Trying to validate simulation of flow over a tractor-trailer

kartikarora111 · February 11, 2015, 11:04

Okay so i took up this project for finding drag force over a truck.
I have liitle to no knowledge of cfd.Ihave made the geometry(attached) and have checked out some tutorials on youtube.My mentor said that you need to validate the problem.
Can you help me understand what is validation and how do i do it?
links would also be appreciated.

CeesH · February 11, 2015, 11:11

Hello Kartik,
Validation is checking if your CFD solution is feasible.
Qualitatively, you can see if the vector field looks good (the flow around the vehicle looks like you would expect it to). But, of course a quantitative validation is much, much more desirable. This means looking up experimental data, and see if your calculations are compatible with experiments.

I am by no means knowledgable about the literature related to trucks, but what I can imagine is that there is data available on the drag coefficient (which you can compute with FLUENT as well), perhaps on the flow velocity at some points around the geometry, and about pressure fields in the wake or so. If you can find those you should be able to compare them with your work, hence validate it.

kartikarora111 · February 11, 2015, 11:48

Thanks for the quick reply.
The geometry i have made is completely random. Won't it be a problem to match exp. data with my model because of difference in geometry?

CeesH · February 11, 2015, 13:01

If you need an exact match, yes. But in any way CFD will never give you the exact answer, since the model is full of approximations and assumptions.

So to an order of magnitude you may validate your results.
For example, engineering toolbox gives, for a tractor trailed truck, a value of 0.96 (and a truck in general 0.8-1.0) for Cd. So, if you find a value of 0.5 or 2 your model is probably not very accurate. But 0.9 would be fine.
Also, see if the results depend on your mesh! Try different mesh densities, and see if the values of cd agree (they should converge to a certain value with increasing mesh density)

kartikarora111 · February 11, 2015, 14:07

Thank you so very much.you have been really helpful to me.
Surprisingly CFD sounds interesting now

February 11, 2015, 14:07		#5
kartikarora111 New Member Kartik Arora Join Date: Feb 2015 Posts: 5 Rep Power: 11	Thank you so very much.you have been really helpful to me. Surprisingly CFD sounds interesting now CeesH likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
About Some Concepts:Laminar flow, turbulent flow, steady flow and time-dependent flow	Jing	Main CFD Forum	8	October 5, 2018 17:02
Domain format problem on airfoil flow simulation	andrenonaka	CFX	14	December 7, 2015 00:42
Coupling of drying process and flow simulation	franzdrs	Main CFD Forum	0	April 29, 2013 14:11
Query regarding temperature distribution in Solidworks flow simulation	syarif	FloEFD, FloWorks & FloTHERM	3	February 27, 2013 05:22
Natural convection - Inlet boundary condition	max91	CFX	1	July 29, 2008 20:28

February 11, 2015, 11:11		#2
CeesH Senior Member Cees Haringa Join Date: May 2013 Location: Delft Posts: 607 Rep Power: 0	Hello Kartik, Validation is checking if your CFD solution is feasible. Qualitatively, you can see if the vector field looks good (the flow around the vehicle looks like you would expect it to). But, of course a quantitative validation is much, much more desirable. This means looking up experimental data, and see if your calculations are compatible with experiments. I am by no means knowledgable about the literature related to trucks, but what I can imagine is that there is data available on the drag coefficient (which you can compute with FLUENT as well), perhaps on the flow velocity at some points around the geometry, and about pressure fields in the wake or so. If you can find those you should be able to compare them with your work, hence validate it.

February 11, 2015, 11:48		#3
kartikarora111 New Member Kartik Arora Join Date: Feb 2015 Posts: 5 Rep Power: 11	Thanks for the quick reply. The geometry i have made is completely random. Won't it be a problem to match exp. data with my model because of difference in geometry?

February 11, 2015, 13:01		#4
CeesH Senior Member Cees Haringa Join Date: May 2013 Location: Delft Posts: 607 Rep Power: 0	If you need an exact match, yes. But in any way CFD will never give you the exact answer, since the model is full of approximations and assumptions. So to an order of magnitude you may validate your results. For example, engineering toolbox gives, for a tractor trailed truck, a value of 0.96 (and a truck in general 0.8-1.0) for Cd. So, if you find a value of 0.5 or 2 your model is probably not very accurate. But 0.9 would be fine. Also, see if the results depend on your mesh! Try different mesh densities, and see if the values of cd agree (they should converge to a certain value with increasing mesh density)