has been developed for hot slab on-line surface inspection and installed in Aceralia (Spain). This innovative inspection system, integrating novel Conoscopic Holography, CCD cameras, a complex mechanical system and intelligent tools, operates in the real environment of a continuous casting facility, allows on-line hot slab surface inspection of 100 % of the production, detecting cracks -without removing the surface scaleand inclusions (peeling a narrow strip on the surface). ■ INTRODUCTION AND OBJECTIVES

This paper describes the development of a flatness measurement system, integrated in the control process of a hot strip mill in the steel industry. The objective of the system is to calculate flatness indexes for every rolled strip, comparing the length of its lateral profiles with the central length. The reconstruction of the profiles is based on a nonlinear triangulation technique. Images of laser spots over the steel strip, at high temperature and high speed, are sampled every two milliseconds at five different points and are processed online in order to calculate height displacement values of the strip, so that final flatness indexes for the steel strip can be calculated. The measurement method developed introduces an innovative geometry in the disposition of the optic elements which increases the measurement range of the system without reducing its accuracy. It also includes a tracking system to compensate for the effects of lateral displacements of the strip. The flatness measurement system has been implemented using a heterogeneous distributed computer system.

A new, advanced and breakthrough technology has been developed for hot slab on-line surface inspection and installed in Aceralia (Spain). This innovative inspection system, integrating novel Conoscopic Holography, CCD cameras, a complex mechanical system and intelligent tools, operates in the real environment of a continuous casting facility, allows on-line hot slab surface inspection of 100 % of the production, detecting cracks -without removing the surface scale- and inclusions (peeling a narrow strip on the surface).

Steel processes are often of a complex nature and difficult to model. All information that we have at hand usually consists of more or less precise models of different parts of the process, some rules obtained on the basis of experience, and typically a great amount of high-dimensional data coming from numerous sensors and variables of process computers which convey a lot of information about the process state. We suggest in this paper the use of a continuous version of the Self-Organizing Map (SOM) to project a high dimensional vector of process data on a 2D visualization space in which different process conditions are represented by different regions. Later, all sorts of information resulting from the fusion of knowledge obtained from data, mathematical models and fuzzy rules can be described in a graphical way in this visualization space.

The paper presents the work done to assess the usefulness of the thermographic analysis in the design of a detection system for longitudinal defects in coils manufactured by a tandem cold mill. The approach started with a first phase devoted to the development of a computer system for the acquisition of high resolution thermal maps of the whole strip. In the second phase the thermal maps are classified and related to the process variables.

Hot rolling products are under an increasing demand of better quality features as thickness profile, strip width and flatness, and material mechanical properties. The project goal is the improvement of the production quality in an existing hot rolling mill at Aceralia Steel Corporation. In this paper the design and implementation of a supervisory system for the real-time compensation of uneven thickness on both sides of a rolled strip is described. The design is based on a multivariable process model, whose parameters are estimated on-line using measurement data from the mill. As a result, a computer system was developed that automatically corrects the automatic gauge control (AGC) output from on-line acquired mill operation data. Previously this task was made manually and periodically by a human operator. This manual correction is a tedious and error prone task, as it is based on visual inspection. Moreover, the proposed method leads to an improvement in the output strip quality, as the correction signal can be continuously applied before the uneven thickness effect is visible.

This paper describes the development of a flatness inspection system, integrated in the control process of a hot strip mill in the steel industry. The objective of the system is to calculate flatness indexes for every strip, comparing the length of its lateral profiles with the central length. The reconstruction of the profiles is based on a nonlinear triangulation technique. Images of the steel strip, at high temperature and high speed, are sampled every 2 ms at five different points and are processed on-line in order to calculate height displacement values of the strip, which allows the calculation of final flatness indexes for the steel strip. The measurement method developed introduces an innovative geometry in the disposition of the optical elements which increases the measurement range without reducing precision. It also includes a tracking system to compensate for the effects of lateral displacements of the strip. The flatness inspection system has been implemented using a heterogeneous distributed computer system.

This paper proposes a method for the diagnosis of roll eccentricity influence on the strip thickness at the exit of a finishing hot strip mill (FHSM). Each exit strip thickness defect is related to one roll of the FHSM, allowing the implementation of an optimal policy for the substitution and maintenance of the rolls, while maintaining the required quality level of the strip. This policy allows the minimization of roll changes and the concentration of several changes at the same time, reducing production costs. Fuzzy logic is used to compare spectra, looking for common patterns, which enables a totally-automated diagnostic system. Likewise, as part of the solution, a novel estimate of roll eccentricity, based on a least-squares algorithm has been developed, which provides higher accuracy than classical algorithms as well as a drastic reduction in the time required to perform eccentricity tests.

