Frequency Response Function and FFT for Modal Analysis

I am experimentally testing a material to retrieve its natural frequencies through modal analysis. The type of testing is based around impulse response. My sampling frequency was 10kHz. My acccelerometer's sensisitivty was 100mV/g and my impact hammer's sensitivity was 2.25mV/N. My aim is to obtain an accurate FRF graph. My entire code:

%reading excel file

dataset = xlsread('BrassFRF.xlsx','Sheet1','A1:C50000');

%%Creating and filling data variables

a=dataset(:,2); 
n=dataset(:,3);

%%Converting acceleraltion voltages into acceleration (mm/sec^2)

a=a.*((1/0.1)*9.81);

%%Converting force voltages into Newton (N)

n=n.*(1/0.00225);

%%Creating FFT using acceleration

L=length(a);
NFFT=2.^nextpow2(L);
V=fft(a,NFFT)/L;
Fs=10000;
freq=Fs/2*linspace(0,1,NFFT/2+1);
subplot(3,1,1)
plot(freq,2*abs(V(1:NFFT/2+1)))
xlabel('Frequency (Hz)');
title('Velocity FFT ')

%%Creating FFT for using force

F=fft(n,NFFT)/L;
subplot(3,1,2)
plot(freq,2*abs(F(1:NFFT/2+1)))
xlabel('Frequency (Hz)');
title('Impulse FFT ')

%%FRF

FRF=V./F;
subplot(3,1,3)
plot(abs(FRF))
xlabel('Frequency (Hz)');
title('FRF')

Please tell me if you find any errors that can be the cause of an incorrect FRF. My suspect is the impulse FFT.

[Lingyuan Kong]

The last section is wrong, it should be:

FRF=V./F;

subplot(3,1,3)

plot(freq, abs(FRF(1:NFFT/2+1)))

xlabel('Frequency (Hz)');

title('FRF')

Answer 1

[durukan dilek]

did you obtain frf of acceleration/force? why did you write V/F? why did you use velocity?

L=length(a); NFFT=2.^nextpow2(L); V=fft(a,NFFT)/L; Fs=10000; freq=Fs/2*linspace(0,1,NFFT/2+1); subplot(3,1,1) plot(freq,2*abs(V(1:NFFT/2+1))) xlabel('Frequency (Hz)'); title('Velocity FFT ') set(gca, 'YScale','log')

Answer 2

[Deniz Kny]

how did you know your sampling frequency was 10khz? I have a accelerometer data and, trying to obtain fft like you. i need a Fs but i dont know how can i get this info.

[imthiyas manarikkal]

You can actually check your resonance frequency or natural frequency. The sampling frequency must be twice of your natural frequency (Nyquiste Theorem)

Answer 3

[Jose Ordaz]

Hello,

i'm working in the same way, trying to obtain the natural frequency and damping of some materiales using the information from a hammer and a accelerometer. Can you please help me with this?, how you solve your problem with the FRF?.

thanks