CAJ | 학술논문

为了降低能耗，提高经济效益，在Aspen Plus平台上建立了分隔壁精馏塔(DWC)分离松节油中蒎烯的四塔等效模拟流程，采用灵敏度分析确定了对能耗和分离效果影响较大的设计变量及其取值范围，以预分离塔塔板数、主塔塔板数以及能耗最小为目标，建立了DWC分离松节油中蒎烯的多目标优化模型，并利用约束多目标微粒群优化(CMOPSO)算法对模型进行了求解。结果表明：CMOPSO算法能很好地解得DWC的Pareto最优解集，为决策者提供了多种可供选择的DWC优化设计方案；经多目标优化后，在总塔板数(或设备投资费)相近时，与DWC分离松节油的单目标优化结果相比，多目标优化结果可进一步节能21.7 kW；气、液相分配比是DWC特有的，且非常重要的设计变量，采用的双变量灵敏度分析方法能够比较准确地得到两者的适宜取值范围，优化时在该范围内搜索气、液相分配比可望进一步缩短寻优时间。
위료강저능모，제고경제효익，재Aspen Plus평태상건립료분격벽정류탑(DWC)분리송절유중파희적사탑등효모의류정，채용령민도분석학정료대능모화분리효과영향교대적설계변량급기취치범위，이예분리탑탑판수、주탑탑판수이급능모최소위목표，건립료DWC분리송절유중파희적다목표우화모형，병이용약속다목표미립군우화(CMOPSO)산법대모형진행료구해。결과표명：CMOPSO산법능흔호지해득DWC적Pareto최우해집，위결책자제공료다충가공선택적DWC우화설계방안；경다목표우화후，재총탑판수(혹설비투자비)상근시，여DWC분리송절유적단목표우화결과상비，다목표우화결과가진일보절능21.7 kW；기、액상분배비시DWC특유적，차비상중요적설계변량，채용적쌍변량령민도분석방법능구비교준학지득도량자적괄의취치범위，우화시재해범위내수색기、액상분배비가망진일보축단심우시간。
In order to reduce energy consumption and enhance economic benefit , on the basis of building a simulation process of four columns equivalent model of dividing wall column (DWC) for separating pinene from turpentine by Aspen Plus, and confirming the design variables that influence energy consumption and separation effect strongly, as well as their search scopes by sensitivity analysis, the multi-objective optimization model of dividing wall column for separating pinene from turpentine was established where the objectives were to minimize the total number of stages in prefractionator and main column as well as the heat duty. Constrained MOPSO (CMOPSO) algorithm was used to solve the multi-objective optimization model. The results indicated that CMOPSO algorithm could obtain good Pareto front of DWC which provided decision makers with a variety of alternative optimization design schemes of DWC. Compared with single objective optimization results, the multi-objective optimization of DWC separation process could further save energy 21.7 kW under the similar total number of stages (or equipment investment cost). Vapor and liquid split fraction were special and important design variables of DWC, and by the use of bivariate sensitivity analysis, their suitable value ranges were confirmed accurately, which was able to further reduce the searching time when vapor and liquid split fraction values were searched in these ranges.