乙烯装置裂解炉群不确定运行周期下的优化方法研究

Ethylene is the most fundamental plants in petrochemical process, and cracking process is the most critical unit of ethylene plant. With the diversification of hydrocarbon feedstocks and the large scale production of ethylene plant, some new problems and challenges have been brought forward to the operation optimization of the pyrolysis process. This project will focus on the uncertain cyclic operation optimization problems of ethylene cracking furnaces, which is characterized by parallel scheduling and semi-continuous operation. Based on the understanding and facilities of the cracking reaction mechanism, the following researches will be carried out, that is, the problem statement and description of the optimization of cracking furnace with uncertain operation cycle, the optimal allocation of raw materials to cracking furnaces considering multiple constrains, the multi-objective dynamic coordination of cracking furnace group. The quantitative relationship of the optimal allocation of raw materials and the efficiency of cracking furnace will be analyzed, and the dynamic optimization methods especially fit for the uncertain cyclic operation characters of ethylene cracking furnace will be carried out. Thus to provide some considerations and results from multi-evaluation indexes, which integrated pyrolysis yield, equipment efficiency and long cycle run of furnaces, for the optimal operation of cracking furnaces.

乙烯是石油化工的龙头，裂解炉则是乙烯装置的关键生产单元。随着乙烯生产原料的多样化和装置规模的大型化发展，为裂解过程的运行优化带来了新问题与挑战。本项目将围绕乙烯装置裂解炉群不确定运行周期下的优化问题，从其并行调度、半连续运行特点出发，结合裂解反应过程机理特性，开展乙烯裂解炉不确定周期运行的优化问题描述、烃裂解原料的多约束优化配置、裂解炉群运行优化的多目标动态协调等方面的研究，着重分析原料的优化配置与裂解炉高效运行的定量关系，针对性地开展适合乙烯裂解炉群运行特点的动态不确定周期下动态优化方法的研究，从多评价指标体系上开展裂解生产过程运行性能最优化的分析与探讨，为综合协调集成裂解收率、裂解效率与长周期运行的多目标决策提供优化方法和技术基础。

本课题针对乙烯裂解炉群并行调度中不确定运行周期等带来的复杂优化问题，首先在原有裂解过程机理模型基础上，进一步拓展了裂解炉传热模型的研究，提出了炉膛与炉管耦合的裂解炉辐射传热建模、裂解炉对流室与炉管传热和传质的耦合模拟等模型，以便为后续炉群调度优化提供不同操作工况下的能耗模型；其次，为兼顾炉群调度优化问题对模型实时性与可靠性的要求，提出了基于自适应采样的神经网络代理模型、基于神经网络-状态空间模型的裂解炉全周期模拟代理模型，以及基于迁移学习方法的代理模型等，建立了原料、操作条件等与乙烯裂解炉关键性能指标（如收率、能耗、排放等）的动态性能表征。针对裂解炉群调度中面临的动态不确定周期下的非线性、多目标优化问题，提出了基于控制向量参数化的动态优化方法、基于混合寻优策略的动态多目标优化方法以及复杂约束支配处理的混合整数非线性规划多目标优化方法等，并针对实际工业背景，对乙烯裂解炉炉群负荷与原料的调度优化、乙烯裂解炉炉群调度与操作优化一体化问题进行了方法有效性验证。相关结果已在工业过程中进行应用验证，获得2016年教育部科技进步一等奖。课题发表/录用论文35篇，其中SCI论文23篇，EI论文5篇，申请发明专利10项，其中已授权3项，登记计算机软件著作权4项。培养博士和硕士研究生17名，其中毕业15名。