Encore has cooperated with the University of La Rioja in the modelling of the amount of sulfur obtained during the manufacture of steel using an electric arc furnace. The development has been made using artificial intelligence techniques and data mining.
The objective of the project was to predict the percentage of sulfur obtained in the steel manufactured using an electric arc furnace. The process is complex and several variables take part in it such as the kind of scrap used, the consumption of energy or oxygen, additives, etc. Usually the workers are in charge of the control of the process applying an open loop control according to the state of the process and to their own experience. To obtain models that can improve the existing knowledge and the quality control of this process is at the moment under investigation.
For the development of the project, data obtained from a electric arc furnace manufacturing plant were used. Besides, the techniques used were the ones also used in data mining processes: preprocessing techniques, clustering and regression. Other techniques such as neural networks, vector model machines or decision trees were used to reach the desired aims.
Open Source R software was used. It offers a great variety of statistical and graphic techniques.