energy spectrum - FFT - velocity fluctuations

Hi guys,

I've got some really serious issues. Maybe for you it' s just about basics, but for me it's like being in Hell.:)I've tried to solve this problem over a long time but obviously I have no clue how to do so. I'm a student and for the time beeing at home and bored I kill some time in how to use Matlab. Currenty I'm working on a flow simulation task that drives me somehow a little bit crazy. But that's another story. What I'm looking for is an algorithm that helps me to get the energy spectrum by using the fft function. So far, I've coded some few lines but I'dont get a decent solution. I hope there is anybody out there who can help me. Below you will find the code and a pic:

clear all; close all; clc;

        st=1;
        et=9000;

%flow data

flow_data=importdata('point_velocity_x.dat'); point_velocity_x=flow_data.data(st:1:et,2); clear flow_data;

%mean values, fluctuations

mean_velocity_x=mean(point_velocity_x); velocity_x_f=point_velocity_x-mean_velocity_x; velocity_x_f2=velocity_x_f.*velocity_x_f;

%Time domain samples=et-st+1; % T=4.088E-07; T_max=(samples-1)*4.088E-07; t=0:T:T_max; % % figure(1); % plot(t,velocity_x_f2);

%Frequency domain f_T=1/T; T_f=f_T/(samples-1); f=0:T_f:f_T;

%Spectrum

Spectrum_x=fft(velocity_x_f2);

%Graph figure(2); loglog(f,Spectrum_x);

ylabel('E(f)'); xlabel('f');

Greetings, Uwe

[Star Strider]

Because it has been Answered. Accept my Answer, then I’ll delete it, and then you or I can delete your Question.

[Rena Berman]

(Answers Dev) Restored edit

Answer 1

[Star Strider]

I don’t follow everything you’re doing. The problem with the Fourier transform plot is that you have to plot the absolute value. Other problems are that you have to normalise the fft by the length of the original data vector, and multiply the absolute by 2 in the plot. See the documentation on the fft (link) function for help on using it correctly. The code between the first (top) two plot figures is important, especially the calculation of the frequency vector and plotting the appropriate part of the fft.

Spectrum_x = fft(velocity_x_f2)/length(velocity_x_f2);

%Graph
figure(2);
loglog(f, abs(Spectrum_x)*2);