- 上一篇：布局优化模型为坚实地固化快速成型过程英文文献翻译
- 下一篇：计算机数字控制机床英文文献和中文翻译
the finishing mill (the next mill in the production line). Such un-
evenness has to be manually and periodically compensated by
an operator. Corrections are made by adjusting a tilting signal
(Fig. 2) to produce a differential reference signal for the hy-
draulic actuators on each side of the mill. Manual correction is
a tedious task as well as being error prone, since it is based on
visual inspection to compensate large visible deviations (by an
operator whose workplace is 60 m from the mill).This paper describes a project aimed at designing and imple-
menting a computer-based system that will automatically gen-
erate the tilting signal from the mill operation data. The goal is
to improve output strip quality, as the correction signal can be
continuously applied before the uneven thickness effect is vis-
ible. Section II describes the actual process and proposed com-
pensation system. In Section III, the mathematical model of the
rolling process is described. Section IV presents its adjustment
and real-time validation. Section V explains implementation de-
tails, experimental results, and conclusions.
II. AGC AND UNEVENNESS COMPENSATION
In a hot rolling mill, the plate thickness is reduced by pulling
the plate between two parallel rolls. The upper roll position
is controlled, and the lower backup roll position is only cali-
brated by means of a set of wedges, in order to assure a hor-
izontal rolling line. Before reaching the rolling mill, the plate
achieves a uniform temperature (around 1250 C) in reheating
furnaces. The plate thickness reduction is achieved in a series
of passes, two spray headers clean and cool the plate after each
pass. Cooling may induce a nonuniform temperature along the
plate, leading to variable hardness.
While the material is being rolled, AGC systems are applied
to keep output thickness target range. Deviations which the
AGC tries to correct can be classified in two groups: irregu-
larities in the material (input thickness, slab temperature) or
irregularities in the mechanical equipment (roll eccentricity,
rolls thermal deformation, backup rolls lubrication). The rolls
position reference calculated by the AGC signals two fast
hydraulic positioning systems on both sides of the load roll
(north and south sides) [2] (Fig. 2).
Most operating AGCs in industry have been designed by
using single-input–single-output (SISO) models, considering
mean thickness values at both ends. The AGC corrections are
implemented by using hydraulic positioning systems regulated
by a stand controller, which adjusts a servo-valve position
according to the position of each roll as calculated by the AGC
[3] (Fig. 3).
To compensate for asymmetries in the roll pack and mill
frame, a tilting base signal is calibrated offline. However,
there are other causes of unevenness, such as a nonuniform
incoming thickness, or hardness along the transversal rolling
direction, that produce visible thickness deviations. They are
manually compensated online, by means of the tilting signal,which produces a differential reference signal to the hydraulic
actuators on each side of the mill.
In developing a system to automatically take into account the
effect of unevenness, two options appear.
1) Replace all AGCs. Existing AGCs fail in compensating
for plate thickness variationwhen they are based on scalar
models (using themean values of the variables at the north
and south sides). It is suggested to use a multivariable
stand controller, based on a multivariable model of the
rolling process, which takes into account thickness varia-
tions along the transversal rolling direction [4]. It implies
the setup of a new complete experimental AGC, a com-