Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

FFT Vibration Data 2

Status
Not open for further replies.

Croco76

Mechanical
Aug 9, 2010
2
thread384-257416
Am new to vibration and any help is greatly appreciated. I have around 1.4 million points of data taken over every 0.002,0.003 secs. and g's (Acceleration) in both X,Y&Z. I need to develop a vibration profile.
How do I transform these (g's vs Time) to Amplitude vs. Freq and eventually to get a PSD vs Freq?

Am familiar with matlab but don't know much about FFT and vibration profiles.
 
Replies continue below

Recommended for you

Your sample rate was only 500 samples/sec.. The F-max would be 250 Hertz or less. Was there a good anti-alias filter ahead of the A/D converter? If not, then don't waste you time converting to the frequency domain.

Walt
 
Thanks Walt. I'm a Thermal Analysis/CFD guy and very naive in mechanical/vibration stuff. Our vibration guy is away and a client needs this data asap. I have no clue as to what anti-alias means..... so if I dont have to go thru. a freq. domain, how do i get this data (Time, g's_x, g's_y & g's_z) into a vibration profile?
1. Do I just plot g^2/freq vs. freq?
2. Is freq. just 1/time?
3. With over 1million points of data, how do i deal with the frequencies.
5. How did you determine that F-max < 250 Hz?
 
Croco,

You don't have any easy task. Here is a link to FFT references:
Q1: Yes after FFT
Q2: Yes, but you can't simply take reciprocal of waveform to get spectrum
Q3: Typically time data is divided into data blacks like 1024 points
Q4: Nyquist, another term you need to know

I use DasyLab for general signal analysis. Perhaps you could do your job with a download demo of that or similar software. You must be ready for an education, either way.

Walt
 
vibrationdata.com will have all the SW (Standalone) you need, but Walt is right, you need to know what you're doing.
 
"How did you determine that F-max < 250 Hz? "
page 29 in the HP/Agilent Application note 243 is a start.

For sample rate I think about the aliens that are monitoring my behaviour. If they take a sample once every 24 hours, and the first sample happens to be at about 2 in the morning, in their report they will conclude that I sleep all the time. They need to sample at a much faster rate to capture the details of my exciting, fast paced existence.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor