Significant "very low frequency" data near DC

1 view (last 30 days)
Tyler
Tyler on 24 Sep 2018
Hi, i am performing an FFT on some strain gauge data to try and separate out the DC and vibratory components
I have Time function data (array of two (or more) columns, column 1 is time, columns 2+ are different strain gauge microstrains)
A piece of my code, most relevant to the FFT is below:
nstep = length(time_i);
L = nstep;
Fs = (L-1)/(time_i(end)-time_i(1)); % Sampling frequency
NFFT = 2^nextpow2(L); % Next power of 2 from length of y
Y = fft(data_i(:,sgnameind),NFFT)/L;
f = Fs/2*linspace(0,1,NFFT/2+1);
Y(1,:) = Y(1,:)/2;
sgspec(:,:,1,ifile) = repmat(f(1:NFFT/2+1)',1,nsg);
sgspec(:,:,2,ifile) = 2*abs(Y(1:NFFT/2+1,:));
The results i get are a bit strange to me in the low frequency range (i am plotting sgspec(:,1,1,1) ~time on x axis and sgspec(:,1,SG,1) ~ microstrain on Y axis). I notice peaks at the engine order harmonics, but at DC, it looks like there are "very low frequency" data that is higher than the DC result (see attached image).
Is this a potential mischaracterization of DC data as "very low frequency" data? Does it make sense to somehow include it in the DC data, and how would that be done in a way that isn't somehow double accounting (i.e. if i have more sample points i might count up and down the curve)
Any help would be greatly appreciated!
Thanks!

Answers (0)

Communities

More Answers in the  Power Electronics Control

Categories

Find more on Fourier Analysis and Filtering in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!