Académique Documents
Professionnel Documents
Culture Documents
e-ISSN: 2455-5703
I. INTRODUCTION
The BBS mimics the real control problems namely stabilization of an airplane during landing, rocket and aircraft vertical takeoff. The BBS has two degrees of freedom namely rolling of the ball back and forth on the beam, and the other is beam rotation. It
has nonlinear dynamics and under-actuated characteristics. The nonlinear property becomes insignificant when the beam deflects
a small angle from the horizontal position which leads the mathematical modelling simpler. It is an ideal example for examining
the performance of classical, modern and advanced control theory. Many researchers have done work on this laboratory test
bench. Hauser et al proposed feedback linearization controller [1]. Pang et al designed a Linear Quadratic Regulator (LQR) for
stabilization [2]. H.Verrelst et al used Neural Network for stabilization of BBS [3],S.K Oh et al proposed Fuzzy based cascade
control [4] and Chang et al designed a tracking control strategy using a fuzzy sliding-mode controllers [5]. Further all the
intelligent controllers are knowledge based and the stability of the system is not guaranteed. In this proposed work controller is
designed for positioning and stabilization of the ball along the beam by manipulating the voltage applied to the servomotor. Thus
the system under consideration is a single input single output system (SISO).
The paper is organized as follows. Section II gives details about the mathematical modelling of BBS and servomotor. In
Section III the concepts of 2 DOF pole placement controllers is presented. In Section IV the design of pole placement controller
for a desired specification is discussed. The simulated results are given in section V. Finally conclusion is given in section VI.
479
Design of Digital Pole Placement Controller for a Ball & Beam System
(GRDJE / CONFERENCE / ICIET - 2016 / 078)
The Ball and Beam system (BBS) is driven by the servomotor and the schematic is shown in Fig. 1. The beam consists of a steel
rod in parallel with a nickel-chromium wire-wound resistor forming a track upon which a metal ball slides. One end of the beams
is connected to the servomotor through a lever arm and gear train while the other end is fixed.
The two forces influencing the motion of the ball are
Ftx
Force due to translational motion
Frx
Ftx m
d2x
mx
dt 2
2x
R R dt
R
dt
R
J
2
mR 2
5
Frx
2
mx
5
Applying the Newtons second law for forces along the inclination, we have
2
mx mx mg sin
5
or
5
g sin
7
5
g
7
(1)
(2)
The arc distance traversed by the gear at radius r is equated to the arc distance traversed by the beam at radius L , hence
r L or
r
L
5 gr
x
K
7L
X (s) K
(s) S 2
(3)
B. Modelling of Servomotor
The armature voltage equation is,
Va t La
dia t
dt
Ra ia t K m m
J L BL
K g Km
480
Design of Digital Pole Placement Controller for a Ball & Beam System
(GRDJE / CONFERENCE / ICIET - 2016 / 078)
J BL
Va t Ra L
K m Kg
K g K m
The transfer function of the servo motor including inertia and friction of load shaft is
Km K g
(s)
Va ( s)
Ra J eq
2 Beq K m2 K g2
S
J eq
Ra J eq
(4)
Or equivalently,
(s)
Va ( s )
am
S ( S bm )
(5)
Where
am
Km K g
Ra J eq
bm
Beq
J eq
K m2 K g2
Ra J eq
Parameter
Beam length (L) in meter
Lever arm offset (r) in meter
Acceleration due to Gravity ( g ) in m/S2
Value
0.1675
0.0254
9.84
2.6
0.00767
gear ratio ( K g )
70
Efficiency ( )
Equivalent moment of inertia ( J eq )Kg
m2
0.8
0.0023
410-3
The transfer function of the BBS connected with servo motor is,
Kam
X (s)
73
3
3
Va ( s ) S ( S bm ) S ( S 40)
Substituting the system parameters tabulated in Table I, the overall transfer function of the system is given as below:
X (s)
73
(6)
3
Va ( s ) S ( S 40)
The transfer function indicates that the open loop system is unstable due to the presence of multiple poles at origin.
) H (z
)U ( z
1
B( z )
H ( z 1 ) z k
A( z 1 )
(7.a)
(7.b)
It is desired to design a controller such that the plant output y tracks the set point ysp in the following manner:
Y (z
)z
k Br
1
Ysp ( z )
cl
(7.c)
481
Design of Digital Pole Placement Controller for a Ball & Beam System
(GRDJE / CONFERENCE / ICIET - 2016 / 078)
cl
Where
is the characteristic polynomial obtained based on desired closed loop specification. And
trajectory tracking is achieved.
Referring to Fig.2,
R( z 1 )U ( z 1 ) T ( z 1 )Ysp ( z 1 ) S ( z 1 )Y ( z 1 )
(7.d)
cl
Br
so that
The two DOF controller shown in Fig.2 has two parts namely, a feedback component R that helps in ensuring internal stability
T
Y
and a feed forward component R that helps in tracking set point sp .
1
1
Where R ( z ) , S ( z ) and T ( z ) are polynomials in
S
The one DOF controller has only the component R in the forward path as shown in Fig.3.
Y z k
TB
Ysp
AR z k BS
(7.e)
For set point tracking equation (2.e) is equated to equation 2.c) hence we arrive at,
B
BT
r
AR z k BS cl
(7.f)
In general degree of Br < deg of B, so that the desired closed loop transfer function is of order lower than given by BT.
