Miscellaneous Exercises

in
MATLAB

Ulf Carlsson

Stockholm 2009

KTH Aeronautical and Vehicle Engineering, Teknikringen 8, 100 44 Stockhom, www.ave.kth.se

© Department of Aeronautical and Vehicle Engineering, KTH
Preface to the 2009 edition
This edition is prepared for the course SD1000 Perspectives on Vehicle Engineering, the
academic year 2009/2010. Compared to the previous edition
- a couple of errors have been corrected,
- 6 new “Level 1”-exercises have been added,
- some of the exercises have been slightly modified.

Stockholm, 2009-08-11

Ulf Carlsson

Preface
This collection of MATLAB™ exercises has been developed for use in the course 4B1052
Perspective on Vehicle Engineering, for the first year students, at the Vehicle Engineering
programme at KTH, Stockholm, Sweden. The main material used in the course is “An
Introduction to MATLAB” by David F Griffiths. This material can be downloaded from the
homepage of the mathematics department at the University of Dundee,
http://www.maths.dundee.ac.uk/. The aim is to exercise the student’s skills in using
MATLAB so that they can use it as a tool for learning and understanding the material taught
in the courses they meet during their studies at KTH. A milestone in becoming a skilled
MATLAB user is to reach the level where the student can search, find and use information
useful for implementing a code on her/his own. The best way to reach this level is to start
working with a tutorial focussing on the basic elements of MATLAB. Then exercise these
steps with exercises with gradually increasing complexity and finally to start using MATLAB
as a supplementary tool in other courses.
The collection starts with exercises practising simple expressions and arithmetic
operators, how to create vectors and matrices, the colon operator, the built-in functions and
plotting. Then the focus is turned to reading and writing data files, script files (M-files),
function M-files, for – while loops, if – then constructs and animation. Then, finally, a couple
of exercises on creating Graphic User Interfaces (GUI) are given.
Another aim with this collection of exercises is to demonstrate how MATLAB is
implemented to solve engineering problems. There is no doubt that the best way to learn
mathematics and programming at a technical university is to use the methods for solving
engineering problems. For this reason a number of engineering problems are provided in the
collection. Several of these applied engineering exercises use concepts and methods that the
students will meet during their later studies. Since these concepts and methods have to be
introduced and explained in a qualitative way, some problem formulations are fairly long.
This will, no doubt, prove a valuable experience for the students in the problem solving
process: find relevant information – formulate a model – implement the model in a MATLAB
code and finally – evaluate the model for specific parameter values in order to solve a realistic
engineering problem.
The examples are formulated in the spirit of the CDIO (Conceive Design Implement
and Operate products and systems) concept. CDIO is a joint project between Chalmers, LiU
(Linköpings Universitet), MIT (Massachusett’s Institute of Technology) and KTH (Kungl
Tekniska Högskolan) that aims to develop new ways of teaching engineering science.
Answers and solutions are provided for most problems. Note that there are several
possible ways to solve a problem and implement a code. The solutions provided herein are by
no means claimed to be optimal in any way.
 As course examination each student has to solve a problem, formulated as an
engineering project, using MATLAB. Up to days date approximately 10 different examination
problems have been developed.
This material may be downloaded and used freely in its present form by anyone who
wishes. Any changes of its contents are not accepted. Please notify the author of all errors
found and possible ways to improve the material. The material is developed in MATLAB 6.

Stockholm 2003-08-28

Ulf Carlsson
MWL, Dept of Vehicle Engineering, KTH

Recommendations for reading and learning

This collection of exercises is used in the MATLAB part of the course 4B1052 Perspektiv på
Farkosttekniken, as a complement to the basic material “An Introduction to MATLAB” by
David F Griffiths. In this course the examination of the MATLAB part is performed at three
levels.
The aims of the 1st level is to provide the student with skills enabling use of
MATLAB as
- a calculator,
- a tool for simpler processing of numbers stored in vectors and matrices,
- a tool for plotting graphs of scientific formulae and
- code the instructions in a simple MATLAB script file.
In addition to the aims for level 1 the aims of the 2nd level is to provide skills sufficient for
- more advanced matrix computations such as eigen-values and eigen-vectors and
- manipulating subvectors and submatrices.
Finally level 3 adds
- developing codes including for loops and logical if statements,
- function subroutines and
- animation.
The exercises in this collection are labelled Level 1–3 corresponding to this list.
To fulfil the requirements for the examination levels we recommend to follow the
learning procedure given below.
1) Depending on your programming skills you should start working 10 – 20 hours with the
basic material “An Introduction to MATLAB” by David F Griffiths.
2) Exercise your skills using some of the recommended (see course home page) exercises
labelled “Level 1”.
3) If necessary, select some additional “Level 1”-exercises.
4) If you do not want to go for level 2 perform your examination exercise for “Level 1”.
5) Repeat step 2)-4) for “Level 2”.
6) Repeat step 2)-4) for “Level 3”.
It should be noted that the step from level 1 to level 2 is not very big. For this reason
we strongly recommend everyone to at least give level 2 a try before choosing to examine
level 1.

Stockholm 2004-08-12

Ulf Carlsson
MWL, KTH Aeronautical and Vehicle Engineering
e-mail: [email protected]
EXERCISES
1 (Level 1)
Evaluate the following MATLAB statements:
a) 9-(3+5)*6/4,
b) 9-3+5*6/4,
c) 0.5+4.1-3^2*2,
d) 0.5+4.1-3^(2*2).

2 (Level 1)
Evaluate the following expressions:
2+3
a) 8+3 ,
4
1 + 4(1 + 3)
b) −5+7 ,
3(2 + 5)
2(3 + 4) 2 (1 + 3) 3
c) 3+ .
3(1 − 3) 2 8

3 (Level 1)
Evaluate the following expressions:
1
a) 3 ⋅ (−4) + (−2) ⋅ (1 − 3) ⋅ ,
4
2 5
(1 − )
1
b) − (−3) − ⋅ 3 4 ,
5 (4 − )(4 + 1 )
1
3 3
4 9
c) (5 − ) 2 ⋅ − (5 + 3)(5 − 3) .
3 8

4 (Level 1)
Evaluate the following expressions:
a) 2.6 ⋅103 ⋅ 3.3 ⋅104 ⋅ 7 ⋅10−5 ,
3.7 ⋅103
b) ⋅ 5.3 ⋅102 .
7.1 ⋅10−5 ⋅ 2.1 ⋅102

5 (Level 1)
Evaluate the following expressions:
2+2
a) 3 ,
2−2
2−2
b) 3 .
2−2

What is the difference between the answers obtained in a) and b) respectively?

1
6 (Level 1)
Evaluate the following expressions:
a) (1 + i ) + (2 − i ) ,
b) (1 + i )(2 − i ) ,
1+ i
c) ,
2−i
d) (1 + i )(1 − i ) ,

7 (Level 1)
Evaluate the following expressions:
π
a) y = 3 sin ,
4
π π
b) y = 3 cos( + ) ,
4 2
π
c) y = 3 tan ,
4
π π
d) y = 3 cot( + ) .
4 2

8 (Level 1)
Evaluate the following expressions:
π
a) sin ,
4
b) cos 1.6,
c) arcsin 1,
d) arctan 0.5.

9 (Level 1)
Evaluate the following expressions:
a) e1 ,
b) ln 2 ,
c) 2.5 ln(1.5) ,
d) log(2) ,
e) 10 ⋅ log 4.5 .

10 (Level 1)
Evaluate the following expressions:
a) y = 5 sin(π 3) ,
b) y = e 2+ 3 ,
c) y = ln(7 + e 2 ) ,
d) y = log e 3 ,
e iπ 2 − e − iπ 2
e) y = arcsin . (How can this expression be reformulated?)
2i

2
11 (Level 1)
Calculate the angles, v, in degrees if,
a) sin v = 0.9 ,
b) cos v = −0.42 ,
c) tan v = −0.13 .
d) Are there other possible solutions to these equations?

12 (Level 1)
The following code is intended to calculate the radius r of a circle from its circumference L.

radius.m
L = input('Enter the circumference L: ');
r = L/2*pi

Use MATLAB Editor to create a new M-file, type these statements and save the file with the
name radius.m.

Unfortunately the code contains an error. Analyse the code and correct it. Finally, try it on, for
example for L = 2π.

13 (Level 1)
The strength of sound is measured in sound pressure level, Lp [dB], defined as,
~
p2
L p = 10 ⋅ log 2 [dB], pref = 2 ⋅10 −5 Pa.
pref
where log is the base 10 logarithm, the rms (root mean square) sound pressure, ~
p [Pa], is the
measured output from the microphone and pref is a reference sound pressure. A definition of
the root mean square value is given below in exercise 52.
Calculate the sound pressure level for the following measured rms sound pressures.
a) A normal conversation, ~ p = 35 ⋅10 −3 Pa.
b) A noisy workshop, ~
p = 1.1 Pa.
c) A jet during take-off, ~
p = 200 Pa.

14 (Level 1)
The following MATLAB code is supposed to evaluate the expression z = x + y for,
a) x = 5, y = 2, and for
b) x = y = 5,

% xplusy.m
clear
% Part a)
x = 5;
y = 2;
z = x + y
% Part b)
x = y;
y = 5;
z = x + y

Use the MATLAB Editor to create the M-file and execute it. Locate the error and correct it.

3
15 (Level 1)
Two equally strong but uncorrelated sound sources do not, as one may think, cause a doubling
of the sound pressure level. To calculate the total sound pressure level, Lp,tot, caused by N
uncorrelated sound sources contributing with individual sound pressure levels, Lp,n, the
following formula is applicable,
N
L p ,tot = 10 ⋅ log ∑ 10
L p ,n 10
[dB].
n =1
Calculate the total sound pressure level in the following cases.
a) Two simultaneously running engines with equal sound pressure levels at 85 dB.
b) Four jets, each with sound pressure level 140 dB, starting at the same time.

16 (Level 1)
Create the following vectors in MATLAB.
a) a = [1 2 1] ,
1 
b) b = 2 ,
1 
c) c = [2 3 4 L 25] ,
d = [20 18 L − 18 − 20] ,
T
d)
 0 .1 
 0 .2 
 
e) e =  0 .3  .
 
 M 
100

17 (Level 1)
Evaluate the vector y in the following cases:
T
10π 20π 30π 
a) y = sin x , x =  L 10π  .
 200 200 200 
What is the value of component 24 of y?

1
b) y = sin x , x = [1 2 3 L 1000] .
T

π
What is the value of component 524 of y?

18 (Level 1)
The equation x3 + x – 1 = 0, is a polynomial equation.
a) Use the MATLAB Help browser to find the command that solves, i.e. calculates the
roots of, a polynomial equation.
b) Solve the polynomial equation given above.

4
19 (Level 1)
Build a time vector t with 100 equidistant components on the interval [0, 1], that is t(1) = 0
and t(100) = 1.

a) What is the value of the 50th component of t?

b) What are the values of the components in the range 51 to 60?

c) Build a time vector t1 containing every fifth component of t starting from t(5). What
is the value of the 17th component of this vector. Plot sin(2πt) and sin(2πt1) as
functions of t and t1 in a diagram. What is the difference between the curves?

20 (Level 1)
Sound pressure values pn [Pa] are measured at frequencies fn [Hz]. The data points (fn, pn), n =
1, 2, 3, … N, are stored on the binary data file soundpress.bin. The values are stored on
format (precision) 'float32' in the following sequence: f1, p1, f2, p2, f3, p3, … fN, pN.

a) Use the MATLAB Help browser to find out how to use commands fopen, fread
and fclose to read the data file soundpress.bin. Read the data into
MATLAB.

b) Use a MATLAB command to find the number of data values in the file.

c) Separate the data into one vector f with all frequencies fn and one vector p with all
sound pressures pn.

d) Plot the sound pressure p versus the frequency f. Add axis labels with proper
quantities and units.

21 (Level 1)
In a mechanics laboratory exercise you measure the position s [m] of a car at time t [s]. The
measurement data is available on the MATLAB data file meclab.mat.

a) Use a MATLAB command and read the data into the MATLAB Workspace.

b) Plot the car’s position s versus the time t. Use markers instead of a line. Add axis
labels with proper quantities and units.

From basic mechanics it is known that the relationship between position s and acceleration a
[m/s2] is,
d 2s
a= 2 . (1)
dt
If the acceleration is constant the position s can be shown to be (integrate Equation (1) twice
with respect to time),
t2
s = a ⋅ + v0 ⋅ t + s0 , (2)
2
where s0 [m] and v0 [m/s] are the position and speed at time t = 0 respectively.

5
c) Use the MATLAB Help and identify the command that fits a polynomial function to
a set of data points. Assume that the car’s acceleration is constant and find the
coefficients in Equation (2). What is the value of acceleration?

d) Use the MATLAB Help and identify the command to evaluate the polynomial at
prescribed time values. What is the position at time t = 0 and t = 2.3 s?

e) Compare the estimated polynomial relationship with the original measured data by
plotting them in the same diagram. Note that you can use a higher time-resolution in
the polynomial plot.

f) Store the polynomial coefficients determined in c) on a MATLAB data file named

accpoly.mat.

22 (Level 1)
a) Build the function in Figure 1.

 sin x, 0≤ x <π
y ( x) = 
repeat for, π ≤ x ≤ 6π

(Hint: Construct a half period sine and use it as a building block.)

b) Find a simpler alternative method to build the function.

1.5

0.5

0
π 2π 3π 4π 5π 6π

-0.5
0 2 4 6 8 10 12 14 16 18

Figure 1 Rectified sine function.

23 (Level 1)
Plot the function you built in exercise 22. Try to get the same axis scaling as in Figure 1.

6
24 (Level 1)
Build the periodic triangle wave, in Figure 1, defined by,
 1 − x, 0 ≤ x <1

y =  x − 1, 1≤ x < 2.
repeat for 2 ≤ x ≤ 6

1.5

0.5

-0.5
0 1 2 3 4 5 6

Figure 1 Periodic triangle wave.

25 (Level 1)
Plot the rectified sine function from exercise 22 in a multi-plot together with the triangle wave
from exercise 24. Use solid line for sine and dashed line for triangle.

26 (Level 1)
Quick estimates of solutions to transcendent equations can be found graphically by plotting
suitable functions and locate their intersections.
Find approximate values for the three lowest non-zero positive solutions to the
equation, x = cot x , by plotting the functions in a multi-plot and zooming the area around the
intersections.

27 (Level 1)
Enter the vector u, u = [1 2 3 L 10] , and calculate,
a) the scalar product, u ⋅u ,
b) the product, uT ⋅ u ,
c) the scalar product using the command dot,
d) the element-wise product, u.*u,

28 (Level 1)
Define an equidistant vector x with first value 0, last value 5 and increment 0.05. Evaluate the
following functions.

a) y = 10 – x. What is the value of the 17th component of y?

b) y = 10 – x – sinx. What is the value of the 17th component of y?
c) y = x2 – 3x +1. What is the value of the 17th component of y?
d) f (x) = sin(x)/x. What is the value of the 17th component of f ?

7
29 (Level 1)
Modulation is important in many fields of technology. Two interacting sinuousoidal signals,
one with low and one with high frequency, produce an amplitude modulated signal. Enter the
2π
vector, x = [1 2 3 L 1000] , and plot the amplitude modulated function,
100
y = sin( 4 x)(1.8 + sin(0.1x)) .

(If you have a computer with a sound card, you can listen to the modulated signal by using the
command sound(y,1000).)

30 (Level 1)
In mechanics the work performed by a force is defined as the scalar product between the
force, F, and its displacement, s. Both F and s are in general vectors with components in all
three spatial (room) directions xyz. Calculate the work performed in the following cases.
a) When the force is, F = [100 0 0] N, and the displacement is, s = [0 3 2] m.
b) When the force is, F = [50 30 -10] N, and the displacement is, s = [1 2 –0.5] m.

31 (Level 1)
The following code is intended to evaluate the function, y ( x) = x sin x for values of x given in
the vector x:

x = (0:1/100:1)*10*pi;
y = x*sin(x);

Check the code and correct it if necessary.

32 (Level 1)
1
The following code is intended to evaluate the function, y ( x) = for values of x
1 + sin 2 x
given in the vector x.

x = (0:1/100:1)*10*pi;
y = 1/(1+sin(x).^2);

What is wrong with the code? Correct the code.

33 (Level 1)
Estimate the real-valued roots to the polynomial equation, x4 + 5x3 – 64x2 – x + 225 = 0, using
a graphical technique.

34 (Level 1)
In acoustics a silencer is used as a means to reduce the acoustic energy radiated from for
instance engine exhaust outlets. The effect of inserting a silencer can be specified in terms of
the transmission loss, DTL [dB], which is defined as,
W
DTL = 10 ⋅ log in ,
Wout
where Win and Wout are the incoming and transmitted (passing) acoustic power respectively. A
reactive silencer is in principle a sudden area expansion followed by a sudden area
contraction, see Figure 1.

8
 L

S1 S1
Win Wout
S2

Figure 1 An expansion chamber is simple realisation of a reactive silencer.

For low frequencies, f, where only plane acoustic waves propagates in the pipe (with cross
section area S1) the following formula can be used to calculate the transmission loss for a
silencer with cross section area S2 and length L,
S S 2πf
DTL ( f ) = 10 ⋅ log(1 + ( 1 − 2 ) 2 sin 2 ( L)) ,
2S 2 2 S1 c
where c is the speed of sound (for air at normal temperature and pressure conditions (NTP),
c = 340 m/s).
In a specific case the pipe has cross section area 0.5 ⋅10 −4 m2, the silencer cross
section area 5 ⋅10 −4 m2 and length 1 m. Plot the transmission loss as a function of frequency
for frequencies in the interval f ∈ [0,500] Hz. Make sure the frequency resolution is
sufficiently high.

35 (Level 1)
a) Define an equidistant frequency vector f with the command
f = linspace(-1,1,200);
What is the value of the 41st component of the frequency vector?

b) Introduce a vector for the angular frequency, ω = 2π f .

