CN1257367C - Intelligent optimizing method for optimal synthesis of heat exchange network - Google Patents

Abstract

The present invention relates to a comprehensive optimizing method of heat exchange networks, which belongs to the technical field of intelligent information processing. Based on system optimizing concepts, the optimal comprehensive problem of heat exchange networks is divided into two partial problems of a system structure and a heat exchange unit in an equivalent method; by a cooperative evolution method, the system structure and the heat exchange unit cooperate for optimization to obtain the optimal compromise of the system structure and the heat exchange unit; then, according to on-site practical situations, optimal optimization schemes which comprises problem decomposition, cooperative optimization of networks, and optimum prioritization scheme selection are further selected. The present invention better solves the optimal comprehensive problem of heat exchange networks in larger scales, and can avoid practical problems that the convergence speed is slower, local minimum can be easily obtained, objective functions must be guided, etc. which exist in the traditional optimal comprehensive problem of heat exchange networks; the present invention can be applied to the network optimization of energy recovery in industrial processes of chemical industries, oil refining, etc.; the present invention is used for the comprehensive optimization of energy use in the overall process of an ethylene device in a certain factory, and is approved by production departments.

Description

Intelligent Optimization Method for Optimal Synthesis of Heat Exchange Network

technical field

The invention relates to an optimization method for the optimal synthesis of a heat exchange network, in particular to an intelligent optimization method for the optimal synthesis of a heat exchange network. It belongs to the technical field of intelligent information processing.

Background technique

Heat exchange network is the main component of energy recovery in process industries such as chemical industry and oil refining. The optimal comprehensive problem of heat exchange network is to recover the effective energy of all cold and hot process streams as economically as possible under various conditions. Reduce the consumption of public works to achieve the purpose of energy saving. Its optimization goal is based on the analysis of the first law of thermodynamics, while considering all investment cost factors (heat exchange area, number of heat exchange units, and structural materials, etc.) and all operating cost factors (public works consumption, etc.) The form of expression is a mixed integer nonlinear programming problem.

The optimal comprehensive optimization method of heat exchange network is basically carried out along three directions, namely, thermodynamic rules, mathematical methods and expert system application. In the past 10 years, two types of methods have basically been formed: pinch point design method and mathematical programming method. The pinch point design method is to determine the pinch point of the network according to the thermodynamic method, and then determine the heat exchange network by weighing the number of heat exchange unit equipment and public works; the mathematical programming method is to establish a superstructure model of the network and use mathematical methods solve. After literature search, it was found that Niniek FP wrote an article "Synthesis of Heat Exchanger Networks Considering Location of Process Stream" ( "Heat Exchange Network Synthesis Problem Considering Process Logistics Location"), this paper studies the optimal synthesis problem of heat exchange network. The research shows that pinching point design method cannot consider network investment cost and operation cost at the same time; although mathematical programming method can be used in The investment cost and operating cost are considered in the model at the same time, but if a deterministic method is used, it can only solve the problem limited to a dozen process streams; if a random method, such as genetic algorithm, is robust and can be parallelized It has the characteristics of processing and high efficiency, but it can only solve problems smaller than dozens of streams, and it is easy to fall into local optimum and slow convergence speed. Because for a heat exchange network with N _H hot streams and N _C cold streams, and the number of stages is N _K , each stage has at most R=N _H ×N _C heat exchangers, if each matching sequence corresponds to A heat exchange network structure, there are $N=\underset{i=0}{\overset{N_h\×N_C}{\mathrm{\Σ}}}{C}_{N_h\×N_C}^{i}i!$ For a heat exchange network structure, as the number of cold and hot streams increases, the topology of the network will increase sharply, which is obviously a combination explosion, which belongs to the NP-Hard problem.

Contents of the invention

The purpose of the present invention is to overcome the deficiencies in the prior art and provide an intelligent optimization method for the optimal synthesis of heat exchange networks to solve large-scale heat exchange network synthesis problems, which can speed up the solution and effectively avoid local optimum excellent.

