Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Journal of ELECTRICAL ENGINEERING, VOL. 56, NO.

3-4, 2005, 78–83

ROBUST CONTROL OF REAL


EXPERIMENTAL BRIDGE CRANE
∗ ∗∗
Marek Hičár — Juraj Ritók

The drive control of a bridge crane includes a crane crab, bridge and uplift control. At first, a mathematic and physical
analysis of the main crane components was performed and variable system parameters were defined. Robust control by
Ackermann ensures crane robustness against the burden weight and rope length variation, which provides sharp positioning
and forbidden swinging of the burden in the final position. The real crane will be connected to a distributed control system
(DSR). An analysis of the working place is performed with accent on the measurable units of burden swinging, and the
results of robust controlling are described.
K e y w o r d s: bridge crane, crane crab, bridge and uplift, robust control, burden swinging

1 INTRODUCTION

The main problem of the crane drive control is to keep


almost zero burden deviation in the final crane crab or
bridge position. The main variable parameters are the
burden weight mG (detected by a tensometer) and hang-
ing rope length l (measured by an incremental sensor). A
robust design by Ackermann’s method is preferably used
for changeable systems or systems with variable parame-
ters. The range of robustness is checked and areas related
to expected variation values are defined. Knowing the val-
ues of changing parameters, a robust switch table is con-
figured for their whole range of variance that satisfies the
condition of forbidden burden swinging. In the future we
will work on eliminating the varying parameters by an Fig. 1. Experimental bridge crane.
observer of the burden weight and burden swinging.
The real crane was completely designed and connected mk
to the distributed control system (DSR). Communication Fk
0
interfaces between the technological, system operator and xk
information levels were created. The bridge crane was
identified by ARX model through experimental identifica-
tion and we obtained the crab, bridge and uplift transfer l
functions and their transformation to the state descrip-
tion. OE model identified burden swinging in the direc- α
tion of the crane crab and bridge. The crane position was
scanned by incremental sensors. For scanning the burden mG β
swinging we first used Hall sensors. Because of interfer- yk xG,yG
ence and output error we eventually measured the swing- FG
ing by rotary rheostats. The frequency converter NORD
Fig. 2. Model of the crane crab (bridge).
is used for the crab, bridge and uplift drive control.
The burden weight, respecting the designed crane con-
struction, ranges from 0 to 100 kg and the rope length 2 ANALYSIS OF CRANE
from 0 to 2.5 m. When trying to keep zero burden swing- CRAB, BRIDGE AND UPLIFT
ing, the worst situation occurs in the case of the shortest
rope at the minimum weight because the frequency of The same model of crab and bridge is considered and
periodical vibrations increases. their motions are perpendicular. For the crab with weight


Technical University of Košice, Letná 9/B, 042 00 Košice, Slovakia, Department of Electric Drives and Mechatronics, E-mail:
∗∗
hicarm@hron.fei.tuke.sk, Department of Constructions, Transport and Logistics, E-mail: juraj.ritok@tuke.sk

c 2005 FEI STU


ISSN 1335-3632
Journal of ELECTRICAL ENGINEERING 56, NO. 3–4, 2005 79

Transmission
Winding drum Im
D1Z
IbZ s - plain ω
Motor D2Z
γ
JmZ,w mZ 0

JpZ,jZ,hpZ Imax b
2aGZ G - plain
Re
2vGZ -a 0 σ
-R
0.5mG
Uplift tackle

vGZ
Burden mG aGZ
mG

Fig. 3. Detail of the crane crab. Fig. 5. Definition of Γ -plane

Uplift motor
CRANE CRAB where l is the length of the hanging rope, Jc is the total
Uplift
Electric
moment of inertia, p is the number of motor poles, Lh is
Crab case tackle
L′
trasmission No.1,2 the main motor inductance, coefficient σ2 = L2σ h
, mz is
the load moment, r is the radius of the drive wheel and
j is the transmission.
A

r37 Burden mass up to 100 kg


Crane bridge 0.000
ELECTRIC MOTOR WITH TRANSMISSION
(It has hollow shaft and is overthrusted
on the wheel shaft) -0.005

