International Conference on Computer Information Systems and Industrial Applications (CISIA 2015)

The Research on Flatness Control Simulation for

Cold Tandem Rolling Mills
X.M. Zhou, X.X. Yue
School of Mechanical Engineering
University of Science and Technology Beijing
Beijing, China

Abstract—The cold tandem rolling of metal strip presents a diameter of the work roll is [Link] roll finite element
significant control challenge because of nonlinearities and process deformation model is shown as Figure 1.
complexities. Flatness control system is a hot topic on cold
tandem rolling process. The process parameters and state
variables cannot be measured due to the rolling process
complexities. So it is important how to indicate the rolling process
parameters change during the whole rolling process. The rolling
model is established by considering the elastic deformation of the
roll and plastic deformation of the strip as a whole. The verified
results represent that the flatness control integration simulation
system can reflect the rolling process parameters and flatness
effectively.

Keywords-flatness control; simulation; cold tandem rolling Mill

I. INTRODUCTION
The flatness and the gauge accuracy are the most important FIGURE I. ROLL DEFORMATION MODEL
factors in evaluating the quality of a strip [1]. Customers’ needs
for product quality are higher and higher. To meet the demand For flatness control process, the most important role of the
for the high quality strip, steel makers and researchers pay strip plastic deformation calculation is to decide the rolling
attention to the flatness control significantly. But the cold force distribution and tension distribution. It is hard to achieve
tandem rolling of metal strip presents a significant control due to the physical and geometry nonlinearity and the
challenge because of nonlinearities and process complexities complexity of the boundary condition for the rolling process.
[2][3]. Nowadays, there are a lot of problems to resolve for the
The model research of the cold rolling process is based on
high precision flatness control because of the rising rolling
the traditional theory of Kalman equation and Orowan equation
speed and the more improving flatness control actuators
under the condition of the two-dimension plane strain
[4][5][6]. The traditional model establishment is hard to deal
assumption. The strip deformation can be considered be
with the more and more complex control process [7][8]. And
produced on the thickness and the rolling direction. The
the measurement of the process parameters online is also
deformation in the strip width direction can be ignored. But this
difficult to achieve. With the rapid improving of the computer
kind of model is clearly not appropriate for studying the strip
technology, the simulation system based on the experimental
edge drop characteristics. Because the edge drop generation
data can be used to analyse and debug the control tragedy and
depends not only on the work roll deflection and flatten, but
model for rolling process [9][10][11]. It is very important to the
also depends on the strip three-dimensional plastic deformation
actual process and research.
caused by the lateral flow of metal near the strip edge. If the
In this paper, the rolling model is established by strip deformation in the width direction is ignored because of
considering the elastic deformation of the roll and plastic using the two-dimensional model to solve this problem, the real
deformation of the strip as a whole. The verified results situation of the strip flatness cannot be reflected actually. So
represent that the flatness control integration simulation system the three-dimensional plastic deformation model must be
can reflect the rolling process parameters and flatness established to consider the strip transverse flow during the
effectively. rolling process.

II. THE ESTABLISHMENT OF THE SIMULATION SYSTEM The three dimensional differential method is used to
calculate the strip plastic deformation in this paper. The vertical
The simulation system is established in which elastic and horizontal balance differential equations of the deformation
deformation of the roll and plastic deformationofthestripare zone are transferred to the differential form. Then these
considered as a whole. The simulation is for five stands 6-high equations are united with the plasticity condition, plastic flow
tandem cold rolling mill. The diameter of the backup roll is equation, constant volume conditions and boundary conditions,
1276mm. The diameter of the intermediate roll is 467mm. The the three-dimensional stress distribution of the deformation

© 2015. The authors - Published by Atlantis Press 946

zone and the edge curve of the strip can be calculated using the The model can analyse the flatness influence factors from
numerical and iterative method. the roll and strip such as roll shape including the initial roll
shape, thermal roll shape and wear roll shape, roll bending, roll
The entry and exit longitudinal stress σx are known shift, strip material, entry strip profile and tension etc. So this
parameters of the differential calculation. The assumed simulation system is the comprehensive means for the shape
transverse deformation values of the edge units are the initial characteristics research. And it can simulate the strip edge
values of iterative computation. The differential calculation of deformation precisely, so it is important method to analyse the
the whole deformation zone is from the first unit of the edge. edge drop for flatness control.
The integration model of the roll and strip combines the
strip plastic deformation model and the roll elastic deformation III. THE VERIFICATION OF THE MODEL BY EXPERIMENTAL
model. The exit strip profile is the output of the roll elastic DATA
deformation model. And it is the input of the strip plastic To verify the model reliability and precise, the sampling
deformation model. The calculated rolling force distribution experiment was carried out in the cold tandem rolling mill. To
from the strip plastic deformation model compared with the retain the whole deformation zone and ensure the maximum
assumed rolling force distribution. If the deviation meets the production capacity, the sampling experiment is carried out
deviation requirement, the calculated results are the final output. before the backup roll changing. The whole deformation zones
Otherwise the rolling force distribution is updated to repeat the of the five stands can be obtained by lifting rolls during the
above iterated steps until the condition is satisfied. The outputs normal rolling process. And the corresponding process
include the rolling force distribution, the exit strip profile and parameters and rolling parameters are recorded at the same
the front tension distribution. The calculation flow chart of the time. The integration model is verified by these valuable
integration model is shown as Figure 2. experimental data.
Start The spiral micrometre (measurement precision of 0.001
mm) is used to measure the entry and exit transverse thickness
Input initial parameters
distribution of the strip near the deformation zone. The
sampling measurement positions are shown as Figure 3.
Assume roll force distribution