The present invention is realized through the following technical solutions. Based on the idea of system optimization, the present invention decomposes the optimal comprehensive problem of the heat exchange network into two parts: system structure optimization and heat exchange unit optimization. Partial collaborative optimization finally obtains the optimal compromise between the system structure and the heat exchange unit, and then further selects the most suitable optimization scheme according to the actual situation on site. The method specifically includes three basic steps: problem decomposition, network cooperative optimization and optimal optimization scheme selection.

System optimization is to consider the optimization of the entire network from the perspective of system optimization, including two parts: system structure optimization and parameter optimization. Generally speaking, the system structure determines the performance that the entire network can achieve; and the quality of a system structure usually needs to be judged by optimizing the parameters of the heat exchange units in the structure according to the objective function. Based on this idea, the cooperative cooperative evolution method can be used to solve the optimal synthesis problem of heat exchange network. In the evolution process, there are two kinds of species evolution: one is the system structure of species evolution network, which is called "dominant species"; the other is the evolution of heat exchange unit parameters, which is called "evaluated species". The purpose is to use the evolution of the dominant species to guide the evolution direction of the evaluated species; use the evolution of the evaluated species to evaluate the advantages and disadvantages of the evolved system structure. the best compromise. Because there may be multiple optimal compromises, the most suitable optimization scheme can also be selected according to the actual situation of the factory site that has not been integrated into the objective function, combined with long-term work experience.

The inventive method is described further below, and specific content is as follows:

1. Break down the problem

From the perspective of system optimization, the general problem of typical heat exchange network is equivalently decomposed into two parts: system structure and heat exchange unit optimization. Among them, the system structure includes the required amount of public works, the matching of streams and the number of heat exchange units; the heat exchange units include the heat load, operating temperature and heat exchange area of each heat exchanger. Compared with the traditional heat exchange network comprehensive problem optimization method, this method can better deal with large-scale heat exchange network optimization problems, and can speed up the solution.

2. Network cooperation and collaboration optimization

Cooperative and collaborative optimization of the network, including the optimization of the system structure and the heat exchange unit of the heat exchange network, adopts the cooperative collaborative evolution method to collaboratively optimize the system structure and the parameters of the heat exchange unit, as follows:

(1) Encoding method

The individual of the leading species represents the system structure of the network, which is represented by an association matrix, and its coding content is a positive integer; is a real number whose individual length is related to the corresponding phylogenetic structure corresponding to an individual in the dominant species. An individual of the leading species and an individual of the rated species can combine a complete heat exchange network.

(2) Adaptive value function

Only after the system structure is determined, the evolutionary optimization of the heat exchange unit is meaningful, and the evolutionary optimization result of the heat exchange unit is bound to reflect the advantages and disadvantages of the network system structure, and the system structure in turn affects the evolutionary optimization of the heat exchange unit. Therefore, firstly, the evaluation species is evolved, that is, the heat exchange unit parameters of the heat exchange network with a given system structure are evolved, and its individual fitness value function is equal to the comprehensive objective function of the heat exchange network; for the dominant species, it is the system The structure is evolved, and its individual fitness value function is given by the best individual in the evaluation species, that is, the evaluation species is used to evaluate the quality of the evolved system structure.

(3) Evolutionary optimization process

First, determine the population size of the dominant species and the rated species, and initialize the population randomly, where each population of the rated species corresponds to an individual of the dominant species, and each individual in the population represents the heat transfer rate of a given system structure Parameters of all heat exchange units of the network.

Secondly, use the partial evolutionary algorithm (PTA) to evolve the population of the selected species, and select the best individual as the individual representative of the population; use the general evolutionary algorithm to evolve the population of the dominant species, in which each individual corresponds to the selected species The individual representatives in represent a complete heat exchange network, and the best heat exchange network is selected as the optimal comprehensive optimization of the current best heat exchange network.

