Multiobjective Optimization Of Heat Exchanger Network Based On Nondominated Sorting Genetic Algorithm

Heat exchanger networks (HEN) is an important energy recovery system in process industry like chemical industry or oil refining industry and so on. The design level of heat exchanger networks determines the energy consumption and economy of process system. Therefore, heat exchanger networks synthesis gradually becomes a hotspot to study. Meanwhile, the existing methods of heat exchanger networks synthesis have some defects which are urgently required further research.The approaches of HEN synthesis are classified into sequential methods and simultaneous methods. Pinch analysis method which analyzes problem from energy is the representative of sequential methods. Mathematical programming is the mainly used method among simultaneous methods. Both sequential methods and simultaneous methods mainly realize heat exchanger networks synthesis from single-objective optimization. To better take trade-offs into account, from multiobjective optimization, this paper adopts nondominated sorting genetic algorithm (NSGA-II) and considers its environmental impact. Based on this topic, the main research work are as follows:1. Based on superstructure, establishes the mathematical formulations of multiobjective optimization of HEN which simultaneously optimize capital cost, utility cost and environmental impact. The model removes the isothermal mixing assumption and introduces the correction factor of temperature difference, and multiple utilities are allowed, which make the model more general and more complex. To solve this complex model, structure optimization and parameter optimization are separately optimized. In order to reduce the computational complexity and problem size, heuristic rules used in the structure optimization and maximum energy recovery (MER) principle used in the parameter optimization are adopted and the linear programming to get heat load distribution is established in the optimization process. Considering the nonconvex of the established formulations, NSGA-II which has good searching ability is choose to realize the multiobjective optimization of HEN while tradional multiobjective optimization methods may break down. For the HEN problem with multiple utilities, a simple constrained NP model based on the exsiting solving method and strategy can solve it.2. Considering the fact that people pay more attention to environment quality, an environmental assess model based on the emission amount of greenhouse gas is proposed. Thus in the optimization process, energy consumption, economy and environmental impact can be taken into account simultaneously, which is consistent with the present situation.3. Relying on the MATLAB-GUI and the existing HEN synthesis software system HenDesign, the module of multiobjective optimization of HEN based on NSGA-II which riches the HenDesign, is developed. Moreover, the detailed functions and operation procedures are introduced through a case.4. Several cases where one has multiple utilities to choose are optimized. Each example obtained satisfying results. The gain of Pareto fronts makes the relationship between each two objectives clear from multiobjective optimization. Comparing with the results with optimized TAC in the interatures, the results with lower total annual cost are obtained from single-objective optimization. In a word, the results for the cases prove the effectiveness and practicability of the proposed approach.