Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Research ArticleRobust Tension Control of Strip for 5-Stand Tandem Cold Mills
Behrooz Shafiei1 Mohsen Ekramian2 and Khoshnam Shojaei3
1 Department of Automation Engineering Tandem Cold Rolling Mill Mobarakeh Steel Company Isfahan Iran2Department of Electrical Engineering Faculty of Engineering University of Isfahan Isfahan Iran3Department of Electrical Engineering Najafabad Branch Islamic Azad University Isfahan Iran
Correspondence should be addressed to Behrooz Shafiei bh shmoyahoocom
Received 3 July 2014 Accepted 8 October 2014 Published 5 November 2014
Academic Editor Xiao He
Copyright copy 2014 Behrooz Shafiei et alThis is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited
Tandem cold rolling process is a nonlinear complex system with external and internal uncertainties and significant disturbancesThe improvement in the quality of the final output depends on the control strategy of centerline thickness and interstand tensionThis paper focuses on interstand tension control problem in 5-stand tandem cold rolling mills Tension dynamics can be describedby a nominalmodel perturbed by parametric uncertainties In order to overcome themodel uncertainties and external disturbancessuboptimal 119867
infinand 120583 controllers are proposed and the Hankel-norm approximation is used to reduce the order of 120583 controller
The performance of the proposed controllers is demonstrated by some simulations
1 Introduction
The production of cold rolled coils typically consists ofseparate processes such as annealing pickling cold rollingand skin pass rolling The cold rolling of strip is one of themost important processes in the areas of manufacturing andprocessing of metals Tandem cold mill (TCM) is one typeof cold rolling processes that strip thickness reduction isachieved with a number of nonreversing standsThe produc-tion of this process can be applied for various industries suchas food packaging automobile manufacturing and homeappliance [1 2] Thus the improvement and developmentof new technologies of rolling have become important forresearchers engineers and manufacturers
Tension control is one of the most important processparameters in tandem cold mill (TCM) Changing roll veloc-ity stabilizes the interstand tension [3] Interstand tensionshould be maintained at acceptable levels Large transientvariations in tension may result in a dramatic and physicallydamaging failure of the mill through looping where the striptension falls to zero or tearing of the strip [4 5]
Various advanced methods have been developed to ana-lyze and control rolling mills [6ndash11] Among the previousinvestigations some notes focus on tension control problemin cold rollingmills [12 13]There are numerous uncertainties
and disturbances in the model of tandem cold rolling process[14] Thus the control strategy for cold rolling should berobust with respect to the uncertainties and disturbancesLinear quadratic control strategy (119867
2) guarantees the opti-
mality but it has weakness in the presence of uncertain-ties The suboptimal 119867
infinmethod and 120583 synthesisanalysis
method have been adopted as the robust techniques in manyapproaches and provide systematic design procedures ofrobust controllers for linear system [3 15 16]
To promote tension control of cold rolling mills twoadvanced robust control systems are proposed First a state-space description of full stands with parametric uncertaintyis derived for tension system with 5-stand tandem millsTaking into account the variation of strip thickness Youngrsquosmodulus forward slip and speed actuator time constant themodel of structured uncertainties is presented In the nextstep two robust controllers are designed by using suboptimal119867infin
controller method and 120583 synthesis The order of 120583controller is too large and this makes it necessary to applycontroller order reduction Simulation results show that the120583 robust controller and reduced-order controller have betterrobustness and disturbance attenuation performance than119867infin
controllerThis paper is organized as follows In Section 2 interstand
tension and relations between its parameters are introduced
Hindawi Publishing CorporationJournal of EngineeringVolume 2014 Article ID 409014 13 pageshttpdxdoiorg1011552014409014
2 Journal of Engineering
Stand i Stand i + 1
StripTensiometer
Tension reference
Tensioncontroller
Speedcontroller
Backup rollWork roll(driven)
ΔV
Figure 1 Schematic of conventional tension control for TCM
Stand 1 Stand 2 Stand 3 Stand 4 Stand 5
Tension controllerInterstandtensions
Tension references
Strip
Figure 2 The structure of the advanced tension control system
Then a state-space description is derived for tension systemand it is shown that the perturbation can be representedby parametric uncertainty In Section 3 suboptimal 119867
infin
control and 120583 synthesis method are applied to tension controlproblem and then the order of 120583 controller is reduced TheSection 4 is devoted to comparing robust stability robust per-formance and time domain responses of robust controllersSome simulation results are also given to show the proposedmethods in the presence of model uncertainties and externaldisturbances Finally the concluding remarks are presentedin Section 5
2 Interstand Strip Tension Overview
In tandem cold mill (TCM) strip passes through individualpairs of work rolls Each work roll is supported by a backuproll of a large diameter [2 14] In order to achieve goodtension control tensiometer is the reliable instrument tomeasure the tension between stands Figure 1 shows theconventional tension control by speed where subscript 119894
represents themill stand number for 119894th standThe differencebetween the strip speed at the input to stand 119894+1 and the stripspeed at the output of stand 119894 implies tension change
The structure of conventional controllers is usually basedon single-input-single-output (SISO) control loops whichlimits its capability for improvement To overcome thisproblem and improvement in the performance of tensioncontrol conventional methods are broken and two advancedtension control strategies based on robust strip tension areproposed Figure 2 shows the structure of MIMO tensioncontrol system First stand is chosen as a ldquopivot standrdquo andthe speed of this stand is not trimmed
In this section a state-space description of the nominalinterstand tension perturbed by parametric uncertainty ispresented to be adopted in controller design
21 Mathematical Model To obtain the state-space and out-put equations for interstand tension the theoretical systemequations are introduced here [14] Abbreviations sectiondescribes the symbols that appear in these equations
Journal of Engineering 3
1 minus 120593i+1
1 + fi
E Hi+1 BL
d
+ +
+minus
ΔVi+1
ΔVi
1S ΔTii+1
Figure 3 Block diagram of tension system between adjacent stands
120575Hi+1
MHi+1
Figure 4 Representation of uncertain parameter (119867119894+1) as LFTs
(1) Interstand tension is developed by applying Hookersquoslaw to a length of strip between successive mill stands and isgoverned by the following dynamic
119889Δ (119879119894119894+1
)
119889119905equiv Δ119894119894+1
=EBH119894+1(119881in119894+1 minus 119881out119894)
119871 (1)
(2) During rolling the work roll is elastically flattenedAt the entry to work roll the strip speed is less than theperipheral speed of the work roll and at the exit of the workroll speed is greater than the peripheral speed of the workroll The forward slip ratio and the backward slip ratio aredetermined by the following
(a) Forward slip ratio
119891 =119881out minus 119881
119881 (2)
(b) Backward slip ratio
120593 =119881 minus 119881in119881
(3)
(3) The relationship among the entry and exit thicknessand strip velocity at the entry and exit to stand is called vol-ume continuity and is determined by the following equation
ℎin times 119881in = ℎout times 119881out (4)(4)Thework roll drive equation is modeled (closed-loop)
and is approximated in order one based on typical data
119894=119880V119894
120591V119894minus119881119894
120591V119894 (5)
Equations (1) to (3) are combined to yield the followingequation
Δ119894119894+1
=EBH119894+1
119871((1 minus 120593
119894+1) Δ119881119894+1minus (1 + 119891
119894) Δ119881119894) (6)
According to (6) the tension system between two stands isdepicted in Figure 3
22 Uncertainties in Modeling Based on (5) and (6) vari-ations of four coefficients 119867
119894+1 119864 120593
119894 and 120591V119894 are con-
sidered as parametric uncertainty Parametric uncertaintyis represented by variations of certain system parameterswhich may adversely affect the stability and performance ofa control system In Figure 3 the four constant blocks can bereplaced by upper linear fractional transformation (LFT) Forinstance119867
119894+1may be represented as upper LFT in 119875
119867119894+1and
120575119867119894+1
(minus1 lt 120575119867119894+1
lt 1) and further 119867119894+1
is the so-callednominal value of 119867
119894+1 The LFT framework for 119867
119894+1depicts
in Figure 4
119867119894+1= 119867119894+1(1 + 119875
119867119894+1120575119867119894+1
) = 119865119906(119872119867119894+1
120575119867119894+1
)
119872119867119894+1
= [0 119867
119894+1
119875119867119894+1
119867119894+1
]
(7)
Similarly the other parameters can be represented as upperLFT in 120575
120578119894+1 120575120591V119894 and 120575
119864 It is supposed that 120578 = 1 minus 120593 for
simplification The parameters of full stand tension systemand their tolerances are given in Table 1 [13]