When the network structure corresponding to the population of the evaluated species changes, the population of the evaluated species is evolved by using the method of adding and deleting nodes (EAN) and the partial evolutionary algorithm (PTA).

a) If a heat exchange unit is added, under the method of adding and deleting nodes, the local evolution algorithm only evolves the newly added heat exchange unit and the parameters of the heat exchange unit at the level, while the parameters of other heat exchange units remain unchanged; b ) If the heat exchange unit is deleted, under the method of adding and deleting nodes, if the heat exchange unit has been added at the level, then delete the heat exchange unit in the reverse order of the original order of adding units, otherwise delete the unit randomly. Then use the local evolution algorithm to evolve the parameters of the heat exchange unit at the stage; c) If the matching order of the heat exchange stream changes, use the local evolution algorithm to evolve the parameters of the heat exchange unit related to it. The above method can greatly reduce the search space, and prevent damage to the learned behavior of the heat exchange network because the parameters of the existing part of the heat exchange unit remain unchanged.

3. Selection of the most suitable optimization scheme for the heat exchange network

Because there may be multiple optimal optimization schemes, the most suitable optimization scheme can be selected according to the actual situation of the site that has not been integrated into the objective function and combined with long-term work experience.

The present invention has substantive features and remarkable progress. Based on the idea of system optimization, the present invention decomposes the optimal comprehensive problem of the heat exchange network into two parts: system structure and heat exchange unit optimization, and utilizes the new cooperative cooperative evolution method to optimize the two parts, which can better solve the large-scale problem The optimal comprehensive problem of heat exchange network can avoid practical problems such as slow convergence speed, easy to fall into local minimum, and objective function must be differentiable in the traditional optimal comprehensive optimization of heat exchange network. The invention can be applied to the optimization of energy recovery network in process industries such as chemical industry and oil refining, and has been used in the comprehensive problem of energy utilization optimization in the whole process of an ethylene plant in a certain factory, and has been approved by the production unit.

Description of drawings

Fig. 1 is a logical structure diagram of the method of the present invention.

Detailed ways

In order to better understand the technical solution of the present invention, further description will be made below in conjunction with the accompanying drawings and specific embodiments.

As shown in Fig. 1, it is a logical structural diagram of the method of the present invention. The figure is mainly divided into two parts: the evolution of the dominant species and the evolution of the rated species. The evolution of the dominant species and the evolution of the rated species are carried out alternately until the algorithm is terminated. Its purpose is to use the evolution of the evaluated species to evaluate the advantages and disadvantages of the network system structure that has evolved; in turn, use the evolution of the dominant species to guide the evolution direction of the evaluated species, so that it may have the optimal structure in the search space. The search is carried out in the area of the network system structure, and they are in the relationship of mutual cooperation. Among them, when evaluating and evaluating species, the method of adding and deleting nodes and the local evolution algorithm are used to maintain the behavioral connection between the parent and offspring individuals in the evaluated species and improve the search efficiency.

Example:

In order to verify the effect of the present invention in the optimal comprehensive optimization of the heat exchange network, it is compared with the existing domestic and foreign methods mentioned above. Taking a typical 10spl problem as an example, it includes 5 hot flows and 5 cold flows. The initial temperatures of the hot flows are 416, 526, 544, 501 and 472 degrees respectively, and the initial temperatures of the cold flows are 355, 366, 311 and 333 degrees respectively. and 389 degrees. The optimal solution to solve this problem using the traditional method is the annual cost of 49,658 US dollars, and the cost calculation formula for all heat exchangers is 145.63·S ^0.6 (S is the heat exchange area required to exchange specific heat). A group of optimization schemes can be obtained by using the present invention. According to the actual situation on the spot, a kind of optimal optimization scheme is selected, and its annual cost is 42,704 US dollars. Seven heat exchangers are adopted, and the stream matching is respectively (H3, C5), (H2 , C3), (H5, C1), (H3, C4), (H1, C3), (H4, C2), (H2, C1); 3 cold flow public heat exchangers, respectively in the hot flow stream H3, H4 and H5. In addition, it is used in the comprehensive problem of optimizing the energy consumption of the whole process of the existing ethylene plant in a factory. The number of hot flow strands is 50, and the number of cold flow strands is 60. The annual cost obtained by the present invention is 893762.71 US dollars, and the result is recognized by the production unit.