Evaluate the complex-valued receptance function h defined as a function of ω by,
1 1
h(ω ) = − ,
ω − 1 − 0,5i ω + 1 − 0,5i
at angular frequency values given by the components of ω. What is the value of the
41st component of h(ω)?

c) Find the maximum absolute value of h(ω) and the corresponding value of the
frequency f.

d) Find the frequency interval f ∈ [fmin , fmax] for which the absolute value of the
receptance function is larger than 1.9.

e) Plot the receptance function h in three different formats:

(i) as a CoQuad plot, i.e. the real and imaginary parts as functions of the
angular frequency ω.
(ii) as a Bode plot, i.e. the absolute value and phase angle (degrees) as functions
of the angular frequency ω.
(iii) as a Nyquist or Argand plot, i.e. the imaginary part as a function of the real
part.

9
36 (Level 1)
Enter the matrices,
 1 2 1
a) a = 0 2 1 ,
0 0 3
1 0 0 0 
0 1 0 0 
b) b=  ,
0 0 1 0 
 
0 0 0 1 
1 0 0 0
0 1 0 0 
c) c=  
0 0 1 1 
 
0 0 1 1 

1 0 1 0 2 2
0 1 0 1 2 2

1 0 3 3 3 1
d) d=  
0 1 3 3 3 1
2 2 3 3 3 1
 
2 2 1 1 1 4

37 (Level 1)
1 0 0  2 1 2 0 
1

Enter the matrices, A = 0 2 0 , B = 2 − 2 − 1 , and the vector b =
 1 ,
 
3
0 0 3 1 2 − 2 0
and calculate,
a) Ab,
b) Bb,
c) AB,
d) BA.

A scaling matrix changes the length of the vector it is multiplied with. A rotation matrix
rotates the vector it is multiplied with.

a) Is A a scaling matrix or a rotation matrix?

b) Is B a scaling matrix or a rotation matrix?

38 (Level 1)
Build the vector, x = {2 1 3 2}.

a) Calculate, A = xTx, where T means transpose.

b) What is the sum of the third element in the second row and the second element in the
third row of A?

10
 a a23 
c) Define the submatrix B =  22 , where akl is the l-th element in row k of A.
 a32 a33 

d) Calculate the sum of the elements in the third row of A?

39 (Level 1)
Solve the equations for x.
2 1 1
a) Ax = b, A =   , b= .
1 2 1
3 0 1  2
b) Ax = b, A = 2 2 2 , b = 0 .

0 1 3 1 

40 (Level 1)
Solve the following systems of complex-valued linear equations.
 x1 + x2 = 1 + i
a)  ,
 x1 + 2 x2 = 2 + i
 x1 + 2ix2 = 1 + i
b)  ,
(1 − i ) x1 + 2 x2 = 2 + i
 x1 + 3ix2 − 2(1 − i ) x3 = 2 − i

c)  x1 + 2 x2 − 3ix3 = −i .
 x + (i + 1) x + 2 x = 2
 1 2 3

41 (Level 2)
An often met eigenvalue problem can be written on the form, (A - λI)x = 0, where I is the
identity matrix. A generalized version of this is given by,
Ax − λBx = 0 or ( A − λB)x = 0 , (1)
where A and B are n×n matrices and λ is a scalar.

a) Use the MATLAB Help browser to find a suitable MATLAB command for solving
eigenvalue problems of the kind given in (1).

Find the solutions to (1), i.e. the eigenvalues and eigenvectors, in the following cases:

2 1  2 − 1
b) A=  and B =  ,
1 2  − 1 2 
 2 −1 0   4 −1 1 
c) A = − 1 2 − 1 and B = − 1 4 − 1 .
 
 0 − 1 2   1 − 1 2 

11
42 (Level 2)
A linear system of equations on the form (compare with exercise 41 when B is the identity
matrix),
Ax − λx = 0 or ( A − λ )x = 0 , (1)
where A is an n×n matrix and λ is a scalar, is an example of an eigenvalue problem. Eigen-
value problems of this type have non-trivial solutions for n specific values of the scalar λ.
These solutions are denoted eigen-vectors and the corresponding values of λ are the eigen-
values.

Solve the eigenvalue problem given by (1), that is determine the eigenvalues and eigen-
vectors, for the following matrices A:
2 1
a) A= ,
1 2 
 2 2
b) A= ,
1 2 
2 1 − i 
c) A= ,
1 2i 
 1 1 − 2
d) A =  2 2 1  .
− 1 2 3 

43 (Level 2)
The solution vectors to eigenvalue problems given by,
( A − λB)x = 0 , (1)
have useful orthogonality properties. Suppose
 2 −1 0   4 −1 1 
A = − 1 2 − 1 and B = − 1 4 − 1 .
 
 0 − 1 2   1 − 1 2 

a) Determine the solution vectors to the eigenvalue problem (1).

b) Let V be a matrix with columns equal to the solution vectors (eigenvectors) and
calculate the matrix products, R = VTAV and S = VTBV. What is the particular
properties of R and S?

44 (Level 3)
Write for-loops that calculate the sums,
100
1
a) ∑ n2 ,
n =1
10
b) ∑e
m=0
− 2πm
,

12
45 (Level 2)
a) Create a vector of random numbers using command rand(100,1) .
b) Transform it to a random integer vector with components in the set [1, 2, 3, … 10].
(Hint: fix, floor, round and ceil are commands useful for this purpose.)
c) Calculate the number of components of this vector larger than 5.

46 (Level 3)
Solve the equation, x = tan x , using the iterative technique suggested in paragraph 20.1
example 20.2 in “An Introduction to MATLAB” by D F Griffiths, i.e. let xn+1 = cotxn, n = 0, 1,
2, … . Continue until the difference between two consequtive iterations are smaller than 10-4.

47 (Level 3)
Write an M-file countval.m that counts the number of components in a vector y with
values in the intervals [0, 1], ]1, 2], ]2, 3] and ]3, 4]. Try the code on the following data:
x = (0:1/100:1)*2*pi ; y = 2+2*sin(x).

48 (Level 1)
A mechanical single degree of freedom system consists of a point mass, a spring and a
damper. The mass is allowed to move in the x-direction only (single degree of freedom). The
spring and damper are both assumed to be massless. The equation of motion for the system,
see Figure 1, reads,
d 2x dx
m 2 + c + kx = F , (1)
dt dt
where m is the system mass, k is the system stiffness or spring constant and c is the system
damping constant. The system displacement is given by the coordinate x and the external
force acting on the system is F. When the system parameters m, k and c are independent of
time t, Equation (1) is a linear, ordinary differential equation of second order with constant
coefficients.
x

m
F(t)
k c

Figure 1 A mechanical single degree of freedom system.

The system displacement when the force is varying sinuousoidally in time with frequency f,
F (t ) = Fˆ sin(2πft ) = Fˆ sin(ωt ) (2)
where F̂ is the size or amplitude of the force and ω is the angular frequency, is of great
practical importance. For this particular time dependence the so called iω-method may be
used to solve the equation of motion. The iω-method assumes that the displacement x and the
force F are replaced with new complex variables x and F,
F(t ) = Fˆ (cos(ωt ) + i sin(ωt )) = Fˆeiωt hence F (t ) = Im(F(t )) (3a)
iωt
x(t ) = xˆ (cos(ωt ) + i sin(ωt )) = xˆe hence x(t ) = Im(x(t )) . (3b)

13
If x and F are inserted in equation 1 the time derivatives are replaced with multiplication with
the factor iω and the differential equation is transformed to a algebraic equation,
(iω ) 2 mxˆ + iωcxˆ + kxˆ = Fˆ , (4)
with the solution,
Fˆ
xˆ = . (5)
− ω 2 m + iω c + k
Usually the solution is given in terms of the ratio between the complex displacement x and the
complex force F, the so called frequency response function,
x(t ) xˆe iωt xˆ 1
= iωt = = . (6)
F (t ) Fˆe Fˆ − ω m + iωc + k
2

Suppose the system parameters are: m = 1kg, c = 1Ns/m and k = 10000 N/m.

a) Introduce a frequency vector f with 500 equidistant components on the interval

[0, 30] Hz. What is the value of the 265th component of f ?

b) Calculate the frequency response function given by (6) as a function of frequency in

frequency points given by the components of f . What is the value of the 265th
component of the frequency response function ?

Obviously the frequency response function is complex-valued. The physically important

properties of a complex-valued entity are its magnitude or absolute value and its phase angle.
The absolute value of the frequency response function tells you the magnitude of the
displacement amplitude relative to the force amplitude. For example, a value larger than 1
implies that the displacement amplitude is larger than the force amplitude. The phase angle,
on the other hand, gives the time shift or lag between the displacement and force. The phase
angle φ can be calculated as,
Im( xˆ Fˆ )
φ = arctan . (7)
Re( xˆ Fˆ )
Using the absolute value and the phase angle the frequency response function given in 6 can
be written as,
xˆ xˆ
= e iφ . (8)
F Fˆ
ˆ

c) Calculate the magnitude and phase angle of the frequency response function as
functions of frequency at frequency values given by the components of f. Plot the
results in two separate diagrams. What is the magnitude of the 265th component of
the frequency response function ? What is the phase (in degrees) of the 265th
component of the frequency response function ?

d) Determine the maximum magnitude of the calculated vector, xˆ / Fˆ . At what

frequency value is the largest value located? What is the phase at this frequency?

The frequency response function provides information on the system as a function of the
frequency of the external force. Suppose the force has a specific (fixed) frequency f how can
the displacement x(t) be calculated ? Using (6) and (3b) yields,

14
 x(t ) xˆ xˆ
x(t ) = ⋅ F(t ) = e iφ ⋅ Fˆe iωt = Fˆe i (ωt +φ ) . (9)
F(t ) Fˆ Fˆ
Finally the real displacement is obtained according to (3b) as,
xˆ xˆ
x(t ) = Im(x(t )) = Im( Fˆei (ωt +φ ) ) = Fˆ sin(ωt + φ ) .
Fˆ Fˆ
e) Define a time vector with 500 equidistant components on the interval [0, 1] s.
Suppose the force amplitude is 10 N and that the frequency is equal to the 265th
component of the frequency vector. Evaluate the displacement, x, as a function of
time at times given by the components of the time vector t. Plot the force and
displacement as functions of time in two separate diagrams. How large is the time
shift between the force and the displacement?

49 (Level 3)
k1 k2
m1 m2

F1 F2
x1 x2

Figure 1 Mechanical 2-degree of freedom system excited by external force.

The displacement x = [x1 x2]T of a linear mechanical 2-degree of freedom system, see
Figure 1, excited with an external force, F = [F1 F2]T, is governed by a system of two
equations of motion,
 d 2 x1
m1 2 + (k1 + k 2 ) x1 − k 2 x2 = F1
dt
 2
 − k 2 x1 + m2 d x2 + k 2 x2 = F2
 dt 2

or in matrix-vector notation,

d 2x m 0 k + k − k2 
M + Kx = F , M =  1  , K= 1 2
dt 2
0 m2   − k2 k 2 

Assume the external force is harmonic, i e sinuousoidal,

 Fˆ sin(2πft )   Fˆ1 
F(t ) =  1
ˆ  =  ˆ  sin(ωt ) = Fˆ sin(ωt ) , ω = 2πf ,
 F2 sin(2πft )  F2 

where F̂1 and F̂2 are force amplitudes. Since the system is linear it can be shown that the
displacement caused by this harmonic force is also harmonic with the same frequency,

 xˆ 
x(t ) =  1  sin(ωt ) = xˆ sin(ωt ) ,
 xˆ2 

15
and that the acceleration vector is,

d 2 x  xˆ1  d 2 (sin(ωt )) xˆ
2 1
2
=   2
= −ω   sin(ωt ) = −ω 2 x .
dt  xˆ2  dt  xˆ2 

The governing equation for the motion of the harmonically excited system can then be
written,
− ω 2 Mxˆ sin(ωt ) + Kxˆ sin(ωt ) = Fˆ sin(ωt ) ,

or since sin(ωt) is not identical to zero,

(−ω 2 M + K )xˆ = Fˆ or A(ω )xˆ = Fˆ .

Thus the coefficient matrix is dependent on the excitation frequency.

In a particular case the system parameters have the values; m1 = 1 kg, m2 = 2 kg, k1 = 100 kN/m
and k2 = 50 kN/m. The external force vector has the amplitude F1 = 10 N and F2 = 10 N for all
frequencies.

a) Use MATLAB to calculate the displacement vector for all frequencies in the
frequency vector f ∈ [0.1 , 100] Hz using frequency step 0.1 Hz.

b) Plot the absolute value of the displacement components in a lin-log diagram.

c) Determine the frequency where the absolute value of the displacement for mass 1,
i.e. x1, is maximum.

d) Determine the displacement amplitude vector x̂ at the frequency determined in c).

Normalize, i.e. multiply with a constant factor, the amplitude vector so that its first
component becomes 1.

50 (Level 2)
Consider the mechanical 2 degree of freedom system in exercise 49. For harmonic excitation
forces, F(t ) = Fˆ sin(2πft ) , the equations of motion governing the motion x of the system can
be written,
(−ω 2M + K )xˆ = Fˆ , (1)
m 0 k + k − k 2   xˆ   Fˆ 
where M =  1  , K= 1 2  , xˆ =  1  , Fˆ =  1  and ω = 2πf .
 0 m2   − k2 k2   xˆ2  ˆ
 F2 
In the absence of exciting forces, i e F = 0, the equations of motion (1) is an eigenvalue
problem,
(−ω 2M + K )xˆ = 0 . (2)
The eigenvalues are interpreted as particular values of the angular frequency squared for
which the homogenous (F = 0) equations of motions have non-trivial solutions. These
frequency values are denoted eigen-frequencies.

Suppose m1 = 1 kg, m2 = 2 kg, k1 = 100 kN/m and k2 = 50 kN/m.

16
a) Solve the eigenvalue problem given by (2) for parameter values as specified above.
Calculate the corresponding eigenfrequencies f. What is the value of the second
eigenfrequency?

b) Normalize the second eigenvector, i.e. the second column vector of the eigenvector
matrix, so that its first component becomes 1. What are the values of the components
of the normalized second eigenvector?

c) Consider the maximum displacement components calculated for the same system in
exercise 49 parts c) and d). Compare the frequency determined in 49 c) with the
second eigenfrequencies determined in part a) above. Also compare the normalized
second eigenvector with the normalized maximum displacement vector determined
in 49 d) by calculating the relative difference between their components. Discuss the
physical interpretation of the eigenvectors.

51 (Level 1)
The transfer function h for a linear system describes the relation between the input signal to
and output signal from the system. For a particular system the transfer function is defined as,
1 1
h(ω ) = − ,
ω − 1 − 0,5i ω + 1 − 0,5i
where ω is a complex-valued angular frequency.

a) Evaluate h on the rectangular domain given by,

Re(ω) ∈ [-2 , 2] and Im(ω) ∈ [-0,5 , 1,5].
Use an equidistant mesh with 40 points along the real axis and 20 points along the
imaginary axis. What is the value of ω in point 24 along the real axis and point 14
along the imaginary axis? What is the corresponding value of the transfer function h?

b) Use the command surf to create a 3-dimensional graph of the transfer function.
(i) Plot the real part of h as a function of ω.
(ii) Plot the imaginary part of h as a function of ω.
(iii) Plot the base 10 logarithm of the absolute value of h as a function of ω.

52 (Level 1)
A signal is a physical quantity varying with time. The sound pressure p, for instance, is he
time varying pressure obtained when the average pressure, i e the static pressure p0 , is
subtracted from the total pressure ptot,
p (t ) = ptot (t ) − p0 . (1)
Calculating the strength of the sound by simply taking the average of the sound pressure gives
strength 0 (from equation 1) which is not a useful result. For this reason the root mean square
or rms value of the signal is introduced as a measure of the signal strength. In terms of the
sound pressure, ~p , it is defined as,
T
~ 1
p =
T ∫
0
p 2 ( t ) dt (2)

where T, the integration time, is the duration of the signal or the measurement time.

17
 0.015

0.01
Sound pressure, [Pa]

0.005

-0.005

-0.01
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Time [s]
Figure 1 In a digital measurement instrument the continuous signal is sampled at discrete times with equidistant
time intervals ∆t.

Modern digital measurement instruments measures or samples, the signal at discrete

equidistant times tn, n = 1, 2, 3, … N, see figure 1. The integral in definition (2), i e the area of
the surface between the squared sound pressure curve and the time axis, is then estimated with
the sum,
T
1 1 N 2 1 N 2
T ∫0
p 2
dt ≈ ∑ n
N∆t n =1
p ∆ t = ∑ pn ,
N n =1
(3)

where ∆t is the time increment between two samples, N is the number of samples taken during
the measurement time T and pn is the sound pressure sampled at time tn , see figure 2. Hence,
the rms value is estimated as,
~ 1 N 2
p≈ ∑ pn .
N n =1
(4)

The static atmospheric pressure is normally 1·105 Pa, whereas the rms value of an ordinary,
not too high, sound pressure is in the order of 0.01 Pa. A harmonic, i.e. purely sinuousoidal,
pressure with frequency 100 Hz and peak sound pressure 0.01 Pa can be written as,

ptot (t ) = p0 + p(t ) = 1 ⋅ 105 + 0.01sin(2π 100t ) (5)

An exact evaluation according to the definition gives, when the integration time becomes
large, the rms value, ~
p = 0.01 2 .

18
 -4
x 10 Squared sound pressure, [Pa*Pa]
1.4

1.2

0.8

0.6 • • •

0.4

0.2

0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Time [s]
∆t
Figure 2 The area of the surface between the squared signal and the time axis is estimated with a sum.

a) Sample the pressure ptot given by (5) at equidistant times tn = (0, 0.0025 , 0.005, …
0.5) s. Plot a single period of the sound pressure p as function of time t. The period T
is the inverse of the frequency. Use Equation (4) and calculate the rms value of the
sound pressure p. Also calculate the relative error in the estimated rms value.

