A planning method, device and configuration system of an integrated energy system

Abstract

The application discloses a kind of comprehensive energy system planning method, device and configuration system, method includes: obtaining system data to be planned;With historical planning scheme library, using different similarity indexes, respectively analyze the scheme similarity of the discrete data to be planned and the continuous data to be planned and the historical planning scheme in historical planning scheme library;Based on the preset scheme similarity threshold, the scheme similarity is judged, and the similar result is obtained;According to the scheme corresponding to the similar result, determine strategy, combine the system data to be planned to plan scheme, give target planning scheme.Through the analysis of system data to be planned, based on scheme similarity, from historical planning scheme library, the historical planning scheme with higher similarity is screened out, according to the screening result, different scheme determination strategies are used to determine target planning scheme, while ensuring the planning effect of comprehensive energy system, the planning efficiency of comprehensive energy system is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of energy system planning, and particularly relates to a planning method, device and configuration system of a comprehensive energy system. BACKGROUND

[0002] A comprehensive energy system refers to a system in which advanced physical information technology and innovative management modes are used to integrate various energies such as coal, oil, natural gas, electric energy and heat energy in a certain region, so as to realize coordinated planning, optimized operation, collaborative management, interactive response and mutual aid between various heterogeneous energy subsystems. The comprehensive energy system is a new integrated energy system that can effectively improve energy utilization efficiency and promote sustainable development of energy while meeting diversified energy demand in the system.

[0003] Current comprehensive energy system planning faces challenges such as multi-data processing, complex scenario optimization and economic trade-off. Energy demand in a regional energy system is affected by construction area, geographical location, natural resources (such as light, wind and geothermal energy) and cold and heat load, and needs to comprehensively integrate various information for efficient planning and design.

[0004] However, the traditional planning method generally constructs a model by analyzing historical planning and re-plans the current situation by using the model. On the one hand, it consumes a lot of time and increases cost, and on the other hand, it is difficult to simultaneously consider data diversity and model optimization complexity in the model construction process, resulting in low efficiency in energy configuration, cost optimization and scheme design.

[0005] Patent CN117810989A discloses a method for establishing and solving an electric grid optimization scheduling model driven by data mechanism, including: 1, establishing a comprehensive energy system physical model with uncertainty injection; 2, generating simulation cases according to the above comprehensive energy system physical model, and generating corresponding solving data as a training set for machine learning; 3, training a machine learning model for learning and solving; 4, fusing the above comprehensive energy system physical model and data model to establish a data mechanism hybrid driven electric grid optimization scheduling model, and proposing an optimization scheduling method driven by data mechanism of the electric grid system. The method improves the mapping accuracy of deep neural networks and greatly reduces the scheduling solving time in a data driven manner; realizes optimal load capacity of new energy and collaborative optimization of controllable equipment planning and operation, and improves the utilization rate of new energy.

[0006] How to improve the planning efficiency of the comprehensive energy system while ensuring the planning effect of the comprehensive energy system is a problem to be solved at present. SUMMARY

[0007] In view of the defects in the prior art, the present application provides a planning method and device for a comprehensive energy system and a configuration system, the method comprising: obtaining to-be-planned system data, wherein the to-be-planned system data comprises to-be-planned discrete data and to-be-planned continuous data; combining a historical planning scheme library, using different similarity indexes, and respectively analyzing the scheme similarity of the to-be-planned discrete data and the to-be-planned continuous data to historical planning schemes in the historical planning scheme library; based on a preset scheme similarity threshold, judging the scheme similarity to obtain a similarity result; determining a strategy according to the scheme corresponding to the similarity result, combining the to-be-planned system data to plan a scheme, and giving a target planning scheme. Through the analysis of the to-be-planned system data, based on the scheme similarity, the historical planning schemes with higher similarity are screened from the historical planning scheme library, according to the screening result, the target planning scheme is determined using different scheme determination strategies, the planning effect of the comprehensive energy system is ensured, and the planning efficiency of the comprehensive energy system is improved.

