Synthesis Of Superstructure Heat Exchanger Networks-Application Of Analytical Solution In Genetic/Simulated Annealing Algorithm

In the past three decades, extensive efforts have been made for heat integration problem and energy recovery technology because of the steadily increasing energy consumption. A heat recovery system consisting of a set of heat exchangers (heat exchanger network (HEN)) is widely used in process industries such as gas processing and petrochemical industries to exchange heat energy among several process streams with different supply temperatures.The task of the synthesis of HENs is to find a HEN which has the minimum total annualized cost of the process. Because of its significant benefits in saving energy consumption and equipment cost, the HEN synthesis has been one of the most important research subjects in process engineering.The synthesis and optimization of heat exchanger networks can be treated as a mixed integer nonlinear programming (MINLP) problem. Since the problem is usually non-linear, non-convex and might have many local optimums, it is hard to be solved with traditional mathematic methods. However, genetic algorithm provides a promising approach for this problem. But the existing heat exchanger networks (HEN) synthesis methods are more applicable for small-size heat exchanger networks; and problems such as the initial solution being infeasible can disturb the synthesis process and make the final solution extremely difficult to find. In this thesis the following studies have been done to solve this problem:(1)A new analytical method was applied to a stage-wise superstructure model for the heat exchanger networks, by which the stream temperatures of the heat exchanger network can be calculated explicitly and the constraints of the MINLP model can be reduced dramatically.(2)Based on the present analytical solution,a genetic algorithm coupled with the simulated annealing algorithm and hill-climbing algorithm was developed. The elite strategy and structure mutation strategy were introduced in the algorithm, which enhance the structural search ability of the algorithm.(3)Three examples from literature were calculated with the proposed procedure,...