4 Journal of Engineering
Table 1 Nominal values and their tolerance
Parameter Nominal value Unit Uncertainty119864119899
21 times 105 Nmm2
plusmn100119879V119899 01 s plusmn2001198672
135 mm plusmn501198673
077 mm plusmn501198674
05 mm plusmn501198675
033 mm plusmn501205932
03 mdash plusmn1001205933
025 mdash plusmn1001205934
02 mdash plusmn1001205935
02 mdash plusmn100119871 45 m 0119861 1 m 0
23 State-Space Representation Based on (4)ndash(6) the state-space description of tension system with 5-stand yields
1198831015840
= 119860119883 + 1198611119880119906+ 1198612119880
119884119906= 1198621119883 + 119863
11119880119906+ 11986312119880
119884 = 1198622119883 + 119863
21119880119906+ 11986322119880
(8)
The state vector 119883 the control inputs 119880 and 119880119906 and the
output vectors 119884 and 119884119906are defined as
119883 = [Δ1198812 Δ12059012 Δ1198813 Δ12059023 Δ1198814 Δ12059034 Δ1198815 Δ12059045]T
119880 = [Δ119880V2 Δ119880V3 Δ119880V4 Δ119880V5]T
119880119906= [119880120591V2 1198801205782 1198801198642 1198801198672 119880120591V3 1198801205783 1198801198643 1198801198673
119880120591V4 1198801205784 1198801198644 1198801198674 119880120591V5 1198801205785 1198801198645 1198801198675]T
119884119906= [119910120591V2 1199101205782 1199101198642 1199101198672 119910120591V3 1199101205783 1199101198643 1199101198673
119910120591V4 1199101205784 1199101198644 1199101198674 119910120591V5 1199101205785 1199101198645 1199101198675]T
119884 = [Δ11987912 Δ11987923 Δ11987934 Δ11987945]T
(9)
119866Ten denotes the inputoutput dynamics of the tension sys-tem which takes into account the uncertainty of parametersThe state-space representation of119866Ten is (the systemmatricesare presented in Appendix A)
119866Ten =[
[
119860 1198611
1198612
11986211198631111986312
11986221198632111986322
]
]
(10)
The uncertain behavior of the original system can bedescribed by an upper LFT representation
119884 = 119865119880(119866Ten Δ)119880 (11)
With diagonal uncertainty matrix Δ = diag(120575V2 1205751205782 12057511986421205751198672 120575V5 1205751205785 1205751198645 1205751198675) as shown in Figure 5 It should
be noted that the unknown matrix Δ has a fixed structureand in general is a block diagonal Such uncertainty is calledstructured uncertainty
Δ
GTen
U Y
Figure 5 LFT representation of the tension system with uncertain-ties
GTension
Δ
d
Tensioncontroller Wp ep
minusWu
euG
Ref
Figure 6 The block diagram of closed system with performancerequirements
3 Robust Tension Controller Design
This section considers the design of tension control system for5-stand tandem mill For this system open loop and closed-loop system interconnections are introduced and robustcontrollers are designed
31 Requirements to Fulfill Design The aim of robust ten-sion controller design is to achieve and maintain the ten-sion between stands in the presence of disturbances anduncertainties The block diagram of the closed-loop systemincluding the feedback and controller as well as the elementsreflecting the model uncertainty and weighting functionsrelated to performance requirements is shown in Figure 6
In Figure 6 the rectangle with dashed line represents thetransfer matrix 119866 Inside the rectangle is the nominal modelof tension system and the uncertainty matrix Δ that includesthe model uncertainties The variables 119889
1 1198892 1198893 and 119889
4are
the disturbances on the system at the system outputs Theoutput weighting performance vector (119890
119901) control weighting
performance (119890119906) and disturbances vector (119889) are defined as
119890119901= [1198901199011 1198901199012 1198901199013 1198901199014]T
119890119906= [1198901199061 1198901199062 1198901199063 1198901199064]T
119889 = [1198891 1198892 1198893 1198894]T
(12)
In general weighting functions would be used to meet thedesign specifications Finding appropriate weighting func-tions is a crucial step in robust controller design and usuallyneeds a few trials The performance weighting function is
Journal of Engineering 5
103
102
101
100
10minus1
10minus2
10minus3
wminus1p1
Inverse performance weighting function
Frequency (rads)
Mag
nitu
de
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Figure 7 Singular values of 1119908119901
101
100
10minus1
10minus5 100 105 1010
wu1
Control weighting function
Frequency (rads)
Mag
nitu
de
Figure 8 Singular values of 119908119906
usually low-pass filter and control weighting function is high-pass filter Design experience will help in choosing weightingfunctions In this design the performanceweighting function119908119901and the control weighting functions 119908
119906119882119901 and119882
119906are
chosen as [13]
119908119901= 0012
1198782
+ 28119878 + 100
051198782 + 40119878 + 001
119908119906= 5
119878 + 10000
119878 + 500000
119882119901= (
1199081199011
0 0 0
0 1199081199012
0 0
0 0 1199081199013
0
0 0 0 1199081199014
)
119882119906= (
1199081199061
0 0 0
0 1199081199062
0 0
0 0 1199081199063
0
0 0 0 1199081199064
)
(13)The singular values of 1119908
119901over the frequency range
[10minus4
104
] are shown in Figure 7 This weighting functionshows that for low frequencies the closed-loop system (thenominal as well as perturbed system) attenuates the outputdisturbance in the ratio of 100 to 001 In other words theeffect of a unit disturbance on the steady-state output shouldbe of the order 10minus2 This performance requirement becomesless stringent with increasing frequency
Also the singular values of 119908119906over the frequency range
[10minus4 104] are shown in Figure 8
32 Open Loop and Closed-Loop System InterconnectionBefore designing the tension controller it is necessary to build
6 Journal of Engineering
Pertin1-16 Pertout1-16
sum
+
+
minus
sum
G
yc1-4
Wp
minusWu
ep1-4
eu1-4
ΔT1-4
Dist1-4
Ref = 0
Figure 9 Structure of the open loop system
Pertin1-16 Pertout1-16
Control1-4
ep1-4
eu1-4
ΔT1-4
Dist1-4 tension ic
Figure 10 Block diagram of open loop system
Pertin1-16 Pertout1-16
T1-4
U1-4
yc1-4
G
K
+
+
+minus
sum
sum
Dist1-4
Ref1-4
Figure 11 Structure of the closed-loop system
interconnection for open-loop and closed-loop system Thestructure of the open-loop system is represented in Figure 9In Figure 10 tension ic denotes the transfer function matrixof open loop system with 24 inputs and 28 outputs
The simulation of closed-loop system with designedcontrollers is based on the structure shown in Figure 11 Theweighting functions 119908
119901and 119908
119906are not included in the block
diagram for obvious reasons
33 119867infin
Controller The interconnection of nominal openloop system is denoted by hin ic as shown in Figure 12
The suboptimal 119867infin
controller minimizes the infinite-normof 119865119897(ℎ119894119899 119894119888 119896) overall stabilizing controller 119870 Note that
119865119897(ℎ119894119899 119894119888 119896) is the transfer function matrix of the nominal
closed-loop system from the disturbances to the errorsTo compute 119867
infinoptimal controller for LTI plant some
conditions are satisfied by plant Sometimes one of theseconditions is not satisfied and the algorithm for computingdoes not work [15 17] When this situation accrued otherrobust control methods are used although some tricks canbe used to satisfy conditions For example adding or sub-tracting epsilon may solve the problem in these conditions
Journal of Engineering 7
K
eDist hin ic
Figure 12 Closed-loop LFTs in119867infindesign
10minus1 100 101 102 103 10401
02
03
04
05
06
Frequency (rads)10minus1 100 101 102 103 104
Frequency (rads)
010
020304050607
Closed-loop complex 120583 controller number 3
Com
plex120583
Closed-loop mixed 120583 controller number 3
Mix
ed120583
Figure 13 120583 plot after third iteration
100
10minus1
10minus2
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Robust stability
Frequency (rads)
120583-controllerReduced-order controller
Upp
er b
ound
of120583
Hinfin controller
Figure 14 Comparison of robust stability for three controllers
For this plant 119867infin
controller is numerically computed andfinally the obtained controller is of 20th orderTheminimumvalue of 120574 is 0101
34 120583 Controller The 120583 synthesis is introduced by twoiterative methods 119863-119870 and 120583-119870 iteration For this plant120583 synthesis is executed by the M-file dkit from the RobustControl Toolbox of Matlab which automates the procedureby using 119863-119870 iterations [15] After the third iteration
the maximum value of 120583 is 0567 which indicates that therobust performance has been achieved The final controllerobtained is of 104th order The 120583 plot of the closed-loopsystem is shown in Figure 13
The order of this controller is 104 which makes it difficultin implementation On the other hand high order controllerswill lead to high cost difficult commissioning poor relia-bility and potential problems in maintenance This is theweakness of 120583 controller in this design Usually the order
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
2 Journal of Engineering
Stand i Stand i + 1
StripTensiometer
Tension reference
Tensioncontroller
Speedcontroller
Backup rollWork roll(driven)
ΔV
Figure 1 Schematic of conventional tension control for TCM
Stand 1 Stand 2 Stand 3 Stand 4 Stand 5
Tension controllerInterstandtensions
Tension references
Strip
Figure 2 The structure of the advanced tension control system
Then a state-space description is derived for tension systemand it is shown that the perturbation can be representedby parametric uncertainty In Section 3 suboptimal 119867
infin
