摘要
在回转窑实验台上对5种不同物性参数的固体废弃物料和窑内风速、回转窑转速、回转窑倾角改变的情况下,获得了物料MRT(Mean Residual Time)的变化规律:随着回转窑转速的提高、倾角增大和窑内风速增加,MRT减小;物料的物性参数中,休止角对MRT影响较大,休止角大的MRT小,而密度的变化影响相对较小。以上各影响因素对于MRT的敏感性差别较大:回转窑转速对MRT比较敏感;转窑倾角的敏感性则比较均匀;窑内风速在小回转窑转速、低窑内风速时对MRT较为敏感。针对回转窑内物料传输过程中影响因数众多、非线性机理强烈的特性,运用多层BP神经网络模拟了MRT与各因素之间的映射关系,建立了非线性传输模型,对模型中40组实验数据的验证结果显示,该模型预测值与实验结果较好吻合,平均相对误差为4.1%,能正确反应物料在回转窑内的传输过程。
Through experiments conducted on a rotary-kiln test rig and under the circumstances of rotary kiln speed and inclination angle changes the following law governing the variation of MRT(mean residence time) of materials for five kinds of solid waste materials with different physical-property parameters has been revealed: with an increase in rotary-kiln rotating speed and inclination angle as well as an air speed increase inside the kiln,the MRT will decrease.Among the physical-property parameters of the materials,the repose angle has a relatively great impact on the MRT.A greater repose angle will lead to a shorter MRT.The change of density,however,has a relatively minor effect.The various influencing factors mentioned above will give rise to a relatively big difference in sensitivity to the MRT.The rotary kiln speed is comparatively sensitive to the MRT,while the sensitivity of the kiln inclination angle assumes a relatively uniform character.As regards air speed in the kiln,a combination of low kiln speed and low air speed in the kiln is rather sensitive to the MRT.Taking account of the characteristics that there exist numerous influencing factors in the transmission process of materials in the rotary kiln and an intense nonlinear mechanism,a multi-layer BP neural network has been used to simulate the mapping relationship between the MRT and the various factors,establishing a nonlinear transmission model.The results predicted for 40 groups of experimental data in the model show that the values predicted by the model are in relatively good agreement with experimental results with an average relative error being assessed at 4.1%.This indicates that the model can correctly reflect the material transmission process in the rotary kiln.
出处
《热能动力工程》
CAS
CSCD
北大核心
2006年第4期409-413,共5页
Journal of Engineering for Thermal Energy and Power
基金
广东省自然科学研究团队基金资助项目(003045)
关键词
回转窑
BP神经网络
传输模型
平均停留时间
非线性
rotary kiln,BP neural network,transmission model,mean residence time(MRT),nonlinear
