摘要
双循环流化床生物质气化装置稳定运行的关键是合理控制颗粒循环流率。在双循环流化床冷态试验台上就鼓泡床风速、提升管风速、静床高和物料平均粒径几方面因素对颗粒循环流率的影响进行了系统的试验研究,并建立了加入动量的BP神经网络预测模型,对双循环流化床颗粒循环流率进行了有效模拟并得到了预测结果。定义了平均偏离度来评价模型预测值相对于试验值的平均偏离情况,通过对比分析试验数据与神经网络模型预测值,表明测试样本神经网络模型预测值相对于试验值偏差不超过0.8 kg·m-2·s-1,相对误差在±8%以内,平均偏离度仅为3.56%。结果表明建立的神经网络模型具有较好的预测效果。
The key to keep the double fluidized bed running steadily for the use of gasifying biomass is to make sure the control of solids circulation rate is appropriate.This paper made a systematic test on the double fluidized bed cold bench.It was mainly about the effect of several factors concerning gas velocity in the bubbling fluidized bed and riser,static bed height and particle size on the solids circulation rate and also established BP neural network with momentum added which gives an efficient simulation in the rate with a forecasting value received.For the purpose of giving a criterion to assess the average diversion of forecasting value from the tested,Mean Diversion Extent was defined.It shows that the diversion of forecasting value from the tested is not more than 0.8kg·m-2·s-1 with a relative error within ±8%,Mean Diversion Extent no more than 3.56% by means of making a comparison between the forecasting value and the tested.It proves that the BP neural network model has a better forecasting ability.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2010年第32期25-29,共5页
Proceedings of the CSEE
基金
国家自然科学基金项目(50876030)
高校博士点基金项目(20090036110008)~~
关键词
双循环流化床
提升管
鼓泡床
颗粒循环流率
BP神经网络
double fluidized bed
riser
bubbling fluidized bed
solid circulation rate
BP neural network