[0008] In a first aspect, the present application provides a planning method for a comprehensive energy system, specifically comprising the following steps:

[0009] Obtaining to-be-planned system data, wherein the to-be-planned system data comprises to-be-planned discrete data and to-be-planned continuous data;

[0010] Combining a historical planning scheme library, using different similarity indexes, and respectively analyzing the scheme similarity of the to-be-planned discrete data and the to-be-planned continuous data to historical planning schemes in the historical planning scheme library;

[0011] Based on a preset scheme similarity threshold, judging the scheme similarity to obtain a similarity result;

[0012] Determining a strategy according to the scheme corresponding to the similarity result, combining the to-be-planned system data to plan a scheme, and giving a target planning scheme.

[0013] Further, combining a historical planning scheme library, using different similarity indexes, and respectively analyzing the scheme similarity of the to-be-planned discrete data and the to-be-planned continuous data to historical planning schemes in the historical planning scheme library, specifically comprising:

[0014] Using a first similarity index to analyze the discrete similarity of the to-be-planned discrete data and the discrete data in the historical planning scheme;

[0015] Using a second similarity index to analyze the continuous similarity of the to-be-planned continuous data and the continuous data in the historical planning scheme;

[0016] Combining a preset discrete similarity weight and a preset continuous similarity weight, fusing the discrete similarity and the continuous similarity, and obtaining the scheme similarity.

[0017] Further, the first similarity index is a Jaccard coefficient, and the second similarity index is an Euclidean distance.

[0018] Further, the discrete similarity and the continuous similarity are fused to obtain the scheme similarity by combining the preset discrete similarity weight and the preset continuous similarity weight, and the scheme similarity specifically includes:

[0019] The discrete similarity and the continuous similarity are normalized to obtain a standard discrete similarity and a standard continuous similarity, respectively.

[0020] The standard discrete similarity and the standard continuous similarity are weighted and summed to obtain the scheme similarity by combining the discrete similarity weight and the continuous similarity weight.

[0021] Further, the scheme similarity is specifically represented as:

[0022] ;

[0023] ;

[0024] wherein, is a scheme similarity corresponding to the to-be-planned system data i and a historical planning scheme j in a historical planning scheme library, is a standard discrete similarity of to-be-planned discrete data in the to-be-planned system data i and discrete data in the historical planning scheme j, is a standard continuous similarity of to-be-planned continuous data in the to-be-planned system data i and continuous data in the historical planning scheme j, is a continuous similarity weight, is a discrete similarity weight.

[0025] Further, the standard continuous similarity is specifically represented as:

[0026] ;

[0027] wherein, is a standard continuous similarity corresponding to to-be-planned continuous data in the to-be-planned system data i and continuous data in the historical planning scheme j, is a maximum value in all , is a continuous similarity of to-be-planned continuous data in the to-be-planned system data i and continuous data in the historical planning scheme j.

[0028] Furthermore, continuous similarity is specifically represented as:

[0029] ;

[0030] in, For the continuous data to be planned in system data i Continuous data in historical planning scheme j The continuous similarity, where p is the dimension of the continuous data to be planned. For continuous data to be planned The k-th element in For continuous data in historical planning scheme j The k-th element in the data, where n is the continuous data to be planned in system data i. The number of groups, m is the number of historical planning schemes in the historical planning schemes.

[0031] Furthermore, discrete similarity is specifically expressed as:

[0032] ;

[0033] in, For the discrete data to be planned in system data i Discrete data in historical planning scheme j The discrete similarity.

[0034] Furthermore, based on the solutions corresponding to similar results, strategies are determined, and solutions are planned in conjunction with the data of the system to be planned, resulting in a target planning solution, specifically including:

[0035] If at least one historical planning scheme in the historical planning scheme database has a similarity result, the historical planning schemes in the historical planning scheme database are sorted based on the scheme similarity, and pending historical schemes are given according to the sorting results.

[0036] A target planning scheme is obtained by using a pre-acquired target decision tree model and combining it with the data of the system to be planned. The pre-acquired target decision tree model is obtained by analyzing and correcting the historical schemes to be planned.

[0037] If all similar results of historical planning schemes in the historical planning scheme database are dissimilar, a target planning scheme is given by combining the data of the system to be planned and the pre-built scheme generation model.

[0038] Furthermore, the pre-acquired target decision tree model is determined through the following steps:

[0039] Based on the similar results of various historical planning schemes, the historical planning schemes in the historical planning scheme library are screened to obtain the model training dataset.