Crab case -0.010


Wheel shaft with springs A
Wheel bearings -0.015
Detail A Wheel
-0.020

Fig. 4. Model of the crane uplift -0.025

-0.030
-0.10 -0.09 -0.08 -0.07 -0.06 -0.05 -0.04 -0.03 -0.02
r37
mK (bridge weight mKM ) and for its speed, angular Fig. 6. Three robust areas for the crane crab.
deviation α of the rope with respect to the vertical line
occurs (Fig. 2). The angle α changes in dependence on
the rope length and can be recalculated to the arc β The uplift drive includes the motor, transmission,
followed by the burden. The actual position of the crab winding drum and uplift tackle. The torque formula for
is xK (or of the bridge xKM ). The crab motion formulas crane uplift analysis is
(also for the bridge) are [1, 4]:
β dωmZ
mK ẍK − mG g = FK , (1) MmZ − MGZ = JCZ , (7)
d dt
β
mK β̈ + (mK + mG ) g = −FK . (2)
d where MmZ is the motor torque, MGZ is the burden
We can modify equations (1, 2) and use the motion equa- torque and JCZ is the total moment of inertia.
tion (3) of the induction motor: Conversion between the shifting burden speed vGZ to
Jc dω 3p Lh the angular motor speed ωmZ is
= i2m i1y − mz , (3)
p dt 2 1 + σ2
a11 ωmZ D2Z
ẍK = a31 c2 c1 i2m i1y + c1 β , (4) vGZ = . (8)
n jZ 4
a12
β̈ = −a31 c2 i2m i1y + c3 ẍK − β. (5)
n If r2Z = 2v ωmZ , then the adjusted burden moment of
GZ

In formulas (4, 5) were used the following notation:


inertia is JCGZ = m2G r2Z
2
. The potential burden torque
 J j2 1 j mG
at uplift when M1Z = 2 gr2Z is
−1
c
c1 = 2
+1 , c2 = ,
p r mK mK r
(6)
Jc j 2 mG mG gD2Z
c3 = , a11 = , M1GZ = , (9)
p r 2 mK mK 4jZ ηpZ
80 M. Hičár — J. Ritók: ROBUST CONTROL OF REAL EXPERIMENTAL BRIDGE CRANE

Imag.
r37M Burden mass up to 100 kg -5
0.00 4
-0.02 AM 3

-0.04 2
1
-0.06
0
-0.08 -1
-0.10 -2
-3
-0.12
-4
-0.14 -5
-0.60 -0.55 -0.50 -0.45 -0.40 -0.35 -0.30 -0.25 -0.20 -11 -10 -9 -8 -7 -6 -5
-4 -3 -2 -1 0
r37M Real.
Fig. 7. Two robust areas for the crane bridge. Fig. 8. Verification of robustness by poles motion at burden weight
variation.

Imag. d0 (α) = 1, d1 (α) = 2σ(α), α — frequency (1, ∞),


6
di+1 = 2σ(α)di (α) − σ 2 (α) + ω 2 (α) di−1 (α)
 
4
for i = 1, 2, . . . , n − 1 .
2

0 Laborec.fei.tuke.sk
Windows 2000
MSSQL server
-2 in touch
Information
level
-4
Ethernet
-5
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0
Real.
Windows 2000 SCADA
Fig. 9. Verification of robustness by poles motion at rope length OPC server level
Rockwell software
variation. in touch 7.1
Matlab 5.2 +
DH

RS-232C
and at burden lowering Technological
PLC 5/25
level
mG gD2Z
M2Z = ηpZ . (10)
4jZ

The total moment of inertia for uplift is


Bridge
JmZ crane
JCZ = + JCbZ + JCGZ , (11)
ηmZ
where JmZ is the motor moment of inertia, JCbZ is the
drum moment of inertia, JCGZ is the burden moment of Fig. 10. Distributed crane control system.
inertia.
All poles of the characteristic polynomial have to be
located in the left part of Γ -area (Fig. 5) and to dis-
3 ROBUST CONTROL DESIGN
tance −a from the imaginary axis and inside the sector
The property of robust control is in system robustness assigned by the damping value d = sin γ . Then stability
and its correct control against variation of parameters. and desired damping are ensured (Fig. 5). n−2 back-feed
In formula (12) the polynomial a(p) includes a variable parameters have to be known for the design to be robust.
characteristic polynomial with robust controllers. Solu- A graphical computing method can locate the values of
tions are feedback parameters suitable for uncertain pa- the rest two controllers r1 , r2 which are included in a(p).
rameters and in this meaning required system robustness Then curves A for the minimum and maximum values of
[1, 2, 3]. variable parameters plot the possible selection of robust
controllers [5].
   
d0 (α) d1 (α) . . . dn (α) 0 We solved the system robustness against the burden
a(p) = , where: (12)
0 d0 (α) . . . dn−1 (α) 0 weight and the rope length. The created robust areas are
Journal of ELECTRICAL ENGINEERING 56, NO. 3–4, 2005 81

3.0
x5Z
2.5

2.0
w5
1.3

1.0
x5
0.5

0.0

-0.5
0 5 10 15 20 25 30
Time (s)
Fig. 11. Incremental sensor for measuring the bridge position. Fig. 12. Sensor of swinging in direction of crane crab and bridge.

