PDPGA优化方法及其在炼厂气脱硫优化中应用

发布时间：2019-01-17 10:43

【摘要】：化工生产过程,已从过去的单纯考虑经济利益最大化单一目标,转向兼顾生态环境、能量消耗等多个目标的评价上,多目标优化问题在化工中越来越受到重视。流程模拟软件的出现为人们模拟实际工业过程提供了方便,在模拟的基础上将其与进化算法结合来解决化工过程多目标优化问题,成为近些年来各方学者研究的一个热门课题。但是,因流程模拟软件存在模型复杂性高、收敛性弱等问题,导致了应用该方法对化工多目标问题进行优化求解的时间消耗大且优化效率不高。而工程与科学研究领域中常用到的并行算法将一个复杂问题交由多个处理器同时处理,可以在有效提高收敛性的同时,降低运算时间提高效率。并行算法的这些特点为人们解决化工多目标优化问题提供了一种高效可行的解决方法。基于对上述问题的考虑,本文对并行算法和多目标优化方法进行系统的研究。通过比较各种进化算法的优劣,选取NSGA-II作为本文的多目标优化研究方法,对其进行并行化改进,提出了一种基于种群的分布式并行遗传算法。该方法与流程模拟软件结合优化能有效解决化工多目标优化中计算速度慢和优化效果不理想的问题。在应用经典测试函数验证该并行算法可靠性的基础上,将该并行优化方法应用到干气脱硫和富液集中回收系统的多目标优化中,在合理的优化时间内取得了降低系统能耗、减少排放的优化目标。本文的研究主要包括以下内容:(1)学习和总结近几十年来国内外的优秀算法,选取带精英策略的非支配排序遗传算法作为研究对象,对其可并行的部分进行研究。(2)通过选取主从式模型,考虑对NSGA-II的适应度评价部分进行并行计算,提出种群分布式并行遗传算法(PDPGA),并与流程模拟软件结合对脱硫溶剂再生塔这一简单实例进行优化求解,结果表明该方法能有效提高优化效率降低时间消耗。(3)选取适宜的吸收剂对炼厂气体脱硫及溶剂再生的工业过程进行研究,并对溶剂集中再生的改进方案进行全流程模拟,通过调节溶剂MDEA浓度、再生塔回流比和塔顶采出率等操作变量,使用PDPGA与Aspen Plus结合的优化方法对上述过程进行了全流程优化。(4)对本文的工作进行系统总结,并对下一阶段的工作进行了展望。
[Abstract]:In the process of chemical production, the problem of multi-objective optimization has been paid more and more attention in chemical industry, from considering the single objective of maximizing economic benefits to the evaluation of ecological environment, energy consumption and so on. The emergence of process simulation software provides convenience for people to simulate the actual industrial process. Combining it with evolutionary algorithm to solve the multi-objective optimization problem of chemical process has become a hot topic for scholars in recent years. However, due to the high complexity of the model and the weak convergence of the process simulation software, the application of this method to solve the multi-objective chemical problems is time-consuming and inefficient. The parallel algorithm, which is often used in the field of engineering and scientific research, can deal with a complex problem by multiple processors at the same time, which can not only improve convergence but also reduce computational time and improve efficiency. These characteristics of parallel algorithms provide an efficient and feasible method for solving multi-objective optimization problems in chemical engineering. Based on the above problems, parallel algorithms and multi-objective optimization methods are systematically studied in this paper. By comparing the advantages and disadvantages of various evolutionary algorithms, NSGA-II is chosen as the multi-objective optimization research method in this paper, and the parallel genetic algorithm based on population is proposed. The method combined with process simulation software can effectively solve the problems of slow calculation speed and unsatisfactory optimization effect in multi-objective optimization of chemical industry. On the basis of applying the classical test function to verify the reliability of the parallel algorithm, the parallel optimization method is applied to the multi-objective optimization of dry gas desulfurization and concentrated recovery system, and the energy consumption of the system is reduced in a reasonable optimization time. Optimized emission reduction targets. The main contents of this paper are as follows: (1) studying and summarizing the excellent algorithms at home and abroad in recent decades, selecting the non-dominated sorting genetic algorithm with elitist strategy as the research object. The parallelism part is studied. (2) by selecting the master-slave model and considering the parallel computation of the fitness evaluation part of NSGA-II, a population distributed parallel genetic algorithm (PDPGA),) is proposed. A simple example of desulfurization solvent regeneration tower is optimized by combining with process simulation software. The results show that this method can effectively improve the optimization efficiency and reduce the time consumption. (3) the industrial process of gas desulfurization and solvent regeneration in refinery is studied by selecting suitable absorbent, and the whole process simulation of the improved scheme of concentrated solvent regeneration is carried out. By adjusting the operating variables such as concentration of solvent MDEA, reflux ratio of regenerator and recovery rate of tower top, the whole process is optimized by combining PDPGA with Aspen Plus. (4) the work of this paper is systematically summarized. The work of the next stage is prospected.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X742

【参考文献】