control and 120583 synthesis method are applied to tension controlproblem and then the order of 120583 controller is reduced TheSection 4 is devoted to comparing robust stability robust per-formance and time domain responses of robust controllersSome simulation results are also given to show the proposedmethods in the presence of model uncertainties and externaldisturbances Finally the concluding remarks are presentedin Section 5
2 Interstand Strip Tension Overview
In tandem cold mill (TCM) strip passes through individualpairs of work rolls Each work roll is supported by a backuproll of a large diameter [2 14] In order to achieve goodtension control tensiometer is the reliable instrument tomeasure the tension between stands Figure 1 shows theconventional tension control by speed where subscript 119894
represents themill stand number for 119894th standThe differencebetween the strip speed at the input to stand 119894+1 and the stripspeed at the output of stand 119894 implies tension change
The structure of conventional controllers is usually basedon single-input-single-output (SISO) control loops whichlimits its capability for improvement To overcome thisproblem and improvement in the performance of tensioncontrol conventional methods are broken and two advancedtension control strategies based on robust strip tension areproposed Figure 2 shows the structure of MIMO tensioncontrol system First stand is chosen as a ldquopivot standrdquo andthe speed of this stand is not trimmed
In this section a state-space description of the nominalinterstand tension perturbed by parametric uncertainty ispresented to be adopted in controller design
21 Mathematical Model To obtain the state-space and out-put equations for interstand tension the theoretical systemequations are introduced here [14] Abbreviations sectiondescribes the symbols that appear in these equations
Journal of Engineering 3
1 minus 120593i+1
1 + fi
E Hi+1 BL
d
+ +
+minus
ΔVi+1
ΔVi
1S ΔTii+1
Figure 3 Block diagram of tension system between adjacent stands
120575Hi+1
MHi+1
Figure 4 Representation of uncertain parameter (119867119894+1) as LFTs
(1) Interstand tension is developed by applying Hookersquoslaw to a length of strip between successive mill stands and isgoverned by the following dynamic
119889Δ (119879119894119894+1
)
119889119905equiv Δ119894119894+1
=EBH119894+1(119881in119894+1 minus 119881out119894)
119871 (1)
(2) During rolling the work roll is elastically flattenedAt the entry to work roll the strip speed is less than theperipheral speed of the work roll and at the exit of the workroll speed is greater than the peripheral speed of the workroll The forward slip ratio and the backward slip ratio aredetermined by the following
(a) Forward slip ratio
119891 =119881out minus 119881
119881 (2)
(b) Backward slip ratio
120593 =119881 minus 119881in119881
(3)
(3) The relationship among the entry and exit thicknessand strip velocity at the entry and exit to stand is called vol-ume continuity and is determined by the following equation
ℎin times 119881in = ℎout times 119881out (4)(4)Thework roll drive equation is modeled (closed-loop)
and is approximated in order one based on typical data
119894=119880V119894
120591V119894minus119881119894
120591V119894 (5)
Equations (1) to (3) are combined to yield the followingequation
Δ119894119894+1
=EBH119894+1
119871((1 minus 120593
119894+1) Δ119881119894+1minus (1 + 119891
119894) Δ119881119894) (6)
According to (6) the tension system between two stands isdepicted in Figure 3
22 Uncertainties in Modeling Based on (5) and (6) vari-ations of four coefficients 119867
119894+1 119864 120593
119894 and 120591V119894 are con-
sidered as parametric uncertainty Parametric uncertaintyis represented by variations of certain system parameterswhich may adversely affect the stability and performance ofa control system In Figure 3 the four constant blocks can bereplaced by upper linear fractional transformation (LFT) Forinstance119867
119894+1may be represented as upper LFT in 119875
119867119894+1and
120575119867119894+1
(minus1 lt 120575119867119894+1
lt 1) and further 119867119894+1
is the so-callednominal value of 119867
119894+1 The LFT framework for 119867
119894+1depicts
in Figure 4
119867119894+1= 119867119894+1(1 + 119875
119867119894+1120575119867119894+1
) = 119865119906(119872119867119894+1
120575119867119894+1
)
119872119867119894+1
= [0 119867
119894+1
119875119867119894+1
119867119894+1
]
(7)
Similarly the other parameters can be represented as upperLFT in 120575
120578119894+1 120575120591V119894 and 120575
119864 It is supposed that 120578 = 1 minus 120593 for
simplification The parameters of full stand tension systemand their tolerances are given in Table 1 [13]
4 Journal of Engineering
Table 1 Nominal values and their tolerance
Parameter Nominal value Unit Uncertainty119864119899
21 times 105 Nmm2
plusmn100119879V119899 01 s plusmn2001198672
135 mm plusmn501198673
077 mm plusmn501198674
05 mm plusmn501198675
033 mm plusmn501205932
03 mdash plusmn1001205933
025 mdash plusmn1001205934
02 mdash plusmn1001205935
02 mdash plusmn100119871 45 m 0119861 1 m 0
23 State-Space Representation Based on (4)ndash(6) the state-space description of tension system with 5-stand yields
1198831015840
= 119860119883 + 1198611119880119906+ 1198612119880
119884119906= 1198621119883 + 119863
11119880119906+ 11986312119880
119884 = 1198622119883 + 119863
21119880119906+ 11986322119880
(8)
The state vector 119883 the control inputs 119880 and 119880119906 and the
output vectors 119884 and 119884119906are defined as
119883 = [Δ1198812 Δ12059012 Δ1198813 Δ12059023 Δ1198814 Δ12059034 Δ1198815 Δ12059045]T
119880 = [Δ119880V2 Δ119880V3 Δ119880V4 Δ119880V5]T
119880119906= [119880120591V2 1198801205782 1198801198642 1198801198672 119880120591V3 1198801205783 1198801198643 1198801198673
119880120591V4 1198801205784 1198801198644 1198801198674 119880120591V5 1198801205785 1198801198645 1198801198675]T
119884119906= [119910120591V2 1199101205782 1199101198642 1199101198672 119910120591V3 1199101205783 1199101198643 1199101198673
119910120591V4 1199101205784 1199101198644 1199101198674 119910120591V5 1199101205785 1199101198645 1199101198675]T
119884 = [Δ11987912 Δ11987923 Δ11987934 Δ11987945]T
(9)
119866Ten denotes the inputoutput dynamics of the tension sys-tem which takes into account the uncertainty of parametersThe state-space representation of119866Ten is (the systemmatricesare presented in Appendix A)
119866Ten =[
[
119860 1198611
1198612
11986211198631111986312
11986221198632111986322
]
]
(10)
The uncertain behavior of the original system can bedescribed by an upper LFT representation
119884 = 119865119880(119866Ten Δ)119880 (11)
With diagonal uncertainty matrix Δ = diag(120575V2 1205751205782 12057511986421205751198672 120575V5 1205751205785 1205751198645 1205751198675) as shown in Figure 5 It should
be noted that the unknown matrix Δ has a fixed structureand in general is a block diagonal Such uncertainty is calledstructured uncertainty
Δ
GTen
U Y
Figure 5 LFT representation of the tension system with uncertain-ties
GTension
Δ
d
Tensioncontroller Wp ep
minusWu
euG
Ref
Figure 6 The block diagram of closed system with performancerequirements
3 Robust Tension Controller Design
This section considers the design of tension control system for5-stand tandem mill For this system open loop and closed-loop system interconnections are introduced and robustcontrollers are designed
31 Requirements to Fulfill Design The aim of robust ten-sion controller design is to achieve and maintain the ten-sion between stands in the presence of disturbances anduncertainties The block diagram of the closed-loop systemincluding the feedback and controller as well as the elementsreflecting the model uncertainty and weighting functionsrelated to performance requirements is shown in Figure 6
In Figure 6 the rectangle with dashed line represents thetransfer matrix 119866 Inside the rectangle is the nominal modelof tension system and the uncertainty matrix Δ that includesthe model uncertainties The variables 119889
1 1198892 1198893 and 119889
4are
the disturbances on the system at the system outputs Theoutput weighting performance vector (119890
119901) control weighting
performance (119890119906) and disturbances vector (119889) are defined as
119890119901= [1198901199011 1198901199012 1198901199013 1198901199014]T
119890119906= [1198901199061 1198901199062 1198901199063 1198901199064]T
119889 = [1198891 1198892 1198893 1198894]T
(12)
In general weighting functions would be used to meet thedesign specifications Finding appropriate weighting func-tions is a crucial step in robust controller design and usuallyneeds a few trials The performance weighting function is
Journal of Engineering 5
103
102
101
100
10minus1
10minus2
10minus3
wminus1p1
Inverse performance weighting function
Frequency (rads)
Mag
nitu
de
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Figure 7 Singular values of 1119908119901
101
100
10minus1
10minus5 100 105 1010
wu1
Control weighting function
Frequency (rads)
Mag
nitu
de
Figure 8 Singular values of 119908119906
usually low-pass filter and control weighting function is high-pass filter Design experience will help in choosing weightingfunctions In this design the performanceweighting function119908119901and the control weighting functions 119908
119906119882119901 and119882
119906are
chosen as [13]
119908119901= 0012
1198782
+ 28119878 + 100
051198782 + 40119878 + 001
119908119906= 5
119878 + 10000
119878 + 500000
119882119901= (
1199081199011
0 0 0
0 1199081199012
0 0
0 0 1199081199013
0
0 0 0 1199081199014
)
119882119906= (
