摘要
通过对磨矿过程原理的分析,提出了基于径向基(RBF)网络的软测量方法,对以往无法在线检测的球磨机关键参数———介质填充率进行实时估计.针对现有软测量辅助变量选择方法存在的问题,结合灰关联分析理论,提出了一致关联度算法,并应用该算法实现了介质填充率软测量建模中辅助变量的选择.仿真结果表明,该软测量方法估计的平均相对误差保持在0.7%以下,可满足实际需要.同时,一致关联度算法可以准确的提取变量间的相关性,是一种有效的辅助变量选择方法.
Through analyzing the ball milling process, and based on radiate basis function （RBF） network, a soft sensor method was presented to estimate the ball mill＇s charge ratio of media （CRM） on line, which couldn＇t be measured on line in the past. Combining grey relation analysis, a uniform-incidence-degree arithmetic was proposed to determine the secondary variables of soft sensor. Simulation results show that the mean relative error of the soft sensing method is less than 0.7% and can meet the practical demands. Meanwhile, uniform incidence degree arithmetic can exactly obtain the correlation between variables, which is an effective way to select secondary variables.
出处
《中国矿业大学学报》
EI
CAS
CSCD
北大核心
2006年第4期549-554,共6页
Journal of China University of Mining & Technology
基金
国家科技攻关计划项目(2001BA204B01)
关键词
介质填充率
径向基网络
软测量
一致关联度
charge ratio of media
radiate basis network
soft sensor
uniform incidence degree
