矿井提升机罐笼纵向载荷冲击及其方位的辨识研究  

Research on identification of longitudinal load impact on mine hoist cage and its position

作　　者：杨超[1] 谭建平[1] 石理想[1] 薛少华[1] YANG Chao;TAN Jianping;SHI Lixiang;XUE Shaohua(School of Mechanical ＆ Electrical Engineering,Central South University,Changsha 410083,Hunan,China)

机构地区：[1]中南大学机电工程学院,湖南长沙410083

出　　处：《矿山机械》2018年第9期22-27,共6页Mining & Processing Equipment

基　　金：国家重点基础研究发展计划(973计划)项目"超深井大型提升装备设计制造及安全运行的基础研究"(2014CB 049400)之课题五"超深矿井提升系统全状态信息映射规律及智能感知研究"(2014CB049405)

摘　　要：为了实现对矿井提升机罐笼纵向载荷冲击及其方位的有效辨识,结合悬挂液压缸油压、罐笼提升张力、罐笼姿态等运行参数,提出了一种基于奇异值分解(SVD)与求方差比、相关系数、脉冲因子等特征参数的特征值提取方法,并以支持向量机(SVM)作为辨识模型。在SVD提取罐笼载荷冲击特征的基础上,针对冲击方位的辨识,加入罐笼提升张力的方差比、相关系数与罐笼姿态的脉冲因子,得到改进的特征向量,并将其作为特征值输入到SVM进行辨识。为提高SVM判别准确性,采用人工蜂群算法实现SVM参数寻优。试验结果表明,该方法能够有效判别矿井提升机罐笼纵向载荷冲击及其发生方位,平均准确率达到92%。In order to realize the effective identification of the longitudinal load impact on the mine hoist cage and its position, in combination with such operating parameters as the oil pressure of the suspended cylinder, the hoisting tension of the cage and the cage posture etc, the paper proposed an eigenvalue extraction method based on singular value decomposition（SVD）, variance ratio, correlation coefficient and pulse factor etc, and support vector machine（SVM） served as the identification model. On the basis of extracting the eigenvalue of the load impact on cage by SVD, in order to identify the impact position, the variance ratio and the correlation coefficient of the cage hoisting tension as well as the pulse factor of the cage posture were added to obtain the improved eigenvector that input into SVM. Meanwhile, to improve the identification accuracy of SVM, the artificial bee colony algorithm was applied to optimize the SVM parameters. Test results showed the method effectively identify the longitudinal load impact on the mine hoist cage and its position with average accuracy of 92%.

关 键 词：矿井提升机 载荷冲击 方位辨识 特征值提取 支持向量机 

分 类 号：TD531[矿业工程—矿山机电] TP391[自动化与计算机技术—计算机应用技术]