1199081199061
0 0 0
0 1199081199062
0 0
0 0 1199081199063
0
0 0 0 1199081199064
)
(13)The singular values of 1119908
119901over the frequency range
[10minus4
104
] are shown in Figure 7 This weighting functionshows that for low frequencies the closed-loop system (thenominal as well as perturbed system) attenuates the outputdisturbance in the ratio of 100 to 001 In other words theeffect of a unit disturbance on the steady-state output shouldbe of the order 10minus2 This performance requirement becomesless stringent with increasing frequency
Also the singular values of 119908119906over the frequency range
[10minus4 104] are shown in Figure 8
32 Open Loop and Closed-Loop System InterconnectionBefore designing the tension controller it is necessary to build
6 Journal of Engineering
Pertin1-16 Pertout1-16
sum
+
+
minus
sum
G
yc1-4
Wp
minusWu
ep1-4
eu1-4
ΔT1-4
Dist1-4
Ref = 0
Figure 9 Structure of the open loop system
Pertin1-16 Pertout1-16
Control1-4
ep1-4
eu1-4
ΔT1-4
Dist1-4 tension ic
Figure 10 Block diagram of open loop system
Pertin1-16 Pertout1-16
T1-4
U1-4
yc1-4
G
K
+
+
+minus
sum
sum
Dist1-4
Ref1-4
Figure 11 Structure of the closed-loop system
interconnection for open-loop and closed-loop system Thestructure of the open-loop system is represented in Figure 9In Figure 10 tension ic denotes the transfer function matrixof open loop system with 24 inputs and 28 outputs
The simulation of closed-loop system with designedcontrollers is based on the structure shown in Figure 11 Theweighting functions 119908
119901and 119908
119906are not included in the block
diagram for obvious reasons
33 119867infin
Controller The interconnection of nominal openloop system is denoted by hin ic as shown in Figure 12
The suboptimal 119867infin
controller minimizes the infinite-normof 119865119897(ℎ119894119899 119894119888 119896) overall stabilizing controller 119870 Note that
119865119897(ℎ119894119899 119894119888 119896) is the transfer function matrix of the nominal
closed-loop system from the disturbances to the errorsTo compute 119867
infinoptimal controller for LTI plant some
conditions are satisfied by plant Sometimes one of theseconditions is not satisfied and the algorithm for computingdoes not work [15 17] When this situation accrued otherrobust control methods are used although some tricks canbe used to satisfy conditions For example adding or sub-tracting epsilon may solve the problem in these conditions
Journal of Engineering 7
K
eDist hin ic
Figure 12 Closed-loop LFTs in119867infindesign
10minus1 100 101 102 103 10401
02
03
04
05
06
Frequency (rads)10minus1 100 101 102 103 104
Frequency (rads)
010
020304050607
Closed-loop complex 120583 controller number 3
Com
plex120583
Closed-loop mixed 120583 controller number 3
Mix
ed120583
Figure 13 120583 plot after third iteration
100
10minus1
10minus2
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Robust stability
Frequency (rads)
120583-controllerReduced-order controller
Upp
er b
ound
of120583
Hinfin controller
Figure 14 Comparison of robust stability for three controllers
For this plant 119867infin
controller is numerically computed andfinally the obtained controller is of 20th orderTheminimumvalue of 120574 is 0101
34 120583 Controller The 120583 synthesis is introduced by twoiterative methods 119863-119870 and 120583-119870 iteration For this plant120583 synthesis is executed by the M-file dkit from the RobustControl Toolbox of Matlab which automates the procedureby using 119863-119870 iterations [15] After the third iteration
the maximum value of 120583 is 0567 which indicates that therobust performance has been achieved The final controllerobtained is of 104th order The 120583 plot of the closed-loopsystem is shown in Figure 13
The order of this controller is 104 which makes it difficultin implementation On the other hand high order controllerswill lead to high cost difficult commissioning poor relia-bility and potential problems in maintenance This is theweakness of 120583 controller in this design Usually the order
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 3
1 minus 120593i+1
1 + fi
E Hi+1 BL
d
+ +
+minus
ΔVi+1
ΔVi
1S ΔTii+1
Figure 3 Block diagram of tension system between adjacent stands
120575Hi+1
MHi+1
Figure 4 Representation of uncertain parameter (119867119894+1) as LFTs
(1) Interstand tension is developed by applying Hookersquoslaw to a length of strip between successive mill stands and isgoverned by the following dynamic
119889Δ (119879119894119894+1
)
119889119905equiv Δ119894119894+1
=EBH119894+1(119881in119894+1 minus 119881out119894)
119871 (1)
(2) During rolling the work roll is elastically flattenedAt the entry to work roll the strip speed is less than theperipheral speed of the work roll and at the exit of the workroll speed is greater than the peripheral speed of the workroll The forward slip ratio and the backward slip ratio aredetermined by the following
(a) Forward slip ratio
119891 =119881out minus 119881
119881 (2)
(b) Backward slip ratio
120593 =119881 minus 119881in119881
(3)
(3) The relationship among the entry and exit thicknessand strip velocity at the entry and exit to stand is called vol-ume continuity and is determined by the following equation
ℎin times 119881in = ℎout times 119881out (4)(4)Thework roll drive equation is modeled (closed-loop)
and is approximated in order one based on typical data
119894=119880V119894
120591V119894minus119881119894
120591V119894 (5)
Equations (1) to (3) are combined to yield the followingequation
Δ119894119894+1
=EBH119894+1
119871((1 minus 120593
119894+1) Δ119881119894+1minus (1 + 119891
119894) Δ119881119894) (6)
According to (6) the tension system between two stands isdepicted in Figure 3
22 Uncertainties in Modeling Based on (5) and (6) vari-ations of four coefficients 119867
119894+1 119864 120593
119894 and 120591V119894 are con-
sidered as parametric uncertainty Parametric uncertaintyis represented by variations of certain system parameterswhich may adversely affect the stability and performance ofa control system In Figure 3 the four constant blocks can bereplaced by upper linear fractional transformation (LFT) Forinstance119867
119894+1may be represented as upper LFT in 119875
119867119894+1and
120575119867119894+1
(minus1 lt 120575119867119894+1
lt 1) and further 119867119894+1
is the so-callednominal value of 119867
119894+1 The LFT framework for 119867
119894+1depicts
in Figure 4
119867119894+1= 119867119894+1(1 + 119875
119867119894+1120575119867119894+1
) = 119865119906(119872119867119894+1
120575119867119894+1
)
119872119867119894+1
= [0 119867
119894+1
119875119867119894+1
119867119894+1
]
(7)
Similarly the other parameters can be represented as upperLFT in 120575
120578119894+1 120575120591V119894 and 120575
119864 It is supposed that 120578 = 1 minus 120593 for
simplification The parameters of full stand tension systemand their tolerances are given in Table 1 [13]
4 Journal of Engineering
Table 1 Nominal values and their tolerance
Parameter Nominal value Unit Uncertainty119864119899
21 times 105 Nmm2
plusmn100119879V119899 01 s plusmn2001198672
135 mm plusmn501198673
077 mm plusmn501198674
05 mm plusmn501198675
033 mm plusmn501205932
03 mdash plusmn1001205933
025 mdash plusmn1001205934
02 mdash plusmn1001205935
02 mdash plusmn100119871 45 m 0119861 1 m 0
23 State-Space Representation Based on (4)ndash(6) the state-space description of tension system with 5-stand yields
1198831015840
= 119860119883 + 1198611119880119906+ 1198612119880
119884119906= 1198621119883 + 119863
11119880119906+ 11986312119880
119884 = 1198622119883 + 119863
21119880119906+ 11986322119880
(8)
The state vector 119883 the control inputs 119880 and 119880119906 and the
output vectors 119884 and 119884119906are defined as
119883 = [Δ1198812 Δ12059012 Δ1198813 Δ12059023 Δ1198814 Δ12059034 Δ1198815 Δ12059045]T
119880 = [Δ119880V2 Δ119880V3 Δ119880V4 Δ119880V5]T
119880119906= [119880120591V2 1198801205782 1198801198642 1198801198672 119880120591V3 1198801205783 1198801198643 1198801198673
119880120591V4 1198801205784 1198801198644 1198801198674 119880120591V5 1198801205785 1198801198645 1198801198675]T
119884119906= [119910120591V2 1199101205782 1199101198642 1199101198672 119910120591V3 1199101205783 1199101198643 1199101198673
119910120591V4 1199101205784 1199101198644 1199101198674 119910120591V5 1199101205785 1199101198645 1199101198675]T
119884 = [Δ11987912 Δ11987923 Δ11987934 Δ11987945]T
(9)
119866Ten denotes the inputoutput dynamics of the tension sys-tem which takes into account the uncertainty of parametersThe state-space representation of119866Ten is (the systemmatricesare presented in Appendix A)
119866Ten =[
[
119860 1198611
1198612
11986211198631111986312
11986221198632111986322
]
]
(10)
The uncertain behavior of the original system can bedescribed by an upper LFT representation
119884 = 119865119880(119866Ten Δ)119880 (11)
With diagonal uncertainty matrix Δ = diag(120575V2 1205751205782 12057511986421205751198672 120575V5 1205751205785 1205751198645 1205751198675) as shown in Figure 5 It should
be noted that the unknown matrix Δ has a fixed structureand in general is a block diagonal Such uncertainty is calledstructured uncertainty
Δ
GTen
U Y
Figure 5 LFT representation of the tension system with uncertain-ties
GTension
Δ
d
Tensioncontroller Wp ep
minusWu
euG
Ref
Figure 6 The block diagram of closed system with performancerequirements
3 Robust Tension Controller Design
This section considers the design of tension control system for5-stand tandem mill For this system open loop and closed-loop system interconnections are introduced and robustcontrollers are designed
