双循环流化床颗粒流率实验研究及模型预测  

Experimental Study and Model Predict of Particle Flow Rate in a Dual Circulating Fluidized Bed

作　　者：于伟锋 

机构地区：[1]航天长征化学工程股份有限公司兰州分公司,甘肃兰州730010

出　　处：《锅炉技术》2016年第1期39-44,共6页Boiler Technology

摘　　要：自行设计搭建了双循环流化床冷态装置,通过单一变量法实验,研究气化室风速、提升管风速、物料粒径及物料量对锥形布风板双流化床颗粒循环流率的影响,得出改良后的锥形布风板颗粒流率要高于水平布风板,并且颗粒循环流率随着气化室风速、提升管风速、物料量的增加有所增大,随着物料粒径的增大而减小。利用Matlab建立了3种改进BP神经网络模型分别预测双流化床颗粒循环流率,得出含有1个隐含层,26个神经元节点数的LM-BP神经网络对锥形布风板双循环流化床颗粒循环流率具有较好的预测效果。A cold dual circulation fluidized bed gasification system was build, and a conical gas distributor was used in the gasification chamber, and effects of gas velocity to the gasifi- cation chamber and the riser, material weight, particle size on solids circulation rate was systematically researched, and the result was compared with that under a planar distributor. The results indicate that solids circulation rate increases with increasing gas velocity in the two beds, With the rise of material weight and particle size, solids circulation rate increases and decreases respectively. By comparison, results show that solids circulation rate in a coni- cal distributor is higher than that in a planar distributor. Using Matlab, the paper established 3 improved BP neutral network prediction models, and the best model was found by comparison. When LM-BP neural networks contains a hidden layer, 26 neurons nodes can better predict the solids circulation rate.

关 键 词：双流化床 颗粒循环流率 BP神经网络 

分 类 号：TK229.66[动力工程及工程热物理—动力机械及工程]