Convert into FFT solution

Question

nur yusof 2015-2-16

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/178603-convert-into-fft-solution

Hi all,

I got the time domain signal from experimental work as per below. The blue is x-direction,green is z-direction and red is y-direction.

When i convert the signal into FFT using matlab, I got as per below

The problem is, the peak should happen in z-direction(Ch3) but happen in x-direction(ch1). The coding as per below:-

clc; y1 = Ch1; y2= Ch2; y3 = Ch3; N=6678; Fs=512;

%Plot FFT Moving Vehicle signal Ch1 NFFT1 = 2^nextpow2(N); Y1 = fft(y1,NFFT1)/N; % the division by N is to scale the amplitude f_plot = Fs/2*linspace(0,1,NFFT1/2+1); % vector of frequencies for the plot subplot(311), plot(f_plot,2*abs(Y1(1:NFFT1/2+1))) % multiplicated by 2 to recover the energy related % to the negative frequencies since in this way only half spectrum is plotted title('FFT Moving Vehicle Ch1'); xlabel('Frequency [Hertz]');

%Plot FFT Moving Vehicle signal Ch2 NFFT2 = 2^nextpow2(N); Y2 = fft(y2,NFFT2)/N; % the division by N is to scale the amplitude f_plot2 = Fs/2*linspace(0,1,NFFT1/2+1); % vector of frequencies for the plot subplot(312), plot(f_plot2,2*abs(Y2(1:NFFT2/2+1))) % multiplicated by 2 to recover the energy related % to the negative frequencies since in this way only half spectrum is plotted title('FFT Moving Vehicle Ch2'); xlabel('Frequency [Hertz]');

%Plot FFT Moving Vehicle signal Ch3 NFFT3 = 2^nextpow2(N); Y3 = fft(y2,NFFT3)/N; % the division by N is to scale the amplitude f_plot3 = Fs/2*linspace(0,1,NFFT3/2+1); % vector of frequencies for the plot subplot(313), plot(f_plot3,2*abs(Y3(1:NFFT3/2+1))) % multiplicated by 2 to recover the energy related % to the negative frequencies since in this way only half spectrum is plotted title('FFT Moving Vehicle Ch3'); xlabel('Frequency [Hertz]');

For information,when the sensor do the measurement,the reading for y-direction will fall at-1 to 1 G due to the gravity.

Is there anything i need to do before convert it into FFT?shift maybe?

Thank you in advance.