31 Requirements to Fulfill Design The aim of robust ten-sion controller design is to achieve and maintain the ten-sion between stands in the presence of disturbances anduncertainties The block diagram of the closed-loop systemincluding the feedback and controller as well as the elementsreflecting the model uncertainty and weighting functionsrelated to performance requirements is shown in Figure 6
In Figure 6 the rectangle with dashed line represents thetransfer matrix 119866 Inside the rectangle is the nominal modelof tension system and the uncertainty matrix Δ that includesthe model uncertainties The variables 119889
1 1198892 1198893 and 119889
4are
the disturbances on the system at the system outputs Theoutput weighting performance vector (119890
119901) control weighting
performance (119890119906) and disturbances vector (119889) are defined as
119890119901= [1198901199011 1198901199012 1198901199013 1198901199014]T
119890119906= [1198901199061 1198901199062 1198901199063 1198901199064]T
119889 = [1198891 1198892 1198893 1198894]T
(12)
In general weighting functions would be used to meet thedesign specifications Finding appropriate weighting func-tions is a crucial step in robust controller design and usuallyneeds a few trials The performance weighting function is
Journal of Engineering 5
103
102
101
100
10minus1
10minus2
10minus3
wminus1p1
Inverse performance weighting function
Frequency (rads)
Mag
nitu
de
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Figure 7 Singular values of 1119908119901
101
100
10minus1
10minus5 100 105 1010
wu1
Control weighting function
Frequency (rads)
Mag
nitu
de
Figure 8 Singular values of 119908119906
usually low-pass filter and control weighting function is high-pass filter Design experience will help in choosing weightingfunctions In this design the performanceweighting function119908119901and the control weighting functions 119908
119906119882119901 and119882
119906are
chosen as [13]
119908119901= 0012
1198782
+ 28119878 + 100
051198782 + 40119878 + 001
119908119906= 5
119878 + 10000
119878 + 500000
119882119901= (
1199081199011
0 0 0
0 1199081199012
0 0
0 0 1199081199013
0
0 0 0 1199081199014
)
119882119906= (
1199081199061
0 0 0
0 1199081199062
0 0
0 0 1199081199063
0
0 0 0 1199081199064
)
(13)The singular values of 1119908
119901over the frequency range
[10minus4
104
] are shown in Figure 7 This weighting functionshows that for low frequencies the closed-loop system (thenominal as well as perturbed system) attenuates the outputdisturbance in the ratio of 100 to 001 In other words theeffect of a unit disturbance on the steady-state output shouldbe of the order 10minus2 This performance requirement becomesless stringent with increasing frequency
Also the singular values of 119908119906over the frequency range
[10minus4 104] are shown in Figure 8
32 Open Loop and Closed-Loop System InterconnectionBefore designing the tension controller it is necessary to build
6 Journal of Engineering
Pertin1-16 Pertout1-16
sum
+
+
minus
sum
G
yc1-4
Wp
minusWu
ep1-4
eu1-4
ΔT1-4
Dist1-4
Ref = 0
Figure 9 Structure of the open loop system
Pertin1-16 Pertout1-16
Control1-4
ep1-4
eu1-4
ΔT1-4
Dist1-4 tension ic
Figure 10 Block diagram of open loop system
Pertin1-16 Pertout1-16
T1-4
U1-4
yc1-4
G
K
+
+
+minus
sum
sum
Dist1-4
Ref1-4
Figure 11 Structure of the closed-loop system
interconnection for open-loop and closed-loop system Thestructure of the open-loop system is represented in Figure 9In Figure 10 tension ic denotes the transfer function matrixof open loop system with 24 inputs and 28 outputs
The simulation of closed-loop system with designedcontrollers is based on the structure shown in Figure 11 Theweighting functions 119908
119901and 119908
119906are not included in the block
diagram for obvious reasons
33 119867infin
Controller The interconnection of nominal openloop system is denoted by hin ic as shown in Figure 12
The suboptimal 119867infin
controller minimizes the infinite-normof 119865119897(ℎ119894119899 119894119888 119896) overall stabilizing controller 119870 Note that
119865119897(ℎ119894119899 119894119888 119896) is the transfer function matrix of the nominal
closed-loop system from the disturbances to the errorsTo compute 119867
infinoptimal controller for LTI plant some
conditions are satisfied by plant Sometimes one of theseconditions is not satisfied and the algorithm for computingdoes not work [15 17] When this situation accrued otherrobust control methods are used although some tricks canbe used to satisfy conditions For example adding or sub-tracting epsilon may solve the problem in these conditions
Journal of Engineering 7
K
eDist hin ic
Figure 12 Closed-loop LFTs in119867infindesign
10minus1 100 101 102 103 10401
02
03
04
05
06
Frequency (rads)10minus1 100 101 102 103 104
Frequency (rads)
010
020304050607
Closed-loop complex 120583 controller number 3
Com
plex120583
Closed-loop mixed 120583 controller number 3
Mix
ed120583
Figure 13 120583 plot after third iteration
100
10minus1
10minus2
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Robust stability
Frequency (rads)
120583-controllerReduced-order controller
Upp
er b
ound
of120583
Hinfin controller
Figure 14 Comparison of robust stability for three controllers
For this plant 119867infin
controller is numerically computed andfinally the obtained controller is of 20th orderTheminimumvalue of 120574 is 0101
34 120583 Controller The 120583 synthesis is introduced by twoiterative methods 119863-119870 and 120583-119870 iteration For this plant120583 synthesis is executed by the M-file dkit from the RobustControl Toolbox of Matlab which automates the procedureby using 119863-119870 iterations [15] After the third iteration
the maximum value of 120583 is 0567 which indicates that therobust performance has been achieved The final controllerobtained is of 104th order The 120583 plot of the closed-loopsystem is shown in Figure 13
The order of this controller is 104 which makes it difficultin implementation On the other hand high order controllerswill lead to high cost difficult commissioning poor relia-bility and potential problems in maintenance This is theweakness of 120583 controller in this design Usually the order
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
4 Journal of Engineering
Table 1 Nominal values and their tolerance
Parameter Nominal value Unit Uncertainty119864119899
21 times 105 Nmm2
plusmn100119879V119899 01 s plusmn2001198672
135 mm plusmn501198673
077 mm plusmn501198674
05 mm plusmn501198675
033 mm plusmn501205932
03 mdash plusmn1001205933
025 mdash plusmn1001205934
02 mdash plusmn1001205935
02 mdash plusmn100119871 45 m 0119861 1 m 0
23 State-Space Representation Based on (4)ndash(6) the state-space description of tension system with 5-stand yields
1198831015840
= 119860119883 + 1198611119880119906+ 1198612119880
119884119906= 1198621119883 + 119863
11119880119906+ 11986312119880
119884 = 1198622119883 + 119863
21119880119906+ 11986322119880
(8)
The state vector 119883 the control inputs 119880 and 119880119906 and the
output vectors 119884 and 119884119906are defined as
119883 = [Δ1198812 Δ12059012 Δ1198813 Δ12059023 Δ1198814 Δ12059034 Δ1198815 Δ12059045]T
119880 = [Δ119880V2 Δ119880V3 Δ119880V4 Δ119880V5]T
119880119906= [119880120591V2 1198801205782 1198801198642 1198801198672 119880120591V3 1198801205783 1198801198643 1198801198673
119880120591V4 1198801205784 1198801198644 1198801198674 119880120591V5 1198801205785 1198801198645 1198801198675]T
119884119906= [119910120591V2 1199101205782 1199101198642 1199101198672 119910120591V3 1199101205783 1199101198643 1199101198673
119910120591V4 1199101205784 1199101198644 1199101198674 119910120591V5 1199101205785 1199101198645 1199101198675]T
119884 = [Δ11987912 Δ11987923 Δ11987934 Δ11987945]T
(9)
119866Ten denotes the inputoutput dynamics of the tension sys-tem which takes into account the uncertainty of parametersThe state-space representation of119866Ten is (the systemmatricesare presented in Appendix A)
119866Ten =[
[
119860 1198611
1198612
11986211198631111986312
11986221198632111986322
]
]
(10)
The uncertain behavior of the original system can bedescribed by an upper LFT representation
119884 = 119865119880(119866Ten Δ)119880 (11)
With diagonal uncertainty matrix Δ = diag(120575V2 1205751205782 12057511986421205751198672 120575V5 1205751205785 1205751198645 1205751198675) as shown in Figure 5 It should
be noted that the unknown matrix Δ has a fixed structureand in general is a block diagonal Such uncertainty is calledstructured uncertainty
Δ
GTen
U Y
Figure 5 LFT representation of the tension system with uncertain-ties
GTension
Δ
d
Tensioncontroller Wp ep
minusWu
euG
Ref
Figure 6 The block diagram of closed system with performancerequirements
3 Robust Tension Controller Design
This section considers the design of tension control system for5-stand tandem mill For this system open loop and closed-loop system interconnections are introduced and robustcontrollers are designed
31 Requirements to Fulfill Design The aim of robust ten-sion controller design is to achieve and maintain the ten-sion between stands in the presence of disturbances anduncertainties The block diagram of the closed-loop systemincluding the feedback and controller as well as the elementsreflecting the model uncertainty and weighting functionsrelated to performance requirements is shown in Figure 6
In Figure 6 the rectangle with dashed line represents thetransfer matrix 119866 Inside the rectangle is the nominal modelof tension system and the uncertainty matrix Δ that includesthe model uncertainties The variables 119889
