Dimensions of a CFD computational domain

samiahmed · July 8, 2019, 19:04

Hello,
I have a question about the dimensions of the computational domain in a CFD simulation.

I want to compare the results of a CFD simulation with an experiment done in a wind tunnel where the test section is (2m 1m x 1.5m) but now I cannot proceed further before creating my computational domain.

My question is very simple, in general when we want to compare the results with experiments (wind tunnel) do I have to use the same dimensions of the experimental setup for my computational domain? or do I have to conduct a parametric study to determine the proper dimensions of my CFD simulation?

I hope my question is clear

FMDenaro · July 9, 2019, 04:24

Quote:

Originally Posted by samiahmed

Hello,
I have a question about the dimensions of the computational domain in a CFD simulation.

I want to compare the results of a CFD simulation with an experiment done in a wind tunnel where the test section is (2m 1m x 1.5m) but now I cannot proceed further before creating my computational domain.

My question is very simple, in general when we want to compare the results with experiments (wind tunnel) do I have to use the same dimensions of the experimental setup for my computational domain? or do I have to conduct a parametric study to determine the proper dimensions of my CFD simulation?

I hope my question is clear

First of all, you should compute the Reynolds number and using the same value in your CFD simulation. Then, the boundary conditions should reply the experiment.

samiahmed · July 9, 2019, 05:33

I want to reproduce the results of the experiment numerically for the full scale geometry.
The test section of the wind tunnel is (2m x 1m x 1.5m). The chord of the airfoil is 0.05m.

The velocity and other parameters are known (Thus Reynolds number ...etc are known for the experiment). my question is: my computational domain should be (2m x1m x1.5) like the test section of the wind tunnel. Or do I have to use large dimensions for my computational domain

Thanks

sbaffini · July 9, 2019, 09:43

It depends, most of the times experimental results are already presented with all the due corrections to be actually representative of reality. Yet, there might be ambiguities as in my sentence itself (like, what's reality in the first place?).

In this case (data with all due corrections), you should try to replicate the conditions that the experiment was trying to replicate.

My favourite route, however, is typically the opposite, but it is only possible if you have the raw data from the experiment (i.e., no corrections for the tunnel blockage, etc.). In this case you try to exactly replicate the experiment, inlcuding the test section, etc.

In more general terms it kind of depends from the reason you want to compare to experiment and, of course, from the kind of data you have available, as you might not actually have a choice.

CFD practitioners typically rely on well known experimental databases (i.e., ERCOFTAC) which are specifically compiled for the comparison with CFD, so you find all the details on how you should actually perform your simulation.

Expect troubles if the data you are using were not meant to support CFD validation. Otherwise, it is typically clear from the experiment description what you have to do to actually simulate it. In some cases you might be able to infer the correct strategy, but that's not always the case.

The very point here is how much you know about the domain where your computation is going to be performed and the relative boundary conditions. It is the experiment description that should be clear on this. If it is not than it is not validation material for CFD.

July 8, 2019, 19:04	Dimensions of a CFD computational domain	#1
samiahmed New Member Join Date: Jun 2019 Posts: 15 Rep Power: 7	Hello, I have a question about the dimensions of the computational domain in a CFD simulation. I want to compare the results of a CFD simulation with an experiment done in a wind tunnel where the test section is (2m 1m x 1.5m) but now I cannot proceed further before creating my computational domain. My question is very simple, in general when we want to compare the results with experiments (wind tunnel) do I have to use the same dimensions of the experimental setup for my computational domain? or do I have to conduct a parametric study to determine the proper dimensions of my CFD simulation? I hope my question is clear fredericgaillard likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Periodic Pressure drop	cfd_begin	CFX	10	May 25, 2017 08:09
Monte Carlo Simulation: H-Energy is not convergating & high Incident Radiation	volleyHC	CFX	5	April 3, 2016 06:41
Error finding variable "THERMX"	sunilpatil	CFX	8	April 26, 2013 08:00
Gruber computational domain	k.anand	Main CFD Forum	0	April 30, 2011 03:07
Relation between CFD and computational chemistry	Marc Segovia	Main CFD Forum	3	May 26, 1999 06:13

July 9, 2019, 05:33		#3
samiahmed New Member Join Date: Jun 2019 Posts: 15 Rep Power: 7	I want to reproduce the results of the experiment numerically for the full scale geometry. The test section of the wind tunnel is (2m x 1m x 1.5m). The chord of the airfoil is 0.05m. The velocity and other parameters are known (Thus Reynolds number ...etc are known for the experiment). my question is: my computational domain should be (2m x1m x1.5) like the test section of the wind tunnel. Or do I have to use large dimensions for my computational domain Thanks

July 9, 2019, 09:43		#4
sbaffini Senior Member Paolo Lampitella Join Date: Mar 2009 Location: Italy Posts: 2,195 Blog Entries: 29 Rep Power: 39	It depends, most of the times experimental results are already presented with all the due corrections to be actually representative of reality. Yet, there might be ambiguities as in my sentence itself (like, what's reality in the first place?). In this case (data with all due corrections), you should try to replicate the conditions that the experiment was trying to replicate. My favourite route, however, is typically the opposite, but it is only possible if you have the raw data from the experiment (i.e., no corrections for the tunnel blockage, etc.). In this case you try to exactly replicate the experiment, inlcuding the test section, etc. In more general terms it kind of depends from the reason you want to compare to experiment and, of course, from the kind of data you have available, as you might not actually have a choice. CFD practitioners typically rely on well known experimental databases (i.e., ERCOFTAC) which are specifically compiled for the comparison with CFD, so you find all the details on how you should actually perform your simulation. Expect troubles if the data you are using were not meant to support CFD validation. Otherwise, it is typically clear from the experiment description what you have to do to actually simulate it. In some cases you might be able to infer the correct strategy, but that's not always the case. The very point here is how much you know about the domain where your computation is going to be performed and the relative boundary conditions. It is the experiment description that should be clear on this. If it is not than it is not validation material for CFD.