A method to increase the accuracy of digital measurements is to increase the sampling rate,
i.e. decrease the time increment between the samples.

b) Sample the pressure at equidistant times tn = (0, 0.00025 , 0.0005, … 0.5) s. Plot a
single period of the sound pressure as function of time t. Calculate the rms value of
the sound pressure and the relative error. Compare with results in a).

53 (Level 1)
The sound pressure level in dB is defined,
~
p2
L p = 10 ⋅ log , (1)
(2 ⋅ 10 − 5 ) 2
where ~p is the root mean square value (see the definition in exercise 52) of the sound
pressure.

a) Write an M-file that reads the rms sound pressure from the keyboard (Hint: Use the
command input.) and calculates the sound pressure level in dB. What is the sound
pressure level when the rms sound pressure is 0.1 Pa?

19
Usually the sound pressure varies with the frequency (tone) of the sound. Suppose the rms
sound pressures for different frequencies are collected in a sound pressure vector p and that
the corresponding frequency values are collected in the frequency vector f. The sound
pressure level in dB at a specific frequency can then be calculated according to (1) where p is
the rms sound pressure at the specific frequency.
The binary file soundpress.bin contains sound pressure data measured in the
exhaust system of a Diesel engine. The data is stored on the floating point format
'float32' in the sequence: f1, p1, f2, p2, f3, p3, … .

b) Read the data in the binary file soundpress.bin and calculate the sound pressure level
as a function of frequency. Use a dialog box to specify the name and path of the file.
First use the command uigetfile then use fopen, fread and finally fclose.
Modify the code so that it calculates the sound pressure level as a function of
frequency, i.e. it evaluates the sound pressure level for all sound pressure values in
the vector p.

c) Modify the code so that it locates the maximum value of the sound pressure level and
the frequency for which it occurs. What is the maximum level and at what frequency
does it occur?

d) Plot the sound pressure level as a function of frequency in a lin-lin diagram. Do not
forget to assign proper labels to the axis.

The total sound pressure level including the energy at all frequencies is defined as,
∑n pn2
L p ,tot = 10 ⋅ log , (2)
(2 ⋅ 10 − 5 ) 2
where pn is the nth component of the sound pressure vector p.

e) Modify the M-file so that it also calculates the total sound pressure level for the
sound pressure vector p. What is the total measured sound pressure level for the
Diesel engine?

f) Use the commands text and num2str to add text giving the maximum and total
sound pressure levels to the diagram.

54 (Level 3)
Modify the M-file written in the previous exercise so that the calculations of the sound
pressure level, its maximum and its total value are calculated by a separate function M-file.

55 (Level 3)
Frequency analysers used in acoustics and vibrations for measuring frequency spectra usually
provide results in terms of narrow-band spectra. In a power spectrum the energy of a physical
entity is measured as a function of frequency. The sound power spectrum, for instance, is the
square of the rms sound pressure as a function of frequency, ~
p2( f ) .

a) Design a function M-file yband = band(fl,fu,f,y) that sums the energy of

the lines within the interval [fl, fu].

20
b) Design a function M-file [throctf,throcty] = throct(f,y) that converts
the narrow-band spectrum defined by the frequency vector f and the power spectrum
y to a third octave band spectrum and labels each band with its center frequency. The
band limits are given in the table below:

Band Center frequency [Hz] Lower limit [Hz] Upper limit [Hz]
number
1 1,25 1,12 1,41
2 1,6 1,41 1,78
3 2 1,78 2,24
4 2,5 2,24 2,82
5 3,15 2,82 3,55
6 4 3,55 4,47
7 5 4,47 5,62
8 6,3 5,62 7,08
9 8 7,08 8,91
10 10 8,91 11,2
11 12,5 11,2 14,1
12 16 14,1 17,8
13 20 17,8 22,4
14 25 22,4 28,2
15 31,5 28,2 35,5
16 40 35,5 44,7
17 50 44,7 56,2
18 63 56,2 70,8
19 80 70,8 89,1
20 100 89,1 112
21 125 112 141
22 160 141 178
23 200 178 224
24 250 224 282
25 315 282 355
26 400 355 447
27 500 447 562
28 630 562 708
29 800 708 891
30 1000 891 1120
31 1250 1120 1410
32 1600 1410 1780
33 2000 1780 2240
34 2500 2240 2820
35 3150 2820 3550
36 4000 3550 4470
37 5000 4470 5620
38 6300 5620 7080
39 8000 7080 8910
40 10000 8910 11200

c) Try the codes developed in a) and b) on the sound pressure data stored in the binary
file soundpress.bin that contains sound pressure data measured in the exhaust
system of a Diesel engine. The data is stored on the floating point format
'float32' in the sequence: f1, p1, f2, p2, f3, p3, … .

21
 (i) Calculate the sound pressure level, Lp, defined as,
~
p2
L p = 10 ⋅ log .
(2 ⋅ 10 − 5 ) 2
(ii) Plot the spectrum, i e the sound pressure level as function of frequency in a
graph.

56 (Level 3)
The standard trigonometric functions require angular input values in radians.

a) Write a function M-file that converts an angle, specified in radians, to degrees.

b) Write a function M-file that calculates the inverse tangent (arcustangent) to an input
argument and then calls the function M-file written in a) to convert the resulting
angle from radians to degrees.

c) Try the written code on the following input values,

(i) x = 1,
(ii) x = 1/0,
(iii) x = 0.5.

57 (Level 3)
Vibration measurement data, i e acceleration data, are stored in 5 different files: acc1.dat,
acc2.dat, … and acc5.dat. Suppose the number assigned to the variable
n ∈ [1, 2, 3, 4, 5] determines the name of the file that is to be read, accn.dat.

a) Write MATLAB commands that, for a fixed numeric value of n, assigns the correct
file name to string variable filename.

If a file name is to be assigned repeatedly several times it may be worthwhile to write a

function procedure that performs the task whenever called.

b) Write a function M-file,

function filename = name(n)

% n – integer in range [1,5]
%
...

that produces the correct file name string as output variable filename.

58 (Level 1)
The displacement field, w, of a vibrating rectangular membrane with fixed edges, see Figure
1, can be approximated with a sum of characteristic components (vibration modes),
πx πy πx 2πy nπ x mπy
w( x, y ) = c11 sin ⋅ sin + c12 sin ⋅ sin + L + cmn sin ⋅ sin + L,
Lx Ly Lx Ly Lx Ly
where Lx and Ly are the length of the edges and cmn is a coefficient specifying the contribution
from mode mn to the displacement field.

22
 y
Lx

Ly

x
Figure 1 Rectangular membrane.

Consider a membrane with dimensions 1 × 2 m2. Plot the following vibration modes using
the commands mesh and contour. Make sure the resolution of the mesh is sufficiently
high.
πy
a) sin(πx) ⋅ sin ,
2
πy
b) sin( 2πx) ⋅ sin ,
2
2πy
c) sin(3πx) ⋅ sin .
2

59 (Level 3)
Vibration testing aims to test whether or not a product can operate in its supposed
environment. During the test the test object is mounted to a shaker, see Figure 1, used to
enforce a motion specified in the test schedule as a power density spectrum for the
acceleration.
Shaker
Test object

Sliding table

Foundation
Figure 1 In a vibration test the test object is mounted to a shaker that enforces a predefined motion to the test
object.

Suppose a microelectronic system developed for aeronautic applications is subject to a

vibration test aiming to investigate its ability to operate in a vibrating environment. The
vibration environment is specified in terms of a acceleration power density spectrum, i.e. the
mean square value, see exercise 52, of the acceleration is given as function of frequency, see
Figure 2.

23
 a~ 2 [m2/s4/Hz]
23,4

8,4
5,0
f [Hz]

15 75 300 800 1000

Figure 2 Acceleration power density spectrum used during test.

A parameter of interest in vibration testing is the total energy supplied to the test object. Since
the energy of a system vibrating at a specific frequency is proportional to the square of the
acceleration amplitude, the total vibration energy is proportional to the acceleration amplitude
squares summed over all frequencies. Mathematically then the total energy is proportional to
the area of the surface below the acceleration power density spectrum, i.e. the integral of the
power density spectrum,
f max

W ∝ ∫ a~
2
df .
f min

In MATLAB numeric integration or quadrature is performed using the command quad . For
information on the use of quad see the MATLAB Help browser.

a) The quad command requires the integrand to be specified as a function providing a

vectorial output y containing integrand values for the vectorial input x. Write a
function M-file that calculates the values of an arbitrary piece-wise linear function,
see for instance the power density spectrum function in figure 2 above.

b) Design a MATLAB code for calculating the integral of a piece-wise linear power
density spectrum. Use MATLAB function for numeric integration quad and the
function developed in part a).

c) Try the code on the spectrum in Figure 2.

60 (Level 3)
Consider a structure consisting of three rods, denoted 1, 2 and 3, mounted to a rigid base
according to the figure 1. Assume the rods have lengths L1, L2 and L3, cross section areas A1,
A2 and A3 and that they are made from materials with Young’s moduli E1, E2 and E3.

24
 Py P = Pxex + Pyey
γ
B Px

L3
L2 E 3 , A3
p
n A2
E1 , A1 L1 o E2

α β

Figure 1 Structure built-up from three rods. An external force, P = Pxex + Pyey , is applied to the
junction, node B.

The structure is subjected to an external load P, with components Px and Py, which causes the
structure to deform. The displacement vector δ, with components δx and δy, in the loading
point B is related to the load P by the matrix equation,
P = K ⋅δ ,
or in component form,
 Px   k1 cos 2 β + k 3 sin 2 β k1 cos α cos β − k 3 sin β cos β  δ x 
 = ⋅ ,
 Py  k1 cos α cos β − k 3 sin β cos β k1 cos 2 α + k 2 + k 3 cos 2 β  δ y 

where, k1 = E1 A1 / L1 , k2 = E2 A2 / L2 and k3 = E3 A3 / L3 , are the longitudinal stiffnesses of the

individual rods. The matrix K, which relates the displacement to the load, is denoted the
stiffness matrix.

a) Write an M-file that calculates the displacement vector δ at point B for a known
force with components Px and Py.

Suppose the following data is valid in a particular case,

L1 = 11 m, L2 = 10 m, L3 = 14 m, A1 = 0.001 m2, A2 = 0.001 m2, A3 = 0.001 m2,
E1 = E2 = E2 = 73 GN/m2 (Aluminium) and P = 10000 N

b) Use the M-file developed in part a) to calculate the displacement at point B:

(i) when the load applied to point B is horizontal and
(ii) when the load applied to point B is vertical.

The code can be made more effective by performing repetitive calculations in a function m-
file.

c) Write a function M-file displ = invK_p(gamma,P,L) that for a given set of

input data, gamma (γ), P (P) and L (L1, L2 and L3), calculates the stiffness matrix K,

25
 the right-hand side vector P and finally solves the system of equations for the node
displacement displ (δ). Validate the code by repeating the calculations performed
in part b).

d) Suppose the load is applied to point B at angle γ with respect to the horizontal axis
see Figure 1. Modify the m-file so that it calculates the displacement components δx
and δy as functions of the load angle γ in the interval γ ∈ [0, 2π]. Plot the
displacement components as functions of the load angle γ in a lin-lin diagram. For
which γ values is the horizontal displacement component zero?

61 (Level 3)
Vibrating machines are important sources of noise in vehicles and buildings. The noise can
either be transmitted directly from the machine as noise radiated from its surfaces or
indirectly as vibrations via the machine mounts. Successful noise abatement in vehicles
therefore requires knowledge information on the amount of acoustic energy transmitted via
the various transmission parts. Noise reducing actions is then aimed for the most important of
the transmission parts. In this exercise method to determine the vibration energy transmitted
from a machine to its foundation is demonstrated.
Suppose a machine is vibrating with a dominant frequency at 90 Hz. The vibrations
is transmitted to the foundation via four mounts and then finally radiated as sound from the
vibrating foundation. The vibration energy transmitted via a mounting can be determined by
simultaneously measuring the force, F, and the velocity, v, in the mounting. In practise the
velocity is measured indirectly by measuring the acceleration, see Figure 1.

Machine

Machine mount
a(t)
F(t) Accelerometer
Force transducer

a
F Foundation

Figure 1 In each of the four mounting points the force and the velocity (acceleration) is
measured using force transducers and accelerometers.

When the vibrations are dominated by a single frequency the following formula may be used
to calculate the mechanical power transmitted over the mounting,
1
W = Re(Fv*) ,
2
where W is the time average of the power, F is the complex force, v is the complex velocity
and * denotes complex conjugation. When a single frequency dominates the measured force
and acceleration signals can be written,
F (t ) = Fˆ cos(2πt + φ F ) = Re( Fˆe i ( 2πft +φ F ) ) ,

26
 a (t ) = aˆ cos(2πt + φ v ) = Re(aˆe i ( 2πft +φv ) ) .
The complex force and acceleration are then introduced as,
F(t ) = Fˆe i ( 2πft +φ F ) ,
a(t ) = aˆe i ( 2πft +φa ) ,
where ^ denotes amplitude or peak value, f is the frequency and φF and φa are phase angles for
the force and acceleration.

a) Write a function M-file, W = powcalc(a,F,f) , that calculates the mechanical

power transmitted via a single mounting for a specified acceleration, force and
frequency.

b) Develop a MATLAB code that calculates the total mechanical power transmitted
over N mounting points.

c) Try the code on a case where vibration energy at 90 Hz is transmitted via four points
with measured data according to table 1 below.

Table 1 Measured acceleration and force data in four mounting points.

Point 1 2 3 4

Force Amplitude Phase Amplitude Phase Amplitude Phase Amplitude Phase

[N] [degrees] [N] [degrees] [N] [degrees] [N] [degrees]
13 0 17 33 8 70 21 85
Acc Amplitude Phase Amplitude Phase Amplitude Phase Amplitude Phase
[N] [degrees] [N] [degrees] [N] [degrees] [N] [degrees]
1,4 192 1,1 160 2,1 212 3,4 263

62 (Level 3)
The stress state in a loaded solid body is given by the stress matrix S. If a cut, with normal n,
is made in the body (Figure 1) then the following stresses are found on the cut surface.
σ

τ
z cut surface
solid body

Figure 1 Stresses in a solid body.

Stress in the direction of n, i.e.the normal stress,

σ = nTSn . (1)
Stress perpendicular to n, i.e.the shear stress,

27
 2
τ = ( Sn − σ 2 )1 2 . (2)
This means that σ and τ will be different for each n. For a given plane the stresses σ and τ can
be plotted as a point in a diagram with σ on the x-axis and τ on the y-axis. Let,
10 4 8 
S =  4 3 1  MPa.
 
 8 1 25
In this exercise we shall use randomly oriented cuts.

a) The easiest way to achieve a random normal to the cut n = [ nx ny nz ]T is the

following. Write a function M-file snorm = surfnorm that determines nx , ny
and nz randomly in the interval ]-1, 1[. Then scale the length of the normal vector so
that its length, i.e. its absolute value, is 1.

b) Write a function M-file [nstress,shear] = stress(S,n) that determines a

random normal direction n by calling the function M-file snorm and then calculates
the normal stress σ (nstress) and the shear stress τ (shear) for n by using
Equations (1) and (2).

c) Write a code nrandstress.m that, for a specified stress matrix S, calculates the
normal stress (nstressn) and shear (shearn) for N random cuts. Save the
corresponding shear and normal stress values in two vectors. Plot the shear as a
function of the normal stress in a diagram with σ on the x-axis and τ on the y-axis.
Enforce equal scaling on the axis. Also, use * as the plot symbol (marker). Execute
the code for N = 500. The final plot will show the possible combinations of σ and τ
that can be found by cutting in different orientations. What is the shape of the graph?

d) Modify the code to find the eigenvalues and eigenvectors of S. Calculate the normal
stress and shear for n being equal to the three eigenvectors. Comment on the results.
Try to find out how the eigen-values relate to the plot in c).

e) On an unloaded surface of a body all components in a row and the corresponding

column of S is zero. Pick a row and the corresponding column in S given above, and
set the components to zero. Calculate the eigenvalues and plot the σ(τ) graph for this
case.

f) Redo a) and b) using another symmetric stress matrix S = ST, for example,
10 2 0 
S =  2 10 1  MPa.
 
 0 1 10

63 (Level 3)
Parametric or semi-empiric models are often used to describe physical phenomena. A
parametric model is usually based on a theoretical model that prescribes a relationship
between two variables on a certain mathematical form. For example, a parametric model of a
linear relationship between variables x and y can be written,
y = c1x + c2 ,

28
where the model parameters c1 and c2 are constants.
In physics the model parameters are often determined experimentally. In principle
two independent measurements are needed to determine the parameters. In practice, however,
the influence of measurement errors must be reduced by using more than two independent
measurements. A least squares approximation of the model parameters can then be estimated
by solving an overdetermined system of equations,
 x1 1  y1 
 x 1 c  
 2   1  =  y2  .
 M M c2   M 
   y N 
 xN 1
Suppose that you have N independent measured data pairs (xn, yn), n = 1, 2, 3 …, N, stored on
floating point format 'float32' in a binary file xydata.bin. The storing sequence is
x1, y1, x2, y2, x3, y3, … .

a) Write a M-file that reads the data file, calls for a function M-file that solves the
overdetermined system for the model parameters and finally plots the measured
points in a graph together with the estimated model curve.

b) Try the code on the data given in xydata.bin.

64 (Level 3)
Transducers are used in mechanical measurements for transforming a mechanical quantity to
a measurable electrical quantity. Important examples are strain gauges for strain or
deformation measurements, force transducers, accelerometers and microphones. In order to
use a transducer the relation between the mechanical input quantity and the electrical output
quantity must be known.
Consider a force transducer that transforms an input force, F [N], to an output
voltage, U [V]. In its linear range the relation between the input F and the output U is
assumed to be,
U = CF + U 0 ,
where C [V/N] is the transducer sensitivity and U0 is a bias voltage.
To determine the transducer relation, i.e.the model parameters C and U0, a series of
five independent measurements was performed. In each measurement a known force was
applied to the transducer whose output voltage was measured with a voltmeter. The results are
tabulated in table 1.
Table 1 Results from calibration measurements.
Force [N] 1 5 10 20 50
Voltage [V] 0.08 0.11 0.14 0.28 0.55

Determine the optimum transducer model in the least squares sense.

