液氮洗工艺的多目标遗传算法优化  

Multi-Objective Genetic Algorithm Optimization for Liquid Nitrogen Wash Process

作　　者：张旭[1] 颜立伟[1] 李云[1] 张早校[1] 

机构地区：[1]西安交通大学能源与动力工程学院,西安710049

出　　处：《西安交通大学学报》2011年第5期114-118,共5页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(50976090;20936004);中国石油科技创新基金资助项目(2010D-5006-0705)

摘　　要：为了优化大化肥中的液氮洗关键流程以提高能量效率、控制CO2排放,在使用Aspen Plus软件对某低温液氮洗工艺进行模拟分析的基础上,结合多目标遗传算法和二次规划法(MOGA-QP)来解决局部极值和大规模优化问题.根据超结构方法,构建了液氮洗工艺流程总结构,即在原流程中加入了一个减压阀和一个分离器,由此可以增补29.546 kJ/kg冷量,使得液氮洗涤塔出口净化气中的CO体积分数为0.55×10-6.利用MOGA-QP求解了超结构,即在进化过程中用二次规划法来模拟流程表单的各个子模块,在最终收敛的环节用二次规划法求解子问题,从而对遗传算法产生的Pareto解做进一步的精细优化.实际应用显示:在满足工艺指标的前提下,优化后的液氮入口流量减至0.131 2 kmol/s,出口CO体积分数降至0.25×10-6,液氮入口流量减少了1.7%,火用损失减少了15.7%,表明改进流程更佳,MOGA-QP算法有效.A multi-objective genetic algorithm with quadratic programming（MOGA-QP） was proposed to optimize the liquid-nitrogen washing process in large fertilizer plants based on the simulation by using Aspen Plus software.According to the superstructure method,the flowsheet of liquid-nitrogen washing process was constructed by adding a throttle valve and a separator.The proposed algorithm MOGA-QP was employed to solve the superstructure of the liquid-nitrogen washing process.The results of the case study show that the optimization of the liquid-nitrogen washing process reduced the inlet flow rate of liquid-nitrogen and the CO volume fraction at the exit to 0.131 2 kmol/s and 0.25×10-6,respectively.Moreover,the inlet flow rate of liquid nitrogen and the exergy loss decreased by 1.7% and 15.7%,respectively.It appears that the improved process is much better and the MOGA-QP algorithm is effective.

关 键 词：液氮洗 模拟分析 多目标遗传算法 过程优化 

分 类 号：TQ441.5[化学工程—化学肥料工业]