[0040] initialize the decision tree model, set model parameters in the decision tree model;

[0041] train the decision tree model using the model training data set until convergence;

[0042] evaluate the trained decision tree model using the model evaluation indicators to obtain the model evaluation results;

[0043] based on the model evaluation results, correct the training of the decision tree model to obtain the target decision tree model.

[0044] Further, the model evaluation indicators are at least one of mean square error, root mean square error, mean absolute error, and R² score.

[0045] Further, the trained decision tree model is evaluated using the model evaluation indicators to obtain the model evaluation results, specifically including:

[0046] using multiple model evaluation indicators, the trained decision tree model is evaluated to obtain multiple evaluation results;

[0047] based on the index weights corresponding to different model evaluation indicators, the multiple evaluation results are fused to obtain a model comprehensive index;

[0048] The model comprehensive index is compared with the preset model training threshold to obtain the model evaluation results.

[0049] Further, the model comprehensive index is specifically represented as:

[0050] ;

[0051] wherein, the model comprehensive index, the weight of the mean absolute error, the weight of the mean square error, MAE is the mean absolute error of the decision tree model, and MSE is the mean square error of the decision tree model.

[0052] Further, the model training data set includes demand training data and scheme training data.

[0053] Based on the model evaluation results, the decision tree model is corrected and trained to obtain the target decision tree model, specifically including:

[0054] If the model evaluation results are qualified, the demand training data and the scheme training data are integrated to obtain integrated features.

[0055] Based on the integrated features, the trained decision tree model is domain adapted to obtain a mapping function containing the mapping relationship between the demand training data and the scheme training data.

[0056] The trained decision tree model is trained in combination with the mapping function to obtain a target decision tree model.

[0057] Further, the decision tree model is an adaptTree model.

[0058] Further, the pre-constructed scheme generation model is determined through the following steps:

[0059] A bi-level programming model is constructed, and an upper target function and a lower target function are set.

[0060] The number of iterations is determined, and the positions and speeds of the particles are initialized.

[0061] The bi-level programming model is iteratively solved by combining the upper target function and the lower target function and using a particle swarm algorithm.

[0062] If the number of iterations reaches the threshold, the bi-level programming model is solved, and a scheme generation model is obtained.

[0063] Further, the upper target function is used to limit the energy loss of each pipe network in the comprehensive energy system, and the lower target function is used to limit the total economic cost of the comprehensive energy system in the life cycle, wherein the total economic cost includes the installation cost of each device and pipe network in the comprehensive energy system, the operation and maintenance cost of the comprehensive energy system, and the new energy subsidy income in the comprehensive energy system.

[0064] Further, the upper target function is specifically represented as:

[0065] ;

[0066] The lower target function is specifically represented as:

[0067] ;

[0068] wherein, is the upper target function, i is the pipe network number of the comprehensive energy system, is the resistance loss of the pipe network i, is the inductance and capacitance loss of the pipe network i, is the fixed loss of the pipe network i, is the variable loss of the pipe network i, is the other loss of the pipe network i, is the lower target function, is the installation cost of the comprehensive energy system, is the operation and maintenance cost of the comprehensive energy system, is the new energy subsidy income of the comprehensive energy system.

[0069] Further, a target decision tree model is adopted to obtain a target planning scheme in combination with the to-be-planned system data, and the target planning scheme specifically includes:

[0070] The mapping function obtained in the target decision tree model is used to map the to-be-planned system data to obtain mapping features;

[0071] The mapping features are input into the decision tree model to obtain the target planning scheme.

[0072] In a second aspect, the present application further provides a planning device for a comprehensive energy system, which adopts the planning method for the comprehensive energy system according to any one of the above aspects, and includes:

[0073] A data acquisition module is configured to acquire to-be-planned system data, wherein the to-be-planned system data includes to-be-planned discrete data and to-be-planned continuous data;

[0074] A similarity analysis module is configured to analyze the scheme similarity between the to-be-planned discrete data and the to-be-planned continuous data and historical planning schemes in a historical planning scheme library by using different similarity indexes in combination with the historical planning scheme library;

[0075] A similarity judgment module is configured to judge the scheme similarity based on a preset scheme similarity threshold to obtain a similarity result;

[0076] A scheme determination module is configured to determine a scheme according to the similarity result, perform scheme planning in combination with the to-be-planned system data, and give a target planning scheme.

