Professional Documents
Culture Documents
Contents
1 Topics of the Lecture – 11/03/2005 (Part I) 2
10 Impulse Response 31
1
41614 – Dynamics of Machinery 18/03/2005 IFS
• Fourier Analysis
• Fourier Series
• Fourier Integral
• Complex Form of Fourier Transform
• Examples: Square Wave and Triangular Wave
• Discrete Form of Fourier Transform
• Example: Sinus with 16 points
• Frequency Response Analyzers
• Digital Signal Processing
• Aliasing
• Anti-Aliasing Filtering
• Leakage
• Windowing
• Averaging
• Overlapping
1 References
• Bendat, J. S. (1993) ”Engineering Applications of Correlation and Spectral Analysis”, John Wiley & Sons, Inc.,
New York.
• Carlson, G. E. (1998) ”Signal and Linear System Analysis with Matlab”, John Wiley & Sons, Inc., New York.
• Ewins, D. J. (1984) ”Modal Testing: theory and practice”, Letchworth: Research Studies Press Ltd., England.
2
2 Fourier Analysis – Fourier Series, Integral and Complex
Form
3
4
5
6
7
3 Fourier Analysis – Example with a Square Wave Function
8
9
10
4 Fourier Analysis – Example with a Triangular Wave Func-
tion
11
12
13
14
5 Discrete Form of Fourier Transform
15
16
17
6 Discrete Fourier Analysis – Examples
• EXAMPLE 1: function x(t) = cos 4πt – It is important to see in table 1 that the spectral
coefficients ak and ·bk provides the information about the frequency of the function, i.e
4π rad/s or 2 Hz. Moreover, all coefficients bk are zero. Analyzing equations (??) and
(??) you can easily see that the spectral coefficients ak are related to the real part of
Xk or cos(k · ∆ω · t) and bk to the imaginary part of Xk or sin(k · ∆ω · t). Obviously, all
coefficients bk shall be zero.
tk [s] xk fk [Hz] ak bk
0 1.0000 0 0.0000 0.0000
0.0625 0.7071 1.0000 0.0000 0.0000
0.1250 0.0000 2.0000 8.0000 0.0000
0.1875 -0.7071 3.0000 0.0000 0.0000
0.2500 -1.0000 4.0000 0.0000 0.0000
0.3125 -0.7071 5.0000 0.0000 0.0000
0.3750 -0.0000 6.0000 0.0000 0.0000
0.4375 0.7071 7.0000 0.0000 0.0000
0.5000 1.0000 8.0000 0.0000 0.0000
0.5625 0.7071
0.6250 0.0000
0.6875 -0.7071
0.7500 -1.0000
0.8125 -0.7071
0.8750 -0.0000
0.9375 0.7071
1
x(t)
−1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1
r
0
x
−1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
tr [s]
10
ak
0
0 1 2 3 4 5 6 7 8
0
bk
−5
−10
0 1 2 3 4 5 6 7 8
fk=ωk/2π [Hz]
Table 1: Data of a digitalized cosine function x(t) = cos 4πt composed of N = 16 points in the
time domain, where T = 1 s, ∆T = T /N = 0.0625 s, ∆ωk = 2π/T = 2π [rad/s] = 1 [Hz],
fk = k · ∆ω/2π = k · 1 [Hz] (k = 0, 1, 2, ..., (N − 1)) and spectral coefficients Xk = ak − j · bk .
18
• EXAMPLE 2: function x(t) = sin 4πt – It is important to see in table 2 that the spectral
coefficients ak and ·bk provides the information about the frequency of the function, i.e
4π rad/s or 2 Hz. Moreover, all coefficients ak are zero. Analyzing equations (??) and
(??) you can easily see that the spectral coefficients ak are related to the real part of
Xk or cos(k · ∆ω · t) and bk to the imaginary part of Xk or sin(k · ∆ω · t). Obviously, all
coefficients ak shall be zero.
tk [s] xk fk [Hz] ak bk
0 0 0 0.0000 0.0000
0.0625 0.7071 1.0000 0.0000 0.0000
0.1250 1.0000 2.0000 0.0000 8.0000
0.1875 0.7071 3.0000 0.0000 0.0000
0.2500 0.0000 4.0000 0.0000 0.0000
0.3125 -0.7071 5.0000 0.0000 0.0000
0.3750 -1.0000 6.0000 0.0000 0.0000
0.4375 -0.7071 7.0000 0.0000 0.0000
0.5000 -0.0000 8.0000 0.0000 0.0000
0.5625 0.7071
0.6250 1.0000
0.6875 0.7071
0.7500 0.0000
0.8125 -0.7071
0.8750 -1.0000
0.9375 -0.7071
1
x(t)
−1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1
r
0
x
−1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
tr [s]
10
ak
0
0 1 2 3 4 5 6 7 8
0
bk
−5
−10
0 1 2 3 4 5 6 7 8
fk=ωk/2π [Hz]
Table 2: Data of a digitalized cosine function x(t) = cos 4πt composed of N = 16 points in the
time domain, where T = 1 s, ∆T = T /N = 0.0625 s, ∆ωk = 2π/T = 2π [rad/s] = 1 [Hz],
fk = k · ∆ω/2π = k · 1 [Hz] (k = 0, 1, 2, ..., (N − 1)) and spectral coefficients Xk = ak − j · bk .
