煤气化炉洗涤冷却室液滴夹带的研究及GA-BP预测模型  

Research and GA-BP neural network prediction on the entraining liquid problem in coal-gasifier scrubbing-cooling chamber

作　　者：赵晓辉[1] 王亦飞[1] 刘霞[1] 柴雪琴[1] 

机构地区：[1]华东理工大学大学煤气化教育部重点实验室,上海200237

出　　处：《计算机与应用化学》2010年第5期565-569,共5页Computers and Applied Chemistry

基　　金：国家重点基础研究发展计划(973计划)(2004CB217700);教育部"长江学者与创新团队发展计划"(PCSIRT-IRT0620)

摘　　要：新型洗涤冷却室是对置式多喷嘴新型水煤浆气化炉的重要组成部分。在工程运行中,可能出现带水问题,影响后续系统的稳定运行。本文研究水煤浆气化炉洗涤冷却室液滴夹带的问题,证明气速与分离空间的高度是影响液滴夹带量的2个重要因素。气速相同,分离空间越矮,液滴夹带量越大,分离空间的高度不变,气速大液滴夹带量越大。遗传算法和BP神经网络模型都适合解决非线性问题,但各有优缺点。实验数据与遗传算法优化BP网络技术,构造出新型洗涤冷却室内液滴夹带的预测模型,使网络的训练步数从186步减小到8步。测试样本检验相对误差较小在5%以内,预测能力较强,能满足工程需求,可作为工业中,液滴夹带的在线监测和自动控制。The entraining liquid problem in new type of scrubbing-cooling chamber had been experimentally researched, and the reaults indicated that gas velocity and height of separate space were the main factors that would influence the quantity of entrainment. In the condition of constant gas velocity, the quantity of entrainment increased with the separate space decreasing, in the condition of constant separate space it also increased with the gas velocity increasing. On the base of experimental data, a GA-BP neural network prediction model was established. Through the examination of test example, it could be proved that the error of the model was small and the ability to predict was strong. Therefore this model could be applied to online supervision and automatic control in industry operation.

关 键 词：洗涤冷却室 液滴夹带 GA-BP技术 神经网络 

分 类 号：TQ015.9[化学工程] O6-39[理学—化学]