混合粒子群算法用于共沸精馏塔的最优设计  

The Application of MPSO Algorithm to the Optimal Design of Azeotropic Distillation Column

作　　者：李红海[1] 姜奕[1] 陶少辉[1] 

机构地区：[1]青岛科技大学化工学院,山东青岛266042

出　　处：《化学工业与工程》2015年第5期113-117,共5页Chemical Industry and Engineering

摘　　要：提出了一种适于求解混合整数非线性规划问题的混合粒子群优化算法,并将其与化工过程模拟软件相结合,用于共沸精馏塔的最优设计。优化模型以年度总费用最小为目标,以精馏段板数、提馏段板数和回流比为优化变量,并引入流体力学约束使得优化结果更具实际价值,并以效率更高的"轮盘赌"式策略处理整数变量,约束处理采用Deb方法。最终以C++实现优化算法,C#编制界面,通过商业模拟软件Aspen Plus计算粒子适应度,将本方法用于一个醋酸甲酯/甲醇/水三元共沸组成的分离案例,所获最优年度总费用优于文献结果。A mixed particle swarm optimization( MPSO) algorithm is proposed to solve mixed integer nonlinear programming problem and to apply to the optimal design of azeotropic distillation column. By adjusting the number of rectifying tray,stripping tray and reflux ratio,the total annual cost of the column is minimized; the hydraulic constraints are added in the optimization model to make the optimal solution more practical. The integer variables are preceded by a roulette way,while the constraints are satisfied with Deb's method. The MPSO algorithm is implemented with C ++ on a C# platform and the particle's fitness is calculated via ASPEN Plus. At last the proposed MPSO is applied to the design of methyl acetatemethanol-water an azeotropic distillation column for the separation of methyl acetate-methanol-water ternary azeotropic system,and a better solution is obtained compared to the reference result.

关 键 词：混合粒子群优化算法 共沸精馏塔 最优设计 流体力学约束 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TQ053.5[自动化与计算机技术—控制科学与工程]