Study On Large-scale Flexible Heat Exchanger Networks Synthesis

As the heat exchanger network (HEN) is intimately related to the energy consumption, the synthesis and optimization of HEN are always hot and key contents in chemical process system engineering. Aiming at the problem in the existing HENs synthesis methods, such as the computational complexity, poor practicability and the restriction of application, the research of large-scale flexible HEN synthesis are carried out, and the main contents and results of this thesis are as follows:(1) An approach for synthesizing large-scale multi-stream HEN with certain conditions is proposed, which yields the initial HEN with the pseudo-temperature enthalpy diagram method. The merge array is introduced for dividing the initial HEN into several sub-networks and optimizing the division ways. To synthesize the sub-networks, an improved super-structure method and a heuristic method are both presented, where the heuristic rules can be applied to both dual-stream HEN and multi-stream HEN. The sub-network can be further optimized via branch coefficient of common stream. With both high optimization quality and high efficiency, the presented approach can effectively solve large-scale HEN synthesis problems, and it is a basic of the flexible HEN synthesis.(2) A flexibility analysis method of HEN is proposed based on the flexibility index, a quantitative performance indicator for a flexible process. The direction matrix is proposed to describe the deviation direction of uncertain parameters, and the critical direction can be searched accordingly. The problems with convex and nonconvex feasible regions are solved by searching matrix and parallel simulated annealing algorithm respectively. The presented method can not only solve the flexibility index efficiently, but also directly show the critical point, which can be employed to guide the structure modification in flexible HEN synthesis.(3) A stepwise synthesis method for flexible HEN is proposed, which is sequentially implemented by structure synthesis and area optimization. An iteration strategy based on topological union is presented to find the flexibility-qualified structure. According to influence trend of the heat exchanger areas on flexibility and the total annual cost, an iteration formula for area optimization is proposed. This formula has high efficiency and strong robustness. Besides, since the presented method uses the flexibility test instead of multi period assumption to ensure that the HEN meets the flexibility requirement, it can be widely applied to flexible HEN synthesis problems with both convex and nonconvex feasible regions.(4) A detailed study on the simulated annealing algorithm, the genetic algorithm and the hybrid simulated annealing genetic algorithm is carried out in this thesis. The algorithm parameters and operators are designed according to the optimization problems in HEN. The selection strategy of individuals for crossover is improved by adding some individuals with large gene difference and high fitness value based on the original randomized crossover. This strategy significantly promotes the computational efficiency.(5) In the last part of this thesis, a large-scale flexible HEN in an ethylene plant is synthesized with the methods proposed in this thesis. The total annual cost of the synthesis solution has a great decrease, and the new HEN also features a great anti disturbance ability. The result has demonstrated the application of our methods for the guidance to the practical engineering problems.