Improving the production scheduling in a steel mill

Abstract

The aim of this thesis was to improve the production scheduling at the SSAB Raahe steel mill by identifying the causes of steelmaking scheduling issues. Scheduling issues can cause a reduction in casting speed, which can negatively impact the quality of the final product. Since the timing relating to the entire steel mill is a broad subject, the focus of this work was limited based on operator interviews. As a result, the accuracy and functionality of the recently implemented scheduling model were assessed. In addition, the current scheduling practices on the CAS-OB production route were analysed.

The accuracy of a scheduling model was analysed using certain error metrics and was improved in various ways. A BOF blow duration model was updated and an issue causing systematic error was identified and solved. The mean absolute error of the model has decreased by 56%. Also, a change was made into the casting time predictions. In addition, a proposal was made to change the operational approach of one of the most important timing-related parameters. This change was successfully implemented into production.

It was observed in this thesis that the scheduling-related issues on the CAS-OB production route are caused by limited processing time, due to short casting times and outdated work instructions that were insufficient to cover the current need for time buffering. Options for improving the CAS-OB scheduling are proposed and some of these options were tested in production environment. Based on this thesis, measures have been taken to implement one of the proposed options into production.