65 (Level 2)
Reduction of vibrations transmitted from a machine to a building is achieved by placing the
machine on soft elastic mounts. The achieved reduction is highly dependent on the mount
elasticity or stiffness. Therefore producers of mounts should provide measured dynamic
stiffness data for the mounts. The dynamic stiffness, k, is defined as the ratio of the force
amplitude, F̂ , over the displacement amplitude, x̂ , both as a functions of frequency,

29
 Fˆ ( f )
k( f ) = .
xˆ ( f )
The measurement set-up is explained in Figure 1 below.

Fˆ sin( 2πft )

Mount
xˆ sin( 2πft + ϕ )

Rigid foundation
Figure 1 When the mount stiffness is measured the mount is fixed to a rigid foundation and loaded with a
harmonic force F̂ sin(2πft). The loading force and the corresponding spring deformation is measured
and their amplitude ratio is the spring stiffness.

One specific producer uses a measurement system that measures the frequency-force-
displacement triplets and stores them on a binary file stiff.dat. The storage precision is
'float32' and the storage sequence is freq1-force1-disp1-freq2-force2-disp2-freq3-… .

a) Read the frequency-force-displacement triplets into a data vector.

b) Partition the data vector in a frequency vector, a force vector and a displacement
vector.

c) Calculate the dynamic stiffness, k, as defined above and plot it as a function of

frequency in a lin-log graph.

66 (Level 3)
The figure shows a mechanical single degree of freedom model used to simulate the motion of
a car running on a road with a bump.
z Vehicle
z1
m1 V
z0 ( x )
k1 c1

x
Road

Figure 1 Mechanical single degree of freedom (SDOF) system for simulating a car running over a road bump.

The equations of motion for the mass and the spring relation yields the following second order
differential equation governing the displacement z of the mass m.
d 2z dz
m 2 + c + kz = kz0 (t ) , (1)
dt dt

30
where m is the mass, k is the spring constant, c is the damping constant and t is the time. The
right-hand side of (1) is the spring force acting on the mass when the road surface profile is z0
at time t.
MATLAB provides routines, e.g. ode45 for solving this type of differential
equations. Use of ode45 requires the equation to be formulated as a first order equation. By
introducing a vector y as,
 dz 
 
y =  dt  , (2a)
 z 
and by using,
d 2z 
dy  dt 2 
= , (2b)
dt  dz 
 dt 
the second order equation (1) can be reformulated to the a first order system of differential
equations. On matrix form this system reads,

 1 0  dy  k m  c m dz0 c m k m
0 1 ⋅ dt =  0  ⋅ z0 +  0  ⋅ dt −  − 1 0 
⋅y , (3)
      

dz0 dz
where = {x = Vt} = V 0 .
dt dx

The equation solving routine ode45 requires the right-hand side of (3) to be calculated at
time t in a function m-file specified in the parameter list. In this case, as can be seen in (3),
this function must contain a calculation of the position of the car at time t, the road profile z0
at the position of the car and finally the right-hand side of (3).

a) Suppose the road surface can be described by the following function,

 cos 2 (πx), − 0 .5 ≤ x ≤ 0 .5
z0 ( x) = zˆ0 ⋅ f ( x) = zˆ0 ⋅  , (4)
0 . 001 ⋅ random ( −1,1), x < − 0 . 5 ∨ x > 0 . 5
where random(-1,1) means a random number on the interval ]-1, 1[. Write a function
M-file f = profile(x) that calculates the road profile f as a function of the car
position x according to (4).
Also, write a function M-file fprim = profileprim(x) that calculates the x-
derivative of the road profile f as a function of the car position x according to (4).
You may assume that the derivative of the random part of the profile is zero.

b) Write a function M-file funcbump(t,y) that calculates the right-hand side of (3).
Note that the function output should be a column vector with the two components of
the right-hand side vector.

In its ordinary use the equation solver ode45 should be called as,
[T, displ] = ode45('funcbump',[tmin, tmax], ystart)
where T is the time vector for which the displacement displ have been calculated by the
solver, tmin and tmax are the lower and upper time limits of the simulation time interval and

31
finally ystart are the starting conditions in terms of the displacement and its time
derivative at tmin. If the solution is to be evaluated at specific time values the time interval
specification above can be replaced with a time vector with components equal to the times for
evaluation.

c) Write an M-file for simulating the motion of a car governed by Equation (1). Make
use of the command ode45 and the function M-files funcbump, profile and
profileprim. Note that parameter values needed in more than one M-file can be
made available by using global declaration.

Suppose the model parameters have the values: m = 250 kg, c = 200 Ns/m and k = 2000 N/m.

d) Simulate the motion of the car running at 50 km/h over a road profile with size, ẑ0 ,
equal to 0.3 m, for time in the interval [-5, 20] seconds. Assume starting conditions
equal to 0. Plot the simulated displacement as function of time in a diagram.

67 (Level 3)
Animated movies can be created using sequential matrix multiplication. In this exercise you
will create a triangle that is used to demonstrate animated sequences.

a) First use the figure and axis commands to create a figure window with fixed
corners in the points {(-10,-10) (10,-10) (10,10) (-10,10)}.

Line figures or wire frames are drawings created by connecting the corner points of a
geometrical object with straight lines. In MATLAB the command line(x,y) is designed to
draw straight lines between the corner point coordinates defined by x and y.

b) A triangle has corner points (-5, -5), (5, -5) and (0, 5). Define a row vector x and a
row vector y containing the x- and y-coordinates of the triangle corner points. Note
that in order to create a closed triangle you have to end the line in the starting corner.
Finally, use the command line to draw the triangle in the figure window. Make the
triangle lines black.

Figure motion can be animated using basic matrix operations. To illustrate this define a corner
point matrix with the x-coordinate vector as first row and the y-coordinates as second row.
Pure translation means that all corner points are moved along the same displacement
vector. Hence, the displacement in x-direction should be added to all x-coordinates and the
displacement in the y-direction should be added to all y-coordinates in the corner point matrix.
To do this, first introduce a unit rigid body translation matrix having the same dimension as
the corner point matrix and then a 2 by 2 point displacement matrix. The elements of the unit
rigid body translation matrix are all 1 and the point displacement matrix is a diagonal matrix
with the diagonal elements equal to the point displacements in x- and y-directions.
The corner point matrix of the translated matrix is calculated by multiplying the unit
rigid body translation matrix from the left with the point displacement matrix and, finally, add
the product to the original corner point matrix.

c) Translate the original triangle 2 units in the positive horizontal direction, i.e. in the
positive x-direction.
(i) Define a unit rigid body translation matrix for the triangle.

32
 (ii) Define a point displacement matrix displacing the point 2 units in the
horizontal (x) direction and 0 units in the vertical (y) direction.
(iii) Calculate corner point matrix of the translated triangle.
(iv) Draw the translated triangle with blue colour.

Zooming a figure can be accomplished by multiplying the corner point matrix with a 2 by 2
diagonal scaling matrix (see exercise 37). Uniform, isotropic, scaling is obtained when the
diagonal elements are equal. The elements on the diagonal represent the scaling factors with
respect to the x- and y-directions. The corner point matrix of the zoomed figure is obtained by
multiplying the original corner point matrix from the left with the scaling matrix.

d) Define an isotropic scaling matrix with scaling factor 1.2. Calculate the corner point
matrix of the zoomed triangle and draw it using green colour.

Rotating a figure is performed using a rotation matrix (see exercise 37) defined as,
cos φ − sin φ 
B= ,
 sin φ cos φ 
where φ is the rotation angle. Multiplying a vector with B from the left rotates the vector φ
radians in the positive (counter-clock wise in a xy-diagram) direction. A figure is rotated by
multiplying its corner point matrix from the left with a rotation matrix.

e) Define a rotation matrix rotating the figure 15 degrees in positive direction. Note that
the trigonometric functions assume angles specified in radians. Calculate the corner
point matrix for the rotated triangle and plot the rotated triangle using red colour.

68 (Level 3)
Motion sequences can be animated using matrix multiplication and graphic commands in a
loop

a) Create the triangle P defined in exercise 67 and a matrix B that rotates the triangle 1
degree in positive, i.e. counter-clockwise direction.

b) Animate the triangle rotating 360 degrees in counter-clockwise direction.

69 (Level 3)
Consider the mechanical single degree of freedom system with mass m, spring constant k and
damping constant c depicted in Figure 1.
x

m
F( t)
k c

Figure 1 A mechanical single degree of freedom system.

33
For a sinuousoidal exciting force F oscillating at frequency f the displacement x as function of
time t can be written,
F 2πfc
x(t ) = sin(2πft + arctan ).
(k − (2πf ) m) + (2πf ) c
2 2 2 2 k − (2πf ) 2 m
In a particular case the system parameters are:
m = 1 kg,
k = 1·104 N/m,
c = 1 Ns/m,
F = 50 N
The displacement was calculated at 500 equidistant frequency points on the interval [0,30] Hz

Write an M-file that animates the motion at the frequency given by the 265th frequency
component. In the animation the system given in figure 1 should be animated.

70 (Level 3)
Consider the car running over a road bump in exercise 66. Write an M-file that animates the
motion as function of time as the car passes the bump.

71 (Level 3)
The purpose of a Graphic User Interface (GUI) is to make it possible to use the program
without detailed knowledge of the code. This exercise shows a step by step design of a simple
GUI. In this exercise a GUI for calculating and plotting the Frequency Response Function,
FRF, for a Single Degree Of Freedom system, SDOF, is developed.
A mechanical single degree of freedom system with mass m, damping c and spring
constant k excited by an external force F is shown in Figure 1.
x
Mass m
F

Spring κ c Damper

Ground
Figure 1 Mechanical single degree of freedom system.

The equation of motion for the system reads,

d 2x dx
m 2 + c + kx = F . (1)
dt dt
Lots of valuable information on the system is provided in the frequency response function
which can be calculated by assuming that the exciting force is harmonic, i.e. it can be written,
F (t ) = Fe iωt , (2)
where ω is the angular frequency. It can be shown that the displacement x caused by a force
given by (2) is,
x(t ) = xeiϕ e iωt , (3)
where ϕ is the phase angle of the displacement relative to the exciting force. The frequency
response function is defined as the ratio between the displacement x and the force F, that is,
x(t ) x iϕ
H (ω ) = = ⋅e . (4)
F (t ) F
After inserting (2) and (3) into (1) the frequency response function (4) becomes,

34
 ic
1 1 i arctan
H (ω ) = = e k −ω 2 m
. (5)
k + iω c − ω m
2
( k − ω m) + ω c
2 2 2 2

This expression is the frequency response function for the mechanical single degree of
freedom system.
The first action taken when a GUI is created is to decide what is to be done and the
graphical appearance of the interface on the screen. In this case the interface should contain
one part where the system parameters are specified, one part with buttons for starting
calculations and exiting the interface and finally one part for graphical presentation of the
results. Figure 2 shows a schematic picture of one possible screen interface.
Specification
of system
parameters
Plotting region
Specification
of calculation
parameters

Action
buttons

Figure 2 Schematic picture of suggested screen interface for calculating mechanical single degree of freedom
systems.

There are several possible methods to implement a GUI. An often used method is based on
recursive calling of a function M-file, meaning that the code consists of a function m-file that
calls itself for performing various tasks. The basic idea of this programming technique is to
use a function m-file that calls itself with an input string variable whose value controls the
program flow. When a button in the user interface window is pushed the variable should be
set to a specific value that when the function is called again allows the expected set of
commands to be performed.
The first time the function is called immediately after starting the code. The user
interface window, its text parts and various buttons must then be initialised. One way to solve
this is to make use of the nargin command which, when used inside a function M-file
returns the number of input arguments. Hence, at the first call, when there are no input
parameters (no buttons in the user interface are pushed) and the user interface should be
initialised, start with the following commands,

function Sdofgui(action)

% ...

if nargin < 1
action = 'initialise';
end

% Here follows an "if...elseif...end" construct that controls % the

actions taken when the buttons in the user interface are % pushed.
The values of two strings may be compared with the
% command strcmp which returns the value 1, i.e. becomes
% true when the strings are equal.

if strcmp(action,’initialise’)

35
 % Initialise user interface window
...
elseif strcmp(action,’ButtonAction1’)
% Commands executed when button number 1 is pushed:
...
end

a) Enter the commands above in a function m-file ‘sdofgui.m’. Use the command
sdofFig = figure('position',[ ... ],...) to create a interface
window named sdofFig with suitable size, name and position etc.

Now it is time to initialise the first pushbutton. Pushbuttons of different kinds are created with
the command

uicontrol('style','pushbutton','position',[...],'callback',...'set
of commands;',...)

The callback variable is a string variable that contains the MATLAB commands that will
be executed when the button is pushed is entered.

b) Initialise a pushbutton named closeBtn which, when pushed, exits the program
and closes the interface window.

The next step is to initialize the various parts of the interface window. In this GUI it is
convenient to have one part where the model parameters, the mass m, the damping c and the
spring constant k are entered. In a second part the frequency range for the FRF calculation
must be entered. In a third part the calculations are performed and finally in a fourth part the
calculated FRF is plotted in diagrams.

c) Begin with creating a header MODEL PARAMETERS placed at a suitable position in

the interface window. This may be done with the command
uicontrol('style','text','string','MODEL ...';...).
Also create headers for entering frequency parameters, executing calculations and presenting
results, placed at suitable positions in the interface. Name the headers Frequency
range:, Calculation: and Diagrams: Frequency Response Function.

Below the header named Model Parameters: the numeric values of the model
parameters should be entered from the keyboard. To accomplish this some initial default
values of m, c and k must be defined.

d) Define initial default values of all model parameters in the interface initialization
part. Also initialize values of the minimum frequency, the maximum frequency and
the frequency increment for the FRF calculation. Then place a lead texts 'Mass m
[kg]:' etc at suitable locations. Finally continue in the same manner with
initialization of the frequency range parameters.

Numeric values entered from the keyboard can be introduced by using an editable field. A
field named massField where you can enter the mass from the keyboard during the
execution is created with the command

36
massField = uicontrol('style','edit','string',num2str(m),'callback',
'sdofgui(''newmass'');'...)

Note that the numeric value is converted to a string variable using the num2str command.
The change of value of the mass from the old to the new is accomplished by the callback
command where the function sdofgui calls itself with the action variable set to the string
'newmass'. The commands actually changing the value of the mass are then placed in the
elseif strcmp(action,'newmass')
...
elseif strcmp(action,…)
...
end
part of the program.

e) Change the program so that the model parameters and the frequency parameters can
be changed from the keyboard. Note that in order to obtain a changed numeric value
of the mass m you must first read the contents in the 'string' parameter of the
variable named massField. This is done with the
get(massField,'string') command which will return a string which is the
string equivalent to the numeric value of m. Then you must convert this string to its
numeric equivalent using the str2num command. Note that both the massField
variable and the mass m must be declared as global.

When all parameters have been updated to their correct values, the FRF calculation starts by
pushing a button.

f) Initialise a pushbutton named calcBtn which when pushed calculates the

frequency response function of the system according to the equations given above.
Label the button with the string 'FRF'.

Finally the calculated frequency response function should be plotted in a diagram. The
plotting may be performed with a Plot button, but it is more convenient to introduce the plot
commands in the same action part and directly after the calculation commands.
Since a frequency response function is complex valued two graphs are needed, one
showing the absolute value and one showing the phase angle. Areas for plotting are created
with the command axes('position',...) .

g) Create in the initialisation part two suitably positioned plotting areas named
absAxes and phaseAxes. Add lines for plotting the absolute value and the phase
to the action FRFcalc part. Note that the plotting areas are activated with the
subplot('...') command.

72 (Level 3)
The following graphic user interface lets the user select a measured signal and plot in a lin-lin
diagram. Add a pushbutton Sound to the graphic user interface so that the user can listen to
the signal. Try the code on the data given in the files gearvibr.bin (containing
acceleration signal measured on a gearbox), railvibr.bin (containing acceleration signal
measured on a rail during a train passage), heart_1.bin (containing pressure signal

37
measured on a healthy heart) and heart_2.bin (containing pressure signal measured on a
heart with a heart condition).
% sigplot.m
% GUI for plotting and listening to a user selected signal.
%
% Author: U Carlsson, 2001-08-23

% Create figure window for graph.

figWindow = figure('Name','Plot alternatives');

% Create file input button. Press 'File' calls the function 'firstplot'.

fileinpBtn = uicontrol('Style','pushbutton','string','File','position',...
[5,395,40,20],'callback','[tdat,pdat,fsampl] = firstplot;');

% Create 'Sound' button.

fileinpBtn = uicontrol('Style','pushbutton','string','Sound','position',...
[75,395,45,20],'callback','soundsc(pdat,fsampl);');

% Create exit pushbutton with red text.

exitBtn = uicontrol('Style','pushbutton','string','EXIT','position',...
[510,395,40,20],'foregroundcolor', [1 0 0],'callback','close;');
% End of sigplot.m

function [tdat,pdat,fsampl] = firstplot

% Brings template for file selection. Reads selected filename and path and plots
% spectrum in a lin-lin diagram. Output data are frequency and pressure amplitude
% vectors, fdat and pdat.

% Author: U Carlsson, 2001-08-23

% Call MATLAB function 'uigetfile' that brings file selection template.

