for, and why it's not the only thing you should be checking in the mesh. 

[Ashen Perera]

I actually check the meshing parameters , residual convergence and the monitor points.

just confused about why I am easily able to get it down to low values but I am unable to increase the y plus much. 

[Rob]

Have a look at velocity contours. 

[Ashen Perera]

Basically what I identify is the regions of high y plus are associated with high velocities among baffles but those with low y plus are with low velocity , but my doubt is because most courses mention about having y plus above 30 to use wall function - standard but the high y plus isn't obtain through out surface but only at regions of high velocity while at other points y plus is low 

[Rob]

 

 

The aim really is to avoid having y+ between 5 and 30 so you either resolve the boundary layer or use a wall function. 

[Rob]

And to repost for the umpteenth time.... 

You're using CFD to answer a question. So, simulation is used to find temperature, pressure, bendiness, electrickery charge etc on a device. Does not quite meeting the y+ requirements offset against slight variations in geometry (omitted gaps or features), approximations in material properties and boundary conditions (do some reading on what the heat transfer coefficient really is) etc? Do those errors/omissions alter the end result sufficiently that you can't use it?

[Ashen Perera]

Thanks for the reply to the query, my only doubt is in Y Plus so I will try to increase it by increasing the number of inflation layers