Scholarlink

The preparation of raw slurry is the first process in aluminaAlumina production, and the composition of bauxite determines its material ratio. Real-time and online analysisOnline analysis of bauxite compositionBauxite composition can guide the material ratio in a timely manner, reduce the fluctuation of process indicators in aluminaAlumina production, and thereby reduce costs and improve product quality. In this study, based on laser-induced breakdown spectroscopySpectroscopy (LIBSLIBS) technology, spectral data of bauxite was collected, pretreated, and modeled into a dataset for element analysis. An analysis model was then established using machine learningMachine learning. The mean relative error of composition analysis of Al2O3, Fe2O3, and SiO2 was less than 10%. The results showed that the real-time and online analysisOnline analysis of bauxite compositionBauxite composition can be realized by combining LIBSLIBS and machine learningMachine learning technology, which laid the foundation for the realization of intelligent controlIntelligent control of charge mixture, and accelerated the intelligence process of aluminaAlumina production process.

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

Combine LIBS and Machine Learning to Realize Real-Time and Online Analysis of Bauxite Composition

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yanfang Zhang,
  • Lijuan Qi,
  • Long Duan,
  • Qiaoyun Liu,
  • Jun Wu,
  • Shuai Shao,
  • Hui Pan

摘要

The preparation of raw slurry is the first process in aluminaAlumina production, and the composition of bauxite determines its material ratio. Real-time and online analysisOnline analysis of bauxite compositionBauxite composition can guide the material ratio in a timely manner, reduce the fluctuation of process indicators in aluminaAlumina production, and thereby reduce costs and improve product quality. In this study, based on laser-induced breakdown spectroscopySpectroscopy (LIBSLIBS) technology, spectral data of bauxite was collected, pretreated, and modeled into a dataset for element analysis. An analysis model was then established using machine learningMachine learning. The mean relative error of composition analysis of Al2O3, Fe2O3, and SiO2 was less than 10%. The results showed that the real-time and online analysisOnline analysis of bauxite compositionBauxite composition can be realized by combining LIBSLIBS and machine learningMachine learning technology, which laid the foundation for the realization of intelligent controlIntelligent control of charge mixture, and accelerated the intelligence process of aluminaAlumina production process.