Anda di halaman 1dari 7

DEPARTMENT OF MECHANICAL ENGINEERING

UNIVERSITY COLLEGE LONDON

MECH3005 Automatic Control


MATLAB Tutorial
Introduction
The aim of this class is to introduce MATLAB to those who may not have used it before, and to
demonstrate specific functions from the control systems toolbox that are particularly useful in control
engineering.
MATLAB 6.1 is found as an icon on the
MENG desktop – double click to start…

Type your commands at the


prompt…

Previous commands are here –use the cursor


keys in the command window to reuse them.
You can also drag and drop commands from
the history to the command window

Course Notes for BEng/MEng Mechanical Engineering


B M Hanson, email b.hanson@ucl.ac.uk
Course web page: http://www.meng.ucl.ac.uk/~b_hanson/MECH3005.htm
Simple Arithmetic
Work through the following exercises to give you a taster of some of the MATLAB commands
and to get you used to the user interfaces. This exercise will help you when you come to do the
examples on which this course is assessed.
Start with some basic commands and record what you see in the boxes below.
1. MATLAB can do sums like a calculator

Your Input: 5+3 5*sin(30*pi/180 ) 3^2 – 6*3


MATLAB output:

2. Notice that the output says ans= 8 (for the first example). What MATLAB does is assign the
answer to the variable ans (by default). This is very useful as you can now use ans for further
calculations. Try the following and record the output (note that spaces are ignored here):

Your Input: 5+3 ans + 3 ans + 5


MATLAB output:

3. So you get the first answer, add 3 with the second command and add 5 more with the
third command. As well as the default variable ans, you can assign your own variables. Try out
the following sequence (note 6.3 x 106 is entered using “e” for exponent):

Your Input: grav =9.8 radius =6.3e6 escape_velocity=2*grav*radius


MATLAB
output:

So the escape velocity you have calculated (in m/s incidentally) is stored in the variable
escape_velocity. You do need to be careful with variable names – you wouldn’t call a variable
pi as MATLAB would get confused with its internal value of pi. MATLAB has also pre-
assigned i, j, (as the square root of -1) and some other constants.

Vectors
Enter each element of the vector separated by a space, between square brackets, and set it
equal to a variable. For example, to create a vector “a”, enter into the Matlab command window:

a = [1 2 3 4 5 6 9 8 7]
Matlab should return:

a =
1 2 3 4 5 6 9 8 7
Let's say you want to create a vector with elements between 0 and 20 evenly spaced in
increments of 2 (this method is frequently used to create a time vector):

t = 0:2:20

t =
0 2 4 6 8 10 12 14 16 18 20
Manipulating vectors is almost as easy as creating them. First, suppose you would like to add 2
to each of the elements in vector 'a':
Here we’re adding a scalar to a vector- try
b = a + 2 a similar method to multiply the elements
of b by a scalar (eg 3.75)
b =
3 4 5 6 7 8 11 10 9

Now suppose, you would like to add two vectors together. If the two vectors are the same length,
it is easy. Simply add the two as shown below:

c = a + b

c =
4 6 8 10 12 14 20 18 16
Subtraction of vectors of the same length works exactly the same way.

Functions
Matlab includes many standard functions. Each function is a block of code that accomplishes a
specific task. Matlab contains all of the standard functions such as sin(), cos(), ln(),
log10(), exp(), sqrt(), as well as many others.
sin(pi/4)
Note: MATLAB performs trigonometry in
ans = radians only

0.7071

For more information on the usage of any function, type help [function name] at the
Matlab command window. For example help sqrt

Polynomials
In Matlab, a polynomial is represented by a vector. To create a polynomial in Matlab, simply
enter each coefficient of the polynomial into the vector in descending order. For instance, let's
say you have the following polynomial:
s 4 + 3s 3 − 15s 2 − 2 s + 9
To enter this into Matlab, just enter it as a vector in the following manner

x = [1 3 -15 -2 9] Note how the elements of the vector


correspond to the polynomial coefficients
x =
1 3 -15 -2 9

If your polynomial is missing any coefficients, you must enter zeros in the appropriate place in
the vector. For example,

s4 + 1 would be represented in Matlab as: y = [1 0 0 0 1]

Also, just s2 on its own would be represented by y = [1 0 0]


You can find the value of a polynomial using the polyval() function. For example, to find the
value of s 4 + 3s 3 − 15s 2 − 2 s + 9 at s = 2:

z = polyval([1 3 -15 -2 9], 2) Note: this function takes two inputs, or


“arguments”: one vector and one
scalar, separated by a comma.
z =
-15
You can also extract the roots of a polynomial using the function roots(). This is useful when
you have a high-order polynomial, for example s 4 + 3s 3 − 15s 2 − 2 s + 9 :
roots([1 3 -15 -2 9])

ans =
-5.5745 Note: this also finds complex roots, so
2.5836 it’s very handy for plotting root
-0.7951 positions, or root locus diagrams
0.7860

So the factorised polynomial would be (s + 5.57)(s – 2.58)(s + 0.80)(s – 0.79)

Very often, we want to multiply two polynomials together. The product of two polynomials is
found by taking the “convolution” of their coefficients. Matlab's function conv() that will do
( )
this for you. To perform this multiplication: (s + 2 ) s 2 + 4 s + 8 = s 3 + 6 s 2 + 16 s + 16 we’d enter:
x = [1 2];
y = [1 4 8]; Note: z = conv([1 2],[1 4 8])
z = conv(x,y) would achieve the same result…

z = You can check the answer by multiplying out


1 6 16 16 the brackets by hand.