[0077] In a third aspect, the present application further provides a configuration system for a comprehensive energy system, which includes a memory, a processor, and a computer program stored in the memory, and the computer program is executed by the processor to execute the instructions according to the above method.

[0078] The planning method, device and configuration system for a comprehensive energy system provided by the present application at least have the following beneficial effects:

[0079] (1) By analyzing the to-be-planned system data, the historical planning schemes with higher similarity are screened from the historical planning scheme library based on the scheme similarity, and the target planning scheme is determined by using different scheme determination strategies according to the screening result, so that the planning effect of the comprehensive energy system is ensured and the planning efficiency of the comprehensive energy system is improved.

[0080] (2) By dividing the to-be-planned system data into to-be-planned discrete data and to-be-planned continuous data, the similarity analysis is performed by using different similarity indexes, so that the accuracy of the scheme similarity between the to-be-planned system data and the historical planning scheme library is improved, and a data basis is provided for better determination of the target planning scheme.

[0081] (3) The TCA technology is used for domain adaptation of the decision tree model, the decision tree model is mapped to a new subspace through a mapping function for further training, so that the obtained target decision tree model can more accurately analyze the data of the system to be planned, guarantee the accuracy of the obtained target planning scheme, and guarantee the planning effect of the comprehensive energy system to a certain extent. BRIEF DESCRIPTION OF DRAWINGS

[0082] Figure 1 A flowchart of the planning method of the comprehensive energy system provided by the embodiment of the present application is shown in the figure.

[0083] Figure 2 A flowchart of the method for determining the similarity of the schemes provided by the embodiment of the present application is shown in the figure.

[0084] Figure 3 A flowchart of the method for fusing the similarity provided by the embodiment of the present application is shown in the figure.

[0085] Figure 4 A flowchart of the method for selecting the scheme determination strategy provided by the embodiment of the present application is shown in the figure.

[0086] Figure 5 A flowchart of the method for determining the target decision tree model provided by the embodiment of the present application is shown in the figure.

[0087] Figure 6 A flowchart of the method for training the decision tree model provided by the embodiment of the present application is shown in the figure.

[0088] Figure 7 A flowchart of the method for determining the scheme generation model provided by the embodiment of the present application is shown in the figure.

[0089] Figure 8 A structural block diagram of the planning device of the comprehensive energy system provided by the embodiment of the present application is shown in the figure.

[0090] Among them, 201, data acquisition module; 202, similarity analysis module; 203, similarity judgment module; 204, scheme determination module. DETAILED DESCRIPTION

[0091] In order to better understand the above technical solutions, the above technical solutions will be described in detail below in combination with the drawings and specific embodiments in the specification. Obviously, the described embodiments are only part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0092] The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in the description of the application and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0093] It is also to be noted that the terms "comprising", "including", and "having" or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises, includes, or has a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The materials, methods, and examples provided herein are illustrative only and not intended to be limiting.

[0094] The development of current integrated energy systems faces the problems of diversified demand and multi-objective optimization, and needs to be customized according to the regional characteristics and energy characteristics. Due to the differences in geographical location, resource distribution and energy load, the traditional single optimization method is difficult to meet the actual engineering demand. Therefore, intelligent and multi-objective planning has become an important direction of integrated energy system research.

[0095] With the development of intelligent technology, improving planning efficiency through multi-dimensional data analysis, intelligent optimization algorithm and transfer learning has become a research direction to solve the complex problems of integrated energy systems.

[0096] The current integrated energy system generally analyzes the historical planning data or uses the historical planning data to train the model, and then generates a new planning scheme. This way has limited effect on integrated energy system planning, and the planning efficiency is low.

