基于结构导向集总的延迟焦化绝热反应过程模型研究

基于结构导向集总方法，设计了包含烃类结构和杂原子结构的21个结构单元，构建了代表延迟焦化原料分子组成的55类共2 791种典型分子的结构向量。基于延迟焦化反应机理，制定了包含碳链断裂、脱氢、开环、双烯合成、脱硫化氢、脱二氧化碳、脱一氧化碳和加氢脱氮等10类38条反应规则，用于描述延迟焦化过程的反应网络。结合相应的反应速率常数和反应热数据，建立了反应动力学微分方程组，通过改进的Runge-Kutta法求解，构建起基于结构导向集总的延迟焦化绝热反应动力学模型，预测延迟焦化过程的典型分子组成和产物分布。延迟焦化绝热反应器模型的预测精度优于未计入反应热效应的等温反应器模型的预测精度。采用延迟焦化结构导向集总绝热反应器模型对不同延迟焦化渣油在不同焦化反应温度和循环比条件下的延迟焦化反应过程进行模拟，焦化产物中气体、汽油、柴油、蜡油和焦炭的产率预测误差均不超过1.7百分点，焦化产物中典型分子含量的预测误差均不超过2.0百分点。

RESEARCH ON MODELING OF DELAYED COKING ADIABATIC REACTION PROCESS BASED ON STRUCTURE-ORIENTED LUMPING

Based on the structure-oriented lumping method (SOL), 21 structural units including hydrocarbon structure and hetero atom structure were designed, and the structural vector of 2 791 kinds of typical molecules representing 55 kinds of delayed coking raw material molecules was constructed. Based on the mechanism of delayed coking reaction, The 38 reaction rules of 10 kinds of reactions, including carbon chain cleavage, dehydrogenation, ring opening, diene synthesis, H2S removal, CO2 and CO removal, as well as hydrodenitrogenation, were developed and used to describe the entire delayed coking reaction network. Based on the corresponding reaction rate constants and reaction heat data, a set of differential equations for reaction kinetics was established. The improved Runge-Kutta method was used to solve equations, building a kinetic model of delayed coking adiabatic reaction based on structure-oriented lumping to predict the typical molecular compositions and product distributions of delayed coking process. The prediction accuracy of the established kinetic model is better than that of the isothermal reactor model without considering the reaction thermal effect. Using the model of lumped adiabatic reactor guided by delayed coking structure to simulate the process of delayed coking reaction under different coking reaction temperature and cycle ratio, the prediction errors of gas, gasoline, diesel, wax oil and coke yields were less than 1.7 percentage points and the prediction errors of typical molecule content in products did not exceed 2.0 percentage points.