1 1198892 1198893 and 119889
4are
the disturbances on the system at the system outputs Theoutput weighting performance vector (119890
119901) control weighting
performance (119890119906) and disturbances vector (119889) are defined as
119890119901= [1198901199011 1198901199012 1198901199013 1198901199014]T
119890119906= [1198901199061 1198901199062 1198901199063 1198901199064]T
119889 = [1198891 1198892 1198893 1198894]T
(12)
In general weighting functions would be used to meet thedesign specifications Finding appropriate weighting func-tions is a crucial step in robust controller design and usuallyneeds a few trials The performance weighting function is
Journal of Engineering 5
103
102
101
100
10minus1
10minus2
10minus3
wminus1p1
Inverse performance weighting function
Frequency (rads)
Mag
nitu
de
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Figure 7 Singular values of 1119908119901
101
100
10minus1
10minus5 100 105 1010
wu1
Control weighting function
Frequency (rads)
Mag
nitu
de
Figure 8 Singular values of 119908119906
usually low-pass filter and control weighting function is high-pass filter Design experience will help in choosing weightingfunctions In this design the performanceweighting function119908119901and the control weighting functions 119908
119906119882119901 and119882
119906are
chosen as [13]
119908119901= 0012
1198782
+ 28119878 + 100
051198782 + 40119878 + 001
119908119906= 5
119878 + 10000
119878 + 500000
119882119901= (
1199081199011
0 0 0
0 1199081199012
0 0
0 0 1199081199013
0
0 0 0 1199081199014
)
119882119906= (
1199081199061
0 0 0
0 1199081199062
0 0
0 0 1199081199063
0
0 0 0 1199081199064
)
(13)The singular values of 1119908
119901over the frequency range
[10minus4
104
] are shown in Figure 7 This weighting functionshows that for low frequencies the closed-loop system (thenominal as well as perturbed system) attenuates the outputdisturbance in the ratio of 100 to 001 In other words theeffect of a unit disturbance on the steady-state output shouldbe of the order 10minus2 This performance requirement becomesless stringent with increasing frequency
Also the singular values of 119908119906over the frequency range
[10minus4 104] are shown in Figure 8
32 Open Loop and Closed-Loop System InterconnectionBefore designing the tension controller it is necessary to build
6 Journal of Engineering
Pertin1-16 Pertout1-16
sum
+
+
minus
sum
G
yc1-4
Wp
minusWu
ep1-4
eu1-4
ΔT1-4
Dist1-4
Ref = 0
Figure 9 Structure of the open loop system
Pertin1-16 Pertout1-16
Control1-4
ep1-4
eu1-4
ΔT1-4
Dist1-4 tension ic
Figure 10 Block diagram of open loop system
Pertin1-16 Pertout1-16
T1-4
U1-4
yc1-4
G
K
+
+
+minus
sum
sum
Dist1-4
Ref1-4
Figure 11 Structure of the closed-loop system
interconnection for open-loop and closed-loop system Thestructure of the open-loop system is represented in Figure 9In Figure 10 tension ic denotes the transfer function matrixof open loop system with 24 inputs and 28 outputs
The simulation of closed-loop system with designedcontrollers is based on the structure shown in Figure 11 Theweighting functions 119908
119901and 119908
119906are not included in the block
diagram for obvious reasons
33 119867infin
Controller The interconnection of nominal openloop system is denoted by hin ic as shown in Figure 12
The suboptimal 119867infin
controller minimizes the infinite-normof 119865119897(ℎ119894119899 119894119888 119896) overall stabilizing controller 119870 Note that
119865119897(ℎ119894119899 119894119888 119896) is the transfer function matrix of the nominal
closed-loop system from the disturbances to the errorsTo compute 119867
infinoptimal controller for LTI plant some
conditions are satisfied by plant Sometimes one of theseconditions is not satisfied and the algorithm for computingdoes not work [15 17] When this situation accrued otherrobust control methods are used although some tricks canbe used to satisfy conditions For example adding or sub-tracting epsilon may solve the problem in these conditions
Journal of Engineering 7
K
eDist hin ic
Figure 12 Closed-loop LFTs in119867infindesign
10minus1 100 101 102 103 10401
02
03
04
05
06
Frequency (rads)10minus1 100 101 102 103 104
Frequency (rads)
010
020304050607
Closed-loop complex 120583 controller number 3
Com
plex120583
Closed-loop mixed 120583 controller number 3
Mix
ed120583
Figure 13 120583 plot after third iteration
100
10minus1
10minus2
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Robust stability
Frequency (rads)
120583-controllerReduced-order controller
Upp
er b
ound
of120583
Hinfin controller
Figure 14 Comparison of robust stability for three controllers
For this plant 119867infin
controller is numerically computed andfinally the obtained controller is of 20th orderTheminimumvalue of 120574 is 0101
34 120583 Controller The 120583 synthesis is introduced by twoiterative methods 119863-119870 and 120583-119870 iteration For this plant120583 synthesis is executed by the M-file dkit from the RobustControl Toolbox of Matlab which automates the procedureby using 119863-119870 iterations [15] After the third iteration
the maximum value of 120583 is 0567 which indicates that therobust performance has been achieved The final controllerobtained is of 104th order The 120583 plot of the closed-loopsystem is shown in Figure 13
The order of this controller is 104 which makes it difficultin implementation On the other hand high order controllerswill lead to high cost difficult commissioning poor relia-bility and potential problems in maintenance This is theweakness of 120583 controller in this design Usually the order
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 5
103
102
101
100
10minus1
10minus2
10minus3
wminus1p1
Inverse performance weighting function
Frequency (rads)
Mag
nitu
de
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Figure 7 Singular values of 1119908119901
101
100
10minus1
10minus5 100 105 1010
wu1
Control weighting function
Frequency (rads)
Mag
nitu
de
Figure 8 Singular values of 119908119906
usually low-pass filter and control weighting function is high-pass filter Design experience will help in choosing weightingfunctions In this design the performanceweighting function119908119901and the control weighting functions 119908
119906119882119901 and119882
119906are
chosen as [13]
119908119901= 0012
1198782
+ 28119878 + 100
051198782 + 40119878 + 001
119908119906= 5
119878 + 10000
119878 + 500000
119882119901= (
1199081199011
0 0 0
0 1199081199012
0 0
0 0 1199081199013
0
0 0 0 1199081199014
)
119882119906= (
1199081199061
0 0 0
0 1199081199062
0 0
0 0 1199081199063
0
0 0 0 1199081199064
)
(13)The singular values of 1119908
119901over the frequency range
[10minus4
104
] are shown in Figure 7 This weighting functionshows that for low frequencies the closed-loop system (thenominal as well as perturbed system) attenuates the outputdisturbance in the ratio of 100 to 001 In other words theeffect of a unit disturbance on the steady-state output shouldbe of the order 10minus2 This performance requirement becomesless stringent with increasing frequency
Also the singular values of 119908119906over the frequency range
[10minus4 104] are shown in Figure 8
32 Open Loop and Closed-Loop System InterconnectionBefore designing the tension controller it is necessary to build
6 Journal of Engineering
Pertin1-16 Pertout1-16
sum
+
+
minus
sum
G
yc1-4
Wp
minusWu
ep1-4
eu1-4
ΔT1-4
Dist1-4
Ref = 0
Figure 9 Structure of the open loop system
Pertin1-16 Pertout1-16
Control1-4
ep1-4
eu1-4
ΔT1-4
Dist1-4 tension ic
Figure 10 Block diagram of open loop system
Pertin1-16 Pertout1-16
T1-4
U1-4
yc1-4
G
K
+
+
+minus
sum
sum
Dist1-4
Ref1-4
Figure 11 Structure of the closed-loop system
interconnection for open-loop and closed-loop system Thestructure of the open-loop system is represented in Figure 9In Figure 10 tension ic denotes the transfer function matrixof open loop system with 24 inputs and 28 outputs
The simulation of closed-loop system with designedcontrollers is based on the structure shown in Figure 11 Theweighting functions 119908
119901and 119908
119906are not included in the block
diagram for obvious reasons
33 119867infin
Controller The interconnection of nominal openloop system is denoted by hin ic as shown in Figure 12
The suboptimal 119867infin
controller minimizes the infinite-normof 119865119897(ℎ119894119899 119894119888 119896) overall stabilizing controller 119870 Note that
119865119897(ℎ119894119899 119894119888 119896) is the transfer function matrix of the nominal
closed-loop system from the disturbances to the errorsTo compute 119867
infinoptimal controller for LTI plant some
conditions are satisfied by plant Sometimes one of theseconditions is not satisfied and the algorithm for computingdoes not work [15 17] When this situation accrued otherrobust control methods are used although some tricks canbe used to satisfy conditions For example adding or sub-tracting epsilon may solve the problem in these conditions
Journal of Engineering 7
K
eDist hin ic
Figure 12 Closed-loop LFTs in119867infindesign
10minus1 100 101 102 103 10401
02
03
04
05
06
Frequency (rads)10minus1 100 101 102 103 104
Frequency (rads)
010
020304050607
Closed-loop complex 120583 controller number 3
Com
plex120583
Closed-loop mixed 120583 controller number 3
Mix
ed120583
Figure 13 120583 plot after third iteration
100
10minus1
10minus2
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Robust stability
Frequency (rads)
120583-controllerReduced-order controller
Upp
er b
ound
of120583
Hinfin controller
