精馏过程的广义析取规划模型及求解方法

Generalized disjunctive programming for distillation processes

针对板式精馏塔的优化设计问题,将广义析取规划(GDP)算法运用到了精馏塔的建模和优化求解过程。本方法是以塔板和进料板的是否存在为逻辑变量,以年费用最小为目标函数,利用GAMS进行优化求解。考虑到该优化问题的复杂性,本文采用了一般预处理过程,为GDP模型的求解提供了恰当的初始值和边界值,降低了求解的难度。最后,将该方法应用到苯、甲苯、邻二甲苯的三元混合物精馏分离过程优化设计中,并优化设计计算结果与HYSYS流程模拟计算结果进行对比研究,验证该数值建模方法求解结果的合理性。

The paper applied a generalized disjunctive programming to the optimal design of distillation columns using tray-by-tray based models. The problems used Boolean variables denoting the existence of tray and the feed tray, minimized the total annualized cost and solved in GAMS. In order to reduce complexity of the optimal problem, a general preprocessing phase was adapted to GDP formulations to initialization and bounding of the variables. Finally, simulators HYSYS were used to verify the rationality of the numerical method results.