19
7 Matlab Routines – 13 Examples with 5 Different Signals
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% MACHINERY DYNAMICS LECTURES (41614) %
% IKS - DEPARTMENT OF CONTROL ENGINEERING DESIGN %
% DTU - TECHNICAL UNIVERSITY OF DENMARK %
% %
% Copenhagen, February 20th, 2002 %
% IFS %
% %
% LECTURE ABOUT SIGNAL ANALYSIS & SIGNAL PROCESSING %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear all close all
% (Case 1)
n = 2048; % number of points in the range of time
freq = 10.0; % [Hz] - frequency of the simulated signal
xi = 0.0; % - damping factor
leak = 1.0; % - leakage factor
% (Case 2)
% n = 2048; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
20
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 3)
% n = 1024; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 4)
% n = 512; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 5)
% n = 256; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 6)
% n = 128; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 7)
% n = 64; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
% * LEAKAGE
% (Case 8)
% n = 128; % number of points in the range of time
% freq = 1.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 9)
% n = 128; % number of points in the range of time
% freq = 1.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.5; % - leakage factor
% (Case 10)
% n = 1024; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.1; % - damping factor
21
% leak = 1.0; % - leakage factor
% (Case 11)
% n = 1024; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.5; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 12)
% n = 1024; % number of points in the range of time
% freq =100.0; % [Hz] - frequency of the simulated signal
% xi = 0.9; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 13)
% n = 512 % number of points in the range of time
% freq =100.0 % [Hz] - frequency of the simulated signal
% xi = 0.1; % - damping factor
% leak = 1.0; % - leakage factor
% (Case 14)
% n = 512; % number of points in the range of time
% freq =10.0; % [Hz] - frequency of the simulated signal
% xi = 0.0; % - damping factor
% leak = 1.0; % - leakage factor
for i=1:n,
t(i)=(i-1)/n*tmax;
f(i)=(i-1)/tmax;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% FIVE DIFFERENT TYPES OF SIGNAL
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
% (1) SINUS SIGNAL WITH 1 FREQUENCY COMPONENT
x(i)=xo*exp(-xi*freqa*t(i))*sin(2*pi*freqa*t(i)*leak);
%
% (2) SINUS SIGNAL WITH 2 FREQUENCY COMPONENT
% x(i)=xo*exp(-xi*freqa*t(i))*sin(2*pi*freqa*t(i)*leak) + ...
% xo*exp(-xi*freqa*t(i))*sin( 4*2*pi*freqa*t(i)*leak);
%
% (3) SINUS SIGNAL WITH 4 FREQUENCY COMPONENT
% x(i)=xo*exp(-xi*freqa*t(i))*sin(0*2*pi*freqa*t(i)*leak) + ...
% xo/2*exp(-xi*freqa*t(i))*sin(2*2*pi*freqa*t(i)*leak) + ...
% xo/4*exp(-xi*freqa*t(i))*sin(4*2*pi*freqa*t(i)*leak) + ...
% xo/6*exp(-xi*freqa*t(i))*sin(6*2*pi*freqa*t(i)*leak);
%
% (4) SINUS SIGNAL + NOISE
% const = 0.5;
% ruido(i) = randn;
22
% x(i)=xo*exp(-xi*freqa*t(i))*sin(2*pi*freqa*t(i)*leak)+const*ruido(i);
%
% (5) STEP WAVE
% x(i)=xo*exp(-xi*freqa*t(i))*sign(sin(2*pi*freqa*t(i)*leak));
end
y=fft(x);
figure(1)
title(’Signal Analysis in Time and Frequency Domain’,’FontSize’,18)
subplot(3,1,1), plot(t,x,’r-’)
title(’(a) Time Domain - (b)
and (c) Frequency Domain with N and N/2 points’,’FontSize’,18)
xlabel(’time [s]’,’FontSize’,18)
ylabel(’(a)’,’FontSize’,18)
grid
subplot(3,1,2), plot(f(1:n),abs(y(1:n)),’r-’)
xlabel(’frequency [Hz]’,’FontSize’,18)
ylabel(’(b)’,’FontSize’,18)
grid
subplot(3,1,3), plot(f(1:n/2),abs(y(1:n/2)),’r-’)
xlabel(’frequency [Hz]’,’FontSize’,18)
ylabel(’(c)’,’FontSize’,18)
grid
23
8 Analyzers and Signal Processing
24
25
26
27
28
29