Scholarlink

<p>The concentration of Al<sub>2</sub>O<sub>3</sub> in the aluminum electrolysis cell was a key factor affecting the safe and stable production of aluminum electrolysis. In this work, the parameters were screened based on the relationship between the Al<sub>2</sub>O<sub>3</sub> concentration of the 400 kA aluminum electrolytic cell and the related process parameter data firstly, and then a Back Propagation neural network prediction model optimized by Improved Quantum Genetic Algorithm was developed for Al<sub>2</sub>O<sub>3</sub> concentration in a 400 kA aluminum electrolysis cell. By using process parameters such as cell current, voltage, electrolyte temperature, aluminum level, and electrolyte level, the model's complexity was reduced, thereby enhancing operational efficiency. Compared to traditional Genetic Algorithm, the Improved Quantum Genetic Algorithm optimization improved the model's accuracy, with the coefficient of determination increasing from 0.8130 to 0.9274 and the relative mean error decreasing from 2.01% to 1.73%.</p> Graphical Abstract <p></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

Prediction of Al2O3 concentration in aluminum electrolytic cells by BP neural network based on improved quantum genetic algorithm optimization

  • 期刊论文

  • 发表日期：

    2025年2月3日

查看全文
  • Hesong Li,
  • Jianwen Wang,
  • Zixun Cao,
  • Shilin Zhao,
  • Zhiqiang Sun

摘要

The concentration of Al2O3 in the aluminum electrolysis cell was a key factor affecting the safe and stable production of aluminum electrolysis. In this work, the parameters were screened based on the relationship between the Al2O3 concentration of the 400 kA aluminum electrolytic cell and the related process parameter data firstly, and then a Back Propagation neural network prediction model optimized by Improved Quantum Genetic Algorithm was developed for Al2O3 concentration in a 400 kA aluminum electrolysis cell. By using process parameters such as cell current, voltage, electrolyte temperature, aluminum level, and electrolyte level, the model's complexity was reduced, thereby enhancing operational efficiency. Compared to traditional Genetic Algorithm, the Improved Quantum Genetic Algorithm optimization improved the model's accuracy, with the coefficient of determination increasing from 0.8130 to 0.9274 and the relative mean error decreasing from 2.01% to 1.73%.

Graphical Abstract