集成筒体振动信号的球磨机出力在线监测  

On-line monitoring method of ball mill pulverizing capacity integrated mill shell vibration

作　　者：王恒[1] 贾民平[2] 黄鹏[2] 陈左亮[3] 

机构地区：[1]南通大学机械工程学院,江苏南通226019 [2]东南大学机械工程学院,江苏南京211189 [3]大唐南京下关发电厂,江苏南京210011

出　　处：《江苏大学学报（自然科学版）》2012年第2期134-138,共5页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(50775035);江苏省自然科学基金资助项目(BK2011391);南通市应用研究计划资助项目(BK2011036)

摘　　要：针对球磨机出力在线监测的不足,提出了一种基于机理模型和黑箱模型相结合的球磨机出力软测量算法.首先,分析了制粉出力的影响因素,并通过与传统的轴承振动法比较,证实筒体振动信号更能反映料位,在此基础上,确定了出力混合模型的辅助变量;其次,基于机理模型获得球磨机出力方程和理论值,并基于最小二乘支持向量机算法对机理模型进行校正,弥补它的不精确性;最后,制粉出力的建模结果表明:该算法能够有效地反映出实际运行中出力的变化趋势和动态特性,比单纯机理模型具有更高的精度和适用性.To solve the disadvantage of on-line monitoring method for ball mill pulverizing capacity, a hy- brid soft sensor model was proposed based on mechanism knowledge and black-box model. The key fac- tors of pulverizing capacity were analyzed to compare vibration signals of mill shell with those of conven- tional mill bearing. The comparison reveals that the vibration signals of mill shell presents high change sensitivity on fill level. Pulverizing capacity equation and theoretic value were obtained based on mecha- nism models and corrected by LS-SVM model to improve accuracy of mechanism models. The modeling results show that the proposed hybrid model can effectively represent the change tendency and dynamic performances of pulverizing capacity with higher accuracy than that of single mechanism models.

关 键 词：球磨机 制粉出力 筒体振动 混合建模 在线监测 

分 类 号：TK223[动力工程及工程热物理—动力机械及工程]