基于人工神经网络的水力旋流器选型及优化  

Select and Optimizate Hydrocyclone Based on Artificial Neural Network

作　　者：冯建成[1] 陈秋波[1] 杜长江[2] 

机构地区：[1]海南大学环境与植物保护学院,海口570228 [2]中国矿业大学(北京)化学与环境工程学院,北京100083

出　　处：《煤矿机械》2012年第11期69-70,共2页Coal Mine Machinery

摘　　要：为了实现对水力旋流器的全面设计,建立了3层BP神经网络模型,该模型可根据分离粒度、生产能力、底流质量浓度等值,选择合适的水力旋流器。经10组数据测试,选型误差为:底流口直径10.43%,溢流口直径7.51%,插入深度17.86%,入料压力20.24%,选型精度高于传统方法。该模型既可用于设备选型,也可用于优化旋流器参数。选择合适的水力旋流器分级加重质,制备得到的粗、细两产品分别满足湿法、干法对加重质要求,对我国选煤业发展有重大意义。In order to design hydrocyclone comprehensively,this paper establishes three-layer BP neural network model, and it can select right hydrocyclone after giving granularity,production capacity and concentration of underflow. After test of 10 samples, result of selection error： underflow diameter is 10.43%, overflow diameter is 7.51%, the insertion depth is 17.86%, feeding pressure is 20.24%, and precision is higher than that of traditional selection method. The network can not only select right hydrocyclone, but also be used to optimize hydrocyclone parameters on site. Select appropriate hydraulic cyclone to prepare magnet powder, and the coarse and fine product can adapt to wet and dry coal preparation, and it is important to development of coal preparation.

关 键 词：BP神经网络 磁铁矿粉 选型 水力旋流器 

分 类 号：TD455[矿业工程—矿山机电]