Strip tracking measurement and control in hot strip rolling

Abstract

It is well known that poor strip tracking can lead to reduced
product quality but also to mill delays. The resulting
costs for internal rejects, customer complaints and yield
losses have historically been significant. Moreover, the severity
of these issues increases dramatically when strips
become wider, thinner and harder. Ultimately the rolling
process becomes completely unstable. Hence, to reduce
cost of poor quality for the current product mix as well as
to enable product development it is vital that strip tracking
is improved.
Most strip tracking issues arise at the head or the tail of
the strip. In the rougher mill the main issue is head camber,
a shape defect of the bar where the head is curved. A
clear example of this shape is shown in Fig 1. Large head
camber of the transfer bar may result in further problems
downstream in the finishing mill and should ideally thus be
prevented.
Another notorious problem closely related to strip tracking
is tail pinching in the finishing mill. This is a phenomenon
where the tail of the strip suddenly moves sideward’s and
gets damaged right after it has left the previous stand. An
Poor strip tracking is one of the notorious problems threatening process stability in a hot strip mill. These
issues often lead to tail pinching and in the worst cases even to cobbles. The main pillars of the strategy set
out to tackle these issues for the Hot Strip Mills in IJmuiden are rougher mill camber control and finishing mill
strip steering and tail control. For such applications, a camera based measurement system has been developed
in-house that is simple, cost-effective and yet both accurate and robust. Moreover, as we show in this paper,
the system has proven its merits both as a finishing mill interstand centerline deviation measurement as
well as a rougher mill camber measurement. In the latter application the measurement data can be used for
automatic levelling in the rougher mill. The results of production tests presented in this paper demonstrate that
the camber measurement in combination with a basic rougher mill tilt set-up model is sufficient to reduce the
transfer bar camber significantly.