Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
98 views

Z-Plane RL Design Using MATLAB: 2.1 Plant

This document describes using MATLAB to design a lead compensator controller for a unit inertia plant with a specified closed-loop response. The plant and specifications are presented, including a desired damping ratio of 0.707 and natural frequency of 100 Hz. A proportional controller results in an unstable system. A lead compensator is designed by placing a zero near the desired closed-loop pole and using root locus analysis to find the compensator pole location. The closed-loop response is analyzed, showing poles near the specified frequency and damping, confirming the successful design of the lead controller.

Uploaded by

pra_zara2637
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
98 views

Z-Plane RL Design Using MATLAB: 2.1 Plant

This document describes using MATLAB to design a lead compensator controller for a unit inertia plant with a specified closed-loop response. The plant and specifications are presented, including a desired damping ratio of 0.707 and natural frequency of 100 Hz. A proportional controller results in an unstable system. A lead compensator is designed by placing a zero near the desired closed-loop pole and using root locus analysis to find the compensator pole location. The closed-loop response is analyzed, showing poles near the specified frequency and damping, confirming the successful design of the lead controller.

Uploaded by

pra_zara2637
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

ME 481/581

Chapter 5

February 29, 2012

Z-Plane RL Design using MATLAB

Introduction.

This document is intended to give you an example of using MATLAB and root locus to design controllers (compensators) in the z-plane.

2
2.1

Plant and Specifications.


Plant.

The plant will be an pure unit inertia; this could be any rotary element in a low-friction environment. The general
transfer function from applied torque to angular position for a pure inertia is
1
1
(s)
=
= 2 (unit inertia)
2
(s)
Js
s

2.2

(1)

Specification.

I arbitrarily selected the closed-loop natural frequency to be 100 Hz and damping of 0.707, along with a sampling
frequency of fs =500 Hz (T =0.002). This corresponds to desired s-plane and z-plane pole locations of
p
s = n jn 1 2 444 j444.
(2)
z = esT = 0.2596 + j0.3192 0.26 + j0.32

(3)

This sampling frequency is quite low; we would typically sample at perhaps 20 the bandwidth (natural frequency).

2.3

Discretized Plant.

With T =0.002, the zoh-discretized model of the plant (found using MATLAB c2d) is:
>> Gc = tf([0 0 1],[1 0 0])
Transfer function:
1
--s^2
>> Gd = c2d(Gc,T)
Transfer function:
2e-06 z + 2e-06
--------------z^2 - 2 z + 1
Sampling time: 0.002
This can be expressed as
G(z) =

(2e 6)(z + 1)
(2e 6)z + (2e 6)
=
z 2 2z + 1
(z 1)2

Note the small numerator constant of 2e 6; that will probably be counteracted by a large forward path gain.

(4)

ME 481/581

Chapter 5

February 29, 2012

Compensator Design.

Proportional control will be tried first, then a lead compensator.

3.1

Proportional Control.

With proportional control the actuating signal sent to the plant is proportional to the error signal, hence the compensator is a pure gain K.
3.1.1

Closed-Loop Transfer Function.

With system input called R(z) and system controlled variable (output) called Y (z) the closed-loop transfer function
W (z) is
KG(z)
Y (z)
=
R(z)
1 + KG(z)

(5)

1 + KG(z) = 0

(6)

hence the system characteristic equation is

This is already in root locus form.


3.1.2

Root Locus of Proportional Control.

Using equation (6) we can draw the MATLAB root locus:


>> rlocus(Gd); axis equal; grid;
Zooming in on the locus near the +1 point we get the figure below:
Root Locus

0.4/T

0.3/T

0.7

0.6
0.2/T

Imaginary Axis

0.5

0.4

0.3

0.1/T

0.2 0.5/T

0.1

0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

0.6/T

0.2

0.4

0.6

0.8
Real Axis

1.2

1.4

The locus (blue line) starts from the double pole at z = 1 and goes outside the unit circlethis system is unstable.

ME 481/581

3.2

Chapter 5

February 29, 2012

Lead Compensator.

We need to improve the stability of this system. A lead compensator (a quasi-differentiator) will do that. This can be
seen from two perspectives:
A lead network will pull the locus to the left, thus keeping it inside the unit circle
A lead network adds positive phase angle, thereby improving the phase margin
The structure of a lead compensator D(z) is
D(z) =

z+b
z+a

(7)

where the zero is closer to the +1 point than the pole.


3.2.1

Compensator Pole/Zero Locations.

My philosophy on lead compensator design is to place the compensator zero in the s-plane at about one-third the
distance of the desired dominant poles. Thus if the desired poles are at 445 j445 Id place the compensator zero
at about 445/3 150. Now in the z-plane this zero location would be
z = esT = e(150)(0.002) 0.74 = D(z) =

z 0.74
z+a

(8)

The location of the compensator pole is found using the root-locus angle condition, in which we use open-loop transfer
function D(z)G(z):
(angles of vectors from poles to desired pole) (angles of vectors from zeros to desired pole) = 180

(9)

The relevant values in this problem are:


The open-loop poles: -1, -1, and the unknown lead compensator pole
The open-loop zeros: 0.74, -1
The desired closed-loop pole: 0.26+j0.32
I suggest you draw the z-plane and draw these vectors. Here are the values that I found:
2(156 ) + 146 14 = 180

(10)

Parameter in (10) is the angle of the vector from the lead compensator pole to the desired closed-loop pole location.
From (10) this angle is
= 28

(11)

This implies that the pole must be located at z = 0.354, thus the lead compensator is
D(z) =

z 0.74
z + 0.35

(12)

The root-locus diagram of D(z)G(z) is shown on the next page; the cross shows where I specified the desired pole
(near = 0.6).
The corresponding K at this point is:
K = 3.14e5 = 314, 000
Remember the small numerator constant in G(z)? This K makes up for it.

(13)

ME 481/581

Chapter 5

February 29, 2012

Root Locus

0.6/T

0.5/T

0.4/T

0.8

0.1 0.3/T
0.7/T

0.2
0.3

0.6

0.2/T

0.4
Imaginary Axis

0.5
0.4 0.8/T

0.6
0.7
0.1/T

0.8
0.2 0.9/T

0.9

/T
0
/T
0.2 0.9/T
0.1/T
0.8/T

0.7/T

0.6/T 0.5/T 0.4/T

0.3/T

0.2/T

0.4
0.5

0.5

Real Axis

Final Closed-Loop System.

With the D(z) and G(z) given previously, the closed-loop transfer function is
Y (z)
KD(z)G(z)
=
R(z)
1 + KD(z)G(z)

(14)

Using MATLAB, this is


>> Wd = feedback(K*Dd*Gd,1)
Transfer function:
0.6287 z^2 + 0.1635 z - 0.4652
----------------------------------z^3 - 1.017 z^2 + 0.4555 z - 0.1112
Sampling time: 0.002
>> damp(Wd)
Eigenvalue

Magnitude

Equiv. Damping

Equiv. Freq. (rad/s)

5.58e-01
2.29e-01 + 3.83e-01i
2.29e-01 - 3.83e-01i

5.58e-01
4.46e-01
4.46e-01

1.00e+00
6.16e-01
6.16e-01

2.91e+02
6.54e+02
6.54e+02

The complex closed-loop poles have a damping ratio of 0.65 and a natural frequency of 654 rad/s (104 Hz). Id say
our design is pretty close.
One could get the step response, etc.
Since this is a Type II system, it should definitely have a DC gain of 1.00; from MATLAB we get
>> dcgain(Wd)
ans = 1

You might also like