Since the present invention does not need to consider constraints when optimizing the network system structure, as long as a suitable system structure expression is found, the task of judging whether the structure is feasible can be given to the heat exchange unit parameter optimization algorithm. The results show that the algorithm is intuitive, easy to implement and has good algorithm performance.

Claims (2)

1, a kind of intelligent optimization method of heat-exchange network optimal synthesis, it is characterized in that, thought based on system optimization, heat-exchange network optimal synthesis problem equivalent is decomposed into system architecture optimization and heat exchange unit optimization two parts, utilizes the team work evolvement method, make this two parts cooperative optimization, finally obtain the optimal compromise between system architecture and the heat exchange unit, according to on-site actual situations, further select optimal prioritization scheme then, method is specific as follows:

Resolution problem at first: from the angle of system optimization, the heat exchanger network synthesis problem equivalent is decomposed into system architecture and heat exchange unit optimization two parts, wherein, system architecture comprises the coupling and the heat exchange unit number of required public work consumption, stream thigh, and heat exchange unit comprises thermic load, operating temperature and the heat exchange area of every heat exchanger;

The team work optimization of network: comprise network architecture and the optimization of heat exchange unit two parts, adopt the team work evolvement method, cooperative optimization system architecture and heat exchange unit parameter;

The suitableeest prioritization scheme is selected: according to the on-site actual situations of object function, from a plurality of optimum prioritization schemes, select optimal prioritization scheme,

The team work optimization of described heat-exchange network comprises following steps:

(1) coding method

The individuality of leading species has been represented the system architecture of network, adopts incidence matrix to represent here, and its encoded content is a positive integer; The individuality of species of appraising and choosing excellent has been represented all heat exchange unit parameters of the heat-exchange network of a given system architecture, and its encoded content is a real number, and its individual lengths is relevant with corresponding system architecture, and this system architecture is corresponding to body one by one in the leading species.The body one by one of leading species can make up a complete heat-exchange network with the body one by one of the species of appraising and choosing excellent;

(2) adaptive value function

At first the species of appraising and choosing excellent are evolved, promptly the heat exchange unit parameter of the heat-exchange network of given system architecture is evolved, its ideal adaptation value function equals the integrated objective function of heat-exchange network; For leading species, be that system architecture is evolved, its ideal adaptation value function is to be provided by best individuality in the species of appraising and choosing excellent, and promptly utilizes the species of appraising and choosing excellent to estimate the quality of the system architecture of having evolved;

(3) evolutionary optimization process

At first, determine the population scale of the leading species and the species of appraising and choosing excellent, and the random initializtion population, each population of the species of wherein appraising and choosing excellent is corresponding to the body one by one of leading species, and the individuality in each population has been represented the parameter of whole heat exchange units of the heat-exchange network of a given system architecture;

Secondly, utilize local evolution algorithm to evolve to appraise and choose excellent the population of species, and select best individuality to represent as the individuality of this population, utilize general evolution algorithm to evolve and dominate the population of species, wherein a complete heat-exchange network is formed in the individuality representative in the species of appraising and choosing excellent that each individuality is corresponding with this individuality, and selects best heat-exchange network as the optimization of current best heat-exchange network optimal synthesis.

2, the intelligent optimization method of heat-exchange network optimal synthesis according to claim 1, it is characterized in that, when the pairing network architecture of the population of the species of appraising and choosing excellent changes, utilize to increase and deletion of node method and the local evolution algorithm population of species of appraising and choosing excellent of evolving: as if having increased heat exchange unit, increase and the deletion of node method under, only the evolve parameter of heat exchange unit of the heat exchange unit that increases newly and place level of local evolution algorithm; If deleted heat exchange unit, under increase and deletion of node method,, then delete heat exchange unit with the backward of original increase unit order if the place level once increased heat exchange unit, otherwise delete cells at random utilizes evolve heat exchange unit parameter with the place level of local evolution algorithm then; If the matching order of heat exchange stream thigh changes, utilize the associated heat exchange unit parameter of local evolution algorithm evolution.

Images