Construction of mathematical models to predict M25 and M10 coke quality indices

In this paper, topological-mathematical models based on multilineal regression analysis have been built as a model of the degradability of 26 alkylphenols through the Chemical Oxygen Demand (COD) and Biochemical Oxigen Demand (BOD5). Two models with three-variable were selected (r2= 0.8793 and q2=0.8075 for log(1/COD) and r2= 0.8928 and q2=0.8327 for log(1/BOD5). The models were validated by cross-validation, internal validation and randomization tests. The results, which stand in good accordance with the obtained results, confirm the robustness of the method.

Download Full-text

Methodology of Developing Mathematical Models with Fuzzy Logic Elements for Quality Indices Control

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.436.374 ◽

2013 ◽

Vol 436 ◽

pp. 374-381 ◽

Cited By ~ 1

Author(s):

Alexey Korchunov ◽

Mikhail Chukin ◽

Aleksandr Lysenin

Keyword(s):

Fuzzy Logic ◽

Mathematical Models ◽

Preference Relation ◽

Fuzzy Relation ◽

Quality Indices ◽

New Models ◽

Acting Process ◽

Metal Products ◽

Process Operations ◽

Fuzzy Logic Theory

Advantages of applying fuzzy logic theory to metal products quality indices control in development of new models and in improvement of acting process operations are shown. It is proved that it is appropriate to determine fuzzy relation as preference relation in process of handling products quality indices in process operations. Elaboration of algorithm of handling mathematical models with fuzzy logic elements to control quality indices is undertaken. Methodology of mathematical models development with fuzzy logic elements for metal products quality indices control is created. Process of metal products quality indices control on the basis of models with fuzzy logic elements is illustrated.

Download Full-text

Forecasting of industrial coke quality at JSC EVRAZ NTMK based on data of passive industrial experiment. Report 1. Forecasting of CSR and CRI of industrial coke

Ferrous Metallurgy Bulletin of Scientific Technical and Economic Information ◽

10.32339/0135-5910-2021-11-1124-1135 ◽

2021 ◽

Vol 77 (11) ◽

pp. 1124-1135

Author(s):

Yu. A. Zolotukhin ◽

N. A. Berkutov ◽

V. V. Kuprygin ◽

S. N. Kupriyanova

Keyword(s):

Mathematical Models ◽

Coke Quality ◽

Influence Factor ◽

Methodological Approach ◽

Coke Production ◽

Industrial Experiment ◽

The Matrix ◽

Report Data ◽

Coke Batteries ◽

Dry Quenching

A forecast of coke quality takes a special place in the coke production, since it enables to increase efficiency of management of batching process of various by composition and properties coals and production of coke of stably high and required quality with minimal costs. Description of a methodological approach to processing of passive industrial experiment data of blends coking at Coke production department of JSC EVRAZ NTMK presented by application selective (general) matrix. The matrix accounts various multilevel values of influence factor CSR and CRI of coke - a complex index of coking ability of blends K.п.к.Vo. It was shown that the proposed approach provides wide variations of response function (CSR/CRI) at symmetrical enough matrix, excluding predominance of any particular area of values of indices K.п.к.Vo and CSR/CRI. By applying the passive industrial experiment, based on processing of actual report data of industrial blends coking at the coke batteries No. 5-6 (wet quenching) and No. 9-10 (dry quenching) by selective matrix, mathematical models of forecast of quality of industrial coke by wet and dry quenching (CSR/CRI) were elaborated depending on coal charges properties (K.п.к.Vo) at the existing modes of their preparation and coking at the coke batteries No. 5-6 and 9-10. Verification of accuracy of the mathematical models of coke quality forecast at wet and dry quenching (CSR/CRI) processing a large actual material of industrial coking (62 coking operations in the coke batteries No. 5-6 and 58 coking operations in the coke batteries No. 9-10 showed accuracy good enough for practical application of forecasting indices CSR and CRI of industrial coke of wet and dry quenching.

Download Full-text

Methodology of Developing Mathematical Models for Quality Indices Control

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.598.643 ◽

2014 ◽

Vol 598 ◽

pp. 643-646 ◽

Cited By ~ 2

Author(s):

Marina Polyakova ◽

Alexey Korchunov

Keyword(s):

Mathematical Models ◽

Production Quality ◽

Control Parameters ◽

Quality Indices ◽

Hierarchical Tree ◽

Input And Output ◽

Structure Of Metal ◽

Different Types ◽

Metal Products ◽

Hardware Industry

It is proposed to set the structure of metal products quality as a hierarchical tree of properties. Input and output variables and also control parameters of technological processes are represented as universal multitudes in formalization of the process of production quality indices handling in hardware industry technologies. Presence of simultaneously diversified information initiating different types of independence, is characteristic of solving problems of metal products quality indices control.

Download Full-text