Y+ values for engine port analysis- SST k omega

pary07 · January 30, 2018, 07:32

Hello,

I am investigation steady-state CFD on engine port flow analysis using SST k omega (Menter). For mesh independence study, what are the values that I need to check? I know mass flow rate is one. Do we look at Y+ values? and what more do I have to account for?

Thank you so much.
Best,

fluid23 · January 30, 2018, 12:38

Y+ values aren't really part of mesh dependency. You should get your y+ values to be on-target for your initial coarse mesh... after all, you want to model the boundary layer accurately. It will really depend on your wall model, but if you used the default 2 layer model then generally you should target 1-5 and 30-100. However, values between 5-30 are OK. They are interpolated from the near wall and far wall models and will be a little less accurate.

Once you have wall y+ where you need it then you can start to refine your mesh for grid independence. Here you will want to look at your figures of merit which is anything you want to get OUT of the model. So mass flow would probably be one. You might look at pressure drop or forces as well... it really depends on what you are modeling. Look at the reports you have generated for your model to know what to track in mesh dependency.

pary07 · January 30, 2018, 15:55

Yes, to model the boundary layers properly, I wanted to look at the Y+ values. So I wanted to know what range should it lie in. Now as I am using the SST k-omega turbulence model with all Y+ treatment.
I was having a look at the Y+ values for a range of mesh sizes. I observed that the values are in the range 0 - 30 for low (0.5 mm) valve lift increasing from 22(fine) to around 27(coarse) . But on the other hand when I simulated a particular base size from these for medium (5 mm) and high (9mm ) valve lift. I observed that the Y+ is around 70 ( for 5mm lift) and 85 (for 9 mm lift).

Now I am not sure what range it should actually fall in. Like is this ok? Or for investigating flow through engine ports through the valve, what is the range that I should target?

Thank you

fluid23 · January 30, 2018, 16:07

The range I quoted is good. You should be between 1 and 5 and 30 and 100, however, from 5 to 30 is still acceptable just slightly less accurate. These values are true regardless of what you are modeling.

It sounds to me like you have your prism layer thickness as a % of your base size. I would NOT recommend doing this. Since y+ is essentially driven by the local velocity outside of the boundary layer, you should set this to be a specific value that will always give you the y+ you want (assuming your local velocity doesn't change drastically with mesh refinement). This way as you refine your core mesh, the properly modeled boundary layer remains unchanged. The only thing you should watch out for here is getting excessively large near wall aspect ratios. Prisms will always be fairly large and are pretty tolerant of aspect ratio, however, you should still manage it as best you can.

January 30, 2018, 07:32	Y+ values for engine port analysis- SST k omega	#1
pary07 New Member Paresh Join Date: Jun 2017 Location: Austria Posts: 27 Rep Power: 9	Hello, I am investigation steady-state CFD on engine port flow analysis using SST k omega (Menter). For mesh independence study, what are the values that I need to check? I know mass flow rate is one. Do we look at Y+ values? and what more do I have to account for? Thank you so much. Best,

January 30, 2018, 12:38		#2
fluid23 Senior Member Matt Join Date: Aug 2014 Posts: 947 Rep Power: 18	Y+ values aren't really part of mesh dependency. You should get your y+ values to be on-target for your initial coarse mesh... after all, you want to model the boundary layer accurately. It will really depend on your wall model, but if you used the default 2 layer model then generally you should target 1-5 and 30-100. However, values between 5-30 are OK. They are interpolated from the near wall and far wall models and will be a little less accurate. Once you have wall y+ where you need it then you can start to refine your mesh for grid independence. Here you will want to look at your figures of merit which is anything you want to get OUT of the model. So mass flow would probably be one. You might look at pressure drop or forces as well... it really depends on what you are modeling. Look at the reports you have generated for your model to know what to track in mesh dependency. lcarasik likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
oscillatingRotatingMotion: simulation time change with omega and amplitude values	decibelle	OpenFOAM Running, Solving & CFD	3	November 14, 2018 07:28
high LHTC values in Coolant jacket analysis	rcm	STAR-CCM+	7	January 15, 2011 01:21
wrong values in thermal analysis	Trofrensis	STAR-CCM+	0	December 7, 2009 06:41
Steady state Inlet port flow analysis	Zamir Rashid	Siemens	1	May 21, 2006 12:47
Y+ Values in 2D analysis	Alan	FLUENT	3	September 29, 2005 12:35

January 30, 2018, 15:55		#3
pary07 New Member Paresh Join Date: Jun 2017 Location: Austria Posts: 27 Rep Power: 9	Yes, to model the boundary layers properly, I wanted to look at the Y+ values. So I wanted to know what range should it lie in. Now as I am using the SST k-omega turbulence model with all Y+ treatment. I was having a look at the Y+ values for a range of mesh sizes. I observed that the values are in the range 0 - 30 for low (0.5 mm) valve lift increasing from 22(fine) to around 27(coarse) . But on the other hand when I simulated a particular base size from these for medium (5 mm) and high (9mm ) valve lift. I observed that the Y+ is around 70 ( for 5mm lift) and 85 (for 9 mm lift). Now I am not sure what range it should actually fall in. Like is this ok? Or for investigating flow through engine ports through the valve, what is the range that I should target? Thank you

January 30, 2018, 16:07		#4
fluid23 Senior Member Matt Join Date: Aug 2014 Posts: 947 Rep Power: 18	The range I quoted is good. You should be between 1 and 5 and 30 and 100, however, from 5 to 30 is still acceptable just slightly less accurate. These values are true regardless of what you are modeling. It sounds to me like you have your prism layer thickness as a % of your base size. I would NOT recommend doing this. Since y+ is essentially driven by the local velocity outside of the boundary layer, you should set this to be a specific value that will always give you the y+ you want (assuming your local velocity doesn't change drastically with mesh refinement). This way as you refine your core mesh, the properly modeled boundary layer remains unchanged. The only thing you should watch out for here is getting excessively large near wall aspect ratios. Prisms will always be fairly large and are pretty tolerant of aspect ratio, however, you should still manage it as best you can.