The Nodes-based Non-structural Model With Stream Splitting And Global Optimization Of Heat Exchanger Networks Research

Under the current situation of increasingly tight energy demand,adopting scientific methods to manage energy and improve energy utilization efficiency can not only reduce fossil energy consumption and pollutant emissions,but also have great significance for companies to improve their economic benefits.As an important system for energy recovery and utilization,heat exchange network can meet the stream processes’ heat exchange requirements,and at the same time,optimize the configuration heat exchangers on streams and achieve the minimum annual overall cost or maximum energy recovery.However,as the system’s scale increases,its heat exchanger sub-system’s non-convex and non-linear characteristics become more and more prominent,the global optimization becomes more difficult.In addition,the setting method of the optimization model also restricts the acquisition efficient of the optimal network configuration.Therefore,the development an optimization model with a higher degree of structural freedom and better matching with algorithms,at the same time,applied it into a higher efficiency optimization model,is an important guarantee for achieving the dual improvement of optimization quality and efficiency.From the perspective of global optimization,this paper conducts in-depth research with the main purpose of realizing the overall improvement of Heat exchange network optimization quality and efficiency.The main research contents and innovations of the thesis are as follows:1)Structuring a nodes-based non-structural model with stream splitting,NNM.This model quantifies the generation position and connection of the heat exchanger by randomly matching the unconnected nodes on the hot and cold streams.The model can include two network structures with and without stream splitting.Under the setting of several nodes,the number of potential structures is increased by the random matching generation mode,and the solution domain is effectively extended.The proposed model overcomes the disadvantages of structural models and lays the the foundataion for the follow-up study.2)Based on the NNM,establishing a random walk evolution with compulsive evolution algorithm and apply it to NNM(NNM-RWCE).The algorithm adopts a differentiated simultaneous optimization strategy to increase the diversity of the structure.The results of calculation examples of different scales show that the combined use of NNM and RWCE can achieve an effective improvement of optimization quality and efficiency.3)Combining the proposed model and algorithm,the influence of the node parameters in NNM on the solution domain,optimization quality and efficiency is studied.After analyzing the limitations of uniformly setting the number of branches and the total number of nodes on the stream,two flexible node configuration strategies are proposed,using stream parameters to customize the positions of the nodes on the stream,which enhances the flexibility of NNM node layout.4)For the continuous and integer variable optimization obstacles in the NNMRWCE and the low efficiency of large-scale cases,proposing corresponding improvement strategies: 1.Analyzing the convective heat transfer capacity of the NNM-RWCE with different optimization steps,a step-size perturbation strategy was proposed.Through the perturbation operation of the optimized step size,the optimization depth of continuous variables was enhanced,and thus the variation of the structure was promoted;2.For the ptimization of integer variables in the optimization obstacles’ s middle and late stages,a structure perturbation strategy is proposed.It uses the competitive relationship between the split structure to enhance the vitality of individuals in the middle and later stages of optimization,and improve the global search ability of the algorithm;3.To achieve the improvement of the efficiency of large-scale cases,the network node dynamical adjustment strategy is established.It constructs a calculation sub-network with only heat exchangers during the optimization process,thereby greatly reducing the time consumed in the calculation cycle.In the same calculation time,the strategy can assist the algorithm to perform more iterative optimizations,which improves Probability of obtaining good results.Finally,starting from the global optimization research,a new optimization idea is proposed—the heat exchange network optimization route’s concept.This method achieves the strategy modules intelligently combination by monitoring the different characteristics and requirement in the process of optimization.It could better coordinate the global and local searching ability and the optimization quality could overall promotion.The proposed optimization route has significant meaning for complex heat exchanger network synthesis,and offers a brand new optimization idea for later research.In a word,this paper takes the heat exchange network’s global optimization as the main research contents,firstly constructing a nodes-based non-structural model with stream splitting,several aspects such as matching mode,parameter setting,and node configuration flexibility confirm that the model can efficiently obtain high-quality results in heat exchange network optimization.Meanwhile,establishing the RWCE suitable for this proposed model,targeting for the characteristics and difficulties of the algorithm and model optimization process,proposing corresponding strategies to enhance the algorithms’ ability.On the basis,an optimization route that can realize the global optimization of complex networks’ global optimization is proposed,overall increasing the heat exchanger network optimization’s whole level.