摘要
在自行搭建的试验台上,对气化室风速、提升管风速、静床层高度和物料粒径对颗粒循环流率的影响进行试验分析,建立了3种优化的BP神经网络模型对颗粒循环流率进行预测,通过比较发现:基于遗传算法优化的BP神经网络预测颗粒循环流率时,其最大误差为6.934%,平均相对误差为1.107%,模型预测值与试验值较吻合,能较好地预测颗粒循环流率。
On the self-made test-bed, some factors influencing particle circulating flow rate were tested and analyzed, including gasification chamber air velocity, riser air velocity, static bed height and particle size. Three optimized BP neural network models were established to predict particle circulating flow rate. By comparison, it was shown that the BP neural network model based on genetic algorithm optimization can predict particle circulating flow rate with maxi- mum error of 6. 934% and average relative error of 1. 107%. The model predicted data well coincides with the tested value, so the model can better predict particle circulating flow rate.
出处
《华东电力》
北大核心
2012年第8期1399-1403,共5页
East China Electric Power
基金
国家自然科学基金项目(50876030)~~
关键词
循环流化床
颗粒循环流率
BP神经网络
遗传算法
circulating fluidized bed （CFB）
particle circulating flow rate
BP neural network
genetic algorithm