Figure 14 Comparison of robust stability for three controllers
For this plant 119867infin
controller is numerically computed andfinally the obtained controller is of 20th orderTheminimumvalue of 120574 is 0101
34 120583 Controller The 120583 synthesis is introduced by twoiterative methods 119863-119870 and 120583-119870 iteration For this plant120583 synthesis is executed by the M-file dkit from the RobustControl Toolbox of Matlab which automates the procedureby using 119863-119870 iterations [15] After the third iteration
the maximum value of 120583 is 0567 which indicates that therobust performance has been achieved The final controllerobtained is of 104th order The 120583 plot of the closed-loopsystem is shown in Figure 13
The order of this controller is 104 which makes it difficultin implementation On the other hand high order controllerswill lead to high cost difficult commissioning poor relia-bility and potential problems in maintenance This is theweakness of 120583 controller in this design Usually the order
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
6 Journal of Engineering
Pertin1-16 Pertout1-16
sum
+
+
minus
sum
G
yc1-4
Wp
minusWu
ep1-4
eu1-4
ΔT1-4
Dist1-4
Ref = 0
Figure 9 Structure of the open loop system
Pertin1-16 Pertout1-16
Control1-4
ep1-4
eu1-4
ΔT1-4
Dist1-4 tension ic
Figure 10 Block diagram of open loop system
Pertin1-16 Pertout1-16
T1-4
U1-4
yc1-4
G
K
+
+
+minus
sum
sum
Dist1-4
Ref1-4
Figure 11 Structure of the closed-loop system
interconnection for open-loop and closed-loop system Thestructure of the open-loop system is represented in Figure 9In Figure 10 tension ic denotes the transfer function matrixof open loop system with 24 inputs and 28 outputs
The simulation of closed-loop system with designedcontrollers is based on the structure shown in Figure 11 Theweighting functions 119908
119901and 119908
119906are not included in the block
diagram for obvious reasons
33 119867infin
Controller The interconnection of nominal openloop system is denoted by hin ic as shown in Figure 12
The suboptimal 119867infin
controller minimizes the infinite-normof 119865119897(ℎ119894119899 119894119888 119896) overall stabilizing controller 119870 Note that
119865119897(ℎ119894119899 119894119888 119896) is the transfer function matrix of the nominal
closed-loop system from the disturbances to the errorsTo compute 119867
infinoptimal controller for LTI plant some
conditions are satisfied by plant Sometimes one of theseconditions is not satisfied and the algorithm for computingdoes not work [15 17] When this situation accrued otherrobust control methods are used although some tricks canbe used to satisfy conditions For example adding or sub-tracting epsilon may solve the problem in these conditions
Journal of Engineering 7
K
eDist hin ic
Figure 12 Closed-loop LFTs in119867infindesign
10minus1 100 101 102 103 10401
02
03
04
05
06
Frequency (rads)10minus1 100 101 102 103 104
Frequency (rads)
010
020304050607
Closed-loop complex 120583 controller number 3
Com
plex120583
Closed-loop mixed 120583 controller number 3
Mix
ed120583
Figure 13 120583 plot after third iteration
100
10minus1
10minus2
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Robust stability
Frequency (rads)
120583-controllerReduced-order controller
Upp
er b
ound
of120583
Hinfin controller
Figure 14 Comparison of robust stability for three controllers
For this plant 119867infin
controller is numerically computed andfinally the obtained controller is of 20th orderTheminimumvalue of 120574 is 0101
34 120583 Controller The 120583 synthesis is introduced by twoiterative methods 119863-119870 and 120583-119870 iteration For this plant120583 synthesis is executed by the M-file dkit from the RobustControl Toolbox of Matlab which automates the procedureby using 119863-119870 iterations [15] After the third iteration
the maximum value of 120583 is 0567 which indicates that therobust performance has been achieved The final controllerobtained is of 104th order The 120583 plot of the closed-loopsystem is shown in Figure 13
The order of this controller is 104 which makes it difficultin implementation On the other hand high order controllerswill lead to high cost difficult commissioning poor relia-bility and potential problems in maintenance This is theweakness of 120583 controller in this design Usually the order
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 7
K
eDist hin ic
Figure 12 Closed-loop LFTs in119867infindesign
10minus1 100 101 102 103 10401
02
03
04
05
06
Frequency (rads)10minus1 100 101 102 103 104
Frequency (rads)
010
020304050607
Closed-loop complex 120583 controller number 3
Com
plex120583
Closed-loop mixed 120583 controller number 3
Mix
ed120583
Figure 13 120583 plot after third iteration
100
10minus1
10minus2
10minus4 10minus3 10minus2 10minus1 100 101 102 103 104
Robust stability
Frequency (rads)
120583-controllerReduced-order controller
Upp
er b
ound
of120583
Hinfin controller
Figure 14 Comparison of robust stability for three controllers
For this plant 119867infin
controller is numerically computed andfinally the obtained controller is of 20th orderTheminimumvalue of 120574 is 0101
34 120583 Controller The 120583 synthesis is introduced by twoiterative methods 119863-119870 and 120583-119870 iteration For this plant120583 synthesis is executed by the M-file dkit from the RobustControl Toolbox of Matlab which automates the procedureby using 119863-119870 iterations [15] After the third iteration
the maximum value of 120583 is 0567 which indicates that therobust performance has been achieved The final controllerobtained is of 104th order The 120583 plot of the closed-loopsystem is shown in Figure 13
The order of this controller is 104 which makes it difficultin implementation On the other hand high order controllerswill lead to high cost difficult commissioning poor relia-bility and potential problems in maintenance This is theweakness of 120583 controller in this design Usually the order
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
8 Journal of Engineering
10minus4 10minus3 10minus2 10minus1 100 101 102 103 1040
01
02
03
04
05
06
07
08
09
1Robust performance
Frequency (rads)Hinfin controller120583-controllerReduced-order controller
Upp
er b
ound
of120583
Figure 15 Comparison of robust performance for three controllers
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref1 (tension ST1-ST2)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref2 (tension ST2-ST3)
Time (s)
0 10 20 30 40minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref3 (tension ST3-ST4)
Time (s)0 10 20 30 40
minus4000
minus2000
0
2000
4000
6000
8000Transient response to ref4 (tension ST4-ST5)
Time (s)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Figure 16 Comparison of transient response to reference input
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 9
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 1
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600Transient response to disturbance 2
Transient response to disturbance 4
Time (s)
0 10 20 30 40minus600
minus400
minus200
0
200
400
600Transient response to disturbance 3
Time (s)0 10 20 30 40
minus600
minus400
minus200
0
200
400
600
Time (s)
Tens
ion1
-2(N
)
Tens
ion2
-3(N
)
Tens
ion3
-4(N
)
Tens
ion4
-5(N
)
Hinfin controller120583-controller
Reduced-order controllerRef
Hinfin controller120583-controller
Reduced-order controllerRef
Figure 17 Comparison of transient response to disturbance signal
can be reduced and reduction of the controller order maylead to deterioration of the closed-loop performance Variousmethods may be used to reduce the order and in this casestudy the Hankel-norm approximation is usedThe reduced-order controller is equal to 13 that leads to a controllermaintaining the robust stability and performance of theclosed-loop system (matrices for reduced-order controllerare presented in Appendix B)
4 Simulation Results
In the previous section robust controllers for tension con-trol system are introduced It is intended here to comparethe robust stability robust performance and time domainresponses of the proposed controllers Since the considereduncertainty is structured verification of the robust stabilityand performance needs the frequency response in terms of120583 values On the other hand controllers should be tested bymeans of 120583 analysis The robust stability and performancedo not maintain if the uncertainty is unstructured whichshows that the 120583 values give less conservative results if furtherinformation is known about the uncertainty
41 Robust Stability The frequency responses of the upperbound of 120583 for 119867
infincontroller and 120583 and reduced-order
controllers are shown in Figure 14Themaximum120583 values for119867infincontroller are equal to 07801 and the maximum 120583 values
for 120583 and reduced-order controllers are 0202 Thereforethe robust stability of control systems is achieved and thesystems with the 120583 controller and reduced-order controllersare characterized with better robust stability
42 Robust Performance The robust performance of controlsystems is plotted in Figure 15 The maximum 120583 valuesfor 119867
infincontroller are 099 for 120583 controller 0597 and for
reduced-order controller 069 which implies the robust 120583controller and reduced-order controller achieve better robustperformance than119867
infincontroller
43 Time Domain Simulation The transient responses of theclosed-loop system to the reference input and the disturbancesignal are shown in Figures 16 and 17 respectively The 120583 andthe reduced-order controllers give a better transient responsewith smaller overshoot than the119867