0.15
x6
0.10

0.05

0.00

-0.05

-0.10

-0.15
0 5 10 15 20 25 30
Time (s)

Fig. 13. Crab trajectory


Fig. 14. Crab swinging without robust control

0.15
acteristic polynomial at burden weight and rope length
no overswingings
0.10
variations. The requirement is to keep the poles in the la-
belled area, which provides stability and desired system
0.05 x6 damping. Based on our experience and consultation with
the crane producer KPK Martin spol. s r.o. we allowed
0.00 the maximum burden swingings at the final position to
be 0.5 cm, this value having no effect upon the quality of
-0.05
control. The poles located at the edge of the sector cause
-0.10 overswinging by 0.5 cm at small rope lengths ( l = 2.5 to
0.5 m).
-0.15
0 5 10 15 20 25 30
Time (s)
4 DISTRIBUTED CONTROL SYSTEM (DSR)
Fig. 15. Crab swinging with robust control

Technological control level — here belongs the model


of technological equipment (crane bridge), technological
divided by these variable parameters. In fact we received nets, control systems and programmable logical automa-
two robust areas in the case of a long rope (l = 2.5 to tons. Data from this level include process data, informa-
1.5 m) and three areas for a short rope (l = 2.5 to 1.5 m). tion about the equipment state, its configuration param-
For robust control of the crane bridge, three areas are eters and so on. Process data continue to a higher level
found at all weights and rope lengths. We have chosen with system state information. Backwards go the control
just one robust pair of speed controllers r36 , r37 type data.
PI (for bridge r36M , r37M ) from each robust area A for SCADA/HMI level — parts of this level are the de-
varying weight and rope length (Figs. 6 and 7). Verifi- vices of SCADA/HMI systems and process database, in
cation of the correct choice of the robust controllers is our cause MS SQL. This level is intended for primary col-
performed by displaying the poles motion of the char- lection of process data, monitoring, controlling and pro-
82 M. Hičár — J. Ritók: ROBUST CONTROL OF REAL EXPERIMENTAL BRIDGE CRANE

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0

-0.1 -0.1
x6M x6M
-0.2 -0.2

-0.3 -0.3
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Time (s) Time (s)

Fig. 16. Incremental sensor for measuring the bridge position. Fig. 17. Sensor of swinging in direction of crane crab and bridge.

0.3 0.3

0.2 0.2
x6M x6M

0.1 0.1

0.0 0.0

x6 x6
-0.1 -0.1

-0.2 -0.2

-0.3 -0.3
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Time (s) Time (s)

Fig. 18. Crab and bridge swinging Fig. 19. Crab and bridge swinging

cess evaluating and visualizing. The data gained can be other one in bridge direction. Burden swinging is trans-
next used in the information control level. mitted through the hanging rope so that chromo bars are
Information level — here are computer servers (Topla, installed on the rope and connected with the correspond-
Laborec, Ondava, Alf, Poprad), database equipment (Or- ing rheostat. For correct robust control it is necessary to
acle, MsSQL) for a higher control level and Internet [6]. know the burden weight and rope length. The load weight
Fig. 10 Distributed crane control system. is gained from the strain-gauge sensor that corresponds
to the resistance value at burden variation.

5 CRANE PARAMETERS MEASUREMENT


6 RESULTS OF CRANE MEASUREMENTS
The crane position and its recalculated speed are mea-
sured by two kinds of incremental sensors (Fig. 11). Two The control program consists of these main parts: cal-
sensors of IRC 120 type are assigned for measuring the ibration and crab, bridge and uplift drive control, cal-
bridge position, next two Hengstler type were used for culation block, setup parameters, identification. We deal
crab and uplift measuring. Their function is to trans- firstly with crab and bridge motion but to reduce the time
form the rotary motion to electrical signals proportional of transport it is very efficient to test simultaneously the
to the passed trajectory. The cables from sensors have to motion of both components.
be shielded otherwise interference and self-graduation of 1. Crab motion: mG = 50 kg; l = 2.5 m (Figs. 13, 14).
incremental sensors will be observed [7, 8]. The experi- Figure 13 displays the actual rope length x5Z (m), ref-
mental crane was assembled with position switches that erence w5 (m) and real x5 (m) crab position. Simulation
protect the crab, bridge and uplift from moving outside of burden deviation x6 (m) is in Fig. 14 without robust
the allowed area. control and shows forbidden periodical oscillation in the
Swinging was scanned by four Hall probes (two for the terminal position (x5 = 1.5 m and back to 0 m).
crab and two for bridge direction). Because of a strong in- Simulation in Fig. 15 presents robust control of burden
terference signal at the start of ac motors this sensor was swinging to the zero value with allowance of 0.5 cm over-
replaced by rotary rheostats. All rheostats are located up- swinging. The burden is damped by the control system to
right, one corresponds to swinging in crab direction and zero oscillation. Simulations of the bridge motion are in
Journal of ELECTRICAL ENGINEERING 56, NO. 3–4, 2005 83

