I have two signals (response and force) and I want to calculate FRF in MATLAB.
And I must calculate it with FFT. but I resulted not good.
I am not that good in explaining things. Hope this helps.
Let x and f be the time histories of response and force respectively, and dt the sampling time step.
Note that FFT results in a 2-fold FRF, that is, you should consider only the first half of the data obtained (this corresponds to the Nyquist frequency = (1/dt)/2).
% code
tempx=fft(x);
tempf=fft(f);
x1=abs(tempx);
f1=abs(tempf);
L=length(x1)/2;
Nq=(1/dt)/2;
X=x1(1:L);
F=f1(1:L);
FRF=X./F;
freq=linspace(0,Nq,L);
figure
semilogy(freq,FRF);
grid on
xlabel('frequency')
ylabel('Displacement FRF')
Hi, you can just follow my paper "Evaluation of Frequency Response Functions of Structural System under Various Base Conditions"
If you need the code just contact me, I can provide it .
I am not that good in explaining things. Hope this helps.
Let x and f be the time histories of response and force respectively, and dt the sampling time step.
Note that FFT results in a 2-fold FRF, that is, you should consider only the first half of the data obtained (this corresponds to the Nyquist frequency = (1/dt)/2).
% code
tempx=fft(x);
tempf=fft(f);
x1=abs(tempx);
f1=abs(tempf);
L=length(x1)/2;
Nq=(1/dt)/2;
X=x1(1:L);
F=f1(1:L);
FRF=X./F;
freq=linspace(0,Nq,L);
figure
semilogy(freq,FRF);
grid on
xlabel('frequency')
ylabel('Displacement FRF')
Thanks. I coding the same.
But I feel it is wrong for estimate of frequency.
I send to you my code with message.
The code from Sundar works when the excitation is harmonic. If it is not harmonic (for example, it contatins more than one harmonic or even is a general function of time), the ratio of FFT of response to FFT of excitation is not its FRF.
For the eq of motion in time domain
m.d2y/dt2 +c.dy/dt+ky=F(t)
The eq in freq domain will be
(-(2pi.f)^2.m+i.2pi.f.c+k)Y=F(f)
If the input F(t)= delta(t) then
(-(2pi.f)^2.m+i.2pi.f.c+k)Y=1
Thus the FRF is
H(f)=(-(2pi.f)^2.m+i.2pi.f.c+k)^-1
Form Prof Ouyang's reply, How could i get FRF if the excitation is not harmonic? thanks
to Prof.Huajiang Ouyang ·
so How can I get the FRF if the excitation force has more than one excitation frequency e.g multisine?
thanks
There is a tool in Matlab, by using it, you can get FRF for any signal.
Regards
- Badges
- Science topic