Saint Petersburg State Technological Institute (Technical University), Saint Petersburg, Russia:

E. G. Egorova, Post-Graduate Student of a Faculty of Information Technologies and Management
L. A. Rusinov, Professor, Head of a Chair of Automation of Technological Processes and Productions, e-mail: lrusinov@yandex.ru

V. A. Trapeznikov Institute of Control Sciences of Russian Academy of Sciences, Moscow, Russia:

M. V. Usachev, Senior Researcher
M. Z. Salikhov, Senior Researcher

Sinter roasting with full or partial desulphurization is used in ore processing of zinc, lead and copper or copper-nickel ores and concentrates. The continuous monitoring and diagnostic system is proposed in order to increasing of high quality ore sinter output. The main concept of the proposed system is based on detection and prevention of early faults during the sinter roasting, reasoning and inference of recommendations to the plant's operator, avoiding further development of dangerous faults and emergencies. Proposed system is based on the two-level artificial neural network model. The upper level network is used for detection and localization of sinter roasting faults and abnormalities, while lower level networks are executed for identification of abnormal situations and their causes. Processing of archive data, containing the dynamics of development of all faults and abnormalities, is carried out at the preliminary stage of system operation. Collected data are filtered and normalized. Main algorithm includes two modes of diagnostic system operation. First mode includes training sets preparation and upper and lower network training. Second mode includes main diagnostic loop. Usage of the proposed diagnostic system includes the following
benefits:
— correct maintenance of technology;
— increasing of sintering output;
— reduction of technology and economic losses, caused by the production of ore sinter with inadequate quality.

1. Zhilkin V. P., Doronin D. N. Proizvodstvo aglomerata. Tekhnologiya, oborudovanie, avtomatizatsiya (Manufacturing of sinter. Technology, equipment, automation). Yekaterinburg : Ural Center of PR and Advertizing, 2004. 292 p.
2. Fan Xiaohui, Xuling Chen, Yi Wang. Expert System for Sintering Process Control. China, 2010. pp. 65–90.
3. Endiyarov S. V., Petrushenko S. Yu. Diagnostika protsessov podgotovki i proizvodstva aglomerata. Metody i modeli iskusstvennogo intellekta (Diagnostics of processes of preparation and manufacturing of sinter. Methods and models of artificial intellect). Germany : LAP LAMBERT Academic Publishing, 2013. 323 p.
4. Isermann K. Supervision, fault detection and fault-diagnosis methods. AIChE Journal. 1989. No. 35 (11). pp. 643–650.