| In order to save energy and enhance the separation efficiency, the dividing wall distillation column (DWC) was adopted to separate the pinene from turpentine. Then the separation process was studied by experiments, simulation and multi-objective optimization. The major contents and results show as below:Firstly, this study carried out in a DWC equipment, in which the influence on the separation process of turpentine were researched under different operation factors such as reflux ratio R, liquid split fraction Rl, feed location NF, side-draw location Ns and divided wall location (Nub and Nlb). Experimental results indicated that reflux ratio and liquid split fraction had an outstanding influence on the purity of a-pinene and P-pinene products. With the increasing of the reflux ratio, the purity of two products increased firstly, and then essentially unchanged. Under a given reflux ratio, there exited a best liquid split fraction to optimum the separating effect of the DWC. Also and NF, Ns, Nub, Blb, had a certain extent influence of the separating effect of DWC. The separation of the a-pinene/β-pinene and β-pinene/heavy components should be balanced when determining the value of NF, Ns, Nub and Nlb. By feeding in the upper divided wall to the left and extracting P-pinene product from the lower divided wall to the right, better separation effect was achieved. In addition, the liquid component distribution in the dividing wall column indicated that the back-mixing effects of the middle component β-pinene was reduced because of the existence of the divided wall.Secondly, four columns equivalent model of the DWC was set up by Aspen Plus software. Simulation of separation process of the turpentine system (a-pinene-camphene-β-pinene-heavy components) in DWC was carried out. The simulation result showed a good agreement with the experimental data, and the average relative error was controlled in 8%. The comparison results proved the stable and reliable of the model, and it laid the foundation for the further optimization.Thirdly, in order to make the multi-objective Cuckoo Search algorithm (MOCS) to solve the constrained multi-objective optimization problems (CMOPs) more efficiently, the constrained multi-objective Cuckoo Search algorithm(CMOCS) was proposed based on the conception of "multi-objective" constrained handling, and several different test functions were used to evaluating the performance of the CMOCS algorithm. The results indicated that the CMOCS algorithm had better global diversity, convergence and distribution of the obtained Pareto optimal solution set when compared with constrained multi-objective Particle Swarm Optimization algorithm (CMOPSO), and it’s worth noting that the CMOCS algorithm is well suited to solving the CMOPs which have multiple feasible regions.Finally, the multi-objective optimization model of DWC for separating pinene from turpentine system was established where the objectives were to minimize the total number of stages in prefractionator and main column as well as the heat duty. Then the Pareto front of DWC was obtained by the use of CMOCS algorithm through solving the multi-objective optimization model, which provided decision makers with a variety of alternative optimization design schemes of DWC. In addition, when compared with two conventional distillation process, the multi-objective optimization results could save energy 26.6%.In conclusion, the four columns equivalent model of the DWC build in this paper was stable and reliable and it can greatly describe the separation process of the turpentine system; using CMOCS algorithm to control the Aspen Plus software for multi-objective optimization design of DWC was a quick and effective design optimization method; adopting DWC to separate turpentine is a new energy-saving technology and has a good prospect for industrial application. |
PDF Full Text Request