项目名称: 石油化工中非均相反应过程建模与优化运行的关键技术研究
项目编号: No.21206037
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 化学工程及工业化学
项目作者: 叶贞成
作者单位: 华东理工大学
项目金额: 25万元
中文摘要: 石油化工中的非均相反应过程具有反应原料组分成百上千、反应数目多、 反应体系具有高度非线性和耦合、催化剂活性时变性强等复杂性。仅采用传统的"三传一反"机理的建模难于满足过程优化和控制对模型精度的要求。与此同时,目前流行的微观研究方法其模型一般都具有大规模和非线性特征,模型求解困难,无法满足工业装置操作优化对模型求解实时性的要求。本课题拟研究能同时兼顾一定计算精度和求解实时性的工业非均相反应器建模关键技术。即,工业运行数据的信息融合技术研究,实现工业数据的去伪存真,为过程建模提供可靠的模型输入数据;研究基于非均相反应机理与工业运行数据多尺度耦合参数的混合智能建模技术,多源信息挖掘的催化剂活性建模方法,实现反应与催化剂失活模型的集成方法。对现有的粒子群算法进行改进,期望找到一种同时具备随机性和启发性优点的混合进化算法,能够实现对模型的全局寻优,将其用于已建的模型进行参数优化,并应用于实际生产。
中文关键词: 非均相催化;数据挖掘;失活;神经网络;智能优化
英文摘要: The complex of heterogeneous reactions in petrochemical productions focus on multicomponent of reaction process, the enormous number of reaction, highly non-linear and high system coupling, the active decrease of catalyst being time dependence respectively. In conventional, the model precisions of heterogeneous reactor which based on the heat-transfer, mass-transfer and momentum-transfer is not satisfied the requirement of process optimization and control. At the same time, the prevalent research based on micro method of the heterogeneous reactor is too large in scale and highly non-linear characteristic to being satisfied the requirement for industrial application, result form time-consuming of solving the model. In this project, we are aim at developing a method which has a balance between the prediction precision and the computation time receivable of the model that being used to optimization the operation parameters of industrial heterogeneous reactor. The research content is organized as follows. Firstly, the technologies of information fusion of industrial data, which will assure the input data for modeling is reliable. Secondly, the hybrid intelligent modeling of heterogeneous reactor will be set up by combing the mechanism with industrial data of multi-scaled for the industrial application. Thirdly, the
英文关键词: heterogeneous catalysis;Data mining;deactivation;neural network;Intelligent optimization