Y-Plus

November 14, 2003, 06:15

hi everybody, i am trying to simulate the surface temperature of a coupling enclosure. thus the heat comes from a rotating coupling (14000 rpm) and now my question is, do i have to use the reynold stress model and which value of Y+ should i have or is it enough when i am only using the k-epsilon turbulence model? what are the values of Y+ for the k-epsilon model? thanks, Florian

November 24, 2003, 00:32

which k-epsilon models do you want to use? high or low-Reynolds number. for high reynolds number: y+ >= 20 and for low-reynolds y+ =< 2.

CFX/CFX-TASCflow guys recommend to use k-omega instead of low-reynolds number model for k-epsilon.

all the best! Check the User manual. email me!

November 25, 2003, 04:30

thanks for your help. now i switched to the sst-model with automatic wall function and with a value of about 100 for y+, that works quite good when i´m using the k-epsilon model for the first 100 iterations.

November 25, 2003, 05:32

The SST model with a Y+ on 100? I had the impression that the SST was a modified k-omega model, which basically means that Y+ should be small also with the scaled wall model ?? Any opinion?

November 25, 2003, 05:43

thats right, the sst model is a mix of k-omega and k-epsilon. in near wall regions the k-omega model is used and in the middle the k-epsilon which can be seen when plotting the first blending factor where 1 means k-omega and 0 for k-epsilon. I have read in an article on the cfx homepage that for sst model and automatic wall function any value for y+ between 2 and about 100 can be used, there will always be the same solution.

November 14, 2003, 06:15	Y-Plus	#1
Florian Guest Posts: n/a	hi everybody, i am trying to simulate the surface temperature of a coupling enclosure. thus the heat comes from a rotating coupling (14000 rpm) and now my question is, do i have to use the reynold stress model and which value of Y+ should i have or is it enough when i am only using the k-epsilon turbulence model? what are the values of Y+ for the k-epsilon model? thanks, Florian

November 24, 2003, 00:32	Re: Y-Plus	#2
KKA Guest Posts: n/a	which k-epsilon models do you want to use? high or low-Reynolds number. for high reynolds number: y+ >= 20 and for low-reynolds y+ =< 2. CFX/CFX-TASCflow guys recommend to use k-omega instead of low-reynolds number model for k-epsilon. all the best! Check the User manual. email me!

November 25, 2003, 04:30	Re: Y-Plus	#3
Florian Guest Posts: n/a	thanks for your help. now i switched to the sst-model with automatic wall function and with a value of about 100 for y+, that works quite good when i´m using the k-epsilon model for the first 100 iterations.

November 25, 2003, 05:32	Re: Y-Plus	#4
Jan Rusås Guest Posts: n/a	The SST model with a Y+ on 100? I had the impression that the SST was a modified k-omega model, which basically means that Y+ should be small also with the scaled wall model ?? Any opinion?

November 25, 2003, 05:43	Re: Y-Plus	#5
Florian Guest Posts: n/a	thats right, the sst model is a mix of k-omega and k-epsilon. in near wall regions the k-omega model is used and in the middle the k-epsilon which can be seen when plotting the first blending factor where 1 means k-omega and 0 for k-epsilon. I have read in an article on the cfx homepage that for sst model and automatic wall function any value for y+ between 2 and about 100 can be used, there will always be the same solution.