|
[Sponsors] |
November 24, 2005, 08:18 |
Data-processing of magnitude & phase data
|
#1 |
Guest
Posts: n/a
|
Hi all,
I would like to know if anyone can recommend useful software tools to analyse flow-field data (magnitude) in 2 spatial directions. I suddenly 'happened' on something during a simulation today & it can only be seen on plots which contains both magnitude & phase information eg. a vector plot. I need to process this information more deeply in order to extract the underlying information. Thanks so much for your assistance. diaw... |
|
November 24, 2005, 23:22 |
Re: Data-processing of magnitude & phase data
|
#2 |
Guest
Posts: n/a
|
Hi, There's plenty of free and GPL tools on Linux that you can use. You can also use MatLab, MathCAD, Mathematica, etc... on Windows. Your dept should have these things. -Edward
|
|
November 24, 2005, 23:32 |
Re: Data-processing of magnitude & phase data
|
#3 |
Guest
Posts: n/a
|
Edward wrote:
Hi, There's plenty of free and GPL tools on Linux that you can use. You can also use MatLab, MathCAD, Mathematica, etc... on Windows. Your dept should have these things. -Edward ---------- diaw's reply: Thanks for your information, Edward. Could you perhaps be more specific in terms of names of the GPL tools, & which Matlab libraries/functions you are refering to? I have Matlab inhouse, but have never used it for this type of work. A pointer in the right direction would be most appreciated. My logic direction is to try & work with 'cutting' (sampling) the data (u,v)@(x,y) in various ways in order to attempt to expose an underlying pattern/s I happened to see whilst viewing that information in another software program (Powerpoint, if you must)... (not at all tailored to my needs, but obviously it just so happens to sample the information at discrete intervals, in just the right way.) This information will provide me (hopefully) a measure of heuristic overview of the flow-field. Thanks for your input. diaw... |
|
November 25, 2005, 23:00 |
Re: Data-processing of magnitude & phase data
|
#4 |
Guest
Posts: n/a
|
Hi, diaw; On MatLab, you use the plotting routines. They will contain what you need. Try it on on the interactive mode, do not program it yet. From what you've said, You may need to write a little bit of code to "window" your data. This is risky on your part because You may accidently erase data that can be crucial to your analysis. What kind of flow is it? All I know it's 2D and that is harmonic in nature. Is it a cylinder flow? or is it an acoustic one.
If you are viewing a plot on powerpoint and you do not have the raw data at all, you may need to use Surfer(graphing software) to get the raw data(Of course, you will have to click on to the points yourself). There is one on Linux called Engauge Digitizer. In this case, what happened to the raw data? Also on Linux, Octave(Matlab compatible Math software); GnuPlot(Plotting software), R(Statistics), Gmsh(Meshing), Paraview and etc... Get a distribution it's will be anice thing to have on your pc. -Edward -Edward |
|
November 26, 2005, 04:55 |
Re: Data-processing of magnitude & phase data
|
#5 |
Guest
Posts: n/a
|
Edward cruz wrote:
What kind of flow is it? All I know it's 2D and that is harmonic in nature. Is it a cylinder flow? or is it an acoustic one. ------------ diaw's reply: Hi Edward, Thanks for your kind reply & excellent information. The flow field is very simple - a 10x10mm square, with inlet west & south, outlets north & east. Entry flow 1mm/s - properties for water. Incompressible fluid. FEM CBS-based solver. Scaling in time & Reynolds number! No convection stabilisation! You will enjoy a rich harvest of acoustic (pressure) & shear wavefields, since the entry flow pushes the flow-field over the onset of instability (predicted theoretically). If you see something, at certain scales that reminds you of flow over a jagged step in a river - then you will be on track. At certain scales, you will see 'lumps & bumps' - what looks like standing wave-forms. There are various groups of these. I have been using the 'wavelet toolbox' as well. This has shown up a few scales of information. We have a classic example of 'chaos'... Welcome to the world of N-S Wave Equations. diaw... |
|
November 27, 2005, 16:30 |
Re: Data-processing of magnitude & phase data
|
#6 |
Guest
Posts: n/a
|
My friend ,
Did u do a fast foutier transform on the data u obtained from the code ? If u need a fft algorithm in fortran I have it. Regards |
|
November 27, 2005, 18:32 |
Re: Data-processing of magnitude & phase data
|
#7 |
Guest
Posts: n/a
|
Vasanth wrote: Did u do a fast foutier transform on the data u obtained from the code ? If u need a fft algorithm in fortran I have it.
