【作者】 唐艳；

【导师】 李建伟；

【作者基本信息】 北京化工大学 ， 化学工程， 2009， 硕士

【摘要】 以MCM-22分子筛为对象,系统考察了挤条成型工艺条件对分子筛催化剂物理化学特性的影响,并以苯与丙烯液相烷基化合成异丙苯为模型反应,在实验研究基础上、结合人工神经网络建模方法,建立了MCM-22分子筛催化剂性能预测的BP（Back Propagation）神经网络模型,该模型关联了催化剂本征性能、工艺条件和反应产物分布之间的相互关系,以其达到较好地预测苯与丙烯烷基化反应性能的功能,为催化剂的开发和工业生产提供应用基础和参考。利用颗粒强度测定、NH₃-TPD酸性表征和氮吸附脱附—压汞方法联合测试等手段对不同条件下成型的MCM-22分子筛物理化学特性进行了表征,系统考察了不同成型工艺条件对催化剂颗粒强度、孔结构和酸性等性质的影响。结果表明:成型过程中适宜助剂的加入,可以调节催化剂颗粒强度、比表面积和孔径分布以及酸性。其中,粘结剂SB粉主要调节催化剂颗粒强度和酸性,扩孔剂PEG20000主要调变孔径分布以及孔容。对不同成型条件下得到的16种织构性能各异的MCM-22分子筛催化剂性能进行实验评价,将所获得的数据用于模型训练和预测结果检验。结果表明,所建立的BP神经网络模型具有进行实验数据拟合和在一定范围内预测未知结果的能力,预测平均相对误差为4.21%。因此,将该BP神经网络模型作为MCM-22分子筛催化剂的性能预测和苯与丙烯液相烷基化过程的定量描述模型,是适宜和可靠的。同时,应用该模型回归分析了温度、空速、苯烯比及分子筛的比表面积、平均孔径与产物选择性之间的相关关系。分析结果表明:丙烯空速、苯烯比、温度及分子筛的比表面积、平均孔径与异丙苯的选择性呈正相关关系,与二异丙苯和三异丙苯的选择性呈负相关关系。更多还原

【Abstract】 The effect of extrusion molding condition on the physical and chemical properties of MCM-22 zeolite had been systematically investigated.And taking the liquid alkylation of benzene with propylene to produce cumene as model reaction,combined with the artificial neural networks（ANN） modeling method,Back-Propagation neural networks model was developed for the performance prediction of MCM-22 zeolite catalyst.The intrinsic performance of MCM-22 zeolite,the operation condition,and the product distribution were considered into this model,in order to predict the catalytic preformance of benzene alkylation with propylene and provide an applied reference for the catalyst development and industrial production.The MCM-22 zeolites prepared under different molding condition were characterized using particle strength apparatus,NH₃-TPD acidity characterization,N₂ adsorption-desorption instrument and mercury penetration instrument to investigate the effects of molding conditions on mechanical strength,pore structure,acidity and other physical-chemical properties of MCM-22 zeolite catalyst using single factor experiment method.The results showed that in the molding process,suitable amount auxiliary materials could adjust the properties of catalyst.SB powder mainly adjusted the partical strength and acidity,PEG20000 regulated pore distribution and pore volume.By means of the extrusion molding method,the catalytic performances of 16 kinds of MCM-22 zeolite catalysts with different textile properties were also experimentally investigated.According to these experimental data,the ANN model developed here was trained and tested.The results showed that the ANN model developed here had a very well simulation capability and the prediction results met with the experimental data,its average relative prediction error was only about 4.21%.Therefore,the ANN model established here was suitable and reliable when it was used to predict the catalytic performance of MCM-22 zeolite and simulate the alkylation result of benzene with propylene over MCM-22 zeolite.At the same time,the relation between some operating parameters and products selectivities had been regression analyzed,such as tempreture,space velocity,Benzene/Propene ratio,specific surface and average pore radius of MCM-22 zeolite etc.The result showed that space velocity,Benzene/Propene ratio,specific surface,average pore radius and tempreture were positive correlation with IPB and negative correlation with DIPB,TIPB.更多还原