[0097] In the related art, a data mechanism hybrid driving power grid optimization scheduling model establishment and solving method, comprising: 1, establishing a comprehensive energy system physical model with uncertainty injection; 2, generating simulation cases according to the above comprehensive energy system physical model, and generating corresponding solving data as a machine learning training set; 3, training a machine learning model for learning and solving; 4, fusing the above comprehensive energy system physical model and data model to establish a data mechanism hybrid driving power grid optimization scheduling model, and proposing a data mechanism hybrid driving optimization scheduling method of the power grid system. The mapping accuracy of the deep neural network is improved, and the data driven method greatly reduces the scheduling solving time; the optimal carrying capacity of new energy and the planning and operation collaborative optimization of controllable equipment are realized, and the utilization rate of new energy is improved. In the related art, the scheduling of the comprehensive energy system is realized by fusing the physical model and the data model, the utilization rate of new energy is realized by optimizing the carrying capacity of new energy and the controllable equipment in the comprehensive energy system, and the utilization rate of the comprehensive energy system is improved. However, the adjustment of traditional energy other than new energy in the comprehensive energy system is not considered, the planning of the comprehensive energy system is limited, and the improvement of the utilization rate of the comprehensive energy system is also limited.

[0098] The application provides a comprehensive energy system planning method, device and configuration system, the method comprising: obtaining to-be-planned system data, wherein the to-be-planned system data comprises to-be-planned discrete data and to-be-planned continuous data; combining a historical planning scheme library, using different similarity indexes to respectively analyze the scheme similarity of the to-be-planned discrete data and the to-be-planned continuous data and historical planning schemes in the historical planning scheme library; judging the scheme similarity based on a preset scheme similarity threshold to obtain a similarity result; determining a strategy according to a scheme corresponding to the similarity result, combining the to-be-planned system data to plan a scheme, and giving a target planning scheme. Through analysis of the to-be-planned system data, based on the scheme similarity, a historical planning scheme with high similarity is selected from the historical planning scheme library, according to the selection result, a target planning scheme is determined using different scheme determination strategies, while ensuring the planning effect of the comprehensive energy system, the planning efficiency of the comprehensive energy system is improved.

[0099] The planning method of the comprehensive energy system fuses multi-type data similarity calculation, a double-layer planning model and a transfer learning method, the similarity of continuous and discrete data is calculated by fusing the Euclidean distance and the Jaccard similarity, and the accuracy of data analysis is improved. In addition, a double-layer planning model is constructed, the upper layer optimizes the energy station and the pipe network layout with the target of reducing the energy loss of the pipe network, and the lower layer focuses on minimizing the life cycle cost of the system. Further, the transfer learning technology is used, the historical data rule is migrated to a new scene through feature distribution alignment and machine learning model training, and the adaptability and intelligent degree of the planning scheme are further improved.

[0100] AsFigure 1 As shown, the embodiment of the present application provides a planning method of a comprehensive energy system, and the specific steps are as follows:

[0101] S101: Obtain system data to be planned.

[0102] Specifically, the system data to be planned is data needed in the planning process of the comprehensive energy system, for example, construction area, geographic location, land use planning, photovoltaic resource, wind power resource, geothermal resource, current energy supply and consumption data, environmental data (for example, sunshine intensity, wind speed, humidity and temperature), energy market situation (for example, energy price), and cold, heat, and electricity load statistics and the like. The cold, heat, and electricity load statistics can be the duration curve of the load of electricity, heat, cold, and gas. It can also include the basic construction situation of the power distribution network and gas network, the load demand of typical days in different seasons, and the specific parameters of energy storage devices, photovoltaic power stations, and transformer substations and the like, and the corresponding data is used according to different emphases. And according to the value mode of the system data to be planned, all the data can be divided into continuous data (i.e., continuous data to be planned) and discrete data (i.e., discrete data to be planned).

[0103] S102: In combination with the historical planning scheme library, different similarity indexes are used to analyze the scheme similarity of the discrete data to be planned and the continuous data to be planned and the historical planning schemes in the historical planning scheme library.

[0104] Referring to Figure 2 , specifically including:

[0105] A first similarity index is used to analyze the discrete similarity of the discrete data to be planned and the discrete data in the historical planning scheme;

[0106] A second similarity index is used to analyze the continuous similarity of the continuous data to be planned and the continuous data in the historical planning scheme;

[0107] The discrete similarity and the continuous similarity are fused by combining the preset discrete similarity weight and the preset continuous similarity weight to obtain the scheme similarity.

