Insights into Structure-Activity Relationships between Y Zeolites and their n-C_(10)Hydrocracking Performances via Machine Learning Approaches  

基于机器学习的加氢裂化分子筛催化材料构效关系研究

作　　者：Qianli Ma Hong Nie Ping Yang Jianqiang Liu Hongyi Gao Wei Wang Songtao Dong 马千里;聂红;杨平;刘建强;高鸿毅;王薇;董松涛(中石化石油化工科学研究院有限公司,北京100083;北京科技大学材料科学与工程学院,北京材料基因工程高精尖创新中心,分子与微结构可控高分子材料技术北京市重点实验室,北京100083)

机构地区：[1]SINOPEC Research Institute of Petroleum Processing Co.,Ltd.,Beijing,100083,China [2]Beijing Advanced Innovation Center for Materials Genome Engineering,Beijing Key Laboratory of Function Materials for Molecule&Structure Construction,School of Materials Science and Engineering,University of Science and Technology Beijing,Beijing 100083,China

出　　处：《Chinese Journal of Catalysis》2025年第4期187-196,共10页催化学报(英文)

基　　金：国家重点研发计划(2021YFA1501204).

摘　　要：Hydrocracking technology represents a crucial position in the conversion of heavy oil and the transformation development from oil refining to the chemical industry.The properties of catalysts are one of the key factors in the hydrocracking process.As the main acidic component of hydrocracking catalyst,the influence of zeolite properties on the reaction performance has been the focus of research.In this study,a series of NiMo/Al_(2)O_(3)-Y catalysts were prepared using different Y zeolites as acidic components,and their performances in the hydrocracking of n-C_(10)were also evaluated.The structure-activity relationship between Y zeolite and the cracking performance of n-C_(10)was investigated with machine learning.First,a database of the physical and chemical properties of Y zeolite and their performance was established,and the correlation analysis was also conducted.Parameters such as the cell constant,acid content,acid strength,B/L ratio,mesopore volume,micropore volume of Y zeolite,and the reaction temperature were selected as independent variables.The conversion of n-C_(10)and the ratios of products C_(3)/C_(7)and i-C_(4)/n-C_(4)were selected as dependent variables.A model was established by the random forest algorithm and a new zeolite was predicted based on it.The results of model prediction were in good agreement with the experimental results.The R^(2)of the n-C_(10)conversion,C_(3)/C_(7)ratio,and i-C_(4)/n-C_(4)ratio were 0.9866,0.9845,and 0.9922,and the minimum root mean square error values were 0.0163,0.101,and 0.0211,respectively.These results can provide reference for the development of high performance hydrocracking catalyst and technology.加氢裂化是重油高效转化和炼油向化工高质量转型的关键支撑技术之一.在氢气(H_(2))、高温和催化剂的协同作用下,可将劣质重油转化为轻质清洁燃料油和有机化工原料.催化剂作为加氢裂化技术的核心,是影响反应活性、选择性和稳定性的关键因素之一.加氢裂化催化剂是典型的双功能催化剂,通常由金属活性中心(加氢功能)、酸性载体(裂化功能)协同作用,并辅以氧化铝和助剂优化性能.分子筛是裂化反应的核心载体,但其复杂结构导致孔分布、B酸量和酸强度等性能参数相互耦合,因此分子筛结构与加氢裂化反应性能的构效关系一直是领域的研究热点.本文利用机器学习揭示了Y型分子筛与正癸烷(n-C_(10))加氢裂化反应性能的构效关系.首先基于23个Y型分子筛样本及对应的92组NiMo/Al_(2)O_(3)-Y催化剂性能数据,构建了物化性质-反应性能关联数据库.通过特征相关性分析,筛选晶胞常数、酸含量、酸强度、B酸/L酸、介孔体积和微孔体积作为结构描述符(自变量),并以n-C_(10)转化率、产物C_(3)/C_(7)比(表征产物二次裂化程度)和i-C_(4)/n-C_(4)比例(表征产物异构化程度)为性能指标(因变量).采用随机森林算法构建预测模型,并通过训练集与预测集R²和均方根误差(RMSE)验证模型可靠性.结果表明,3种因变量预测集和训练集的R^(2)均接近1,相关性良好,训练集的n-C_(10)转化率、C_(3)/C_(7)和i-C_(4)/n-C_(4)的均方根误差(RMSE)均较低.实验发现,转化率主要受晶胞参数、酸量和介孔体积影响;C_(3)/C_(7)比主要受晶胞参数、酸量、介孔体积和微孔体积的影响;对i-C_(4)/n-C_(4)比影响较大的变量为晶胞参数、酸量和微孔体积.基于模型预测一种新型Y分子筛的性能,其n-C_(10)转化率、C_(3)/C_(7)比和i-C_(4)/n-C_(4)比的预测值与实验值R^(2)分别为0.9866,0.9845和0.9922,RMSE值分别为0.0163,0.101和0.0211,表明预测结果与实�

关 键 词：HYDROCRACKING Machine learning Y zeolite N-DECANE ACID Pore structure 

分 类 号：TE624.9[石油与天然气工程—油气加工工程]