This paper presents a computationally scalable real-time observer, which allows the evolution of temperature distribution within slabs reheated in continuous furnaces to be tracked. Accurate tracking of skidmarks requires the use of two-dimensional models, which demand high computational power, generally provided by multiprocessor computers. To evaluate the behaviour of the observer under both typical and extreme operational conditions, a simulator of a complete furnace is presented. The emulation of extreme operational conditions using a real furnace would be impossible, due to the material waste and the cost of delaying normal production. This furnace simulator is the simplest method of generating the variables of the temperatures in the various furnace zones, which are fed to the observer during its operation in real time. Finally, in order to configure the observer, a simple and efficient graphic method is proposed. The method generates an optimum solution, minimizing the number of processors required to fulfil the specifications and maximize their utilization under all operational conditions.

The OMEGA project (Brite-EuRam 5470) intends to solve the problems arising from the lack of real time dimensional control on steel structural mill facility products. The main objective of the project is to assess the feasibility of an on-line, advanced passive non-contact measuring system, controlling complex geometry parts in adverse environment, featuring a high level of reliability and accuracy on measurements, with low cost. Such a measuring system will be based on Conoscopic Holography as the basic technology to be applied. Conoscopic Holography avoids some of the problems of conventional holography, such as the necessity of ultra high stability, coherent (but monochromatic) illumination and ultra high resolution sampling techniques. Thus, it is possible to produce holograms using incoherent light with a typical fringe period compatible with the resolution of normal electronic imaging devices, such as CCD cameras. In the application developed under this project the conoscopic head will have to meet requirements never tried before regarding the adverse working conditions. The paper will summarize the main objectives of the OMEGA project, now in progress, a description of the technology applied and the results available from the lab testing work when the paper is submitted.

We measure all the dimensions of one H profile of a rolling mill, reposed and cold, by means of using two matricial video cameras and a software package developed for this application. The software package develped identifies the intersection of the profile edges with a thin line drawn by a laser beam. We made the integral calibration of the system, and we get, in the worst case, an accuracy of millimetre tenths. The designed prototype has been validated for real applications in Ensidesa company. On the basis of this project we can attempt the measurement of these profiles, when hot and on the rolling mill.

The main objective of the article is to detail the current investigations to improve the control and supervision system of two identical walking-beam furnaces of a hot strip mill of the Ensidesa factory located in Aviles, Spain. The paper is structured as follows: in the first section the problem to be solved is introduced. The second part includes a brief description of the two furnaces. The following section is devoted to the development of a mathematical model for a generic continuous furnace. This represents the dynamic behaviour of the furnace, and it can be also used for obtaining of the steady-state behaviour of the furnace. These models allow the development and approximate evaluation of several control strategies before they are applied for online control. The next part is devoted to the development of observers for the furnace charge, based on simplifications of the dynamic model developed in the third section. At present, a very simple model is being implemented for on-line control. A comparative study between this simple observer and more elaborate ones is presented. Particularly, the on-line use posibilities of a more accurate two-dimensional slab model are analyzed. In the 5th section, the present control system is briefly described, performing an analysis of the possibilities offered by the new observer in order to close a control loop through them. In the next section the implementation of the mathematical model and some results of its application are illustrated. Finally, the conclusions obtained from this work are reviewed.

In finishing hot strip mills (FHSM), roll-eccentricity is one of the most important disturbances affecting the thickness quality of rolled strips. The economic issues related to roll-eccentricity are very important. Often, after the rolling of a strip, the mill operators find the thickness measured by the X-ray meter at the exit of the FHSM is not up to desired quality standards, with the cost that this implies. In this case, the operator must diagnose which stands have produced the thickness problems. The solution is to change the rolls in the problematic stand. The diagnosis of which stand is responsible for the defects must be fast and reliable, because meanwhile the mill must stop production. Furthermore, analyzing the historical values of individual stand influence, it is possible to prevent failures of quality thickness. A novel method of diagnosis based on observers and pattern matching using fuzzy technology has been developed. Nevertheless, the high computational cost of the solution requires multiprocessor-based equipment to accomplish the real-time constraints