Case-based reasoning method based on mechanistic model correction for predicting endpoint sulphur content of molten iron in KR desulphurization

ABSTRACT

In order to improve the control on sulphur content at the endpoint of Kanbara Reactor (KR) desulphurization process, the case-based reasoning (CBR) method based on mechanistic model correction is used to predict the endpoint sulphur content of molten iron. First, the KR desulphurization process is analysed to determine its kinetic mechanistic model, and partial derivatives are obtained for different attributes. Then, according to the analysis ofthe attributes, the corrected model is determined by the method of selecting the attributes, determining the reasonable weights and fitting the calculation results. Finally, the CBR method, the Back Propagation Neural Network (BPNN) and the corrected model are used to predict the endpoint sulphur content at KR desulphurization. The experiment results indicate that the prediction accuracy of corrected model is significantly higher than that of BPNN and mechanistic model.

KEYWORDS:

• Case-based reasoning
• kinetic model
• endpoint sulphur content
• Kanbara Reactor
• desulphurization

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This is financially supported by National Key Research and Development Program (No. 2016YFB0601301), National Natural Science Foundation of China (No. 51574032) and Fundamental Research Funds for the Central Universities (No. FRF-TP-16-081A1).