Development of self-learning intelligent decision support system to control of steel production technological processes

Authors

  • I. Ziborov
  • T. Zheldak

DOI:

https://doi.org/10.34185/1562-9945-3-140-2022-04

Keywords:

metallurgy, information systems, control, optimization, self-learning, integration, decision support

Abstract

Taking to the consideration the current state of converter production and measuring equipment at Ukrainian enterprises, it follows that the smelting process is based on a complex dynamic non-deterministic system. The process is complicated by the large number of param-eters, the inability to accurately identify the state of the system at any time, as well as the dif-ficulty of forecasting system requirements. Preliminary analysis has shown that in the conditions of this production converter manufacturing efficiency increase can be reached at the expense of: - reducing the cost of raw materials, such as iron-containing additives, deoxidizers, non-metallic elements in steel; - reduction of melting time, especially blowing time; - reducing defects and improving product quality. It is proposed the architecture of integrated control DSS in converter steel production based on the principle of minimal interference in the production process. The primary aim of such a system is to predict the behavior of the production process, providing the recommen-dations for its impact in order to optimize the external criterion of efficiency. The source and amount of data required for the database formation and DSS knowledge base are substantiated. The mechanism of self-learning in the course of technological tasks is described. The structural scheme of self-learning DSS, self-learning algorithm, which is mainly featured with modularity, is offered in the paper. The approach allows testing of any number of existing algorithms for learning, forecasting and optimization in order to further select the most effective ones, modifies the system in the future and allows the parallel use of a number of com-peting algorithms. The operator has the opportunity to choose as a control solution one of the proposed systems, or the formation of its own, better by a certain external criterion of result quality. Based on the suggested software structure, a number of tasks are formulated that need to be performed to build a decision support system. It is also considered to apply the mathematical apparatus of fuzzy sets to describe certain pa-rameters of the technological process and quality criteria, fuzzy neural network for modeling reasoning processes and the choice of algorithm for its training.

References

Cherneha D.F. Osnovy metalurhiinoho vyrobnytstva metaliv i splaviv / D.F. Cher-neha, V.S. Bohushevskyi, Yu.Ya. Hotvianskyi, S.H. Hryshchenko, V.M. Shyfrin, V.Ya. Shurkhalo — K.: Vyshcha shk., 2006. — 503 s.

Nesterov M.E. Povyshenie effektivnosti ustarevsheho proizvodstva s pomoshchiu sovremennykh samoobuchaiushchykhsia system podderzhky pryniatyia reshenyi na prymere kyslorodno-konverternoho tsekha DMZ im. Petrovskoho / M.E. Nesterov, T.A. Zheldak // Zbirnyk naukovykh prats NHU – D.: NMU, 2010. – V.34, t.2. –

s. 202-207.

Boichenko B.M. Konverterne vyrobnytstvo stali (teoriia, tekhnolohiia, yakist stali, konstruktsiia ahrehativ, retsyrkuliatsiia materialiv i ekolohiia) / B.M. Boichenko, V.B. Okhotskyi, P.S. Kharlashyn – D.: RVA «Dnipro–VAL», 2004. — 454 s.

Novykova E. Yu., Mykhalev A. Y., Bublykov Yu. A. Nechetkaia identifikatsyia protsessa mykrolehirovaniia staly s karbonitridnym uprochneniem // Suchasni prob-lemy metalurhii: Naukovi pratsi. – D.: Systemni tekhnolohii, 2006.- S. 113-127.

Zheldak T.A. Systema pidtrymky pryiniattia rishen planuvannia vyrobnytstva ta kontroliu perebihu tekhnolohichnoho protsesu / T.A. Zheldak, D.M. Haranzha // 17-ta Mizhn. konf. «Avtomatyka-2010». Tezy dop. T. 1. – Kh: KhNURE, 2010. –

s. 212-214.

Us S. A. Modeli y metody pryiniattia rishen: navch. posib. / S. A. Us, L. S. Koriashkina; M-vo osvity i nauky Ukrainy, Nats. tekhn. un-t «Dniprovska politekhnika». – 2-he vyd. vypr. – D: NTU «DP», 2018. – 302 c.

Slesariev V.V. Intehrovani systemy keruvannia bahatoetapnym metalurhiinym vyrobnytstvom na prykladi prokatky trub / V.V. Sliesariev, T.A. Zheldak // Systemni tekhnolohii. – № 4 (75). – D., 2011. – s. 78–85.

Model keruvannia konverternym protsesom u systemi pryiniattia rishen / V. S. Bohushevskyi, S. V. Zhuk, K. O. Serheieva, M. V. Horbachova // Systemni doslidzhennia ta informatsiini tekhnolohii. – D.,2012. - № 4 - S. 35-44.

Barsehian A.A. Tekhnolohii analiza dannykh: Data Mining, Visual Mining, Text Mining, OLAP / A.A. Barsehian, M.S. Kupryianov, V.V. Stepanenko, Y.Y. Kholod // 2-e yzd., pererab. y dop. – SPb.: BKhV-Peterburh, 2007. – 384 s.

Novikova K. Yu. Matematychne modeliuvannia protsesiv metalurhiinoho vyrob-nytstva z vykorystanniam metodiv nechitkoho vyvodu: avtoref. dys... kand. tekhn. nauk : 01.05.02 / K. Yu. Novikova. – NMetAU. - D., 2008. - 20 s.

Witten, I. H. Data mining: practical machine learning tools and techniques.—3rd ed. / Ian H. Witten, Frank Eibe, Mark A. Hall. // Elsevier, 2011. –630 р.

Published

2022-04-08