[filename,pathname] = uigetfile('*.bin','Select binary data-file:');

% Change directory. Open file for reading binary floating point numbers.
% Read data. Close file.

cd(pathname);
fid = fopen(filename,'rb');
data = fread(fid,'float32');
fclose(fid);

% Partition data vector into time and signal vectors.

tdat = data(1:2:length(data));
pdat = data(2:2:length(data));

% Calculate the sampling frequency for correct sound reproduction.

fsampl = 1/(tdat(2)-tdat(1));

% Plot pressure signal in a lin-lin diagram. Define suitable axis labels.

plot(tdat,pdat);
xlabel('Time, [s]');
ylabel('Signal amplitude');

% End of file firstplot.m

38
 SOLUTIONS AND HINTS TO SOME EXERCISES
1
a) -3
b) 13.5000
c) -13.4000
d) -76.4000

2
a) >> 8+3*(2+3)/4
ans =
11.7500
b) >> -5+7*(1+4*(1+3))/3/(2+5)
ans =
0.6667
c) 68.3333

3
a) -11
b) 3.0021
c) 11.1250

4
a) 6.0060 ⋅103
b) 1.3152 ⋅108

5
a) Inf
b) NaN

6
All expressions in this exercise are complex-valued. The imaginary unit, i, is introduced in
MATLAB as a predefined name i or j.

a) >> (1+i)+(2–i) (or equivalently (1+j)+(2–j)) gives

ans =
3
b) 3.0000 + 1.0000i
c) 0.2000 + 0.6000i
d) 2

7
π is predefined in MATLAB as pi.
a) >> y=3*sin(pi/4)
y =
2.1213
b) -2.1213
c) 3.0000
d) -3.0000

39
8
a) 0.7071
b) -0.0292
c) 1.5708
d) 0.4636

9
a) 2.7183
b) 0.6931
c) 1.0137
d) 0.3010
e) 6.5321

10
Find the base 10 logarithm function using MATLAB Help browser.
a) >> y = sqrt(5*sin(pi/3))
y =
2.0809
b) 41.7647
c) 2.6665
d) 1.3029
e) 1.5708 (Expression is equivalent to arcsin(sin(π/2)) ).

11
Use the inverse trigonometric functions asin, acos, and atan. Remember that
trigonometric equations may have several solutions. Use help facilities to find the functions
domains and ranges in MATLAB.
a) >> v = asin(0.9)*180/pi
v =
64.158
b) 114.83
c) -7.4069
d) Yes, since the trigonometric functions are periodic there are an infinite number of
roots. See basic mathematics textbooks. Also see MATLAB Help.

12
Corrected code:

% radius.m
L = input('Enter the circumference L: ');
r = L/2/pi % Alternatively r = L/(2*pi)

Test run,

>> radius
Enter the circumference L: 2*pi
r =
1

40
13
a) >> pref=2e-5;
>> Lp=10*log10((35e-3/pref)^2)
Lp =
64.8608
b) 94.8073
c) 140

14
Execution with erroneous code:
>> xplusy
z =
7
z =
7

Corrected code:
% xplusy.m
% Part a)
clear
x = 5;
y = 2;
z = x + y
% Part b)
x = 5;
y = x;
z = x + y

Execution with corrected code:

>> xplusy
z =
7
z =
10

15
a) >> Lptot=10*log10(10^(85/10)+10^(85/10))
Lptot =
88.0103
b) 146.0206

16
a) a = [1 2 1] , alternatively a = [1,2,1]
b) b = [1;2;1]
c) c = 2:25;
d) T denotes transposed matrix or vector.
d = [20:-2:-20]';
e) e = [0.1:0.1:100]';

41
17
a) >> x=(1:200)'*10*pi/200;
>> y=sin(x);
>> y(24)
ans =
-0.5878
b) >> x = 1/pi*(1:1000)';
>> y = sin(sqrt(x));
>> y(524)
ans =
0.3415

18
a) Use the keyword roots in the MATLAB Help browser.
b) The roots are found in two steps. First define the coefficient vector c defining the
polynomial.

>> c = [1 0 1 -1];

Then use roots to find the polynomial roots:

>> x = roots(c)
x =
-0.3412 + 1.1615i
-0.3412 - 1.1615i
0.6823

The answer is a vector with three roots, one real-valued and two complex-valued.

19
Build the time vector. There are several alternative solutions.
Alternative 1: x = (0:1/99:1)'; % Gives a column vector
Alternative 2: x = linspace(0,1,100); % Gives a row vwctor
a) >> t(50)
ans =
0.4949
b) >> t(51:60)
ans =
0.5051
0.5152
0.5253
0.5354
0.5455
0.5556
0.5657
0.5758
0.5859
0.5960

42
c) The vector t1 containing every fifth component of t is built by

t1 = t(5:5:length(t)); % Alt t1 = t(5:5:end);

The value of the 17th component of this vector is,

>> t1(17)
ans =
0.8485

The functions sin(2πt) and sin(2πt1) are plotted with the command,

plot(t,sin(2*pi*t),t1,sin(2*pi*t1))

20
d)

4.5

3.5
Sound pressure, [Pa]

2.5

1.5

0.5

0
0 50 100 150 200 250 300 350 400
Frequency, [Hz]

43
21
b)

500

450

400

350
Position, [m]

300

250

200

150

100

50
1 2 3 4 5 6 7 8 9 10
Time, [s]

c) a = 3.0659 [m/s2], v0 = 9.3411 [m/s], s0 = 51.1567 [m].

d) Car position at t = 0 is 51.1567 m ≈ 51.2 m;

at t = 2.3 seconds, s = 88.8598 m ≈ 88.9 m.

e)

500

450

400

350
Position, [m]

300

250

200

150

100

50
0 1 2 3 4 5 6 7 8 9 10
Time, [s]

44
22
a) Start with building the first part of the function in the interval [0, π]. The following
commands generate a column vector x1 with 50 equidistantly spaced components on
the interval [0, 49π/50] and a column vector y1 with the corresponding sine values.

>> x1 = [0:49]’*pi/50;
>> y1 = sin(x1);

Use the vectors x1 and y1 as building blocks and build the rectified sine function.

>> x = [x1;x1+pi;x1+2*pi;x1+3*pi;x1+4*pi;x1+5*pi];
>> y = [y1;y1;y1;y1;y1;y1];

Plot the result in a graph.

>> plot(x,y)

Note that the end point of the interval, 49π/50, is selected so that no boundary points
are doubled when the rectified sine is built.

b) A simpler solution is to use the absolute value of the sine function.

>> x1 = [0:1/50:1-1/50]'*pi;
>> x = [x1;x1+pi;x1+2*pi;x1+3*pi;x1+4*pi;x1+5*pi];
>> y = abs(sin(x));

23
Use the vectors x and y built in exercise 22. Use the command axis to change the plot axis.
The command grid generates grid lines in the graph.
>> axis([0 6*pi -0.5 1.5])
>> grid

24
In the first alternative the curve is constructed using the same technique as in exercise 22.
>> x1 = [0 1];
>> y1 = [1 0];
>> x = [x1 x1+2 x1+4 6];
>> y = [y1 y1 y1 1];
>> plot(x,y);

An alternative method is simply to enter the x- and y-coordinates for the break points in two
vectors.
>> x = [0 1 2 3 4 5 6];
>> y = [1 0 1 0 1 0 1];
>> plot(x,y)

25
Suppose the x- and y-coordinates from exercise 22 are denoted X1 and Y1 and those from
exercise 24 X2 and Y2. The following plot command will produce the required multi-plot:
>> plot(X1,Y1,X2,Y2,'--')

45
26
The first rough estimate of the roots can be obtained by looking at the function cotx in a basic
mathematics book or by plotting cotx together with x in a diagram.
>> x1 = [0:0.1:10];
>> x2 = 1./tan(x1);
>> plot(x1,x1,x1,x2)

The plot of the two functions is given below. Clearly the three lowest positive roots are
located around 1, 3.5 and 6.5. Note that the zero crossings in between these do not represent
roots to the equation. These zero crossings are a consequence of the plot command
connecting each point with a line. In reality the function has discontinuities at these points,
approaching negative infinity from below and positive infinity from above.
60

40

Discontinuity at x = 3pi
Discontinuity at x = pi
20
x, cot(x)

Discontinuity at x = 2pi
-20

-40

-60
0 1 2 3 4 5 6 7 8 9 10
x

The next step is to refine the plot around the roots and to get improved estimates using the
zoom option. Which gives: x1 = 0.86, x2 = 3.43 and x3 = 6.44.
9

5
x, cot(x)

Third root x3 = 6.44

Second root x2 = 3.43
4

2 First root x1 = 0.86

0
0 1 2 3 4 5 6 7
x

46
27
Define the row vector u using the colon operator:
>> u = 1:10;
a) Calculate the scalar product:
>> u*u'
ans =
385
b) Calculate the product:
>> u'*u
ans =
1 2 3 4 5 6 7 8 9 10
2 4 6 8 10 12 14 16 18 20
3 6 9 12 15 18 21 24 27 30
4 8 12 16 20 24 28 32 36 40
5 10 15 20 25 30 35 40 45 50
6 12 18 24 30 36 42 48 54 60
7 14 21 28 35 42 49 56 63 70
8 16 24 32 40 48 56 64 72 80
9 18 27 36 45 54 63 72 81 90
10 20 30 40 50 60 70 80 90 100
c) Calculate the scalar product with the command dot:
>> dot(u,u)
ans =
385
d) Calculate the element-wise product:
>> u.*u
ans =
1 4 9 16 25 36 49 64 81 100

28
a) >> x = 0:0.05:5;
>> y = 10-x;
>> y(17)
ans =
9.2000
b) 8.4826
c) 0.7600
d) 0.8967

29
Create the vector x with the x-values and vector y with the modulated signal. Finally plot the
signal in a graph to view the results. Note that .* must be used when a component-wise
multiplication is to be performed.

>> x = (1:1000)*2*pi/100;
>> y = sin(4*x).*(1.8+sin(0.1*x));
>> plot(x,y)

47
30
a) >> F = [100 0 0];
>> s = [0 3 2];
>> A = F*s'
A =
0
Comment: The work performed by the force is 0 since the displacement is orthogonal, i.e.
perpendicular, to the force.
b) 115.

31
It is not possible to execute the given code, because the dimensions of the vectors x and
sin(x) are incompatible. A correct code should use component-wise multiplication
obtained with the .* operator.
x = (0:1/100:1)*10*pi;
y = x.*sin(x);

32
The matrix dimensions of the numerical constant 1 and the vector (1+sin(x).^2) are
incompatible. A correct code should use component-wise division and reads,
x = (0:1/100:1)*10*pi;
y = 1./(1+sin(x).^2);

33
One possible graphical method is to plot the two functions, y = x − 225 and
y = x 4 + 5 x 3 − 64 x 2 , in the same diagram and find approximate solutions by zooming in the
intersections. Estimates of the four real-valued roots are: x = -10.758, x = -1.803, x = 2.138
and x = 5.423.

34
>> f = linspace(0,500,251);
>> S1 = 0.5e-4;
>> S2 = 5e-4;
>> L = 1;
>> c = 340;
>> DTL = 10*log10(1+(S1/2/S2-S2/2/S1)^2*(sin(2*pi*f*L/c)).^2);
>> plot(f,DTL)
>> xlabel('Frequency, [Hz]');
>> ylabel('Transmission loss, [dB]');

48
 15

10
Transmission loss, [dB]

0
0 50 100 150 200 250 300 350 400 450 500
Frequency, [Hz]

Figure Transmission loss as function of frequency for silencer in exercise 34.

35
>> f=linspace(-1,1,200);
>> f=f';
>> om=2*pi*f;
>> h=1./(om-1-0.5*i)-1./(om+1-0.5*i);

>> f(41)
ans =
-0.5980

>> h(41)
ans =
0.1432 - 0.0418i

>> [maxvalue,maxlin]=max(abs(h))
maxvalue =
1.9995
maxlin =
87

>> f(maxlin)
ans =
-0.1357

49
>> freqint=f(abs(h)>1.9)
freqint =
-0.1558
-0.1457
-0.1357
-0.1256
-0.1156
-0.1055
0.1055
0.1156
0.1256
0.1357
0.1457
0.1558

>> plot(om,real(h),om,imag(h),'--')
>> xlabel('Angular frequency, [rad/s]')
>> ylabel('Receptance')
>> legend('Real part','Imaginary part')
>> title('CoQuad plot')
CoQuad plot
2
Real part
Imaginary part
1.5

0.5
Receptance

-0.5

-1

-1.5

-2
-8 -6 -4 -2 0 2 4 6 8
Angular frequency, [rad/s]

50
>> phaseangle = 180/pi*unwrap(angle(h)); % Degrees
>> figure % New figure window
>> subplot(211)
>> plot(om,abs(h))
>> xlabel('Angular frequency, [rad/s]')
>> ylabel('abs(Receptance)');
>> title('Bode plot')
>> subplot(212)
>> plot(om,phaseangle)
>> xlabel('Angular frequency, [rad/s]')
>> ylabel('phase(Receptance)');

Bode plot
2

1.5
abs(Receptance)

0.5

0
-8 -6 -4 -2 0 2 4 6 8
Angular frequency, [rad/s]

-100
phase(Receptance)

-200

-300

-400
-8 -6 -4 -2 0 2 4 6 8
Angular frequency, [rad/s]

51
>> subplot(111); % Restore to full plot area
>> plot(real(h),imag(h));
>> xlabel('Real part')
>> ylabel('Imaginary part')
>> title('Nyquist plot')
Nyquist plot
2

1.5

0.5
Imaginary part

-0.5

-1

-1.5

-2
-2 -1.5 -1 -0.5 0 0.5 1
Real part

36
a) >> a = [1 2 1;0 2 1;0 0 3]
a =
1 2 1
0 2 1
0 0 3
b) Can be solved using the identity matrix command eye.
>> b = eye(4)
b =
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
c)
 1 0  0 0  
 
 0 1 0 0 
The matrix can be partitioned to, .
 0 0 1 1 
 
 0 0 1 1 

52
>> c = [eye(2) zeros(2);zeros(2) ones(2)]
ans =

1 0 0 0
0 1 0 0
0 0 1 1
0 0 1 1

d) The matrix is most simply built by partitioning the matrix in 2×2 blocks and by
introducing the following two by two matrices.
>> a = eye(2);
>> b = ones(2);
>> c = [3 1; 3 1];
>> e = [3 1; 1 4];

Finally these are combined to the requested matrix.

>> d = [a a 2*b; a 3*b c; 2*b c' e]
d =
1 0 1 0 2 2
0 1 0 1 2 2
1 0 3 3 3 1
0 1 3 3 3 1
2 2 3 3 3 1
2 2 1 1 1 4

37
>> A = [1 0 0;0 2 0;0 0 3]; % Alt A = diag(1:3);
>> B = [2 1 2;2 -2 -1;1 2 -2]/3;
>> b = [0 1 0]';
a) >> A*b
ans =
0
2
0
b) >> B*b
ans =
0.3333
-0.6667
0.6667
c) >> A*B
ans =
0.6667 0.3333 0.6667
1.3333 -1.3333 -0.6667
1.0000 2.0000 -2.0000
d) >> B*A
ans =
0.6667 0.6667 2.0000
0.6667 -1.3333 -1.0000
0.3333 1.3333 -2.0000
Hence, matrix multiplication is not commutative.

53
e) A is a scaling matrix which changes the length but not the orientation of the vector.
For instance the orientation of the vector b is not changed after multiplication with
A. This is verified by calculating the direction cosine before and after the
multiplication.

f) B is a rotation matrix which changes the orientation but not the length of the vector.
For instance the length of b is not changed after multiplication with B.

38
>> x = [2 1 3 2];
a) >> A = x'*x;
b) >> A(2,3) + A(3,2)
ans =
6
c) >> B = A(2:3,2:3)
B =
1 3
3 9
d) >> sum(A(3,:))
ans =
24

39
a) Linear systems of equations are solved using the operator \.
Define the coefficient matrix and the right-hand side vector:
>> A = [2 1;1 2];
>> b = [1;1];
Solve the equation:
>> x = A\b
x =
0.3333
0.3333
b) x =
0.4286
-1.1429
0.7143

40
a) The system can be written, Ax = b, where,
1 1  1 + i   x1 
A=  , b=  and x =   .
1 2 2 + i   x2 
>> A = [1 1;1 2];
>> b = [1+i;2+i];
>> x = A\b
x =
0 + 1.0000i
1.0000

54
b) x =
1.0000 + 2.0000i
-0.5000
c) x =
1.6309 + 1.2416i
-0.9664 - 0.4765i
0.4295 + 0.1007i

41
a) The MATLAB command eig solves eigenvalue problems of the kind specified in
equation (1).
b) >> A=[2 1;1 2];
>> B=[2 -1;-1 2];
>> [v,d]=eig(A,B)
v =
-0.4082 0.7071
0.4082 0.7071
d =
0.3333 0
0 3.0000
c) v =
0.3466 0.2806 -0.3047
0.4234 -0.3331 0.0382
0.2607 0.0416 0.7452
d =
0.2205 0 0
0 0.5974 0
0 0 1.2655

42
a) Let v be the eigenvector matrix with the eigenvectors in each column and d the
eigenvalue matrix with the eigenvalues as diagonal elements.
>> A = [2 1;1 2];
>> [v,d] = eig(A)
v =
-0.7071 0.7071
0.7071 0.7071
d =
1 0
0 3
b) In this case the matrix A is non-symmetric.
v =
0.8165 -0.8165
0.5774 0.5774
d =
3.4142 0
0 0.5858

55
c) In this case the matrix A is non-symmetric and complex-valued.
v =
0.9540 -0.4045 + 0.0364i
0.2301 + 0.1922i 0.9138
d =
2.4426 - 0.0398i 0
0 -0.4426 + 2.0398i
d) v =
0.6848 -0.5181 + 0.2686i -0.5181 - 0.2686i
-0.5775 -0.0524 + 0.3631i -0.0524 - 0.3631i
0.4443 0.7245 0.7245
d =
-1.1409 0 0
0 3.5705 + 0.6317i 0
0 0 3.5705 - 0.6317i

43
Define the matrices A and B:
>> A = [2 -1 0;-1 2 -1;0 -1 2];
>> B = [4 -1 1;-1 4 -1;1 -1 2];

The command [V,D] = eig(A,B) returns the solution vectors as columns in the matrix V
and the eigen-values as diagonal elements in the matrix D.