--------- diaw writes: No, I haven't done an FFT yet, but would certainly like to. I would really appreciate a copy of your FFT. Thanks so much for your very kind offer. The best e-mail to contact me on is <des@adtherm.com>. Once again, thanks so much. diaw... |
|
November 28, 2005, 21:18 |
Re: Data-processing of magnitude & phase data
|
#8 |
Guest
Posts: n/a
|
Hi Diaw; From your description, I now have a basic idea of your flow. Please be very careful trying to attempt to do FFT of your signal. It may give you false information if you're not careful. Please consult a short book on signal analysis before doing this. -Edward
|
|
November 29, 2005, 05:09 |
Re: Data-processing of magnitude & phase data
|
#9 |
Guest
Posts: n/a
|
Hi Edward,
Thanks so much for your wisdom & insight. I have a few books on hand specialising in Control Theory. Do you have any suggestions for books that could be useful in the 'Signal analysis' field? I have also been brushing up on the Matlab tools, as per your direction. Once again, thanks so much for your very kind input. diaw... |
|
November 29, 2005, 21:59 |
Re: Data-processing of magnitude & phase data
|
#10 |
Guest
Posts: n/a
|
Hi, Diaw; No, the control theory book will not help you here. Try the electronic section of your library. Look for books that deal with digital signal analysis.
Anyway, my concern with FFT analysis is the time interval matching, between your data and the time interval you choose for the FFT. All you need to do is make sure your deltaTime(your data) is not n*deltaTime(FFT process). Send me a plot of your initial FFT plot to check if you did it ok. -Edward |
|
November 29, 2005, 22:15 |
Re: Data-processing of magnitude & phase data
|
#11 |
Guest
Posts: n/a
|
Edward wrote:
Anyway, my concern with FFT analysis is the time interval matching, between your data and the time interval you choose for the FFT. All you need to do is make sure your deltaTime(your data) is not n*deltaTime(FFT process). Send me a plot of your initial FFT plot to check if you did it ok --------- diaw writes: Excellent point. I had not thought about that one. I guess you would be refering to the implications of the Sampling Theorem & correct sampling rate - in terms of aliasing & 'leakage' in the FFT... as in 'time-sampling'. (I learned about this in an Experimental Methods class a while back). If that is the case, then I will certainly need to be careful - I had not got that far yet... Am I one the correct path? diaw... |
|
November 30, 2005, 22:37 |
Re: Data-processing of magnitude & phase data
|
#12 |
Guest
Posts: n/a
|
Hi, Diaw; Yes, you are! But be very careful when interpreting your results. -Edward
|
|
December 1, 2005, 00:39 |
Re: Data-processing of magnitude & phase data
|
#13 |
Guest
Posts: n/a
|
Edward Cruz wrote:
Hi, Diaw; Yes, you are! But be very careful when interpreting your results. -Edward ------------ diaw replies: Thanks very much for your very kind assistance on this score. I would like to e-mail you results as they emerge, for your insights. Can e-mail me an e-mail address to send the data to, as this forum doesn't seem to allow attachments to internal e-mails. Once again, thanks very much for your kind assistance. diaw... |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Post Processing Fluent Data | Jonathan Vogt | FLUENT | 2 | August 18, 2017 02:58 |
【Help】"Error: Update_Time_Level: invalid data" | Chen | FLUENT | 2 | August 24, 2014 07:51 |
post processing using fluent data | Ramesh.K | FLUENT | 3 | October 29, 2008 11:20 |
DNS / LES data processing | Carlos | Main CFD Forum | 0 | April 3, 2008 03:29 |
help: data format - post processing | Pei-Ying Hsieh | FLUENT | 1 | January 10, 2001 13:25 |