Research On Thermal Process Optimization Based On Aspen Plus And Genetic Algorithm

Simulation and optimization of a thermal process have been a hot and difficult problem of thermal engineering. The traditional thermal process modeling method is inefficient, and the workload is large. In contrast, the Aspen Plus process modeling software is a more efficient choice in thermal process modeling. However, the optimization module of Aspen Plus is still not perfect, such as being not able to deal with multi objective optimization or different types of decision variable. Its default optimization method, which is sequential quadratic programming algorithm, also has some limitations, such as the bad convergence performance, the results are affected by the initial value and being easy to fall into locally optimal value and so on. In order to guarantee the global thermal process optimization performance and the convergence stability while taking advantage of the powerful process modeling function of Aspen Plus at the same time, this paper uses the ActiveX interface of Aspen Plus and genetic algorithm to optimize the thermal process simulation in Aspen Plus.Firstly, based on the PIKAIA genetic algorithm, referring to the NSGA-II, using the ActiveX interface of Aspen Plus, this paper implements a single objective and a multi objective genetic algorithm combined with Aspen Plus and the correctness of the program is verified. Then the processing method of constraints and different types of decision variables is introduced. This lays a solid foundation of the following example calculations.Secondly, the paper has selected the following 3 representative thermal process examples to introduce.The design optimization of a single device is the basis of process optimization. In this paper, the optimization process of a single thermal equipment is introduced by taking the example of design optimization of shell and tube heat exchanger. After the optimization of this paper, the heat transfer area of the heat exchanger is reduced by 17.34% compared with that of reference.The lithium bromide absorption refrigeration optimization example shows that, under the given conditions such as refrigerating capacity, by changing the parameters like gas-emission scope, with the multi objective optimization based on non-dominated sorting, we could get the relationship (Pareto front) between the COP of the system and the total heat exchanger area (SUMA) by the data fitting of the non-dominated solution, which is approximately "SUMA=148.12*COP-64.474".The integrated gasification combined cycle (IGCC) optimization example shows that, by changing the three flash tank pressures, with the multi objective optimization based on non-dominated sorting, we could get the relationship (Pareto front) between the system efficiency (SE) and CO2 capture rate (CCR) by the data fitting of the non-dominated solution, which is approximately "CCR=-9.9998*SE+4.1682". Moreover, there are some combination of the flash pressures that could make both system efficiency and the capture rate of CO2 better than the original system condition.Three examples of thermal process optimization show that, the optimization idea of the thermal process based on Aspen Plus and genetic algorithm, while at the same time using the Aspen Plus as a powerful process modeling function and the good optimization performance of genetic algorithm, is a more efficient thermal process modeling and optimization method. Moreover, the efficiency of process optimization of the simulation in Aspen Plus will be higher if combined with the Calculator module and Design Spec module of Aspen Plus.Finally, this paper has achieved an Excel workbook that could handle different simulation files in Aspen Plus, decision variables, constraints and objective functions without more programming, which is a useful Excel workbook.