>> [V,D] = eig(A,B)

D =
0.2205 0 0
0 0.5974 0
0 0 1.2655
V =
0.3466 0.2806 -0.3047
0.4234 -0.3331 0.0382
0.2607 0.0416 0.7452
>> R=V'*A*V
R =
0.2205 0.0000 -0.0000
0.0000 0.5974 0.0000
-0.0000 0.0000 1.2655
>> S=V'*B*V
S =
1.0000 -0.0000 -0.0000
-0.0000 1.0000 0.0000
-0.0000 0.0000 1.0000

Hence both matrices R and S are diagonal. The eigenvector matrix V is said to be orthogonal
with respect to R and S.

56
44
a) Begin with initializing the sum to 0, then for every n increase the sum with the value
1/n2.
>> s = 0;
>> for n = 1:100
s = s + 1/n^2;
end;
>> s
s =
1.6350
b) 1.0019

45
a) A random column vector x with 100 components is created by,
x = rand(100,1);
b) The random number generator generates numbers in the range ]0, 1[. Hence, a vector
y with values in the specified set is obtained by multiplying with 10 and then
rounding the components to the nearest integer above.
>> y = ceil(10*x);
c) The number of components larger than a specified value is obtained by using the
length and find commands.
>> length(find(y>5))
ans =
53
The find command extracts all components larger than 5 and the length
command gives the number of components extracted by the find command.
Note that your answer can be different because of the random nature of x.

46
x = 0.8604. Compare the result with exercise 26.

47
One possible code is

% countval.m
% Counts values of vector y in ranges.

% Ulf Carlsson. 2002-03-07.

clear;
x = (0:1/100:1)'*2*pi;
y = 2+2*sin(x);

number01 = 0;
number12 = 0;
number23 = 0;
number34 = 0;

for n = 1:length(y)
if (y(n) >= 0 & y(n) <= 1)
number01 = number01+1;
elseif (y(n) > 1 & y(n) <= 2)

57
 number12 = number12+1;
elseif (y(n) > 2 & y(n) <= 3)
number23 = number23+1;
elseif (y(n) > 3 & y(n) <= 4)
number34 = number34+1;
end;
end

number01
number12
number23
number34

% End of code

An alternative simpler code makes use of commands max, size and find. Replace the
for…if… construction above with the following commands:
number01 = max(size(find(y>=0 & y<=1)));
number12 = max(size(find(y>1 & y<=2)));
number23 = max(size(find(y>2 & y<=3)));
number34 = max(size(find(y>3 & y<=4)));

The result of an execution is:

>> countval
number01 =
33
number12 =
18
number23 =
17
number34 =
33

48
A code solving the exercise is given in the following:

% Exercise48.m

% Ulf Carlsson. 2002-03-05.

% Initialise parameters

m = 1;
c = 1;
k = 1e4;
Ftopp = 10; % Used in part e)

% Part a)
% Define column vector with equidistant frequency values on
% interval

f = (0:30/499:30)'; % Alt f = linspace(0,30,500)’;

frequency = f(265)

58
w = 2*pi*f;
w2 = w.^2;

% Part b)
% Calculate the frequency response function FRF in in points
% given by vector f

xoverF = 1 ./(-w2*m+i*w*c+k);
FRF = xoverF(265)

% Part c)
% Calculate absolute value and phase angle

sizexoverF = abs(xoverF);
phasexoverF = angle(xoverF);
absFRF = sizexoverF(265)
phaseFRF = 180/pi*phasexoverF(265) % Phase angle in degrees

% Part d)
% Determine the largest absolute value and the corresponding
% frequency and phase.

[sizexoverFmax,linmax] = max(sizexoverF);
sizexoverFmax
frequencymax = f(linmax)
phasexoverFmax = 180/pi*phasexoverF(linmax) % Phase in degrees

% Part e)
% Plot versus frequency in separate diagrams. Convert phase
% from radians to degrees in plot.

subplot(4,1,1);
semilogy(f,sizexoverF);
xlabel('Frequency, [Hz]');
ylabel('Abs(x/F) [m/N]');
grid;
subplot(4,1,2);
plot(f,180/pi*phasexoverF);
xlabel('Frequency, [Hz]');
ylabel('Phase(x/F), [deg]');
grid;

% Define equidistant time vector

t=linspace(0,1,500);

% Calculate displacement as function of time

x = sizexoverF(265)*Ftopp*sin(2*pi*f(265)*t+phasexoverF(265));
F = Ftopp*sin(2*pi*f(265)*t);

% Plot x(t) and F(t) in seprate diagrams

subplot(4,1,3);
plot(t,F);
xlabel('Time, [s]');

59
ylabel('Force, [N]');
grid;
subplot(4,1,4);
plot(t,x);
xlabel('Time, [s]');
ylabel('Displacement, [m]');
grid;

% Calculate time shift between force and displacement

lag = -phasexoverF(265)/2/pi/f(265)

% End of code

Running the code yields:

a) frequency = 15.8717
b) FRF = 0.0042 - 0.0077i
c) absFRF = 0.0088
phaseFRF = -61.1652
d) sizexoverFmax = 0.0098
frequencymax = 15.9319
phasexoverFmax = -101.6180
e) lag = 0.0107

The plot is shown below.

-2
10
Displacement, [m] Force, [N] Phase(x/F), [deg]Abs(x/F) [m/N]

-4
10
-6
10
0 5 10 15 20 25 30
0 Frequency, [Hz]

-100

-200
0 5 10 15 20 25 30
10 Frequency, [Hz]

-10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.1 Time, [s]

-0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time, [s]

60
49
% Exercise49.m

% Ulf Carlsson. 2002-04-18.

% Initialise parameters

m1 = 1;
m2 = 2;
k1 = 1e5;
k2 = 5e4;
F1 = 10;
F2 = 10;

% Define column vectors and matrices

M = [m1 0;0 m2];

K = [k1+k2 -k2;-k2 k2];
F = [F1;F2];
f = linspace(0.1,100,1000); % Alt: f = (0.1:0.1:100);
om = 2*pi*f;
om2 = om.*om;

% Part a)
% Calculate the displacement amplitude as a function of frequency.

for n=1:length(f)
A = -om2(n)*M+K;
x = A\F;
xtopp1(n) = x(1);
xtopp2(n) = x(2);
end;

% Part b)
% Plot the absolute value of the displacement amplitude vector.

semilogy(f,abs(xtopp1),f,abs(xtopp2),'--')
xlabel('Frequency, [Hz]');
ylabel('Displacement, [m]');
grid;
legend('x_1','x_2');

% Part c,d)
% Determine the maximum displacement and
% the corresponding frequency.

[xmax,linmax] = max(abs(xtopp1));
fmax = f(linmax);
xtoppmax = [xtopp1(linmax),xtopp2(linmax)]';

% Normalize the maximum displacement vector so that the 1st

% component becomes 1.

xtoppmax = xtoppmax/xtoppmax(1);

% End of code

61
Running the program gives:
c) fmax = 63.5000
d) xtoppmax =
1.0000
-0.1869

The graph produced in part b) is shown below.

0
10
x1
-1
10 x2

-2
10

-3
Displacement, [m]

10

-4
10

-5
10

-6
10

-7
10

-8
10
0 10 20 30 40 50 60 70 80 90 100
Frequency, [Hz]

50
% Exercise50.m

% Ulf Carlsson. 2002-04-17.

% Initialise parameters

m1 = 1;
m2 = 2;
k1 = 1e5;
k2 = 5e4;
F1 = 10;
F2 = 10;

% Define column vectors and matrices

M = [m1 0;0 m2];

K = [k1+k2 -k2;-k2 k2];

62
F = [F1;F2];

% Part a). Solve the eigenvalue problem.

[v,d] = eig(K,M);
feig = diag(sqrt(d))/2/pi;

% Part b). Normalize the second eigenvector, i.e. the second column
% of v, so that its 1st component becomes 1.

v(:,2) = v(:,2)/v(1,2);

% Part c). Compare the second eigenvector and the maximum

% displacement vector calculated in exercise 49 by calculating the
% relative difference between their components.
% xtoppmax is calculated in exercise 49.

reldiff = (v(:,2)-xtoppmax)./v(:,2);

% End of code

Running the program gives:

a) feig =
19.9376
63.5240
(compare with maximum frequency according to exercise 49, which was 63.5 Hz)
b) v(:,2) =
1.0000
-0.1861
Compare with results in exercise 43 part d).
c) reldiff =
0
-0.0041
Hence, the relative difference is only 0.41 %. Obviously the eigenvalues and eigenvectors of a
system provides information on frequencies that are easily excited to large motion amplitudes
and the deformation shape for the system when it vibrates at these large amplitude conditions.

51
>> im=linspace(-0.5,1.5,20);
>> re=linspace(-2,2,40);
>> for n=1:40
>> for m=1:20
>> h1(n,m)=1./(re(n)+i*im(m)-1-0.5*i)-1./(re(n)+i*im(m)+1-0.5*i);
>> end;
>> end;

>> h1(24,14)
ans =
-1.8581 - 0.4881i

>> re(24)+i*im(14)
ans =
0.3590 + 0.8684i

63
>> surf(re,im,real(conj(h1'))); % ' is conjugate transpose
% Hence conj((...)') is transpose
>> xlabel('Re(om)');
>> ylabel('Im(om)');
>> zlabel('Re(h_1)');

10

5
Re(h1 )

-5

-10
1.5
1 2
0.5 1
0
0 -1
Im(om) -0.5 -2
Re(om)

>> surf(re,im,imag(conj(h1')))
>> xlabel('Re(om)');
>> ylabel('Im(om)');
>> zlabel('Im(h_1)');

20

10
Im(h1 )

-10

-20
1.5
1 2
0.5 1
0
0 -1
Im(om) -0.5 -2
Re(om)

64
>> surf(re,im,log10(abs(conj(h1'))))
>> xlabel('Re(om)');
>> ylabel('Im(om)');
>> zlabel('abs(h_1)');

1.5

1
abs(h1 )

0.5

-0.5
1.5
1 2
0.5 1
0
0 -1
Im(om) -0.5 -2
Re(om)

52
The exercise is solved using the following MATLAB code:

% rmscalc.m
% Ulf Carlsson, MWL, KTH, 2002-08-09.

t = (0:0.0025:0.5); % Equidistant sample points

f = 100; % Frequency
ptot = 1e5 + 0.01*sin(2*pi*100*t); % Calculate total pressure
p = ptot - mean(ptot); % Calculate the sound pressure
T = 1/f; % The period of the sound pressure
tper = t(find(t<=T)); % Build a time vector with samples
% covering a single period

% Plot the sound pressure with the sample points overlayed as rings

plot(tper,p(1:length(tper)),tper,p(1:length(tper)),'o');

axis([0 max(tper) -0.012 0.012]); % Set the axis scaling

xlabel('Time, t, [s]'); % Horizontal axis notation
ylabel('Sound pressure, p, [Pa]'); % Vertical axis notation
N = length(t); % Number of samples
prms = sqrt(1/N*sum(p.*p)) % Calculate and write the rms value
prmsex = 0.01/sqrt(2); % Calculate the exact rms value
relerror = (prms-prmsex)/prmsex % Calculate the relative error

% End of code

65
a) Running the code gives the following output:

>> rmscalc
prms =
0.0071
relerror =
-0.0025

Hence, the estimation is 0.25 percent lower than the correct value. The plot of a
period of the sampled function is given in figure 1 below.

0.01

0.005
Sound pressure, p, [Pa]

-0.005

-0.01

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Time, t, [s]
Figure 1 Sampled sound pressure with time increment 0.0025 s. Rings denote sample points.

b) Running the code with modified time increment between samples gives the following
results:

>> rmscalc
prms =
0.0071
relerror =
-2.4991e-004

Hence the relative error is reduced a factor of 10 when the sampling rate is changed. The plot
of a period of the sampled function is given in Figure 2 below.

66
 0.01

0.005
Sound pressure, p, [Pa]

-0.005

-0.01

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Time, t, [s]

Figure 2 Sampled sound pressure with time increment 0.00025 s. Rings denote sample points.

53
a) The following code solves part b) – f) of the exercise

% splevel.m
% Program that calculates the total sound pressure level and
% the sound pressure level as function of frequency.
% Sound pressure data is stored on floating point format
% 'float32' in a binary file specified in a dialog box. The
% storage sequence is f1, p1, f2, p2, ... .

% Ulf Carlsson, 2002-03-07.

clear

% Read name and path of input data file.

[name,path] = uigetfile('*.bin',...
'Specify a binary input data-file');

% Open specified datafile and read binary data.

fid = fopen([path,name],'rb');
data = fread(fid,'float32');
fclose(fid);

67
% Partition data into x and y variables.

f = data(1:2:end);
p = data(2:2:end);

% Calculate sound pressure level as function of frequency.

p2 = p.*p;
Lp = 10*log10(p2/(2e-5)^2);

% Localise the maximum value and its frequency.

[Lpmax,linmax] = max(Lp)
freqmax = f(linmax)

% Calculate the total sound pressure level.

p2sum = sum(p2);
Lptot = 10*log10(p2sum/(2e-5)^2)

% Plot the sound pressure level as function of frequency.

plot(f,Lp);
xlabel('Frequency, [Hz]');
ylabel('Sound pressure level, [dB]');
grid

% Add info to plot

text1 = text(f(length(f))/3,0.9*Lpmax,...
['L_p_t_o_t = ',num2str(Lptot),' dB']);
text2 = text(f(length(f))/3,0.75*Lpmax,...
['L_p_m_a_x = ',num2str(Lpmax),...
' dB at f = ',num2str(freqmax),' Hz']);

% End of code

Execution of the code:

>> splevel
Lpmax =
107.5623
linmax =
69
freqmax =
49.8047
Lptot =
109.1187

A plot of the calculated sound pressure level is given in the figure below.

68
 110

100
Lptot = 109.1187 dB

90
Sound pressure level, [dB]

80 Lpmax = 107.5623 dB at f = 49.8047 Hz

70

60

50

40

30

20
0 50 100 150 200 250 300 350 400
Frequency, [Hz]

54
The modified code is:

% splevelmod.m
% Program that calculates the total sound pressure level and
% the sound pressure level as function of frequency.
% Sound pressure data is stored on floating point format
% 'float32' in a binary file specified in a dialog box.
% The storage sequence is f1, p1, f2, p2, ... .

% Ulf Carlsson, 2002-03-08.

clear

% Read name and path of input data file.

[name,path] = uigetfile('*.bin','Specify a binary input data file');

% Open specified datafile and read binary data.

fid = fopen([path,name],'rb');
data = fread(fid,'float32');
fclose(fid);

% Partition data into x and y variables.

f = data(1:2:length(data));

69
p = data(2:2:length(data));

% Call function m-file soundcalc.m for calculations

[fmax,Lpmax,Lptot,Lp] = soundcalc(f,p);
Lptot
fmax
Lpmax

% Plot the sound pressure level as function of frequency.

plot(f,Lp);
xlabel('Frequency, [Hz]');
ylabel('Sound pressure level, [dB]');
grid;

% Add info to plot

text1 = text(f(length(f))/3,0.9*Lpmax,...
['Lptot = ',num2str(Lptot),' dB']);
text2 = text(f(length(f))/3,0.75*Lpmax,...
['Lpmax = ',num2str(Lpmax),' dB at f = ',num2str(fmax),' Hz']);

% End of main program

function [xmax,ymax,ytot,Ly] = soundcalc(x,y);

% Function for calculating fmax, Lpmax, Lptot. Input data is y.

% Ulf Carlsson, 2003-03-08.

% Calculate sound pressure level as function of frequency.

y2 = y.*y;
Ly = 10*log10(y2/(2e-5)^2);

% Localise the maximum value and its frequency.

[ymax,linmax] = max(Ly);
xmax = x(linmax);

% Calculate the total sound pressure level.

y2sum = sum(y2);
ytot = 10*log10(y2sum/(2e-5)^2);

% End of function soundcalc

70
55
a)
function yband = band(fl,fu,f,y);

% Function m-file for energy summation of spectra.

% Input parameters: fl – lower frequency band limit
% fu – upper frequency band limit
% f – current narrow-band frequency
% y – narrow band energy function
% Output: yband – total energy in band

% Ulf Carlsson, 2002-03-11.

% Determine limit lines for band within given limits.

fband = find(f > fl & f < fu);

% If no lines exist in band the sum is 0.

if isempty(fband)
yband = 0;
return;
end;

% Sum energy of lines inside band.

yband = sum(y(fband));

% Add half of the energy in lines within 1e-3 Hz from limits.

if abs(f(fband(1)-1)-fl) <= 1e-3

yband = yband + y(fband(1)-1)/2;
end; % Lower limit.
if abs(f(fband(length(fband))+1)-fu) <= 1e-3
yband = yband + y(fband(length(fband))+1)/2;
end; % Upper limit.

% End of function band

b)
function [throctf,throcty] = throct(f,y);

% Function m-file for summing narrowband spectra to third

% octave band spectra.
% Input parameters: f – frequency vector
% y – energy vector
% Output parameters: throctf – vector with center frequencies
% throcty – vector with band energies

% Ulf Carlsson, 2002-03-11.

% Define band limits.

limits = [1.12 1.41 1.78 2.24 2.82 3.55 4.47 5.62 7.08 ...
8.91 11.2 14.1 17.8 22.4 28.2 35.5 44.7 56.2 70.8 ...

