| Because the advance of industrialization and the shortage of energy,the heat exchanger network plays a decisive role in the rational use of the whole system of energy production which is significant for reducing the cost of production and energy consumption.The present research work for heat exchanger network optimization are mostly classified as deterministic optimization method and heuristic optimization method.Although the development of the deterministic optimization methods is more mature and it has formed a series of classical theory,the deterministic optimization methods still rely on the original structure too much.Due to the limitations of the optimization method itself,it is impossible to get the global optimal solution.Heuristic optimization method on the optimization objective function is very effective,and it is easy to get the optimal results.However,the method based on the probability statistics principle and simulating some natural phenomenon or process,so the result can not be proved as global optimal.In view of this,this paper respectively give a heat exchanger network global optimization study of deterministic method and heuristic method,and make them jump out of local optimal solution looking for better heat exchanger network structure of reconciliation until the global optimal solved.First of all,this paper comprehensive cost as the objective function and establish mathematical model to optimize heat exchanger networks by newton method which is representative of determining method.Then one-dimensional search is improved,so that it can optmize the network model better.The sequential linear programming algorithm optimization is also applied to the heat network optimization.However it is easy to fall into the local extremum problem for the search strategy.Thus an improved sequential linear programming is presented and it successfully escapes from local minima to find a better optimal design of networks.This paper combines the differential evolution algorithm with sequential linear programming optimization method for the first time adding certainty in stochastic algorithm to improve the efficiency of the algorithm.On this basis,the new algorithm is applied to heat exchanger network optimization,and examples are presented to verify the feasibility of the method.Then two strategies for heat exchanger cooperation evolution is proposed,respectively,using differential evolution principle and minimal heat load constraints for the generation and elimination of cooperation evolution of heat exchanger,in order to find reasonable network by incorporating differential evolution algorithm for thermal unit matching.Optimization results are better than previous literature.Then,dynamic multi-agent differential evolution algorithm is provided to solve the problem,and the algorithm makes use of the sensing capability of the multi-agent with the dynamic update strategy,which improve the formation mechanism of population and mutation mechanism of the differential evolution algorithm and the globle searching ability of the algorithm in the large scale nonlinear system.The proposed algorithm has been applied to cases of heat exchanger network problems,and the results are very encouraging that better total annual cost is obtained,which indicates the better globle searching ability of the algorithm.Finally,based on the field synergy principle,inner utilities network structure of the temperature uniformity is analyzed,and the positive role of inner utility is verified.Because of the inner utilities can improve temperature field distribution of the system,this paper combine the stage-wise superstructure model with inner utilities,and dynamic multi-ntelligent differential evolution algorithm optimized heat exchange network based on new superstructure model.The result of networks with the inner utilities is better than the previous network structure,which is a further evidence of the temperature uniformity factor of feasibility. |
PDF Full Text Request