infincontroller In summary
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
10 Journal of Engineering
the 120583 and the reduced-order controllers lead to better trade-off between the requirements in terms of transient responsedisturbance rejection and robustness
5 Conclusion
In cold rolling the strip tension should be controlled forbetter quality of products For cold rolling with 5 standsthe state-space representation perturbed with parametricuncertainties is derived and further 119867
infincontroller and
120583 controller based on robust control theory are designedTension control is obtained by means of the proposed robustcontrollers in the presence of uncertainties and disturbancesThe order of the 120583 controller is large which requires thecontroller order reduction to preserve the closed-loop per-formance and robustness while making implementation ofthe controller easier andmore reliableThe simulation resultsshow that control systems achieve robust stability and robust
performance The comparison of the robust stability androbust performance for controllers as well as the comparisonof the corresponding transient responses shows that the 120583controller and its reduced-order are preferable although thereduced-order controller would be usually preferred dueto its lower order It should be mentioned that in suchimportant industrial process most commissioning personnelhave very little or no background in advanced control theoryespecially in the theory and application of119867
infinrobust control
techniques and are accustomed to single-input-single-output(SISO) proportional-integral (PI) type control loops Thesesimpler configurations are easy to tune at commissioningduring actual operation
Appendices
A The System Matrices
Consider
119860 =
[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
44200 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
minus44200 0 26000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 minus26800 0 18600 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 minus18400 0 12200 0
]]]]]]]]]]
]
1198612=
[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 10 0
0 0 0 0
0 0 0 10
0 0 0 0
]]]]]]]]]]]]
]
1198611=
[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 6280 003 00011 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 minus6220 0 0 0 3600 0017 00011 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 minus3600 0 0 0 2300 0011 0011 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 minus2300 0 0 0 1540 0073 0011
]]]]]]]]]]
]
1198621=
[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus10 0 0 0 0 0 0 0
07
147000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
199000 0 0 0 0 0 0 0
0 0 minus10 0 0 0 0 0
0 0 075 0 0 0 0 0
minus257000 0 157000 0 0 0 0 0
minus197000 0 120000 0 0 0 0 0
0 0 0 0 minus10 0 0 0
0 0 0 0 08 0 0 0
0 0 minus242000 0 168000 0 0 0
0 0 minus1210000 0 84000 0 0 0
0 0 0 0 0 0 minus10 0
0 0 0 0 0 0 08 0
0 0 0 0 minus253000 0 168000 0
0 0 0 0 minus84000 0 55000 0
]]]]]]]]]]]]]]]]]]]]]]]]]]]]
]
1198622=
[[[
[
0 1 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1
]]]
]
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 11
11986311=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
minus02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 21000 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 28300 0135 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 minus02 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 36700 0 0 0 21000 0 0 0 0 0 0 0 0 0 0
0 minus28200 0 0 0 16700 0077 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 minus02 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 minus32300 0 0 0 21000 0 0 0 0 0 0
0 0 0 0 0 minus16100 0 005 10500 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 minus02 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 minus31700 0 0 0 21000 0 0
0 0 0 0 0 0 0 0 0 minus10400 0 0 0 6930 0033 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
11986312=
[[[[[[[[[[[[[[[[[[[[[[[[[[
[
10 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 10 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 10
0 0 0 0
0 0 0 0
0 0 0 0
]]]]]]]]]]]]]]]]]]]]]]]]]]
]
(A1)
B Matrices for Reduced-Order Controller
Consider
ak =
[[[[[[[[[[[[[[[[[[[[
[
minus13534 9671 2455 minus1674 minus122 155 minus048 299 109 014 minus008 minus019 011
0 minus2734 523 1642 147 minus068 04 minus136 minus065 0005 007 01 minus0048
0 minus2872 minus277 03 minus1267 minus021 005 minus065 003 minus01 009 002 004
0 0 0 minus611 minus184 018 minus029 031 042 minus009 minus002 minus002 0018
0 0 0 145 minus662 minus008 019 minus011 minus006 minus002 0 0003 0
0 0 0 0 0 minus0008 006 minus002 minus0008 minus0001 0 0002 0001
0 0 0 0 0 minus006 minus0004 0008 0008 minus0002 0 0 0001
0 0 0 0 0 0 0 minus002 002 minus0004 minus0001 0 0
0 0 0 0 0 0 0 minus004 minus0008 0002 minus0003 0 0
0 0 0 0 0 0 0 0 0 minus0003 0 0 0
0 0 0 0 0 0 0 0 0 0 minus0002 0 0
0 0 0 0 0 0 0 0 0 0 0 minus0002 0
0 0 0 0 0 0 0 0 0 0 0 0 minus0002
]]]]]]]]]]]]]]]]]]]]
]
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
12 Journal of Engineering
bk =
[[[[[[[[[[[[[[[[[[[[
[
015 008 002 004
minus004 minus003 minus001 minus001
minus005 minus0002 minus002 0
minus0008 001 001 001
minus0006 0008 0003 minus001
0 0 0 0
0 0 0 0
minus0001 0 0 0
0 0 0 0
0 0 0 0
0001 0 0 0
0 0 0 minus0001
0 0 0 minus0001
]]]]]]]]]]]]]]]]]]]]
]
ck =[[[
[
minus0001 minus0006 0 0001 minus0001 0 0 0 0 0 0 0 0
006 minus002 minus0009 0001 0007 0 0 0 0 0 0 0001 0
0009 minus004 minus003 001 0 0 0 0 0 0 0001 0 0001
013 minus008 minus005 001 0 0 0001 minus0002 0 0 0001 minus0001 minus0001
]]]
]
dk =[[[
[
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
]]]
]
(B1)
Abbreviations
119881 Work roll linear velocity119880V Speed actuator input120591V Speed actuator time constant119881in Input strip velocity119881out Output strip velocityℎin Stand input thicknessℎout Stand output thickness119891 Forward slip ratio120593 Backward slip ratio119864 Youngrsquos modulus119861 Width of strip119867119894+1
Strip thickness between stand 119894 and stand 119894 + 1119871 Length between adjacent stands119879119894119894+1
Strip tension between stand 119894 and stand 119894 + 1
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] W S Levine Control System Applications Taylor amp FrancisBoca Raton Fla USA 2011
[2] J Pittner N S Samaras and M A Simaan ldquoA simple rollingmillmodel with linear quadratic optimal controllerrdquo inProceed-ings of the 37th IAS Annual Meeting and World Conference onIndustrial Applications of Electrical Energy pp 142ndash149 October2002
[3] H R Koofigar F Sheikholeslam and S Hosseinnia ldquoUnifiedgauge-tension control in cold rolling mills a robust regulationtechniquerdquo International Journal of Precision Engineering andManufacturing vol 12 no 3 pp 393ndash403 2011
[4] E J M Geddes and I Postlethwaite ldquoImprovements in productquality in tandem cold rolling using robust multivariablecontrolrdquo IEEE Transactions on Control Systems Technology vol6 no 2 pp 257ndash269 1998
[5] E J M Geddes and I Postlethwaite ldquoMultivariable control ofa high performance tandem cold rolling millrdquo in Proceedings ofthe International Conference on Control (CONTROL rsquo94) vol 1pp 202ndash207 Coventry UK March 1994
[6] E Arinton S Caraman and J Korbicz ldquoNeural networks formodelling and fault detection of the inter-stand strip tension ofa cold tandem millrdquo Control Engineering Practice vol 20 no 7pp 684ndash694 2012
[7] KMTakami JMahmoudi andEDahlquist ldquoAdaptive controlof cold rolling system in electrical strips production systemwithonline-offline predictorsrdquo International Journal of AdvancedManufacturing Technology vol 50 no 9ndash12 pp 917ndash930 2010
[8] G Pin V Francesconi F A Cuzzola and T Parisini ldquoAdaptivetask-space metal strip-flatness control in cold multi-roll millstandsrdquo Journal of Process Control vol 23 no 2 pp 108ndash1192013
[9] J Pittner and M A Simaan ldquoAn optimal control method forimprovement in tandem cold metal rollingrdquo in Proceedings ofthe IEEE 42nd AnnualMeeting Industry Applications Conference(IAS rsquo07) pp 382ndash389 September 2007
[10] J R Pittner Pointwise linear quadratic optimal control of atandem cold rolling mill [PhD thesis] School of EngineeringUniversity of Pittsburgh 2006
[11] L E Zarate ldquoPredictive model for the cold rolling processthrough sensitivity factors via neural networksrdquo Journal of the
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Engineering 13
Brazilian Society ofMechanical Sciences and Engineering vol 28no 1 pp 111ndash117 2006
[12] S-H Song and S-K Sul ldquoA new tension controller for con-tinuous strip processing linerdquo IEEE Transactions on IndustryApplications vol 36 no 2 pp 633ndash639 2000
[13] X Zhang and Q Zhang ldquoRobust control of strip tension fortandem cold rolling millrdquo in Proceedings of the 30th ChineseControl Conference (CCC rsquo11) pp 2390ndash2393 Yantai China July2011
[14] J Pittner and M A Simaan Tandem Cold Metal Rolling MillControl Using Practical Advanced Methods Springer LondonUK 2011
[15] DGu PH Petkov andMKonstantinovRobust Control Designwith MATLAB Springer London UK 2005
[16] Z Xiaofeng and ZQingdong ldquoRobust control of strip thicknessfor cold rolling millrdquo in Informatics in Control Automation andRobotics vol 133 of Lecture Notes in Electrical Engineering pp777ndash785 Springer Berlin Germany 2012
[17] K Zohu and J C Doyle Essentials of Robust Control PrenticeHall 1997
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of