71
 89.1 112 141 178 224 282 355 447 562 708 891 1120 ...
1410 1780 2240 2820 3550 4470 5620 7080 8910 11200 ...
14100 17800 22400];
center = [1.25 1.6 2 2.5 3.15 4 5 6.3 8 10 12.5 16 20 25 ...
31.5 40 50 63 80 100 125 160 200 250 315 400 500 630 ...
800 1000 1250 1600 2000 2500 3150 4000 5000 6300 ...
8000 10000 12500 16000 20000];

% Determine the first and last "full" band.

fmin = min(f);
fmax = max(f);
bandmin = min(find(limits > fmin));
bandmax = max(find(limits < fmax))-1;

% Loop through all bands.

for n = bandmin:bandmax
throcty(n-bandmin+1) = band(limits(n),limits(n+1),f,y);
throctf(n-bandmin+1) = center(n);
end;

% End of function throct

c) A code for testing the cods developed in a) and b) is given below.

% exercise55.m
% Program testing the codes developed in part a) and b).

% Ulf Carlsson, 2002-08-15.

clear;

% Read input sound pressure data from file.

[name,path] = uigetfile('*.bin',...
'Specify a binary data input file with sound pressure data.');
fid = fopen([path,name],'rb');
data = fread(fid,'float32');
fclose(fid);

% Partition data

f = data(1:2:end);
y = data(2:2:end);

% Calculate the squared (power) data.

y2 = y.*y;

% Call function m-file throct

[F,Y] = throct(f,y2);

% Calculate and plot sound pressure level.

72
Lp = 10*log10(Y/(2e-5)^2);
semilogx(F,Lp);
title('Third octave band spectrum');
ylabel('Sound pressure level, L_p [dB]');
xlabel('Frequency, f [Hz]');

% End of code

The result when the code is used on the data in soundpress.bin is shown in the figure below.
Third octave band spectrum
110

105

100
Sound pressure level, Lp [dB]

95

90

85

80

75

70

65
0 1 2 3
10 10 10 10
Frequency, f [Hz]

56
a) The function M-file rad2degr.m converts angular values in radians to degrees.

function d = rad2degr(r)

% rad2degr.m
% Function that converts angular values from radians to
% degrees.
% Input parameter: r – angle measured in radians
% Output parameter: d – angle measure in degrees

% Ulf Carlsson. 2002-03-06.

d = 180*r/pi;

% End of code

73
b) The function M-file atandeg.m provides arcustangent function in degrees.

function v = atandeg(x)

% Arcustangent function that provides resulting angles in degrees

% Input parameter: x – angle measured in degrees
% Output parameter: v – value of atan(x)

% Ulf Carlsson. 2002-03-06.

y = atan(x);

% Call function rad2degr for conversion from radians to degrees

v = rad2degr(y);

% End of code

c) Try the programs on different input values:

>> atandeg(1)
ans =
45
>> atandeg(1/0)
ans =
90
>> atandeg(0.5)
ans =
26.5651

57
a) The following MATLAB command assigns the string accn.dat, where n is the
current numeric value of n, to a string variable filename:
filename = [‘acc’,num2str(n),’.dat’]
b)
function filename = name(n);
% Ulf Carlsson, 2002-04-18.

filename = [‘acc’,num2str(n),’.dat’];

% End of function filename

Running the function with input value of n = 3 yields.

>> filename = name(3)

filename =
acc3.dat

74
58
a) Use 11 mesh points in both, x- and y-directions.

>> x = (0:10)'*0.1;
>> y = (0:10)'*sqrt(2)/10;

Calculate the mode function z in the mesh defined by x and y. Finally plot the mode
function in a 3-dimensional plot.

>> z = sin(pi*x)*sin(pi*y/sqrt(2))';
>> figure;
>> mesh(x,y,z); % contour(x,y,z) gives a contour plot
>> title('Mode 1');

The plot of the mode 1 is shown in the figure below.

Mode 1

0.8

0.6

0.4

0.2

0
1.5
1
1
0.8
0.6
0.5 0.4
0.2
0 0

b) and c) Modify the calculation of z.

75
59
a) One possible MATLAB implementation able to treat a function with any number of
linear segments follows below:

function y = powerspec(x,X,Y);

% Evaluation of power density spectrum.

% Input data: vector x
% vector X - break points frequency
% vector Y - break points power density
% Output data: vector y

% Ulf Carlsson, 2002-08-19

global X Y;

% Find the lower interval limit for the minimum x value

m = min(find(X>min(x)))-1;

% Interpolate the y values

for n = 1:length(x)
if (x(n)>=X(m) & x(n)<=X(m+1))
y(n) = Y(m) + (Y(m+1)-Y(m))/(X(m+1)-X(m))*(x(n)-X(m));
else
m = m + 1;
y(n) = Y(m) + (Y(m+1)-Y(m))/(X(m+1)-X(m))*(x(n)-X(m));
end;
end

% End of code

b) A MATLAB code calling the function M-file powerspec.m follows below. The
specific function specified in this case corresponds to the spectrum in Figure 2.

% powerintegr.m
% MATLAB code for integrating the energy below a
% power density spectrum. The power density function
% is assumed to be a series of break points connected
% with straight line elements.

% Ulf Carlsson, 2002-08-19

clear;
global X Y;

% Define break points which specifies the function.

X = [15;75;300;800;1000]; % Frequency values

Y = [5;5;23.4;23.4;8.4]; % Power values

% Plot function as check.

plot(X,Y);

76
title('Power Density Spectrum');
ylabel('Acceleration^2, [m^2/s^4/Hz]');
xlabel('Frequency, [Hz]');

% Use MATLABs integration (quadrature) function quad.

% Integrate between minimum frequency and maximum
% frequency.

W = quad(@powerspec,min(X),max(X))

% End of code

c) Executing the code developed in part a) and part b) gives:

>> powerintegr
W =
1.8375e+004

60
a) The code given below calculates the displacement for a specified force:

% threerodab.m
% Calculation of the node displacement in a simple rod structure.
% Based on exercise suggested by Sören Östlund, Solid Mechanics, KTH

% 2002-11-13, Ulf Carlsson, MWL, KTH

% Initialise parameter data

clear; % Clear workspace

% Young's moduli, cross-section areas and lengths:

E = [2e11;2e11;2e11];
A = [0.1e-2;0.1e-2;0.1e-2];
L = [11;10;14];

% Initialize force magnitude

Force = 1e4;

% Calculate column vector with direction angles for rods

% specified with respect to horizontal plane

alfa = asin(L(2)./L);

% Calculate the axial stiffness of the rods

k = E.*A./L;

% Initialise stiffness matrix coupling the node force with

% the node displacement.

k11 = k(1)*cos(alfa(3))^2+k(3)*sin(alfa(3))^2;
k12 = k(1)*cos(alfa(1))*cos(alfa(3))-k(3)*sin(alfa(3))*cos(alfa(3));
k22 = k(1)*cos(alfa(1))^2+k(2)+k(3)*cos(alfa(3))^2;

77
K = [k11 k12;k12 k22];

% Part b). Calculate the node displacement

% Horizontal force

P = [Force;0];
dispbi = K\P

% Vertical force

P = [0;Force];
dispbii = K\P

% End of program for part a-b)

b) Results obtained when executing the code developed in part a) are the following:
>> threerodab
dispbi =
1.0e-003 *
0.6218
0.0380
dispbii =
1.0e-003 *
0.0380
0.3340

c) A function m-file performing the specified tasks:

function displ = invK_p(gamma,Force,L,E,A)

% Function m-file that calculates the stiffness matrix

% for specified input data rod lengths, L, rod cross
% section areas, A, rod Young's moduli and finally for a
% specified force magnitude, Force, and force angle,
% gamma solves the system for the displacement, displ.

% Ulf Carlsson, MWL, KTH, 2002-11-21,

% Calculate column vector with direction angles for rods

% with respect to horizontal plane.

alfa = asin(L(2)./L);

% Calculate the axial stiffnesses of the rods

k = E.*A./L; % column vector with axial stiffnesses for the rods

% Initialise stiffness matrix coupling the node force

% with the node displacement.

k11 = k(1)*cos(alfa(3))^2+k(3)*sin(alfa(3))^2;
k12 = k(1)*cos(alfa(1))*cos(alfa(3))-k(3)*sin(alfa(3))*cos(alfa(3));
k22 = k(1)*cos(alfa(1))^2+k(2)+k(3)*cos(alfa(3))^2;
K = [k11 k12;k12 k22];

78
% Calculate force vector and solve linear system of
% equations for node displacement.

P = Force*[cos(gamma);sin(gamma)];
displ = K\P;

% End of function displ

When the code is tested on the data given in part b) the results are:

i)
>> gamma = 0;
>> disp = invK_p(gamma,Force,L,E,A)
disp =
1.0e-003 *
0.6218
0.0380
ii)
>> gamma=pi/2;
>> disp=invK_p(gamma,Force,L,E,A)
disp =
1.0e-003 *
0.0380
0.3340

d) A code for evaluating the node displacement as function of the force angle γ is:

% threerodd.m
% MATLAB code for calculating the displacement vector as
% function of the force angle gamma.

% 2002-11-21, Ulf Carlsson, MWL, KTH.

clear;

% Initialise parameter data,

% Young's moduli E, cross section areas A and lengths L.

E = [2e11;2e11;2e11];
A = [0.1e-2;0.1e-2;0.1e-2];
L = [11;10;14];

% Initialize force and force angle, gamma.

Force = 1e4;
gamma = [0:1/100:1-1/100]'*2*pi;

for n = 1:length(gamma)
disp = invK_p(gamma(n),Force,L,E,A);
dispx(n) = disp(1); % Collect displacement components in
dispy(n) = disp(2); % separate vectors
end;

% Plot results

79
subplot(2,1,1)
plot(gamma,dispx)
xlabel('Angle gamma, [rad]');
ylabel('Horizontal displacement, [m]');

subplot(2,1,2)
plot(gamma,dispy)
xlabel('Angle gamma, [rad]');
ylabel('Vertical displacement, [m]');

% End of code

Execution of the code gives graphs showing the displacement components as functions of the
force angle γ , see figure below. Obviously the displacement components are zero for two
different force angles. The horizontal component is zero for angles equal to 1.63 and 4.77
radians. The vertical component is zero for 3.03 and 6.17 radians.
-3
x 10
1
Horizontal displacement, [m]

0.5

-0.5

-1
0 1 2 3 4 5 6 7
Angle gamma, [rad]

-4
x 10
4
Vertical displacement, [m]

-2

-4
0 1 2 3 4 5 6 7
Angle gamma, [rad]

80
61
a) The following function M-file calculates the power transmitted over a single mount:

function W = powcalc(a,F,f);

% Function m-file calculating the power transmitted over a

% mount using measured force and acceleration data.
% Input data:
% a - [peak phase(degrees)], measured acceleration at
% frequency f, [m/s2]
% F - [peak phase(degrees)], measured peak force at %
% frequency f, [N]
% f - frequency, [Hz]
% Output data:
% W - time average of transmitted mechanical power, [W]

% Ulf Carlsson, 2002-08-20

% Calculate complex force and acceleration

apeak = a(1);
fia = a(2);
Fpeak = F(1);
fiF = F(2);
acc = apeak*exp(i*fia*pi/180);
forc = Fpeak*exp(i*fiF*pi/180);

% Calculate complex velocity

v = acc/(i*2*pi*f);

% Calculate transmitted power via the coupling point.

W = real(forc*conj(v))/2

% End of code

b) The following code calls the code developed in part a) for the calculation of the
power transmitted over N mounts.

% powertot.m

% MATLAB code for calculating and summing total mechanical

% power transmitted from a machine over N mounts.

% Ulf Carlsson, 2002-08-20

clear;

% Initialise data

Force = [13 0; 17 33; 8 70; 21 85]; % Amplitude and phase

acc = [1.4 192; 1.1 160; 2.1 212; 3.4 263]; % Amplitude and phase
freq = 90;

81
% Calculate power vector

for n = 1:length(acc)
W(n) = powcalc(acc(n,:),Force(n,:),freq);
end;

% Calculate total transmitted power

Wtot = sum(W)

% End of code

c) Executing the code with input data according to table 1 gives,

>> powertot
W =
-0.00334578419127
W =
0.01320498309531
W =
0.00914531959314
W =
0.00220325620661
Wtot =
0.02120777470379

62
a) The first step is to write a MATLAB function that generates random components of a
unit surface normal vector.

function snorm = surfnorm

% Function m-file for generating random normal surface

% unit vectors.

% 2002-11-22, Ulf Carlsson, MWL, KTH

snorm = 1-2*rand(1,3)'; % Generate random components

snorm(:) = snorm(:)/norm(snorm(:)); % Normalise

% End of code

b) The second step is to write a MATLAB function that for a specified stress matrix
calls the random normal vector generator and calculates the normal and shear
stresses.

function [nstress,shear] = stress(S)

% Function m-file for calculating the normal stress and

% shear for a specified stress matrix S and a cut surface
% with normal n.

% 2002-11-22, Ulf Carlsson, MWL, KTH.

82
n = surfnorm; % Generate random unit surface vector.
Sn = S*n;
nstress = n'*Sn;
shear = sqrt(Sn'*Sn-nstress^2);

% End of code

c) The next step is to write a MATLAB code that calculates and plots the relationship
between the normal and shear stresses using a number N of, say 500, randomly
oriented cuts.

% nrandstress.m
% S - Stress matrix [N/m2]
% N - Number of random cuts
% sigma - normal stress vector calculated for N random cuts
% tau - vector with shear values calculated for N random cuts
% MATLAB file that for a given stress matrix calculates and
% plots the normal stress and shear for N random cuts.

% 2002-11-25, Ulf Carlsson, MWL, KTH

clear;
% Initialise number of random cuts N and stress matrix S.
N = 500;
S = [10 2 0;2 10 1;0 1 10];
for n = 1:N
[sigma(n),tau(n)] = stress(S);
end

% Plot normal stress as function of shear in graph with

% equal scaling and each point represented with a star '*'.
plot(sigma,tau,'*')
axis('equal')
xlabel('Normal stress, [MPa]');
ylabel('Shear stress, [MPa]');

% End of code

Testing the code on the matrix given produces the diagram given in part d) below, with
exception for the three rings at zero shear stress.

d) The final step in developing the code is to include the calculation of the eigenvalues
for the stress tensor.

% nrandstress.m
% S - Stress matrix [N/m2]
% N - Number of random cuts
% sigma - normal stress vector calculated for N random cuts
% tau - vector with shear values calculated for N random cuts
% MATLAB file that for a given stress matrix calculates and
% plots the normal stress and shear for N random cuts.

% 2002-11-25, Ulf Carlsson, MWL, KTH

83
clear;
% Initialise number of random cuts N and stress matrix S.
N = 500;
S = [10 2 0;2 10 1;0 1 10];
for n = 1:N
[sigma(n),tau(n)] = stress(S);
end

% Calculate the eigen-values (val) of the stress matrix

val = eig(S);

% Plot normal stress as function of shear in graph with

% equal scaling and each point represented with stars ('*').
% Add eigenvalues plotted along the normal stress axis.
% Represent the eigenvalues with rings ('o').

plot(sigma,tau,'*',val(1),0,'o',val(2),0,'o',val(3),0,'o')
axis('equal')
xlabel('Normal stress, [MPa]');
ylabel('Shear stress, [MPa]');

% End of code

The stress matrix eigenvalues are:

val =
0.8782
8.4067
28.7151

Testing the code on the matrix given produce the diagram below.

16

14

12

10
Shear stress, [MPa]

-2

-4
5 10 15 20 25
Normal stress, [MPa]

84
Try the code for the stress matrix representing an unloaded surface by letting,
0 0 0 
S = 0 4 8  MPa.
0 1 25
The stress matrix eigen-values are
val =
0
2.9546
25.0454

The relation between the normal shear stresses is shown in the diagram below.

16

14

12

10
Shear stress, [MPa]

-2

0 5 10 15 20 25
Normal stress, [MPa]

Finally try the code on the stress tensor given by,

10 2 0 
S =  2 10 1  MPa.
 
 0 1 10
The stress matrix eigen-values are,
val =
7.7639
10.0000
12.2361

The relationship between the normal and hear stresses is shown in the diagram below.

85
 2.5

1.5
Shear stress, [MPa]

0.5

-0.5

8 8.5 9 9.5 10 10.5 11 11.5 12

Normal stress, [MPa]

63
a) The code consist of two parts: the main program named linmodel.m and the
function M-file linreg.m.

% linmodel.m
% m-file for least squares estimation of a linear relation
% y = c1*x + c2. Measured data (x,y) is provided in sequence
% x1, y1, x2, y2 ... on 'float32' format in binary file.

% Ulf Carlsson. 2002-03-07.

clear;

% Get name and path of input data file.

[name,path] = uigetfile('*.bin','Specify a binary data-file');

% Open specified datafile and read binary data.

fid = fopen([path,name],'rb');
data = fread(fid,'float32');
fclose(fid);

% Partition data into x and y variables.

x = data(1:2:length(data));
y = data(2:2:length(data));

% Call function m-file for model parameter estimation.

c = linreg(x,y)

% Plot characteristic model cuve together with measured data.

86
u = [min(x);max(x)];
v(1) = c(1)*u(1)+c(2); % Calculate characteristic model
v(2) = c(1)*u(2)+c(2);
plot(u,v,x,y,'*');
xlabel('x-values');
ylabel('y-values');
legend('Linear model','Measured data');

% End of linmodel

function [c] = linreg(x,y)

% Function m-file for least squares estimation of linear model

% parameters on the form y = c1*x + c2. The calling program
% should provide input data in two vectors x and y.

% Ulf Carlsson. 2002-03-07.

% Assemble system matrix A.

