A hybrid machine learning framework for coke quality prediction and coal blend optimization

High-quality metallurgical coke production requires accurate quality prediction and efficient coal blend optimization. This study developed a hybrid machine learning framework integrating GBDT, SVR, and WOA for coke quality prediction and coal blend design. GBDT identified influential coal characteristics and reduced input redundancy. Ash content was the dominant factor, contributing more than 65% to all predicted quality parameters. The selected variables were used in SVR models to predict pore number, porosity, equivalent pore diameter, and CCR. The models achieved testing R 2 values of 0.587–0.816. WOA subsequently optimized coal blending ratios under quality and operational constraints. The optimal blend comprised 25% coal A, 45% coal B, and 30% coal C, with a predicted CCR of 31.1%. Experimental validation confirmed model reliability. Coke produced using a 3-mm particle size exhibited the best performance, with a CCR of 28.3%, pore number of 498, and equivalent pore diameter of 82.4 μm. Overall, the proposed GBDT–SVR–WOA framework provides an accurate and practical approach for coke quality prediction and coal blend optimization.

Authors

Institutions

Publication Details

Journal
Mining Technology Transactions of the Institutions of Mining and Metallurgy
Published
2026-11-20
DOI
https://doi.org/10.1177/25726668261509079
Primary Topic
Coal and Coke Industries Research
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A hybrid machine learning framework for coke quality prediction and coal blend optimization

Farhang Sereshki, Mohammad Ataei, Mohammad Reza Ameri Siahuei
3 links
Mining Technology Transactions of the Institutions of Mining and Metallurgy
Coal and Coke Industries Research
article

A hybrid machine learning framework for coke quality prediction and coal blend optimization

Farhang Sereshki, Mohammad Ataei, Mohammad Reza Ameri Siahuei
article en

Abstract

High-quality metallurgical coke production requires accurate quality prediction and efficient coal blend optimization. This study developed a hybrid machine learning framework integrating GBDT, SVR, and WOA for coke quality prediction and coal blend design. GBDT identified influential coal characteristics and reduced input redundancy. Ash content was the dominant factor, contributing more than 65% to all predicted quality parameters. The selected variables were used in SVR models to predict pore number, porosity, equivalent pore diameter, and CCR. The models achieved testing R 2 values of 0.587–0.816. WOA subsequently optimized coal blending ratios under quality and operational constraints. The optimal blend comprised 25% coal A, 45% coal B, and 30% coal C, with a predicted CCR of 31.1%. Experimental validation confirmed model reliability. Coke produced using a 3-mm particle size exhibited the best performance, with a CCR of 28.3%, pore number of 498, and equivalent pore diameter of 82.4 μm. Overall, the proposed GBDT–SVR–WOA framework provides an accurate and practical approach for coke quality prediction and coal blend optimization.

Mining Technology Transactions of the Institutions of Mining and Metallurgy
University of Shahrood (IR)
Affordable and clean energy, Responsible consumption and production, Industry, innovation and infrastructure
Coal and Coke Industries Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon