Industrial park integrated energy system development stage division method based on hybrid algorithm

Abstract

The invention discloses an industrial park integrated energy system development stage division method based on a hybrid algorithm. The method comprises the following steps: establishing an industrial park integrated energy system planning and design platform based on a digital twin model; setting multiple stages of development and evolution of the industrial park comprehensive energy system in a planning and designing period; predicting multi-stage multi-element loads in the planning design period; screening load indexes influencing system development stage division to form a multi-dimensional index matrix X with N ordered sample sequences and M index characteristics; respectively calculating the weight of each index feature by adopting a plurality of mixed weighting algorithms, and carrying out weight fusion to obtain a comprehensive weight coefficient of each index feature; and converting the standardized multi-dimensional index matrix into a one-dimensional feature vector based on the comprehensive weight coefficient of each index feature, taking the one-dimensional feature vector as a sample sequence Y for initial division, inputting the sample sequence Y into an optimal segmentation method, and dividing the planning and design period of the industrial park comprehensive energy system according to the optimal segmentation number.

Description

Technical Field

[0001] The present invention belongs to the technical field of planning and designing integrated energy systems for industrial parks, and specifically relates to a method for dividing development stages of integrated energy systems for industrial parks based on a hybrid algorithm. Background Art

[0002] In the planning and construction of industrial parks, integrated energy systems serve as a crucial hub for ensuring energy supply. Their planning and design should adapt to the evolution of industrial parks, thereby achieving a balance between supply and demand in terms of quantity, quality, time, and space. The investment and construction of integrated energy systems in industrial parks is an extremely complex process. The planning and design process is subject to simultaneous changes in park land planning, user influx, and demand load growth. These changes are influenced by multiple factors, including economic, policy, and technological factors, resulting in multi-stage dynamic changes.

[0003] To ensure unified planning and operational management of the entire industrial park's integrated energy system, the planning and design of the industrial park's integrated energy system should be a multi-stage, interconnected, and progressive process, from planning and design, phased construction, enterprise settlement, and subsequent stable operation. Each development stage has different load characteristics and evolves dynamically over time. How to divide the planning and design period of the industrial park's integrated energy system into multiple stages, ensuring that each stage conforms to its load characteristics and the park's dynamic development and changes, is an urgent issue. Summary of the Invention

[0004] To overcome the shortcomings of the aforementioned prior art, the present invention provides a hybrid algorithm-based method for categorizing the development stages of an industrial park's integrated energy system. This method involves establishing a digital twin-based integrated energy system planning and design platform for industrial parks and performing multi-stage, multivariate load forecasting. This method utilizes multiple hybrid weighting algorithms and optimal segmentation methods to categorize the development stages of the industrial park's integrated energy system. Each stage effectively reflects the load characteristics of the industrial park's integrated energy system, meeting the park's dynamic development and changes.

[0005] In order to solve the above technical problems, the technical solution of the present invention is:

[0006] A method for dividing the development stages of the integrated energy system of industrial parks based on a hybrid algorithm includes:

[0007] S1. Establish an integrated energy system planning and design platform for industrial parks based on digital twin models;

[0008] S2. Divide the development of the integrated energy system of the industrial park into multiple stages within the planning and design period; wherein the planning and design period is N years;

[0009] S3. Using the industrial park integrated energy system planning and design platform based on the digital twin model to predict the multi-factor loads at each development stage during the planning and design period, and obtaining multi-factor load prediction information for each development stage of the industrial park integrated energy system;

[0010] S4. Establish a multidimensional indicator matrix: Based on the multivariate load forecast information of each development stage of the system, the load variation characteristics of the system within N years are obtained, and M load indicators that affect the development stage division of the industrial park integrated energy system are screened. A multidimensional indicator matrix X with N ordered sample sequences and M indicator characteristics is established;

[0011] S5. After standardizing the multidimensional indicator matrix X, a variety of hybrid weighting algorithms are used to calculate the weights of each indicator feature, and the weights are fused to obtain the comprehensive weight coefficients of each indicator feature;

[0012] S6. Based on the comprehensive weight coefficient of each indicator feature, the standardized multidimensional indicator matrix is ​​converted into a one-dimensional feature vector to obtain the ordered sample sequence Y of the initial division;

[0013] S7. According to the ordered sample sequence Y of the initial division, based on the optimal segmentation method, determine the classification diameter, calculate the objective function, obtain the optimal segmentation and the optimal number of segments, divide the industrial park comprehensive energy system planning and design period according to the optimal segmentation and the optimal number of segments, and use the industrial park comprehensive energy system planning and design platform based on the digital twin model to evaluate and confirm the division results.

[0014] Furthermore, the S1, establishing an industrial park integrated energy system planning and design platform based on a digital twin model, includes the following steps:

[0015] Establish an integrated energy system planning process for industrial parks based on the demand model, structural model, behavior and state description model, and parameter model of the integrated energy system; the demand model includes a load forecasting model, a planning and design target model, and an energy supply and demand condition model; the structural model includes cooling, heating, electricity, and gas subsystems and equipment models, and a multi-energy coupling model; the behavior and state description model includes the behavior and state of energy production, transmission, distribution, and supply; the parameter model includes quantitative digital parameter models for meteorological factors, fuel prices, business development changes, policies, and regional planning policies, as well as candidate parameter models for equipment specifications, system performance constraints, and system operation constraints;

[0016] Utilize the industrial park comprehensive energy system planning process, follow the demand analysis, system design, subsystem design, system integration, and evaluation confirmation process, and combine the digital twin model to conduct demand modeling, system modeling, subsystem modeling, system integration modeling, and evaluation confirmation modeling for the industrial park comprehensive energy system, and build an industrial park comprehensive energy system planning and design platform based on the digital twin model;

[0017] The digital twin model uses digital twin technology to expand the industrial park integrated energy system planning process to the entire integrated energy system development process, and constructs the industrial park integrated energy digital twin system in the virtual world through digital means;

[0018] The demand analysis is to analyze the requirements of the system, subsystem, software and hardware of the comprehensive energy system planning and design platform of the industrial park, including resource endowment analysis, energy price analysis, load forecast analysis, system structure analysis, system topology analysis, planning and design optimization analysis, and planning and design scheme analysis; the system design is based on the demand analysis, and each part of the demand analysis is constructed into a corresponding planning module, and the interface between each functional module is clarified; the subsystem design includes the cold, heat, electricity, and gas subsystem equipment structure model design, subsystem operation analysis model design, subsystem simulation model design and multi-energy flow coupling model design; the system integration is to combine different subsystems and software and hardware; the evaluation confirmation is to conduct a trial run in the digital twin virtual environment before the actual operation of the system to determine whether the system planning and design scheme meets user needs and system requirements.

[0019] Furthermore, in step S2, the method for dividing the development of the industrial park comprehensive energy system into multiple stages during the planning and design period is: based on the scale expansion of each functional plot in the industrial park, the occupancy rate of enterprises, industrial structure adjustment, fluctuations in policies and market economic factors, external climate and environmental factors, and actual life span differences of various types of equipment, combined with the dynamic change characteristics of the load of the industrial park comprehensive energy system during the planning and design period, the development and evolution of the industrial park comprehensive energy system during the planning and design period is divided into multiple stages. The stages obtained after division include at least the system formation stage, the growth stage, the maturity stage, and the decline stage. The load characteristics of each stage are different, and each stage can also be subdivided into multiple sub-stages.

[0020] Furthermore, in step S3, the specific method of using the industrial park comprehensive energy system planning and design platform based on the digital twin model to predict the multiple loads in each development stage during the planning and design period is to collect and process the relevant meteorological data and multiple energy consumption data such as cold, heat, electricity, and gas of the industrial park, and use the load forecasting model to carry out load forecasting analysis, and then carry out multiple load forecasting for the loads in each development stage during the planning and design period, and output the multiple forecasting results of the industrial park.

[0021] Furthermore, in step S4, the load variation characteristics during the system planning and design period include the multi-element load density, multi-element load maximum value, multi-element load average value, multi-element load peak-to-valley ratio, multi-element load variation, and multi-element load growth rate of the system within N years; the load variation characteristics match the development stage of the industrial park integrated energy system;

[0022] In the process of screening and obtaining the M load indicators that affect the development stage division of the industrial park comprehensive energy system, the algorithm used is the kernel principal component analysis method KPCA;

[0023] The multidimensional indicator matrix X is expressed as:

[0024]

[0025] Among them, x ij is the jth indicator feature of the i-th ordered sample sequence.

[0026] Furthermore, in step S5, after the multidimensional indicator matrix is ​​standardized, a variety of hybrid weighting algorithms are used to calculate the weights of each indicator feature respectively, and after weight fusion, the comprehensive weight coefficient of each indicator feature is obtained, including:

[0027] The characteristics of each indicator in the multidimensional indicator matrix are standardized to obtain the standardized matrix X′=[x ij ′] NM ; x max,j 、x min,j are the maximum and minimum values ​​of the indicator features in the jth column of the multidimensional indicator matrix X;

[0028] The entropy weight method, CRITIC weighting method and hierarchical analysis method are used to calculate the weight of each indicator feature respectively, and the multiple weights obtained by calculation are fused based on game theory to obtain the comprehensive weight coefficient of each indicator feature.

[0029] Furthermore, the entropy weight method is used to calculate the weight of each indicator feature. The specific method is:

[0030] Calculate the entropy value e of the j-th indicator feature of the i-th ordered sample sequence j , expressed as:

[0031]

[0032] Where, i = 1, 2, ..., N; j = 1, 2, ..., M; Assume that when p ij = 0, p ij lnp ij =0;

[0033] According to the entropy value e j Calculate the weight w of the j-th indicator feature 1j , expressed as:

[0034]

[0035] Where, 0≤w 1j ≤1;

[0036] Then we get the weight set w1 of each indicator feature calculated by entropy weight method, which is expressed as:

[0037] w1=[w 11 ,w 12 ,...,w 1j ,...,w 1M ]

[0038] The CRITIC weighting method is used to calculate the weight of each indicator feature. The specific method is:

[0039] Calculate the mean of the j-th indicator feature Expressed as:

[0040]

[0041] Calculate the standard deviation S of the j-th indicator feature j , expressed as:

[0042]

[0043] Calculate the conflict coefficient R of the j-th indicator feature j , expressed as:

[0044]

[0045] Where r ij is the correlation coefficient, which indicates the conflict between the characteristics of each indicator; r ij The absolute value of r ij The larger the |, the stronger the correlation with other indicator characteristics;

[0046] Calculate the information content C of the j-th indicator feature j , expressed as:

[0047] C j =S j ·R j

[0048] Calculate the weight w of the j-th indicator feature 2j , expressed as:

[0049]

[0050] Then we get the weight set w2 of each indicator feature calculated by the CRITIC weighting method, which is expressed as:

[0051] w2=[w 21 ,w 22 ,...,w 2j ,...,w 2M ]

[0052] The weight of each indicator feature is calculated using the hierarchical analysis method. The specific method is:

[0053] The index characteristics in the standardized matrix X′ are analyzed by the hierarchical analysis method to construct the judgment matrix G=(g ij ) M×M , g ij It is used to characterize the relative importance of the index feature i and the index feature j in X′, and the Satty scaling method is used to measure the relative importance of g ij Scaling is performed, and g ij =1 / g ji , g ii =1;

[0054] After normalizing the judgment matrix G, the normalized judgment matrix A is obtained norm , specifically including:

[0055] Calculate the sum of the jth column of the judgment matrix G

[0056] For each element g of the judgment matrix G ij Perform normalization:

[0057] Furthermore, the normalized judgment matrix A norm It can be expressed as:

[0058]

[0059] Calculate the normalized judgment matrix A norm The average value of each row element w' 3i :

[0060]

[0061] Let i = j, then the weight w' of each indicator feature j is initially obtained 3j =w' 3i ;

[0062] Furthermore, the weight set w3' of each indicator feature is initially obtained, that is, w3'=[w' 31 ,w' 32 ,...,w'3j ,...,w' 3M ];

[0063] Calculate the consistency index CI, expressed as:

[0064]

[0065] λ max is the maximum eigenvalue in the judgment matrix G;

[0066] Calculate the consistency test coefficient CR, expressed as:

[0067]

[0068] RI is a random consistency index, which is a constant obtained based on a random matrix under dimension M. For example, when M = 9, the value of RI is 1.45;

[0069] If CR is less than 0.1, the test is passed; otherwise, the test fails. After re-adjusting the judgment matrix G, recalculate CR and judge whether CR is less than 0.1 until the test is passed. When the CR test is passed, the corresponding w3' is the weight set w3 of each indicator feature under the hierarchical analysis method, w3 = [w 31 ,w 32 ,...,w 3j ,...,w 3M ].

[0070] Furthermore, the multiple weights obtained by calculation based on game theory are integrated to obtain the comprehensive weight coefficient of each indicator feature, including:

[0071] Linearly combine multiple weight vectors at each level to form a preliminary weight set W, expressed as:

[0072]

[0073] α k is the linear combination coefficient, for each linear combination coefficient α k The objective function to be optimized is expressed as:

[0074]

[0075] Construct a system of linear equations, expressed as:

[0076]

[0077] Solve the linear equations to obtain the corresponding α k , and perform standardization to obtain the standardized linear combination coefficient α k ′, expressed as:

[0078]

[0079] Calculate the comprehensive weight coefficient wm of the jth indicator feature in the multidimensional indicator matrix X j , expressed as:

[0080] wm j =α1′w 1j +α′2w 2j +α′3w 3j

[0081] Among them, α1′, α2′, and α3′ are the linear combination coefficients of the standardized entropy weight method, CRITIC weight method, and hierarchical analysis method weights, respectively.

[0082] Furthermore, the comprehensive weight coefficient matrix wm corresponding to the multidimensional indicator matrix X is formed, which is expressed as:

[0083]

[0084] Furthermore, in step S6, the standardized multidimensional indicator matrix is ​​converted into a one-dimensional feature vector to obtain the initial divided sample sequence Y, which is expressed as:

[0085]

[0086] Furthermore, in step S7, the ordered sample sequence Y initially divided is determined based on the optimal segmentation method to determine the classification diameter, calculate the objective function, and obtain the optimal segmentation and the optimal number of segments, which specifically includes the following steps:

[0087] The initially divided ordered sample sequence Y includes N ordered sample sequences. Let the ordered sample segment interval from the sth to the tth in the N ordered sample sequences be represented as {x s ,x s+1 ,…,x t}, each ordered sample is a row vector, and the complete representation of the ordered sample segment interval from the sth to the tth is: {(x s1 ,x s2 ,…,x sM ),(x (s+1)1 ,x (s+1)2 ,…,x (s+1)M ),…,(x t1 ,x t2 ,…,x tM )}, the classification diameter D(s,t) is defined as:

[0088]

[0089] Among them, x ais an ordered sample from the sth to the tth ordered samples.

[0090] The calculation objective function specifically includes:

[0091] Suppose that N ordered sample sequences are divided, there may be 2 N-1 If N ordered sample sequences are randomly divided into K segments, the specific segmentation process can be expressed as:

[0092]

[0093] Among them, para i represents the i-th segment after the N ordered sample sequences are divided into K segments, x i +j represents the j+1th sample in the i-th segment.

[0094] In order to calculate the total sum of squares of deviations of each segment, the objective function G(N,K) is defined as:

[0095]

[0096] Among them, D(k,k-1) represents the classification diameter of the ordered sample segment interval from k-1 to k; the smaller the objective function value G(N,K), the smaller the difference within each segment and the greater the difference between segments.

[0097] The segmentation method that minimizes the objective function value is the optimal segmentation, that is:

[0098]

[0099] Determining the optimal number of segments includes:

[0100] Plot the objective function G * Curve G of (N,K) changing with the number of segments K * (N,K)~K; the absolute value formula of the slope of the curve at each segment number is expressed as:

[0101] γ(K)=|G * (N,K)-G * (N,K-1)|

[0102] According to the absolute value of the slope that changes with the number of segments, a γ(K)~K curve is drawn, and the number of segments K corresponding to the maximum value is the optimal number of segments.

[0103] The beneficial effects of the present invention are:

[0104] (1) The integrated energy planning and design platform based on the digital twin model proposed in this invention, with the help of the system engineering concept and the digital twin model as the medium, scientifically organizes the integrated energy system planning and design according to the system requirements, system design, subsystem design, system integration, evaluation and confirmation, and supports its implementation, realizing the coordinated coupling integrated optimization design of multiple energy flow subsystems in the integrated energy system, improving the quantitative, refined and scientific management level of the planning and design process of complex open energy systems, and providing an intelligent and scientific platform for subsequent load forecasting and development stage division;

[0105] (2) The present invention adopts multiple hybrid weighting algorithms to calculate the weights of each indicator feature separately, which can take into account the advantages of various weighting algorithms, quantify the uncertainty degree of each indicator feature in the development stage division and the differences between different indicator features. At the same time, the weights after game theory synthesis are more realistic than the weights obtained by the traditional single weighting method, and can also better reflect the impact of each indicator feature on the development stage division of the industrial park;

[0106] (3) Based on the optimal segmentation method, the development stages of the industrial park are divided. Each stage obtained by the division can better reflect the development characteristics of the comprehensive energy system of the industrial park and meet the dynamic development and changes of the park.

[0107] (4) During the development of the industrial park, managers can evaluate and confirm the results of the industrial park development stage division based on the industrial park comprehensive energy system planning and design platform based on the digital twin model. If the division results do not match the current actual situation of the park, managers can adjust resource allocation and optimize energy scheduling through feedback from the platform to ensure that the development plan is in line with the actual needs of the park, thereby improving the resource utilization efficiency and development flexibility of the park; if the park development stage divided by the platform is consistent with the current actual situation, managers can confirm the rationality of the existing resource allocation and energy scheduling plan based on this, further revise and adjust the park management and planning measures, ensure the steady progress of the park development, and enhance the accuracy of the prediction of future development stages.

[0108] Other features and advantages will be described in the following description, and in part will become apparent from the description or understood through implementation of the invention. The objects and other advantages of the invention are realized and obtained by the structures particularly pointed out in the description and the drawings.

[0109] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0110] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0111] Figure 1 This is a flow chart of a method for dividing the development stages of an industrial park integrated energy system based on a hybrid algorithm according to the present invention;

[0112] Figure 2 This is a schematic diagram of the demand analysis of the integrated energy system planning and design platform for industrial parks of the present invention;

[0113] Figure 3 Schematic diagram of the γ(K)~K curve for determining the optimal number of segments in the present invention. DETAILED DESCRIPTION

[0114] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0115] Example 1

[0116] Figure 1 This is a flow chart of a method for dividing the development stages of an industrial park integrated energy system based on a hybrid algorithm involved in the present invention.

[0117] like Figure 1 As shown, this embodiment 1 provides a method for dividing the development stages of an industrial park integrated energy system based on a hybrid algorithm, which includes:

[0118] S1. Establish an integrated energy system planning and design platform for industrial parks based on digital twin models;

[0119] S2. Divide the development of the integrated energy system of the industrial park into multiple stages within the planning and design period; wherein the planning and design period is N years;

[0120] S3. Using the industrial park integrated energy system planning and design platform based on the digital twin model to predict the multi-factor loads at each development stage during the planning and design period, and obtaining multi-factor load prediction information for each development stage of the industrial park integrated energy system;

[0121] S4. Establish a multidimensional indicator matrix: Based on the multivariate load forecast information of each development stage of the system, the load variation characteristics of the system within N years are obtained, and M load indicators that affect the development stage division of the industrial park integrated energy system are screened. A multidimensional indicator matrix X with N ordered sample sequences and M indicator characteristics is established;

[0122] S5. After standardizing the multidimensional indicator matrix X, a variety of hybrid weighting algorithms are used to calculate the weights of each indicator feature, and the weights are fused to obtain the comprehensive weight coefficients of each indicator feature;

[0123] S6. Based on the comprehensive weight coefficient of each indicator feature, the standardized multidimensional indicator matrix is ​​converted into a one-dimensional feature vector to obtain the ordered sample sequence Y of the initial division;

[0124] S7. According to the ordered sample sequence Y of the initial division, based on the optimal segmentation method, determine the classification diameter, calculate the objective function, obtain the optimal segmentation and the optimal number of segments, divide the industrial park comprehensive energy system planning and design period according to the optimal segmentation and the optimal number of segments, and use the industrial park comprehensive energy system planning and design platform based on the digital twin model to evaluate and confirm the division results.

[0125] Figure 2 It is a demand analysis diagram of the integrated energy system planning and design platform for industrial parks involved in the present invention.

[0126] In this embodiment, in step S1, an industrial park integrated energy system planning and design platform based on a digital twin model is established, including:

[0127] Establish an integrated energy system planning process for industrial parks based on the demand model, structural model, behavior and state description model, and parameter model of the integrated energy system; the demand model includes a load forecasting model, a planning and design target model, and an energy supply and demand condition model; the structural model includes cooling, heating, electricity, and gas subsystems and equipment models, and a multi-energy coupling model; the behavior and state description model includes the behavior and state of energy production, transmission, distribution, and supply; the parameter model includes quantitative digital parameter models for meteorological factors, fuel prices, business development changes, policies, and regional planning policies, as well as candidate parameter models for equipment specifications, system performance constraints, and system operation constraints;

[0128] Utilize the industrial park comprehensive energy system planning process, follow the demand analysis, system design, subsystem design, system integration, and evaluation confirmation process, and combine the digital twin model to conduct demand modeling, system modeling, subsystem modeling, system integration modeling, and evaluation confirmation modeling for the industrial park comprehensive energy system, and build an industrial park comprehensive energy system planning and design platform based on the digital twin model;

[0129] The digital twin model uses digital twin technology to expand the industrial park integrated energy system planning process to the entire integrated energy system development process, and constructs the industrial park integrated energy digital twin system in the virtual world through digital means;

[0130] The demand analysis is to analyze the system, subsystem, software and hardware requirements of the industrial park comprehensive energy system planning and design platform, and establish the corresponding use case model (such as Figure 2 ), including resource endowment analysis model, energy price model, load forecasting model, system structure analysis model, system topology analysis model, planning and design optimization analysis model, and planning and design scheme analysis model; the system design is based on demand analysis, decomposing the demand analysis into multiple functional modules, and clarifying the interfaces between the functional modules; the subsystems include cold, hot, electric, and gas subsystem equipment structure models, subsystem operation analysis models, subsystem simulation models, and multi-energy flow coupling models; the system integration is the combination of different subsystems and software and hardware; the evaluation and confirmation is a trial run in a digital twin virtual environment before the actual operation of the system to determine whether the system planning and design scheme meets user needs and system requirements.

[0131] It should be noted that with the dynamic development of industrial parks, the scale and complexity of integrated energy systems are constantly increasing. The integrated energy construction of industrial parks has put forward new requirements for the overall and systematic planning and design of energy systems. The planning and design of industrial integrated energy systems has gone beyond the scope of traditional planning and design disciplines, covering all aspects of energy, economy, environment, and industrial collaboration. It is also a complex system planning and design process. Here, by introducing the concept of systems engineering and integrating systems engineering theoretical methods with digital twin technology, a comprehensive energy system planning and design platform based on digital twin models is proposed to guide the planning and design of the next generation of integrated energy systems. This includes multi-stage and multi-dimensional load forecasting of the integrated energy system of industrial parks and the subsequent development stage division, all of which are implemented based on the comprehensive energy system planning and design platform.

[0132] The digital twin-based integrated energy system planning and design approach applies systems engineering to integrated energy system planning and design, centering digital twins throughout demand management, system design, subsystem design, system integration, and evaluation and validation. Digital twins can be shared across multiple planning phases, addressing the increasing complexity of integrated energy systems while reducing the time, cost, and risk of planning and design, solution delivery, and updating and improving system design solutions.

[0133] The digital twin planning and design platform for the integrated energy system of industrial parks also includes basic operation modules and planning modules.

[0134] The basic operation module includes:

[0135] 1) System management module, which includes user addition, password modification and user history operation records;

[0136] 2) Project management module. This module has the functions of creating, opening, deleting and exiting planning and design projects;

[0137] Planning module, including:

[0138] 1) Resource endowment analysis module, which mainly focuses on environmental data information and calls urban environmental data information (temperature, wind speed, humidity, and light radiation information) to support the use of renewable energy equipment;

[0139] 2) Load forecasting module: Load demand information is the basis and prerequisite for the planning and design of integrated energy systems. It includes three types of load data: electricity, heat, and cooling. This module has the function of manually entering typical daily load values ​​and uploading typical daily load data according to a table template;

[0140] 3) Energy price analysis module, which has the functions of inputting energy raw material prices, energy product prices, and setting price curves;

[0141] 4) System structure analysis module, which has the functions of selecting / deleting existing devices, adding new devices, setting parameters, and calling the device digital twin database based on the selected devices;

[0142] 5) Integrated energy system topology analysis module, which has the function of reading the device storage database and simulating the integrated energy system based on the digital twin model of the integrated energy system structure;

[0143] 6) Planning and Design Optimization Analysis Module: This module reads environmental data, load data, energy market data, equipment digital twin models, and integrated energy system structure digital twin models into the database, uses them as simulation constraints, and calls the planning optimization simulation program to perform simulations. The completed data is then stored in the simulation results database.

[0144] 7) Planning and design scheme analysis module: Users can view equipment types, quantities, equipment capacity configuration, and economic and environmental evaluation data of planning and design schemes.

[0145] In this embodiment, in step S2, the development of the comprehensive energy system of the industrial park is divided into multiple stages during the planning and design period. The specific method is: based on the scale expansion of each functional plot in the industrial park, the occupancy rate of enterprises, industrial structure adjustment, fluctuations in policies and market economic factors, external climate and environmental factors, and actual life span differences of various types of equipment, combined with the dynamic change characteristics of the load of the comprehensive energy system of the industrial park during the planning and design period, the development and evolution of the comprehensive energy system of the industrial park during the planning and design period is divided into multiple stages. The stages obtained after division include at least the system formation stage, the growth stage, the maturity stage, and the renewal or decline stage. The load characteristics of each stage are different, and each stage can also be subdivided into multiple sub-stages.

[0146] It should be noted that, usually, the load density of the integrated energy system of an industrial park grows slowly in the early stages of development; as the load density grows faster, it enters a rapid development stage, and in a certain year, the load density growth rate reaches a maximum value; then the load density continues to grow rapidly and enters a post-development stage; when it reaches the saturated development stage, with the restrictions of long-term urban population planning, land resources, energy utilization efficiency, etc., the load density will eventually maintain a steady growth until it reaches the saturation value. Therefore, load indicators are the main basis for dividing the development stages, and the load indicators of this application are obtained based on the multi-stage multi-dimensional load forecast information of the park, and the dynamic development and change factors of the park have been comprehensively considered.

[0147] In this embodiment, in step S3, the specific method of using the industrial park comprehensive energy system planning and design platform based on the digital twin model to predict the multi-dimensional loads in each development stage during the planning and design period is to collect and process the relevant meteorological data and multi-dimensional energy consumption data such as cooling, heating, electricity, and gas of the industrial park, and use the load forecasting model to carry out load forecasting analysis, and then carry out multi-dimensional load forecasting for the loads in each development stage during the planning and design period, and output the multi-dimensional forecasting results of the industrial park.

[0148] In this embodiment, in S4, the load variation characteristics during the system planning and design period include multi-element load density, multi-element load maximum value, multi-element load average value, multi-element load peak-to-valley ratio, multi-element load variation, and multi-element load growth rate within a multi-stage preset time period (N years); the load variation characteristics match the development and evolution stage of the industrial park integrated energy system;

[0149] When screening the load indicators that affect the development stage division of the industrial park's integrated energy system, the algorithm used is the kernel principal component analysis method KPCA;

[0150] The multidimensional indicator matrix X is expressed as:

[0151]

[0152] Among them, x ij is the jth indicator feature of the i-th ordered sample sequence.

[0153] In this embodiment, in step S5, after the multidimensional indicator matrix is ​​standardized, a plurality of hybrid weighting algorithms are used to calculate the weights of the indicator features respectively, and the weights are fused to obtain the comprehensive weight coefficients of the indicator features, including:

[0154] Standardize the characteristics of each indicator in the multidimensional indicator matrix to obtain the standardized matrix x max,j 、x min,j are the maximum and minimum values ​​of the indicator features in the jth column of the multidimensional indicator matrix X;

[0155] The entropy weight method, CRITIC weighting method, and hierarchical analysis method are used to calculate the weights of each indicator feature. Based on game theory, the multiple weights obtained by calculation are integrated to obtain the comprehensive weight coefficient of each indicator feature.

[0156] In this embodiment, the entropy weight method is used to calculate the weight of each indicator feature, including:

[0157] Calculate the entropy value e of the j-th indicator feature of the i-th ordered sample sequence j , expressed as:

[0158]

[0159] Where, i = 1, 2, ..., N; j = 1, 2, ..., M; Assume that when p ij = 0, p ij lnp ij =0;

[0160] According to the entropy value e j Calculate the weight w of the j-th indicator feature 1j , expressed as:

[0161]

[0162] Where, 0≤w 1j ≤1;

[0163] Then we get the weight set w1 of each indicator feature calculated by entropy weight method, which is expressed as:

[0164] w1=[w 11 ,w 12 ,...,w 1j ,...,w 1M ]

[0165] The CRITIC weighting method is used to calculate the weight of each indicator feature. The specific method is:

[0166] Calculate the mean of the j-th indicator feature Expressed as:

[0167]

[0168] Calculate the standard deviation S of the j-th indicator feature j , expressed as:

[0169]

[0170] Calculate the conflict coefficient R of the j-th indicator feature j , expressed as:

[0171]

[0172] Where r ij is the correlation coefficient, which indicates the conflict between the characteristics of each indicator; r ij The absolute value of r ij The larger the |, the stronger the correlation with other indicator characteristics;

[0173] Calculate the information content C of the j-th indicator feature j , expressed as:

[0174] C j =S j ·R j

[0175] Calculate the weight w of the j-th indicator feature 2j , expressed as:

[0176]

[0177] Then we get the weight set w2 of each indicator feature calculated by the CRITIC weighting method, which is expressed as:

[0178] w2=[w 21 ,w 22 ,...,w 2j ,...,w 2M ]

[0179] In this embodiment, the weight of each indicator feature is calculated using the hierarchical analysis method, including:

[0180] The index characteristics in the standardized matrix X′ are analyzed by the hierarchical analysis method to construct the judgment matrix G=(g ij ) M×M , g ijIt is used to characterize the relative importance of the indicator feature i and the indicator feature j in X′, and to use the 1 to 9 scale method to measure g ij Scaling is performed, and g ij =1 / g ji , g ii =1;

[0181] The normalized judgment matrix A obtained by normalizing the judgment matrix G norm , specifically including:

[0182] Calculate the sum of the jth column of the judgment matrix G

[0183] For each element g of the judgment matrix G ij Perform normalization:

[0184] Furthermore, the normalized judgment matrix A norm It can be expressed as:

[0185]

[0186] Calculate the normalized judgment matrix A norm The average value of each row element w' 3i :

[0187]

[0188] Let i = j, then the weight w' of each indicator feature j is initially obtained 3j , that is, w' 3j =w' 3i ;

[0189] Furthermore, the weight set w3' of each indicator feature is initially obtained, that is, w3'=[w' 31 ,w' 32 ,...,w' 3j ,...,w' 3M ];

[0190] Calculate the consistency index CI, expressed as:

[0191]

[0192] λ max is the maximum eigenvalue in the judgment matrix G;

[0193] Calculate the consistency test coefficient CR, expressed as:

[0194]

[0195] RI is a random consistency index, which is a constant obtained based on a random matrix under dimension M. For example, when M = 9, the value of RI is 1.45;

[0196] If CR is less than 0.1, the test is passed; otherwise, the test fails. After re-adjusting the judgment matrix G, recalculate CR and judge whether CR is less than 0.1 until the test is passed. When the CR test is passed, the corresponding w3' is the weight set w3 of each indicator feature under the hierarchical analysis method, w3 = [w 31 ,w 32 ,...,w 3j ,...,w 3M ].

[0197] The game theory-based method is used to fuse the multiple weights obtained by calculation to obtain the comprehensive weight coefficients of each indicator feature, including:

[0198] Linearly combine multiple weight vectors at each level to form a preliminary weight set W, expressed as:

[0199]

[0200] α k is the linear combination coefficient, for each linear combination coefficient α k The objective function to be optimized is expressed as:

[0201]

[0202] Construct a system of linear equations, expressed as:

[0203]

[0204] Solve the linear equations to obtain the corresponding α k , and perform standardization to obtain the standardized linear combination coefficient α k ′, expressed as:

[0205]

[0206] Calculate the comprehensive weight coefficient wm of the jth indicator feature in the multidimensional indicator matrix X j , expressed as:

[0207] wm j =α1′w 1j +α′2w 2j +α′3w 3j

[0208] Among them, α1′, α2′, and α3′ are the linear combination coefficients of the standardized entropy weight method, CRITIC weight method, and hierarchical analysis method weights, respectively.

[0209] Furthermore, the comprehensive weight coefficient matrix wm corresponding to the multidimensional indicator matrix X is formed, which is expressed as:

[0210]

[0211] It should be noted that the weights of each indicator characteristic calculated using a single weighting method are subject to significant volatility, and using a single weighting method can easily result in weighting results that are influenced by human preferences or objective data. The weights derived through game theory are more realistic than those derived using a traditional single weighting method and can better reflect the impact of each indicator characteristic on the development stage classification of industrial parks.

[0212] In this embodiment, in step S6, the standardized multidimensional indicator matrix is ​​converted into a one-dimensional feature vector to obtain the initial divided sample sequence Y, which is expressed as:

[0213]

[0214] In this embodiment, in step S7, the method of determining the classification diameter, calculating the objective function, and obtaining the optimal segmentation and the optimal number of segments based on the initially divided ordered sample sequence Y based on the optimal segmentation method specifically includes the following steps:

[0215] The initially divided ordered sample sequence Y includes N ordered sample sequences. Let the ordered sample segment interval from the sth to the tth in the N ordered sample sequences be represented as {x s ,x s+1 ,…,x t}, each ordered sample is a row vector, and the complete representation of the ordered sample segment interval from the sth to the tth is: {(x s1 ,x s2 ,…,x sM ),(x (s+1)1 ,x (s+1)2 ,…,x (s+1)M ),…,(x t1 ,x t2 ,…,x tM )}, the classification diameter D(s,t) is defined as:

[0216]

[0217] Among them, x a is an ordered sample from the sth to the tth ordered samples.

[0218] The calculation objective function specifically includes:

[0219] Suppose that N ordered sample sequences are divided, there may be 2 N-1If N ordered sample sequences are randomly divided into K segments, the specific segmentation process can be expressed as:

[0220]

[0221] Among them, para i represents the i-th segment after the N ordered sample sequences are divided into K segments, x i +j represents the j+1th sample in the i-th segment.

[0222] In order to calculate the total sum of squares of deviations of each segment, the objective function G(N,K) is defined as:

[0223]

[0224] Among them, D(k,k-1) represents the classification diameter of the ordered sample segment interval from k-1 to k; the smaller the objective function value G(N,K), the smaller the difference within each segment and the greater the difference between segments.

[0225] The segmentation method that minimizes the objective function value is the optimal segmentation, that is:

[0226]

[0227] Determining the optimal number of segments includes:

[0228] Plot the objective function G * Curve G of (N,K) changing with the number of segments K * (N,K)~K; the absolute value formula of the slope of the curve at each segment number is expressed as:

[0229] γ(K)=|G * (N,K)-G * (N,K-1)|

[0230] According to the absolute value of the slope that changes with the number of segments, a curve γ(K)~K curve is drawn, and the number of segments K corresponding to the maximum value is the optimal number of segments. Figure 3 It is a schematic diagram of the γ(K)~K curve for determining the optimal number of segments involved in the present invention.

[0231] like Figure 3 As shown in the figure, the minimum objective function value γ(K) under different segmentation methods (corresponding to different K values) is obtained. To determine the optimal number of segments K, a γ(K)-K curve is plotted. When K = 6, the corresponding γ(K) = 0.28 is the maximum value, and the optimal number of segments is determined to be 6. The corresponding segmentation pattern is [5, 8, 11, 14, 19, 40], which divides the 40-year planning and design period into [1-5, 6-8, 9-11, 12-14, 15-19, 20-40].

[0232] The industrial park comprehensive energy system planning and design platform based on the digital twin model evaluates and confirms the division results. If the division results are inconsistent with the current actual situation of the park, managers can adjust resource allocation and optimize energy scheduling through real-time feedback results on the platform to ensure that the development plan is in line with the actual needs of the park, thereby improving the park's resource utilization efficiency and development flexibility; if the park development stage divided by the platform is consistent with the current actual situation, managers can confirm the rationality of the existing resource allocation and energy scheduling plans based on this, further consolidate optimization measures, ensure the steady progress of park development, and at the same time enhance the accuracy of predictions for future development stages.

[0233] In the several embodiments provided in this application, it should be understood that the proposed system and method can also be implemented in other ways. The system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a portion of code, and the module, program segment or a portion of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or can be implemented using a combination of dedicated hardware and computer instructions.

[0234] In addition, the functional modules in the various embodiments of the present invention may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0235] If the function is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can essentially be understood as a part that contributes to the existing technology or a part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, ReadOnly Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0236] With the above-described preferred embodiments of the present invention as a guide, and with reference to the above description, relevant personnel are fully capable of making various changes and modifications without departing from the technical scope of this invention. The technical scope of this invention is not limited to the contents of the specification and must be determined according to the scope of the claims.

Claims

1. A method for dividing the development stages of the integrated energy system of an industrial park based on a hybrid algorithm, characterized by: It includes: S1. Establish an integrated energy system planning and design platform for industrial parks based on digital twin models; S2. Divide the development of the integrated energy system of the industrial park into multiple stages within the planning and design period; wherein the planning and design period is N years; S3. Using the industrial park integrated energy system planning and design platform based on the digital twin model to predict the multi-factor loads at each development stage during the planning and design period, and obtaining multi-factor load prediction information for each development stage of the industrial park integrated energy system; S4. Establish a multidimensional indicator matrix: Based on the multivariate load forecast information of each development stage of the system, the load variation characteristics of the system within N years are obtained, and M load indicators that affect the development stage division of the industrial park integrated energy system are screened. A multidimensional indicator matrix X with N ordered sample sequences and M indicator characteristics is established; S5. After standardizing the multidimensional indicator matrix X, a variety of hybrid weighting algorithms are used to calculate the weights of each indicator feature, and the weights are fused to obtain the comprehensive weight coefficients of each indicator feature; S6. Based on the comprehensive weight coefficient of each indicator feature, the standardized multidimensional indicator matrix is ​​converted into a one-dimensional feature vector to obtain the ordered sample sequence Y of the initial division; S7. According to the ordered sample sequence Y of the initial division, based on the optimal segmentation method, determine the classification diameter, calculate the objective function, obtain the optimal segmentation and the optimal number of segments, divide the industrial park comprehensive energy system planning and design period according to the optimal segmentation and the optimal number of segments, and use the industrial park comprehensive energy system planning and design platform based on the digital twin model to evaluate and confirm the division results.

2. The method for dividing the development stages of the industrial park integrated energy system based on the hybrid algorithm according to claim 1 is characterized in that: S1, establishes an industrial park integrated energy system planning and design platform based on the digital twin model, the specific method is as follows: Establish an integrated energy system planning process for industrial parks based on the demand model, structural model, behavior and state description model, and parameter model of the integrated energy system; the demand model includes a load forecasting model, a planning and design target model, and an energy supply and demand condition model; the structural model includes cooling, heating, electricity, and gas subsystems and equipment models, and a multi-energy coupling model; the behavior and state description model includes the behavior and state of energy production, transmission, distribution, and supply; the parameter model includes quantitative digital parameter models for meteorological factors, fuel prices, business development changes, policies, and regional planning policies, as well as candidate parameter models for equipment specifications, system performance constraints, and system operation constraints; Based on the industrial park comprehensive energy system planning process, we will conduct demand modeling, system modeling, subsystem modeling, system integration modeling, and evaluation and confirmation modeling for the industrial park comprehensive energy system in combination with the digital twin model, following the demand analysis, system design, subsystem design, system integration, and evaluation and confirmation processes. This will help us build an industrial park comprehensive energy system planning and design platform based on the digital twin model. The digital twin model uses digital twin technology to expand the industrial park integrated energy system planning process to the entire integrated energy system development process, and constructs the industrial park integrated energy digital twin system in the virtual world through digital means; The demand analysis is to analyze the requirements of the system, subsystem, software and hardware of the comprehensive energy system planning and design platform of the industrial park, including resource endowment analysis, energy price analysis, load forecast analysis, system structure analysis, system topology analysis, planning and design optimization analysis, and planning and design scheme analysis; the system design is based on the demand analysis, and each part of the demand analysis is constructed into a corresponding planning module, and the interface between each functional module is clarified; the subsystem design includes the cold, heat, electricity, and gas subsystem equipment structure model design, subsystem operation analysis model design, subsystem simulation model design and multi-energy flow coupling model design; the system integration is to combine different subsystems and software and hardware; the evaluation confirmation is to conduct a trial run in the digital twin virtual environment before the actual operation of the system to determine whether the system planning and design scheme meets user needs and system requirements.

3. The method for dividing the development stages of the industrial park integrated energy system based on the hybrid algorithm according to claim 1 is characterized in that: In the step S2, the development of the comprehensive energy system of the industrial park is divided into multiple stages during the planning and design period. The specific method is: based on the scale expansion of each functional plot in the industrial park, the occupancy rate of enterprises, industrial structure adjustment, fluctuations in policies and market economic factors, external climate and environmental factors, and actual life span differences of various equipment, combined with the dynamic change characteristics of the load of the comprehensive energy system of the industrial park during the planning and design period, the development and evolution of the comprehensive energy system of the industrial park during the planning and design period is divided into multiple stages. The stages obtained after division include at least the system formation stage, the growth stage, the maturity stage, and the renewal or decline stage, and the load characteristics of each stage are different.

4. The method for dividing the development stages of an industrial park integrated energy system based on a hybrid algorithm according to claim 1 is characterized in that: In S4, the load variation characteristics of the system within N years include the multi-load density, multi-load maximum value, multi-load average value, multi-load peak-to-valley ratio, multi-load variation, and multi-load growth rate of the system within N years; the load variation characteristics match the development and evolution stage of the integrated energy system of the industrial park; In the process of screening and obtaining the M load indicators that affect the development stage division of the industrial park comprehensive energy system, the algorithm used is the kernel principal component analysis method KPCA; The multidimensional indicator matrix X is expressed as: x ij is the jth indicator feature of the i-th ordered sample sequence.

5. The method for dividing the development stages of the industrial park integrated energy system based on the hybrid algorithm according to claim 4 is characterized in that: In S5, after the multidimensional indicator matrix is ​​standardized, a variety of hybrid weighting algorithms are used to calculate the weights of each indicator feature respectively, and after weight fusion, the comprehensive weight coefficient of each indicator feature is obtained, including: The characteristics of each indicator in the multidimensional indicator matrix are standardized to obtain the standardized matrix X′=[x ij ′] NM ; x max,j 、x min,j are the maximum and minimum values ​​of the indicator features in the jth column of the multidimensional indicator matrix X; The entropy weight method, CRITIC weighting method and hierarchical analysis method are used to calculate the weight of each indicator feature respectively, and the multiple weights obtained by calculation are fused based on game theory to obtain the comprehensive weight coefficient of each indicator feature.