This is achieved by cancelling common terms between numerator and denominator. In view of this, A and B is factored as good
and bad terms as,
A Ag Ab
and
B Bg Bb
If we let,
S Ag S1
T Ag T1
R Bg R1
Now, equation (7.f) becomes,
Bg Bb Ag T1
Ag Ab Bg R1 z k Bg Bb Ag S1
Br
cl
(8)
Cancelling the good common terms and on equating the numerator and denominator parts in the equation (8) we get,
Br BbT1
(9)
Ab R1 z k Bb S1 cl
(10)
e(n) r n cos n
482
Design of Digital Pole Placement Controller for a Ball & Beam System
(GRDJE / CONFERENCE / ICIET - 2016 / 078)
And from the desired specifications on the transient response the allowable range for r and is determined.
2N
For a small fall time in error signal e(n) the angular frequency
specified overshoot.
Y ( z)
The transfer function relating the set point sp
and the actual output Y ( z ) is given by,
TY ( z )
1 r cos r 2 r cos z 1
1 2 z 1r cos r 2 z 2
is the
( z 1 )
cl ( z 1 )
TY ( z 1 )
Y ( z)
( z 1 )
z 1
Ysp ( z )
cl ( z 1 )
(11)
2.82x10
zk
B(z)
A(z)
(12)
Bb 1
2) Step 2: From the desired specifications on the transient response the allowable range for r and is determined.
The rise time is chosen as at least one third of the settling time i.e, 0.15 sec
N rise time / Ts =15
0.1047
cl ( z 1 ) 1 1.8z 1 0.819z 2
A R z
4) Step 4: The Aryabhattas identity as given below is used to determine the terms S1 and R1. b 1
1 z 1 1 1.01z 1 R1 z 1 S1 1 1.8 z 1 0.819 z 2
Bb S1 cl
V. SIMULATION RESULTS
The performance of the proposed control algorithms are analyzed by simulating the 1 DOF and the 2 DOF pole placement
controller structure shown in Fig.2 and Fig.3 using MATLAB. The discrete plant transfer function is obtained using a zero order
483
Design of Digital Pole Placement Controller for a Ball & Beam System
(GRDJE / CONFERENCE / ICIET - 2016 / 078)
S z
R z
T z
Rz
loop response of the BBS connected with servo motor is shown in Fig.4, which indicates that the steady displacement is not
achieved in finite time and the system dynamics is nonlinear. The performance of 1 DOF and 2 DOF pole placement controllers
is shown in Fig.5 for step change in voltage variation. It indicates that the 1 DOF controller exhibits excess overshoot compared
to 2 DOF controllers. In order to show the disturbance rejection capability a disturbance input is given at 1 & 4 secs respectively
as indicated in Fig.6. The variation of armature voltage is shown in Fig.7 which indicates that the excitation voltage is damped
oscillatory until the ball tracks the desired trajectory. This is because the lever arm rotates in a bidirectional way within 1800
angle variation. The stabilization response of the ball and beam system from its initial position of 2 Cm is shown in Fig.8.
Fig. 4: Open loop step response of BBS connected with servo motor
484
Design of Digital Pole Placement Controller for a Ball & Beam System
(GRDJE / CONFERENCE / ICIET - 2016 / 078)
VI. CONCLUSIONS
The 1 DOF and 2 DOF pole placement controller algorithms are demonstrated for a modelled BBS connected with servo motor.
The time response shows that the settling time, overshoot and rise time constraints are almost met for the 2 DOF controller
compared to the 1 DOF controller. The performance summary given in Table II indicates that, the 2 DOF pole placement
controller offers less ISE and IAE value. Further 1 DOF pole placement controller exhibits more overshoot compared to 2 DOF
pole placement controllers and which is an undesirable phenomenon.
Parameter
1 DOF PID Controller 2 DOF PID Controller
Peak time tp
0.15
0.24
Rise time tr
0.04
0.1
Settling time ts
0.51
0.4
% overshoot
27.8
7.5
ISE
95.67
95.21
IAE
95.60
95.10
Table 2: Comparison of Time domain Specifications and Performance Indices
REFERENCES
[1] J. Hauser, S. Sastry, and P. Kokotovic, Nonlinear control via approximate input-output linearization the ball and beam
example, Institute of Electrical and Electronics Engineers, Transactions on Automatic Control, vol. 37, no. 3, pp. 392398,
1992.
[2] Z.H Pang, G. Zheng and C.X. Luo, Augmented State Estimation and LQR Control for a Ball and Beam System, Proc. of
the 6th IEEE Conference on Industrial Electronics and Applications., pp.1328-1332, June 21-23, 2011.
[3] H. Verrelst, K. Van Acker, J. Suykens, B. Motmans, B. De Moor and J. Vandewalle, Neural Control Theory: Case Study
for a Ball and Beam System, Proc. of the European Control Conference Brussels, Belgium, July 4, 1997.
[4] S.K Oh, H.J. Jang and W. Pedrycz, The Design of a Fuzzy Cascade Controller for Ball and Beam System: A Study in
Optimization with the Use of Parallel Genetic Algorithms, Engineering Applications of Artificial Intelligence, Vol. 22, pp.
261271, 2009.
[5] Y.H. Chang, C.W. Chang, C.W. Tao, H.W. Lin and J.S. Taur, Fuzzy Sliding-mode Control for Ball and Beam System with
Fuzzy ant Colony Optimization, Expert Systems with Applications, Vol. 39, No. 3, pp. 3624-3633, 2012.
[6] S. Nagammai, S.V. Prasanna Kumar, S. Latha, Design of Digital Controller for an Unstable CSTR Process Journal of
Electronic Design Technology, Volume 5, Issue 3, pp.235-240, 2014.
[7] Ball and Beam Experiment and Solution, Quanser Consulting, 1991
485