Different Initialized Velocity Profile vs CFX-Post Values

aero_head · June 23, 2021, 18:26

EDIT: Sorry, I wasn't thinking; I had posted this to the Main CFD Forum but should have posted it here instead.

Hello Everyone,

I performed a simulation of flow over a flat plate where I used an initialized velocity profile (.CSV file with velocity and distance in Y) as my inlet velocity. It was <10 points of data. I noticed that though the values are correctly shown on the inlet plane in CFX-Post (i.e. the value at the top of the domain is the same as the value at the top for my defined velocity profile), the average value reported in CFX-Post with the function calculator does not match the average value of the profile I calculated in Excel. I also note that when I export the velocity data from the plane, there are over 4000 values.

Any ideas as to what could be causing the discrepancy and/or ways of fixing it? I tried both hybrid and conservative values, but they gave me the same result from the function calculator.

ghorrocks · June 23, 2021, 20:55

There are many ways to calculate the average. I assume your mesh is much finer than the <10 points of data you have in your profile, and I also assume that you have a finer mesh at the wall compared to far away.

If my assumptions are correct then you will get lots of points because these are the CFX mesh node values not your profile points, but are your profile interpolated onto your mesh. And secondly, the average probably does not compare because if you do a nodal average with average(variable)@face it will base the average on the node count (which is biased towards the finer mesh at the wall), compared to areaAve(variable)@face which will do an area average. There are also other types of averaging, line, volume and mass flow. They can all give different values for the average.

aero_head · June 23, 2021, 22:32

Hello ghorrocks,

Both of your assumptions are indeed correct. Thank you so much for taking the time to explain how the profile is interpolated as well as how the averaging is performed. It is much clearer in my mind now.

June 23, 2021, 18:26	Different Initialized Velocity Profile vs CFX-Post Values	#1
aero_head Senior Member Kira Join Date: Nov 2020 Location: Canada Posts: 435 Rep Power: 9	EDIT: Sorry, I wasn't thinking; I had posted this to the Main CFD Forum but should have posted it here instead. Hello Everyone, I performed a simulation of flow over a flat plate where I used an initialized velocity profile (.CSV file with velocity and distance in Y) as my inlet velocity. It was <10 points of data. I noticed that though the values are correctly shown on the inlet plane in CFX-Post (i.e. the value at the top of the domain is the same as the value at the top for my defined velocity profile), the average value reported in CFX-Post with the function calculator does not match the average value of the profile I calculated in Excel. I also note that when I export the velocity data from the plane, there are over 4000 values. Any ideas as to what could be causing the discrepancy and/or ways of fixing it? I tried both hybrid and conservative values, but they gave me the same result from the function calculator.

June 23, 2021, 20:55		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	There are many ways to calculate the average. I assume your mesh is much finer than the <10 points of data you have in your profile, and I also assume that you have a finer mesh at the wall compared to far away. If my assumptions are correct then you will get lots of points because these are the CFX mesh node values not your profile points, but are your profile interpolated onto your mesh. And secondly, the average probably does not compare because if you do a nodal average with average(variable)@face it will base the average on the node count (which is biased towards the finer mesh at the wall), compared to areaAve(variable)@face which will do an area average. There are also other types of averaging, line, volume and mass flow. They can all give different values for the average. aero_head likes this. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CFX Post Velocity in Tubes in Shell-and-Tube-Heat-Exchanger	TimNum	CFX	17	March 27, 2019 07:32
CFX Velocity Profile Data	semih ersan	CFX	3	September 7, 2018 14:29
Problem with assigned inlet velocity profile as a boundary condition	Ozgur_	FLUENT	5	August 25, 2015 05:58
Post Processing - Measure Velocity Profile in Particulate flow	yesaswi92	STAR-CCM+	0	October 1, 2014 16:57
Accessing cell values (CFX Pre and CFX Post)	brahmarishiraj	CFX	14	June 24, 2014 20:12

June 23, 2021, 22:32		#3
aero_head Senior Member Kira Join Date: Nov 2020 Location: Canada Posts: 435 Rep Power: 9	Hello ghorrocks, Both of your assumptions are indeed correct. Thank you so much for taking the time to explain how the profile is interpolated as well as how the averaging is performed. It is much clearer in my mind now.