Scholarlink

<p>The distribution of the flow field inside the oxygen-enriched side-blown furnace during the smelting process is complex, and analyzing it using the CFD method is time-consuming. This paper focuses on the gas–liquid two-phase flow in the side-blown smelting furnace. A dataset established through orthogonal experiments was used to extract flow field characteristics, and KNN (K-nearest neighbor) machine learning algorithm and matrix analysis methods were combined to establish a rapid prediction model for the velocity field in the molten bath. The results demonstrate that the liquid height has the greatest influence on the mean velocity magnitude within the high-speed zone, circulation zone, and stable zone, among the five zones of the furnace, accounting for approximately 50 pct of the total effect. The model can accurately predict the flow field, with mean absolute error (MAE) values less than 0.034, root mean square error (RMSE) values less than 0.075, and <i>R</i><sup>2</sup> values greater than 0.96.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Feature Extraction and Rapid Prediction of Multiphase Flow Field in a Copper Side-Blown Smelting Furnace Based on CFD Simulations

  • 期刊论文

  • 发表日期：

    2026年1月5日

查看全文
  • Jie Wang,
  • Wei Wang,
  • Hongliang Zhao,
  • Mingzhuang Xie,
  • Ke Din,
  • Liangjun Huang,
  • Fengqin Liu,
  • Hong Yong Sohn

摘要

The distribution of the flow field inside the oxygen-enriched side-blown furnace during the smelting process is complex, and analyzing it using the CFD method is time-consuming. This paper focuses on the gas–liquid two-phase flow in the side-blown smelting furnace. A dataset established through orthogonal experiments was used to extract flow field characteristics, and KNN (K-nearest neighbor) machine learning algorithm and matrix analysis methods were combined to establish a rapid prediction model for the velocity field in the molten bath. The results demonstrate that the liquid height has the greatest influence on the mean velocity magnitude within the high-speed zone, circulation zone, and stable zone, among the five zones of the furnace, accounting for approximately 50 pct of the total effect. The model can accurately predict the flow field, with mean absolute error (MAE) values less than 0.034, root mean square error (RMSE) values less than 0.075, and R2 values greater than 0.96.