基于BP神经网络的白云母超细磨工艺参数优化
Optimization of muscovite ultrafine grinding process parameters based on BP neural network
田家怡 1彭祥玉 1张帅 1王宇斌 1赵鑫 1肖巍1
作者信息
- 1. 西安建筑科技大学 资源工程学院,陕西 西安 710055
- 折叠
摘要
为提高白云母超细磨的效率,采用皮尔森相关系数分析超细磨效率与各参数之间的敏感性,并建立BP神经网络模型对白云母的超细磨正交试验参数进行优化.研究结果表明:各参数对超细磨效率敏感程度由大到小依次为瓷球级配、搅拌速率、助磨剂种类、超细磨时间和助磨剂用量.利用BP神经网络优化后的工艺参数进行超细磨试验,可获得-13 μm粒级质量分数为 83.04%的白云母,与正交试验最佳点相比提高了 2.19%,所建模型可提高白云母超细磨效率,且预测精度较高.研究结论为超细白云母粉体的高效制备提供参考.
Abstract
To improve the efficiency of ultrafine grinding of muscovite,the sensitivity between the ultrafine grinding efficiency and parameters was analyzed by using Pearson's correlation coefficient and the parameters of ultrafine grinding orthogonal test of muscovite was optimized based on a BP neural network model.The reseach results show that the sensitivity of each parameter to the ultrafine grinding efficiency is in the order of ceramic ball gradation,stirring rate,grinding aid type,ultrafine grinding time and grinding aid dosage.The muscovite with a mass fraction of 83.04%at-13 μm can be obtained using the process parameters optimized by the BP neural network for the ultrafine grinding test,which is increased by 2.19%compared with optimum conditions of orthogonal test.This model can improve the efficiency of ultrafine grinding of muscovite,and the prediction accuracy is high.The research conclusions provide a reference for the efficient preparation of ultrafine muscovite powder.
关键词
白云母/超细磨/正交试验/BP神经网络/参数优化Key words
muscovite/ultra-fine grinding/orthogonal test/BP neural network/parameter optimization引用本文复制引用
基金项目
国家自然科学基金项目(52004197)
中国博士后科学基金项目(2023M732746)
出版年
2024
国家科技期刊平台
NSTL
万方数据