Figs. 16 (switch off control) and 17 with robust control. The next trend in upgrading the experimental crane
We observe again damping of burden deviation x6M (m) control will be made by a weight and deviation observer,
in the direction of the bridge to the zero value. which eliminates sensor measuring of parameters.
2. Bridge motion: mG = 50 kg; l = 2.5 m (Figs. 16, 17).
3. Simultaneous crane crab and bridge motion: mG = Acknowledgements
50 kg; l = 2.5 and 0.5 m (Figs. 18, 19). Project Crane is a result of cooperation of the follow-
Control system simultaneously adjusts burden deviation ing departments: Department of Construction, Transport
in the direction of the crane crab and also of the bridge. and Logistic, Department of Electrical Drives and Mecha-
The reference position of both components is w5 = tronics and, last but not least, Department of Cybernetics
w5M = 1.5 m and backwards. and Artificial Intelligence at the Technical University in
Figure 18 shows the burden deviation x6 , x6M (m) Košice. The main part belongs to Ing. Juraj Ritók, PhD.
with robust control. In the final position of motion the who permitted realization of research at the laboratory
burden will be damped with the allowed tolerance 0.5 cm. crane.
The worst situation of control is met at the minimum rope
length (Fig. 19) where correct designed controllers were
employed to damp burden swinging. References

[1] ACKERMANN, J. : Parameter Space Design of Robust Control


7 CONCLUSIONS Systems, IEEE Trans. on Automatic Control, 1980.
[2] ACKERMANN, J. : Robuste Regelung, Springer Verlag, Berlin,
The aim of the control design of the experimental crane 1993.
was to ensure crane robustness against the burden weight [3] LEONHARD, W. : Control of Electrical Drives Springer Verlag,
and rope length variations. The designed drive control of Berlin, 1985.
the crane crab, bridge and uplift warrants sharp burden [4] HIČÁR, M. : Written Work to the Dissertation Exam, Robust
Control of Crane Crab, Košice, 2001. (in Slovak)
positioning and avoids burden swinging in the final po-
[5] BALARA, L. : Written Work to the Dissertation Exam, Robust
sition. These conditions improve precision and transport
Control of Electrical Drives, Košice, 2001. (in Slovak)
speed of the load. We have verified a correct selection of
[6] FABIAN, M. : Diploma Work, Bridge Crane Control, Visual-
robust controllers by poles movement at changing vari- ization of its Tasks and Arrangement to the DSR KKUI, Košice,
able parameters, which demonstrated stability and sys- 2004. (in Slovak)
tem damping. The distributed control system gives trans- [7] RITÓK, J.—BIGOŠ, P.—DZURŇÁK, P. : Conditions of Put-
parent information on the control of separate compo- ting Automated Crane to the Distributed Control System, In:
nents. The technological level is intended for elaboration Uplifting Equipments in Theory and Practice, Brno, 1999. (in
of process data and full information about the system Slovak)
state. The second level is used for monitoring and con- [8] RITÓK, J.—BIGOŠ, P. : Positioning of Automated Cranes, In.:
trolling the lower level. Project crane is integrated into Logistika 2000, Košice, 2000, pp. 113–116. (in Slovak)
the information level by computer nets, which allows to Received 13 September 2004
control the object from a further station. Measuring of
necessary parameters is realized mostly by incremental
sensors and for measuring of swinging a rotary rheostat Marek Hičár (Ing) was born in Bardejov, Slovakia in
1977. In 2000 has finished engineering education at the Fac-
was effectively used.
ulty of Electrical Engineering and Informatics, Technical Uni-
Simulation results obtained suggest profitable using versity in Košice. Today he is an external doctoral student
of the robust crane control with guaranteed forbidden at the Department of Electrical Drives and Mechatronics of
swinging. Stability and correct damping were kept at dif- the Technical University in Košice. His main work is robust
ferent burden weight and rope length. Simultaneous crab control of crane drives and its application to real cranes.
and bridge motion has increased the effectiveness of trans- Juraj Ritók (PhD) is with the Department of Construc-
port, where individual control systems controlled burden tions, Transport and Logistics of the Technical University of
swinging in specific directions. Košice. Biography not supplied.

You might also like