6. The method for dividing the development stages of the industrial park integrated energy system based on the hybrid algorithm according to claim 5 is characterized in that: The entropy weight method is used to calculate the weight of each indicator feature. The specific method is: Calculate the entropy value e of the j-th indicator feature of the i-th ordered sample sequence j , expressed as: Where, i = 1, 2, ..., N; j = 1, 2, ..., M; Assume that when p ij = 0, p ij lnp ij =0; According to the entropy value e j Calculate the weight w of the j-th indicator feature 1j , expressed as: Where, 0≤w 1j ≤1; Then we get the weight set w1 of each indicator feature under the entropy weight method, which is expressed as: w1 = [w 11 ,w 12 ,...,w 1j ,...,w 1M ] The CRITIC weighting method is used to calculate the weight of each indicator feature. The specific method is: Calculate the mean of the j-th indicator feature Expressed as: Calculate the standard deviation S of the j-th indicator feature j , expressed as: Calculate the conflict coefficient R of the j-th indicator feature j , expressed as: Where r ij is the correlation coefficient, which indicates the conflict between the characteristics of each indicator; r ij The absolute value of r ij The larger the |, the stronger the correlation with other indicator characteristics; Calculate the information content C of the j-th indicator feature j , expressed as: C j =S j ·R j Calculate the weight w of the j-th indicator feature 2j , expressed as: Then we get the weight set w2 of each indicator feature under the CRITIC weighting method, which is expressed as: w2=[w 21 ,In 22 ,...,In 2j ,...,In 2M ] The weight of each indicator feature is calculated using the hierarchical analysis method. The specific method is: The index characteristics in the standardized matrix X′ are analyzed by the hierarchical analysis method to construct the judgment matrix G=(g ij ) M×M , g ij It is used to characterize the relative importance of the index feature i and the index feature j in X′, and the Satty scaling method is used to measure the relative importance of g ij Scaling is performed, and g ij =1 / g ji , g ii =1; The normalized judgment matrix A obtained by normalizing the judgment matrix G norm , specifically including: Calculate the sum of the jth column of the judgment matrix G For each element g of the judgment matrix G ij Perform normalization: Normalized judgment matrix A norm Expressed as: Calculate the normalized judgment matrix A norm The average value of each row element w' 3i : Let i = j, then the weight w' of each indicator feature j is initially obtained 3j , that is, w' 3j =w' 3i ; The weight set w3' of each indicator feature is initially obtained, that is, w3'=[w' 31 ,w' 32 ,...,w' 3j ,...,w' 3M ]; Calculate the consistency index CI, expressed as: λ max is the maximum eigenvalue in the judgment matrix G; Calculate the consistency test coefficient CR, expressed as: RI is the random consistency index, which is a constant obtained based on the random matrix under dimension M; If CR is less than 0.1, the test passes; Otherwise, if the test fails, the judgment matrix G is readjusted, CR is recalculated, and it is judged whether CR is less than 0.1 until the test passes. When the CR test passes, the corresponding w3' is the weight set w3 of each indicator feature under the hierarchical analysis method, w3 = [w 31 ,w 32 ,...,w 3j ,...,w 3M ].

7. The method for dividing the development stages of an industrial park integrated energy system based on a hybrid algorithm according to claim 6 is characterized in that: The game theory-based method is to fuse the multiple weights obtained by calculation to obtain the comprehensive weight coefficient of each indicator feature. The specific method is as follows: Linearly combine multiple weight vectors at each level to form a preliminary weight set W, expressed as: Among them, α k is the linear combination coefficient, for each linear combination coefficient α k The objective function to be optimized is expressed as: Construct a system of linear equations, expressed as: Solve the linear equations to obtain the corresponding α k , and perform standardization to obtain the standardized linear combination coefficient α k ′, expressed as: Calculate the comprehensive weight coefficient wm of the jth indicator feature in the multidimensional indicator matrix X j , expressed as: wm j =α1′w 1j +α2′w 2j +α3′w 3j Among them, α1′, α2′, and α3′ are the linear combination coefficients of the standardized entropy weight method, CRITIC weight method, and hierarchical analysis method weights respectively; Then the comprehensive weight coefficient matrix wm corresponding to the multidimensional indicator matrix X is expressed as:

8. The method for dividing the development stages of an industrial park integrated energy system based on a hybrid algorithm according to claim 7 is characterized in that: In step S6, the standardized multidimensional indicator matrix is ​​converted into a one-dimensional feature vector to obtain the initial divided sample sequence Y, which is specifically expressed as:

9. The method for dividing the development stages of an industrial park integrated energy system based on a hybrid algorithm according to claim 1 is characterized in that: In step S7, the ordered sample sequence Y initially divided is used to determine the classification diameter, calculate the objective function, and obtain the optimal segmentation and the optimal number of segments based on the optimal segmentation method, which specifically includes the following steps: The initially divided ordered sample sequence Y includes N ordered sample sequences. Let the ordered sample segment interval from the sth to the tth in the N ordered sample sequences be represented as {x s ,x s+1 ,…,x t }, each ordered sample is a row vector, and the complete representation of the ordered sample segment interval from the sth to the tth is: {(x s1 ,x s2 ,…,x sM ),(x (s+1)1 ,x (s+1)2 ,…,x (s+1)M ),…,(x t1 ,x t2 ,…,x tM )}, the classification diameter D(s,t) is defined as: Among them, x a is an ordered sample from the sth to tth ordered samples; The specific method for calculating the objective function is: Suppose that N ordered sample sequences are divided, there may be 2 N-1 If N ordered sample sequences are randomly divided into K segments, the specific segmentation process can be expressed as: Among them, para i represents the i-th segment after the N ordered sample sequences are divided into K segments, x i +j represents the j+1th sample in the i-th segment; In order to calculate the total sum of squares of deviations of each segment, the objective function is defined as: Where D(k,k-1) represents the classification diameter of the ordered sample segment interval from k-1 to k; the smaller the objective function value G(N,K), the smaller the difference within each segment and the greater the difference between segments; The segmentation method that minimizes the objective function value is the optimal segmentation, that is: The method for obtaining the optimal number of segments is: Plot the objective function G * Curve G of (N,K) changing with the number of segments K * (N,K)~K; curve G * The absolute value formula of the slope of (N,K)~K in each segment number is expressed as: γ(K)=|G * (N,K)-G * (N,K-1)| According to the absolute value of the slope that changes with the number of segments, a γ(K)~K curve is drawn, and the number of segments K corresponding to the maximum value is the optimal number of segments.

Images