Calculate roll deformation model

Get exit strip profile

Calculate strip deformation model Correct

roll force
distribution
Calculate roll force distribution

,
Not equal
Compare with the assumed roll
force distribution FIGURE III. THE SAMPLING MEASUREMENT POSITIONS
Equal
Output results The strip thickness measurement is from the work side (WS
side) to the drive side (DS side) across the strip width. In order
End
to study the edge drop of the strip, the measurement is fine in
the strip edge. These measurement data describe the entry and
exit strip transverse thicknessdistribution of each stand.
The correspondingrolling parameters, process parameters
FIGURE II. THE CALCULATION BLOCK DIAGRAM OF THE
INTEGRATION MODEL
and the measured entry thickness distribution are the input of
the integration system to calculate. To compare the calculated
The whole model is the integration of the strip plastic exit thickness distribution of each stand with
deformation model and the roll elastic deformation model. The themeasuredvalues of each stand, the accuracy of the model
roll elastic deformation model is calculated by the variable can be analysed.
thickness finite element method which has the high calculation The comparison of the stand 1exit strip measured
speed and precision. The strip plastic deformation model is values and calculated values are shown as Figure 4. The
calculated by the three-dimensional differential method which comparison of the stand 5exit strip measured values and
can reflect the actual strip deformation precisely considering calculated values are shown as Figure 5. It can be seen from
the transverse flow in rolling zone. Fig.9 and Fig.10 thatthe integration model can achieve the
satisfied precision for the flatness control [Link] the
comparison of the measured values and calculated values, the

947
calculated strip transverse thickness distribution are very close [4] ZHOU X. M., WU Y. F., MAO Y. J..The Research on the Coupling
Relation between Flatness and Edge Drop for Tandem Cold Rolling Mill.
to the measured values including the edge zones. Advanced Materials Research. 2008,572(8),pp104-pp109.
[5] C.y. JIA, H. M. LIU. Fuzzy Shape Control Based on Elman
Comparison of stand 1 exit strip profile between DynamicRecursion Network Prediction Model. Journal of Iron and Steel
simulation and measurement Research International,2006,13(1),pp31-pp35.
10 width(mm) [6] Tateno J., Kenmochi K., Yarita I.. Controlling Edge Drop by Tapered –
Crown Work Roll Shifting Mill and Work Roll Crossing Mill in Cold
-600 -400 -200 0 200 400 600 Strip Rolling. Journal of the JSTP[J],1999 ,(7).pp653-pp656.
-10
[7] Lu H. T., Cao J. G., Zhang J.. Edge drop control of a taper roll during
continuous cold rolling[J]. Journal of University of Science and
-30 Technology Beijing, 2006,28(8), pp774-pp777.
exit_simu
exit_meas [8] Zhou X. M., Zhang Q. D., Wang C. S., et al. Edge Drop and Flatness
Decoupling Control on UCMW Cold Mill[J].Iron and Steel,2007,42(5),
-50
pp55-pp57.
(um)

[9] Cavaliere M.A.,Goldschmit M.B.,Dvorkin E.N.. Finiteelement analysis of

-70 steel rolling processes. ComputStruct 2001,79,pp2075–pp2089.
[10] Xiong S. W., Liu X. H., Wang G. D., et al．Three-dimensional finite
FIGURE IV. THE COMPARISON OF THE STAND 1 EXIT STRIP
elementsimulation of the vertical-horizontal rolling process in the width
BETWEEN SIMULATION AND MEASUREMENT reductionof slab[J].Journal of Materials Processing Technology.
2000,10(1),pp146-pp151.
Comparison of stand 5 exit strip profile between [11] S. Chandra, U. S. Dixit. A rigid-plastic finite element analysis of
simulation and measurement temperrolling process[J]. Journal of Materials Processing Technology,
10 width(mm) 2004, 152,pp9-pp16.

-600 -400 -200 0 200 400 600

-10

-30
exit_simu
exit_meas
-50
(um)

-70

FIGURE V. THE COMPARISON OF THE STAND 5 EXIT STRIP

BETWEEN SIMULATION AND MEASUREMENT

IV. CONCLUSION
In this paper the rolling model is established by considering
the elastic deformation of the roll and plastic deformation of
the strip as a whole. The simulation system is verified by
experiment data from the five stands 6-high tandem cold
rolling mill. The verified results represent that the flatness
control simulation system can reflect the rolling process
parameters and flatness effectively. This simulation system can
be used to analyse the flatness control process including the
edge drop control.
ACKNOWLEDGEMENTS
We would like to acknowledge the financial support by
National Natural Science Foundation of China through Grant
No.51204017.
REFERENCES
[1] Bilkhu T.S.. Dynamic Control of Tension, Thickness and Flatness for a
Tandem ColdMill. AISE Steel Technology, 2001,78(10),pp49-pp54.
[2] A. BEMPORAD, D. BERNARDINI, F. ALESSANDRO,et al.
Optimization-based Automatic Flatness Control in Cold
TandemRolling[J]. Journal of Process Control, 2010, 20(4),pp396-pp407.
[3] Zhou X. M., Zhang Q. D., Wu P.. The Summary of the Strip Edge Drop
Control Technology[J]. ShangHai Metals.2007,29(1), pp21-pp24.

948