Go to content Go to navigation Go to search

Prediction of Chemical Composition of Refining Slag With Exploitation of Artificial Neural Networks

Zora Jančíková, Pavel Švec


Contribution deals with application of artificial neural networks for prediction of chemical composition of refining slag with aim to optimize production process and to improve production economics. The first part of the paper contains the evaluation of chemical parameters of the selected kinds of ladle slag that are formed during the ladle processing of steel at the ArcelorMittal Ostrava Company. Main interest is paid to metallurgical parameters of test heats and their comparison with the common technology. The theoretical calculations will enable to save part of metallurgical lime and Al2O3 flux additions. Exploitation of neural networks is advantageous, if it is necessary to express complex mutual relations among sensor-based data. More accurate results of predictions of different metallurgical parameters with exploitation of neural networks follow on the fact that application of neural networks enables assignment of relations among process parameters which are not possible to be traced using common methods due to their mutual interactions, considerable amount of data, dynamics and thus following time demands.


Neural networks, prediction, steel, chemical composition, slag.
Full paper


commenting closed for this article

Grammatical Evolution For Hardware Optimization Planning Smooth Trajectories in the Plane with Obstacles