[0108] It can be understood that the historical planning scheme library includes a plurality of previous planning schemes of the comprehensive energy system, i.e., historical planning schemes, and the historical planning schemes include demand data and scheme data. The demand data is data needed to obtain the scheme data, which corresponds to the above-mentioned system data to be planned, and the scheme data is related data of the planning scheme of the comprehensive energy system based on the demand data. Therefore, the discrete data in the historical planning scheme refers to the discrete data in the demand data, and by analogy, the continuous data in the historical planning scheme refers to the continuous data in the demand data.

[0109] In the embodiments provided by this invention, the first similarity index is the Jaccard coefficient, and the second similarity index is the Euclidean distance. Other similarity indices can be used in other embodiments, such as Manhattan distance, Hamming distance, Mahalanobis distance, and Pearson correlation coefficient, etc., without limitation. The Jaccard coefficient measures the ratio of the intersection to the union of two sets. A larger Jaccard similarity coefficient indicates greater similarity between the two sets. Euclidean distance is a measure of the straight-line distance between two vectors. A smaller Euclidean distance indicates greater similarity between the two vectors. Manhattan distance is the sum of the absolute distances of two vectors on each coordinate axis. A smaller Manhattan distance indicates greater similarity between the two vectors. Hamming distance measures the number of different characters between two strings of equal length. A smaller Hamming distance indicates greater similarity between the two strings. Mahalanobis distance considers the covariance matrix of the data and can better reflect the actual distance between data points. The Pearson correlation coefficient measures the linear relationship between two variables, with a value range of [-1, 1], where 1 indicates a perfect positive correlation and -1 indicates a perfect negative correlation.

[0110] In a specific example, for continuous data to be planned, Euclidean distance is used to measure continuous similarity. The smaller the calculated Euclidean distance, the higher the continuous similarity; conversely, the larger the calculated Euclidean distance, the lower the continuous similarity.

[0111] Assuming each set of continuous data is a vector, the continuous data to be planned in system data i is represented by... This indicates that continuous data in historical planning scheme j is used... This indicates that the continuous data to be planned in system data i is calculated. Continuous data in historical planning scheme j The Euclidean distance is specifically expressed as:

[0112] ;

[0113] in, For the continuous data to be planned in system data i Continuous data in historical planning scheme j The Euclidean distance, where p is the dimension of the continuous data to be planned. For continuous data to be planned The k-th element in For continuous data in historical planning scheme j The k-th element in the data, where n is the continuous data to be planned in system data i. The number of groups, m is the number of historical planning schemes in the historical planning schemes.

[0114] For discrete data to be planned, the Jaccard coefficient is used to measure discrete similarity. The smaller the calculated Jaccard coefficient, the lower the discrete similarity; conversely, the larger the calculated Jaccard coefficient, the higher the continuous similarity.

[0115] The discrete data to be planned in system data i is represented by a set. This indicates that the discrete data in historical planning scheme j is represented by a set. This indicates that the discrete data to be planned in system data i is calculated. Discrete data in historical planning scheme j The Jaccard coefficient is specifically expressed as:

[0116] ;

[0117] in, For the discrete data to be planned in system data i Discrete data in historical planning scheme j The Jaccard coefficient.

[0118] Furthermore, referring to Figure 3 By combining preset discrete similarity weights and preset continuous similarity weights, the discrete similarity and continuous similarity are fused to obtain the scheme similarity, which specifically includes:

[0119] The discrete similarity and continuous similarity are normalized respectively to obtain the standard discrete similarity and standard continuous similarity;

[0120] By combining discrete similarity weights and continuous similarity weights, the standard discrete similarity and standard continuous similarity are weighted and summed to obtain the scheme similarity.

[0121] In one specific implementation, the continuous similarity, i.e., the Euclidean distance, is normalized to the range [0,1].

[0122] ;

[0123] in, For the continuous data to be planned in system data i Continuous data in historical planning scheme j The corresponding standard continuous similarity, For all The maximum value in.

[0124] Since the Jaccard coefficients range from [0,1], there is no need to normalize them.

[0125] Scheme similarity is specifically expressed as:

[0126] ;

[0127] ;

[0128] wherein, is a scheme similarity corresponding to the system data i to be planned and the historical planning scheme j in the historical planning scheme library, is a continuous similarity weight, is a discrete similarity weight.

