基于多神经网络的磨矿粒度软测量模型  

Multiple Neural Networks-based Soft Sensing Model for Particle Size of Grinding Product

作　　者：何桂春[1] 毛益平[2] 倪文[1] 

机构地区：[1]北京科技大学 [2]马鞍山矿山研究院

出　　处：《金属矿山》2005年第2期47-49,共3页Metal Mine

摘　　要：讨论了多神经网络的建模思路及其实现,针对某选矿厂实际一段磨矿分级过程建立了相应的多神经网络粒度软测量模型,并进行了仿真试验。结果表明,模型具有较高的精度,能成功地对磨矿产品粒度进行估计。The thought and implementation of modelling based on multiple neural networks are discussed. A multiple neural networks-based soft sensing model was established and the simulation test was made according to the real primary grinding and classification process in a concentrator. The results showed that the model had a high accuracy in its successful estimation of the particle size of the grinding product.

关 键 词：磨矿粒度 一段磨矿 选矿厂 分级 矿产品 精度 神经网络 软测量模型 建模 仿真试验 

分 类 号：TD951.1[矿业工程—选矿] TD952