中文摘要：

提出了一种机理知识和最小二乘支持向量机（LS-SVM）相结合的球磨机料位混合软测量方法。基于机理模型建立料位理论模型;将料位机理特性融入LS-SVM的建模过程中,利用LS-SVM算法对机理模型进行校正,弥补它的不精确性。通过现场试验和实际运行分析表明,混合模型能够有效地反映出实际运行中料位的变化趋势和动态特性。

英文摘要：

A hybrid soft measurement based on mechanism knowledge and LS-SVM（Least Square Support Vector Machines）is proposed for fill level of ball mill.Theoretic fill level model is established based on the mechanism models and the mechanism knowledge is integrated into the modeling process of LS-SVM,which is used to correct the mechanism models for accuracy improvement.Experimental results and practical operation show that,the hybrid model represents effectively the dynamic performance and change tendency of fill level.