ABSTRACT
With the extensive application of integrated automation system, it can obtain not only the macroscopic parameters which describe the overall quality of steel products, but also the microscopic parameters which indicate the parameter changes in the productive process. In high-precision quality control, we should assess the overall quality parameters whether reach the target as well as each technological parameter whether meets the set requirements in the whole manufacturing process as well. For example, for hot rolling steel products, the overall parameters can be utilized such as average thickness, maximum thickness deviation and meanwhile its thickness tendency along the length direction should not be ignored. Nowadays, the data warehouse model of iron and steel industry is more imitation of other commercial and service
1 INTRODUCTION
Metallurgical industry is a typical batch process oriented industry. Products which meet the requirements of customs will be finally formed due to sequentially processing the products in different procedures. The construction of information system has been conducting one after another by Chinese iron and steel enterprises since 2002. After many years’ development, most iron and steel enterprises have built up their own independent information system (Zhang et al. 2010), which is naturally accompanied by the birth of massive amount of data. However, the information system itself has no way to automatically manage the data, forming lots of so called “Information isolated island” (Huang et al. 2012) among production equipments. How to organize and make full use of these data to analyze and improve product quality is a big problem for iron and steel enterprise. Nevertheless, process oriented enterprise and data in complicated technological procedure are usually multivariable with the characteristics of strong coupling, nonlinear, time-varying and uncertainty, which make it more difficult to build an exact mathematic model to forecast and control products quality.