A(:,1) = x;
A(:,2) = ones(length(x),1);

% Solve overdetermined system of equations.

c = A\y;

% End of linreg

b) The test data in file xydata.bin is generated from the linear model y = 2x + 1, i.e. c1
and c2 are 2 and 1 respectively.
>> linmodel
c =
1.9815
1.0283

64
This problem can be solved using the function linreg.m developed in exercise 63. Before
the function is called the force and voltage values must be collected in two column vectors.
Let F and U denote these two vectors.
>> F = [1 5 10 20 50]';
>> U = [0.08 0.11 0.14 0.28 0.55]’;
>> [c] = linreg(F,V)
c =
0.0098
0.0632

>> y = c(1)*x+c(2);
>> plot(F,V,'*',x,y)
>> xlabel('Force, [N]')
>> ylabel('Voltage, [V]')
>> grid
>> legend('Measurement','Calculated');

87
Hence, the optimum linear model is,
U = 0.0098 F + 0.0632 .
This function is plotted together with the measured data in the graph below.

0.7
Measurement
Calculated

0.6

0.5

0.4
Voltage, [V]

0.3

0.2

0.1

0
0 10 20 30 40 50 60
Force, [N]

Figure Graph showing measured force and voltage together with calculated data according to
optimum linear model.

65
a) Begin with opening the binary file stiff.dat for reading.
>> fid = fopen('stiff.dat','rb');

Read the contents of the file into the vector data.

>> data = fread(fid,'float32');

Close the input data file

>> fclose(fid);

b) Partition the data vector into three vectors: a frequency vector f, a force vector F and
a displacement vector x.
>> f = data(1:3:end);
>> F = data(2:3:end);
>> x = data(3:3:end);

c) Calculate the dynamic stiffness vector k. And finally plot the stiffness in a lin-log
diagram.
>> k = F./x;
>> semilogy(f,k)
>> xlabel('Frequency, [Hz]');
>> ylabel('Dynamic stiffness, [N/m]')

88
66
a)
function f = profile(x)

% Calculates road profile f(x) at position x.

% Ulf Carlsson, 2002-03-14.

global m k c V bumpsize;

% Function m-file that calculates the spatial derivative of

% the road profile. The bump is placed on interval [-0.5, 0.5].

if (x <= 0.5 & x >= -0.5)

f = bumpsize*(cos(pi*x))^2;
else
f = 0.001*(2*rand(1)-1);
end;

% End of function profile

function fprim = profileprim(x)

% Function m-file for road profile derivative.

% Ulf Carlsson, 2002-04-26.

global m k c V bumpsize;

% The bump is placed on interval [-0.5, 0.5].

fprim = 0;
if (x <= 0.5 & x >= -0.5)
fprim = -bumpsize*pi*sin(2*pi*x);
end;

% End of function profileprim.

b)
function bump = funcbump(t,y)

% Function m-file for evaluating the right-hand side of

% governing equations for SDOF car running over bumpy road.
% The road profile is calculated in function profile.

% Ulf Carlsson, 2002-03-14.

global m k c V bumpsize;

% At time t the position is x = Vt.

A = [c/m k/m;-1 0];

bump = [k*profile(V*t)/m;0]+[c*V*profileprim(V*t)/m;0]-A*y;

% End of function funcbump

89
c and d)
% bumproad.m
% m-file for simulating the motion of a car running at speed V
% over a road bump with size bumpsize.

% Ulf Carlsson, 2002-03-14.

clear

% Define as global parameters that are used in function

% m-files.

global m k c V bumpsize;

% Initialize parameter values.

V = 50/3.6;
m = 250;
k = 2000;
c = 200;
bumpsize = 0.3;

ystart = [0;0]; % Initialiaze starting values.

% Call equation solver y'={b}+Ay. Right-hand side is

% calculated in function 'funcbump'. Solution is evaluated in
% 200 equidistant points at time interval [-5,20]. Solution is
% returned in displ. T is the time vector.

[T,displ] = ode45('funcbump',[linspace(-5,20,200)],ystart);

% Plot simulated displacement in a diagram.

plot(T,displ);
xlabel('Time [s]');
ylabel('Displacement [m], Velocity [m/s]');
grid
legend('Velocity','Displacement')

% End of m-file bumproad.m

90
d) The diagram below shows the vertical displacement and velocity as functions of
time.

0.15
Velocity
Displacement

0.1
Displacement [m], Velocity [m/s]

0.05

-0.05

-0.1
-5 0 5 10 15 20
Time [s]

67
The following code contains solution to all parts of the exercise.

% triangle.m
% Code demonstrating how to use matrix
% operations to perform simple animated movies.

% Ulf Carlsson, 2002-08-14.

clear

% Part a).

% Create window with fixed corners for displaying

% the figures.

fig = figure;
axis([-10 10 -10 10]);

% Part b).

% Create two vectors one containing the x-coordinates

% the other one the y-coordinates. Note that the triangle
% is closed by finishing in the starting point.

x = [-5 5 0 -5];
y = [-5 -5 5 -5];

91
% Draw triangle in window using line 'command'.

line(x,y,'color','black');

% Part c).

% Translate the triangle in 2 units along the horizontal axis.

% Do this using matrix operations. First introduce the corner
% matrix triangle with x-coordinates in the first row and
% y-coordinates in the second row.

triang = [x;y];

% Then introduce a unit rigid body translation matrix filled

% with ones and having the same dimension as the corner
% matrix.

rigid = ones(size(triang));

% Then introduce a 2 by 2 diagonal point displacement matrix

% having the coordinate translations on the diagonal.

xdispl = 2;
ydispl = 0;
displace = [xdispl 0;0 ydispl];

% The triangle is translated by adding the matrix product

% between the displacement matrix and the rigid body
% translation matrix to the original triange matrix.

transltri = triang + displace*rigid;

% The triangle is translated by adding the matrix product

% between the translation matrix and the rigid body matrix
% translation to the original triange matrix.

transltri = triang + transl*rigid;

x = transltri(1,:);
y = transltri(2,:);
line(x,y,'color','blue');

% Part d).

% Zooming in on a figure can be accomplished by introducing a

% diagonal 2 by 2 zoom matrix with the zoom factor as diagonal
% elements.Isotropic zoom is obtained if the diagonal elements % are
equal.

xzoom = 1.2; % Zoom factor 1.2 in x-direction

yzoom = 1.2; % Zoom factor 1.2 in y-direction
zoom = diag([xzoom yzoom]);

% The zooming is performed by multiplying the corner matrix

% with the zoom matrix from the left.

92
zoomtria = zoom*triang;
x = zoomtria(1,:);
y = zoomtria(2,:);
line(x,y,'color','green');

% Part e).

% Rotating a figure an angle v in counter-clockwise direction

% can be performed by multiplying the corner matrix with a
% rotation matrix. Introduce the 2 by 2 rotation matrix
% rotate.

v = 15; % Rotation angle [degrees]

fi = v*pi/180; % Rotation angle [radians]
rotate = [cos(fi) sin(fi);-sin(fi) cos(fi)];

% Rotation in the opposite direction is obtained by

% transposing the rotation matrix.

% Rotate the triangle by multiplying the corner matrix with

% the rotation matrix.

rottri = rotate*triang;
x = rottri(1,:);
y = rottri(2,:);
line(x,y,'color','red');

% End of code

68
% triangmov.m

% Ulf Carlsson, 2002-03-12.

clear

% Prepare plot window with DoubleBuffer property set on.

fig = figure;
set(fig,'DoubleBuffer','on');

% Part a). Define triangle.

x = [-5 5 0 -5];
y = [-5 -5 5 -5];
triang = [x;y];

% Define rotation matrix B.

v = 1;
b = v/180*pi;
B = [cos(b) -sin(b);sin(b) cos(b)];

% Animate triangle rotation.

93
newtriang = triang;
N = 360/v;
for n = 1:N
x = newtriang(1,:);
y = newtriang(2,:);
plot(x,y); % Plot the triangle.
axis([-10 10 -10 10]); % Use fixed axis.
axis('square'); % A square axis creates
% an isotropic triangle.
drawnow; % Executes pending graphic
% commands.
newtriang = B*newtriang; % Rotate the triangle 1 degree.
end;

% End of triangmov.

69
The code below generates a plot representing the SDOF system according to the figure below.

fline
mass
force
coupl4
coupl3
damp2
spring

coupl2
coupl1

ground

% sdofanim.m

% Ulf Carlsson, 2002-04-18.

clear;

% Initialise parameters.

m = 1;
k = 1e4;
c = 1;
f = 15.8717; % f(265) from exercise 48.
om = 2*pi*f;
F = 50;

% Generate displacement from results in exercise 48.

t = [0:0.001:1];
y0 = F/sqrt((k-om^2*m)^2+om^2*c^2);

94
phi = atan(om*c/(k-om^2*m));
y = y0*sin(om*t+phi);

% Create circle representing the mass.

s = [0:1:40]*2*pi/40;
xmass = cos(s);
ymass = sin(s)+5;

% Create ground at y = 0.

xground = [-1 1];

yground = [0 0];

% Create connection to ground.

xcoupl1 = [0 0];
ycoupl1 = [0 0.5];

% Create rigid (non-moving) part of damper and spring

xcoupl2 = [-1 -1 1 1];

ycoupl2 = [1 0.5 0.5 1.5];
xdamp1 = [0.2 0.2 1.8 1.8];
ydamp1 = [2.5 1.5 1.5 2.5];

% Create moving part of damper

xdamp2 = [0.2 1.8];

ydamp2 = [2 2];

% Create undeformed spring.

xs = [-0.5 0.5];
xspring = [0 xs xs xs xs 0]-1;
yspring = [0 1:2:15 16]/16;

% Create part of spring and damper connected to mass

ycoupl3 = [3 3.5 3.5 2];

ycoupl4 = [3.5 4];

% Create connecting line between force and mass

xfline = [1 2];
yfline = [5 5];

% Create force vector.

xforce = [1.75 1.75];

forcemagn = 1.5*sin(om*t); % Magnitude force
xarrowhead = [1.6 1.75 1.9];
yarrowhead = [-0.3 0 -0.3];

% Create window for animation. Set DoubleBuffer = 'on' in order

% to prevent unsteady animations.

95
fh = figure;
set(fh,'DoubleBuffer','on');

% Animate.

for n = 1:length(t)
yspringmov = 1+yspring*(2+y(n));
plot(xground,yground,'b',xcoupl1,ycoupl1,'b',xcoupl2,...
ycoupl2,'b',xdamp1,ydamp1,'b',xdamp2,ydamp2+y(n),...
'b',xspring,yspringmov,'b',xcoupl2,ycoupl3+y(n),...
'b',xcoupl1,ycoupl4+y(n),'b',xmass,ymass+y(n),...
'b',xfline,yfline+y(n),'b',xforce,...
[y(n) y(n)+forcemagn(n)]+5,'r',xarrowhead,...
yarrowhead*sign(forcemagn(n))+5+forcemagn(n)+y(n)...
,'r');
axis([-5 5 -2 8]);
axis('square');
axis off;
drawnow;
end;

% End of code sdofanim.m

70
One possibility is to modify the solution to exercise 65. The vertical motion y now represents
the cars vertical displacement. The horizontal motion is accounted for by adding a constant
increment, representing constant speed, to all x-coordinates.

71
The following code contains one possible solution to the problem.

function sdofgui(action)

% Graphic User Interface: sdofgui.m.

% Program for calculating and plotting the frequency response %
function (FRF)
% for a mechanical single degree of freedom system.
% Model parameters: m - mass [kg]
% k - spring constant [N/m]
% c - damping constant [Ns/m].
% The FRF is evaluated at Nf equidistant frequency values on
% the interval [low,high].

% Ulf Carlsson, 2002-03-19.

global massField stiffField dampField m k c

global lowField highField nlinField low high Nf
global absAxes phaseAxes

if nargin < 1,
action = 'initialisation';
end;
if strcmp(action,'initialisation'),

96
% Initialise the interface window

ValFig = figure('Name','SDOF System GUI','NumberTitle',...

'off','position',[20 60 750 480]);

% End of part a).

% Part b). Initialise close button and text

closeBtn = uicontrol('style','pushbutton','string',...
'Close','callback','close;','position',...
[87 30 40 20],'foregroundcolor',[1 0 0]);

% End of part b).

% Part c). Create headers for model parameters, frequency

% parameters, calculation and finally diagrams.

textHeader1 = uicontrol('style','text','position',...
[50 450 118 13],'string','MODEL PARAMETERS:', ...
'BackgroundColor',[0.8 0.8 0.8]);
textHeader2 = uicontrol('style','text','position',...
[50 330 110 13],'string','FREQUENCY RANGE:',...
'BackgroundColor',[0.8 0.8 0.8]);
textHeader3 = uicontrol('style','text','position',...
[66 200 77 13],'string','CALCULATION:',...
'BackgroundColor',[0.8 0.8 0.8]);
textHeader4 = uicontrol('style','text','position',...
[390 450 209 13],'string',...
'DIAGRAMS: Frequency Response Function',
'BackgroundColor',[0.8 0.8 0.8]);

% End of part c).

% Part d). Initialise numeric values for model and frequency

% parameters and place lead texts at suitable positions.

m = 1; % Initialise the mass

k = 1; % Initialise the spring constant
c = 1; % Initialise the damping constant
low = 0; % Initialise the lower frequency limit
high = 100; % Initialise the upper frequency limit
Nf = 101; % Initialise the number of frequency lines
textMass = uicontrol('style','text','position',...
[10 423 66 13],'string','Mass m, [kg]: ', ...
'BackgroundColor',[0.8 0.8 0.8]);
textStiff = uicontrol('style','text','position',...
[10 395 119 13],'string','Spring constant k, [N/m]: ' ...
,'BackgroundColor',[0.8 0.8 0.8]);
textDamp = uicontrol('style','text','position',...
[10 368 136 13],'string','Damping constant c, [Ns/m]:'...
,'BackgroundColor',[0.8 0.8 0.8]);
textLow = uicontrol('style','text','position',...
[10 303 127 13],'string',...
'Lower frequency limit, [Hz]: ','BackgroundColor'...
,[0.8 0.8 0.8]);

97
 textHigh = uicontrol('style','text','position',...
[10 276 127 13],'string',...
'Upper frequency limit, [Hz]: ','BackgroundColor',...
[0.8 0.8 0.8]);
textNlin = uicontrol('style','text','position',...
[10 249 129 13],'string',...
'Number of frequency lines : ','BackgroundColor',...
[0.8 0.8 0.8]);

% End of part d).

% Part e). Initialise fields for editable numeric values.

% Initialise editable mass field

massField = uicontrol('style','edit','position',...
[90 420 60 20],'string',num2str(m),'callback',...
'sdofgui(''newmass'');');
% Initialise editable spring field
stiffField = uicontrol('style','edit','position',...
[145 393 60 20],'string',num2str(k),'callback',...
'sdofgui(''newstiff'');');
% Initialise editable damper field
dampField = uicontrol('style','edit','position',...
[160 366 60 20],'string',num2str(c),'callback',...
'sdofgui(''newdamp'');');
% Initialise frequency parameter fields
lowField = uicontrol('style','edit','position',...
[150 300 60 20],'string',num2str(low),'callback',...
'sdofgui(''newlow'');');
highField = uicontrol('style','edit','position',...
[150 273 60 20],'string',num2str(high),'callback',...
'sdofgui(''newhigh'');');
nlinField = uicontrol('style','edit','position',...
[150 246 60 20],'string',num2str(Nf),'callback',...
'sdofgui(''newNf'');');

% End of part e). Part of solution to e) is located below in

% elseif strcmp(…) command part.

% Part f). Initialise calculation button and text.

calcBtn = uicontrol('style','pushbutton','string',...
'FRF','callback','sdofgui(''FRFcalc'');',...
'position',[83 170 40 20]);

% End of part f). Part of solution to f) is located below.

% Part g). Initialise diagrams for graphic presentation.

absAxes = axes('position',[0.4 0.58 0.5 0.3],'NextPlot',...

'replace');
phaseAxes = axes('position',[0.4 0.18 0.5 0.3],...
'NextPlot','replace');

% End of part g). Part of solution to g) is located below.

98
% Part e) continued.

% Execute whenever editable mass field is hitted

elseif strcmp(action,'newmass'),
% Convert "mass" string to numeric value
m = str2num(get(massField,'string'));
% Execute whenever editable spring field is hitted
elseif strcmp(action,'newstiff'),
% Convert "stiff" string to numeric value
k = str2num(get(stiffField,'string'));
% Execute whenever editable damper field is hitted
elseif strcmp(action,'newdamp'),
% Convert "damp" string to numeric value
c = str2num(get(dampField,'string'));
elseif strcmp(action,'newlow'),
low = str2num(get(lowField,'string'));
elseif strcmp(action,'newhigh'),
high = str2num(get(highField,'string'));
elseif strcmp(action,'newNf'),
Nf = str2num(get(nlinField,'string'));

% End of continued part e).

% Part f) continued.

elseif strcmp(action,'FRFcalc'),
% Define equidistant frequency vector.
f = linspace(low,high,Nf);
om = 2*pi*f; % Calculate om = 2*pi*f
om2 = om.^2;
FRF = 1./(k+i*om*c-om2*m); % Calculate FRF.

% End of continued part f).

% Part g) continued.

% Make absAxes current axis for next plot.

subplot(absAxes);
plot(f,abs(FRF));
xlabel('Frequency, [Hz]')
ylabel('abs(FRF) x/F, [m/N]')
grid
zoom on
% Make phaseAxes current axis for next plot.
subplot(phaseAxes);
plot(f,180/pi*angle(FRF));
xlabel('Frequency, [Hz]')
ylabel('phase(FRF) x/F, [degrees]')
grid
zoom on
% End of continued part g).

end;

% End of code

99
Executing the code with parameter values according to the figure below gives the graphic user
interface shown in the figure below.

Figure Graphic user interface for calculation of mechanical single degree of freedom system frequency response
functions.

72
Bold faced lines are additional lines for Sound button.

...

fileinpBtn = ...
uicontrol('Style','pushbutton','string','File','position',...
[5,395,40,20],'callback','[tdat,pdat,fsampl] = firstplot;');

% Create 'Sound' button.

fileinpBtn =
uicontrol('Style','pushbutton','string','Sound','position',...
[75,395,45,20],'callback','soundsc(pdat,fsampl);');

% Create exit pushbutton with red text.

100