三床串联化学链生物质气化实验分析与预测  

Experimental Analysis and Prediction of Biomass Gasification in Three-bed Series Chemical Chain

作　　者：陈鸿伟[1] 刘博朝 贾建东 刘玉强 梁锦俊 陈哲 冯旭鹏 朱楼 CHEN Hongwei;LIU Bochao;JIA Jiandong;LIU Yuqiang;LIANG Jinjun;CHEN Zhe;FENG Xupeng;ZHU Lou(School of Energy and Power Engineering,North China Electric Power University,Baoding 071003,Hebei Province,China)

机构地区：[1]华北电力大学能源动力与机械工程学院,河北保定071003

出　　处：《动力工程学报》2021年第8期667-673,699,共8页Journal of Chinese Society of Power Engineering

摘　　要：在自行搭建的三床串联化学链(TBS-CLC)生物质气化系统上研究了各控制参数(空气反应器风速、燃料反应器风速、循环物料质量和气化室床料高度)对循环流率的影响,将实验测得的129组数据作为训练样本和测试样本,利用反向传播(BP)神经网络模型、遗传算法(GA)优化的BP神经网络模型即GA-BP模型和支持向量机(SVM)模型预测循环流率。结果表明:循环流率随着燃料反应器风速增大而增大,但增幅逐渐变缓;循环流率随着空气反应器风速的增大、循环物料质量的增加、床料高度的减小而增大;GA-BP模型预测循环流率时,测试样本所得均方根误差为0.84110 kg/(m^(2)·s),平均绝对百分比误差为4.42%,预测结果与实验值较吻合,该模型能较好地预测循环流率。Based on a three-bed series chemical chain(TBS-CLC)biomass gasification experiment system,the effect of the control parameters on circulating flow rate was analyzed,such as air reactor wind velocity,fuel reactor wind velocity,circulating material mass and gasification chamber bed material level.129 test data were used to predict the circulating flow rate using the back propagation neural network model,Genetic algorithm optimized back propagation(GA-BP)neural network model and the support vector machine model.Results show that circulating flow rate increases with the increase of fuel reactor wind velocity,but the increase becomes gradually gentler.Circulating flow rate increases with the increase of air reactor wind velocity and circulating material mass and with the decrease of bed material level.The root mean square error predicted by the test sample is 0.84110 kg/(m^(2)·s),and the average absolute percentage error is 4.42%.The predicted value of the model are in good agreement with the test value,and the particle circulating flow rate can be well predicted by GA-BP model.

关 键 词：生物质 化学链燃烧 循环流率 控制变量法 GA-BP模型 

分 类 号：TK229[动力工程及工程热物理—动力机械及工程]