基于泡沫视觉特征的浮选精矿品位的软测量

摘要

针对矿物浮选在线检测X荧光分析仪缺失、人工检测严重滞后的问题,依据泡沫视觉表面特征对浮选精矿品位具有直接指示作用的原理,采用最小二乘支持向量机(Least Squares Support Vector Machine, LS-SVM)实现浮选精矿品位的软测量。工业运行数据仿真结果表明,建立的软测量模型能够实现精矿品位的预测精度,满足工业要求。%Aimed at absence of on-line detection X fluorescence analyzer for mineral flotation and serious delay by manual detection, based on the principle of visual feature of froth directly indicating the grade of flotation concentrate, this paper give a discussion on using least squares support vector machine to achieve soft-sensing of the grade of flotation concentrate. Industrial operation data simulation result indicates that the established soft-sensing model can satisfy the precision of prediction of concentrate grade to meet the industrial requirement.