Hi, experts..
In ANSYS Fluent CFD, I applied data sampling for time statistics and got time average properties. Contours of time average properties is shown easily. but when I try to draw vectors of time average velocity, there is just a instantaneous vectors colored by average velocity. I mean vectors do not change instantaneous to time average vectors. time average is shown by color only. here is only one option like
Vectors of... velocity. phase.. water.. colored by..Unsteady Statistics.. Mean velocity Magnitude.. phase...water...
But, I need to show original time average vectors.
Please , guide me.. Thank you
Respected Sir,
Thank you very much to spare some of your valuable time to reply my question. Now,
1) Do you mean that when I am applying data sampling for time statistics I must calculate average of X, Y, Z Velocity components too ?( If it is your meaning.. I did not calculate average of these components. but in the next simulation I would apply your expert suggestion.)
After getting time average of velocity magnitude and its x,y,z components.. Should I define Custom Vectors.. to get the overall time average liquid flow field ?
2) Yes, It is also in my mind to explain my question. because I am totally confuse to put the appropriate value of Sampling Interval.. I run 20+ simulation to check myself the value of sampling interval..( i.e., I put 50, 100, 1000, 2000 etc..). In my simulation Time step size is 0.01 and Number of Time Steps is 5000. So, I need to know the appropriate Period Of Time( Time Sampled).. Please
3) Large secondary Oscillations.. I do know how to answer you.. but my system is Air/water. Air velocity-inlet boundary condition velocity magnitude is 0.253 m/s. So, air plume oscillate many times in 50 second simulation. For your kind consideration, I am attaching one picture of flow field here...
I hope you would guide me further.. Thank you again..
Respected Sir,
You can say ! yes.... but experimental work is done by digital high speed camera and PIV ( partial image velocity) not LDA. For comparison previous LDA data also used. Here, I am attaching on graph that is not LDA actually. Experiments on high speed camera and you can easily understand the simulation in bubble column. That is not my simulation, I am overlapping or improving that previously done simulation..
please, give me suggestion for previously asked question part (2), about Time Sampled..
Thank you
very good option you raised..!!!
but How to remove early time steps...?? for example, in 50 seconds simulation, if I need to average after 10 sec, and in between 20 to 50 seconds. how to set data sampling time statistics options.. please guide..
thank you for guidance.. you are technically very strong and informative.. I may contact you again.. thanks..
Dear Muhammad
Interesting question. What about having time averaged velocity component first and then obtain the resultant and its corresponding direction. How long is the period of time that you choose for averaging the speed. Does the flow consist of large secondary flow oscillations?
Abdul Aziz
= = = = = = =
Greetings Muhammad,
Greetings Abdul Aziz,
Let me try to answer your question Mohammad, by answering
Abdul Aziz's questions:
This software you're trying to use, sounds like it has a fault/flaw.
I would suggest you put your data - into Excel. Then you hand calculate it,
using Excel, while using below steps and simple formulas:
- - - - - - - -
First thought, an example - for a very simple case ... for you to follow the math.
A. Assume you have this (X,Y) data in this format: where X = units (of time), and Y = your measured variable.
Now: 1) (1,2), 2)(2,4), 3) (3,6), 4 (4,8), 5 (5,10), 6 (6,12), 7(7,14)
(You should recognize this data - as that of a 'straight line' where y = 2X.
Now, lets assume (for this example) that you want to generate a Simple Moving Average of the first 5 data points.
So, take the value of the "Y's", for the first 5-data points; add them, and divide by '5'.
(2+4+6+8+10)/5 =6
But here's the mistake - a lot of software, and people make. This data "IS NOT"
one-dimentional (in the y direction); but instead it's two-dimentional (because you must have both an x and a y value - to plot the data points.
What this means, simply, is that since you took "the average value" of the "y's"; that you must also do "the exact same average" - of the "Time" values.
So lets show you this step. The first five data points have "X values", (to first sum, and then divide by 5) of:
(1+2+3+4+5)/5 = 3 (units of time.)
This the correct answer, for this data set (of a 5-point moving average).
This: (3,6) is the correct answer - and not the point (6,6)
- - - - - - - -
Now, since this is a moving average - do it for a second one.
You add a new data point to the front ... and subtract off the oldest data point.
2) (2,4), 3) (3,6), 4 (4,8), 5 (5,10), 6 (6,12) .... = (4,8) as the correct answer for the
second set of 5 data points.
- - - - - - -
This moving average has a name. Its called, an "Auto-Regressive, Integrated, Moving-Average") ... for you must move(auto-regress) the time backward ... half a cycle(moving average).
- - - - - - -
Now, this may be enought to have solved your problem for you ... but please be careful. This is a "warning"! You 'have not' generated a 5-point moving average; but instead, you have done a "Filter" of your data. Let me state that one more time. You have "filtered" your data ... by "removing" from it, all signal/data/short cycle components ... having a cyclic length of lest than or equal to 5 units of time!
Now, the good news. You can go back, and 'recover' this data you filtered out of the data set. Its simple to do, to: All you have to do is take your original "Y" value for the third data point ... and subtract from it, your ARIMA value you calculated above.
This second step, described above, has a name. It is called, "Inverse ARIMA".
- - - - - - -
So, you first filtered out 'all signal(s), of 5 or less data points; but you recover the
lost data by subtraction. Meanwhile, the ARIMA value you first calculated - passes thru all signal(s)/cycles having a time-length of greater than (>=) to 5 data points.
- - - - -
Thus, for your problem - you should clearly see the small, quick vector-data (using "Inverse ARIMA" You also, using the ARIMA value - get to see the sum of all time components signal ... of the bulk flow-rate of the water.
I hope I have explained this clearly enought, to make myself understood.
-Bruce Zerr
How to show Time Average velocity Vectors ? - ResearchGate. Available from: httpswww.researchgate.net/post/How_to_show_Time_Average_velocity_Vectors [accessed Sep 12, 2016].
- Badges
- Science topic