Notice how the elements of vector z relate to the coefficients of s in the answer of the
multiplication above. Also remember to watch out for “missing” coefficients, for example
( )
This multiplication: s 3 s 2 + 4 s + 8 = s 5 + 4 s 4 + 8s 3 would have to be calculated as:
(s 3
)( )
+ 0 s + 0 s + 0 s + 4 s + 8 = s + 4 s 4 + 8s 3
2 2 5
( plus 0s 2 + 0s + 0 )
x = [1 0 0 0];
y = [1 4 8];
z = conv(x,y)

z =
1 4 8 0 0 0

Note: using a semicolon at the end of the line (as seen above) stops MATLAB from printing the
answer each time! This is essential for calculations on large arrays.
Transfer Functions
Consider an open loop system which has a transfer function of
Y (s) s+7
H (s) = =
U ( s ) s ( s + 5)( s + 15)( s + 20)

Enter the transfer function by entering the two polynomials (called num and den here just to
indicate to ourselves that they’re numerator and denominator, but we’re free to call them other
things, e.g. A, B, plantnum, plantden):

num=[1 7];

The denominator can be entered as follows (there are other ways to perform the convolution):

den=conv(conv([1 0],[1 5]), conv([1 15],[1 20]));

Note that the “s” term is represented by the vector [1 0] meaning 1 times “s” and 0 times “s0”).
Each “conv” function returns a vector result of a multiplication, and “den” is set to be the result
of the multiplication of those two vectors.
You can check the value of any variable by just typing the name and pressing enter:
>> den

den =
1 40 475 1500 0

Once a system has been entered in this format, MATLAB can do lots of useful things:
Step and impulse responses:
step(num,den)
impulse(num,den)

We can show the positions of the poles and zeros on the s-plane:
pzmap(num,den)

And we can even plot a root locus diagram with one simple command – how useful!:
rlocus(num,den)

Closed-loop Transfer Function


Matlab will even compute a closed-loop transfer function from an open-loop transfer function –
the function cloop() should be sufficient for this course. It assumes negative feedback, with
unity gain. See “help feedback” for more details/options. Here, the result of the function
cloop() is a two-element vector, which we’re assigning to the two element vector
“[numCL, denCL]”. Each element of that array is itself a vector.

[numCL, denCL] = cloop(num, den)

We’re using the terms “numCL” and “denCL” here, but you can call the closed loop numerator
and denominator whatever you like.
Now you can look at the step response of the closed loop transfer function:
step(numCL,denCL)
You might notice that the steady state error to a step input is zero, since the open-loop transfer
function has a pole at the origin…
Frequency Response
MATLAB generates a frequency response with no trouble at all:
bode(num,den)

It’s as simple as that – but you can add more complexity. For example, by default, bode() plots
a graph, but you can also assign the results to variables in the workspace:
[mag,phase,w]=bode(num,den);  you might want a semicolon here

Where w is the list of frequencies used, phase is the phase angle in degrees, and mag is the
magnitude response, but NOT in dB, just to confuse you. You can (of course) convert to dB
using something like:
mag_in_dB = 20*log10(mag);

You can also generate a frequency response over any frequency range you like, not just
MATLAB’s chosen range. In that case, the logspace command is useful (see its help file):
my_freqs = logspace(-2,3,100); Generates 100 log-spaced points between 10-2 and 103
bode(num,den,my_freqs)

To plot in polar format, try the following:


polar(phase*pi/180, mag)  have a look at help polar()
(Here, we’ve needed to convert the phase to radians.)

Phase & Gain margins? No problem at all:


margin(num,den)
Though do make sure you know what you’re entering – i.e. the open loop transfer function!
Example:
Using the Remote Position Control Servo Example.
This example compares two models of the same system – one model ignores the effect of coil impedance
st
in the electric motor – the other includes it, as a 1 order lag.
The two models:

Without coil impedance With coil impedance included


J f J f

motor motor
amplifier amplifier
1
θi V = kθ i T = 1 .0 × V θi V = kθ i T= V
error error Ls + R
signal signal

output output
• Equation of motion: • Equation of motion:
shaft θ shaft θ

(
T = θ 0 Js 2 + fs = kθ i ) position (
T = θ 0 Js 2 + fs = ) (Lsk+ R ) θ i
position

θ o (s ) k θ o (s ) k
= =
• Laplace transfer function: θ i (s ) s( Js + f )
• Laplace transform:
θ i (s ) s( Js + f )(Ls + R )

2 -3
J = 0.01 kg m , f = 0.05 N m s, R = 1 Ohm, L = 5 x 10 Henrys

• With k = 0.1, find the positions of the closed loop roots


Hint: use cloop() to find the closed loop transfer function (numerator, denominator) then use pzmap()

• Hence define the impulse response of the c.l. system in terms of ζ, ωn - What effect does
inclusion of coil impedance have?
Hint: these are calculated from the positions of the poles on the s-plane. If you mouse-click on the poles, MATLAB
displays information about the position of the poles – damping and natural frequency.

• Now plot the full root locus for the each system (as a check, this should pass through your point
at k=0.1)
Hint: use rlocus(), remember to use the numerator and denominator of the open-loop transfer function

• For each system, find the value of K to give ζ = 0.7 and 0.5 and find the damped natural
frequency ωd at those points
Hint: use rlocus(), and by mouse-clicking on the locus, you can identify the gain for a particular point

• Fully describe the stability of each system with k = 0.1 (phase margin, gain margin). Critical gain
can be found using Routh array or Nyquist method – try both, see which you prefer!
Hint: use margin()

• Further: have a look at PM & GM as k varies.


Hint: use margin(), either change the value of “k”, and recalculate “num” and “den” or multiply “num” by a scalar.

Final task:
Copy and paste the contents of the command history (see first page of these notes) into an email
to b.hanson@ucl.ac.uk. Make sure your email contains your name!

Anda mungkin juga menyukai