[0129] It can be understood that the continuous similarity weight is the influence of continuous data on the scheme similarity, and the discrete similarity weight is the influence of discrete data on the scheme similarity, and the values of the continuous similarity weight and the discrete similarity weight are set according to actual conditions, and are not limited.

[0130] S103: judging the scheme similarity based on a preset scheme similarity threshold to obtain a similarity result.

[0131] Specifically, each historical planning scheme in the historical planning scheme library corresponds to a scheme similarity threshold, and each scheme similarity is compared with the scheme similarity threshold. If the scheme similarity is greater than or equal to the scheme similarity threshold, the similarity result is similar, indicating that the system data to be planned has a high similarity with the demand data in the corresponding historical planning scheme, and the corresponding historical planning scheme has a certain reusability, and the historical planning scheme can be modified to obtain a target planning scheme corresponding to the system data to be planned. If the scheme similarity is less than the scheme similarity threshold, the similarity result is dissimilar, indicating that the system data to be planned has a low similarity with the demand data in the corresponding historical planning scheme, and the reusability of the corresponding historical planning scheme is poor, and a target planning scheme corresponding to the system data to be planned needs to be planned again.

[0132] S104: determining a strategy according to the scheme corresponding to the similarity result, combining the system data to be planned to plan a scheme, and giving a target planning scheme.

[0133] Referring to Figure 4 , a target planning scheme is given, specifically including:

[0134] If the similarity result corresponding to at least one historical planning scheme in the historical planning scheme library is similar, the historical planning schemes in the historical planning scheme library are sorted based on the scheme similarity, and a to-be-determined historical scheme is given according to the sorting result;

[0135] A target planning scheme is obtained by combining the system data to be planned by using a target decision tree model obtained in advance, wherein the target decision tree model obtained in advance is obtained by analyzing and modifying the to-be-determined historical scheme;

[0136] If the similar results of the historical planning schemes in the historical planning scheme library are all dissimilar, the target planning scheme is given in combination with the data of the system to be planned and the pre-constructed scheme generation model.

[0137] In a specific embodiment, when there is one or more historical planning schemes in the historical planning scheme library corresponding to similar results being similar, the historical planning schemes in the historical planning scheme library are sorted based on the scheme similarity, and the schemes with a scheme similarity greater than or equal to a scheme similarity threshold are screened out to obtain a pending historical scheme. The target decision tree model is trained based on the pending historical scheme, and the pending historical scheme is corrected using the trained target decision tree model to obtain the target planning scheme. If the similar results of the historical planning schemes in the historical planning scheme library are all dissimilar, it indicates that the reusability of the schemes in the historical planning scheme library is poor, and the target planning scheme needs to be given in combination with the data of the system to be planned and the pre-constructed scheme generation model.

[0138] Further, with reference to Figure 5 , the pre-acquired target decision tree model is determined by the following steps:

[0139] The historical planning schemes in the historical planning scheme library are screened based on the similar results of the historical planning schemes to obtain a model training data set;

[0140] The decision tree model is initialized, and the model parameters in the decision tree model are set;

[0141] The decision tree model is trained using the model training data set until convergence;

[0142] The trained decision tree model is evaluated using a model evaluation index to obtain a model evaluation result;

[0143] The decision tree model is corrected and trained based on the model evaluation result to obtain a target decision tree model.

[0144] In the examples provided by the present application, the above-mentioned decision tree model is an AdaptTree model.

[0145] In a specific embodiment, the historical planning schemes with a scheme similarity reaching a scheme similarity threshold are screened out as the model training data set. At the beginning of the model training, the model parameters of the decision tree model are set to complete the initialization of the decision tree model. The model parameters include the maximum depth and the minimum sample split number, which are used to control the complexity of the decision tree to prevent overfitting. The model parameters are set according to actual requirements, and are not limited.

[0146] The demand data of each historical planning scheme in the model training data set, i.e., the demand training data and scheme data, i.e., scheme training data As a source domain, system data to be planned As a target domain. It can be understood that the demand data of the source domain is highly similar in distribution or feature form to the system data to be planned in the target domain. The distribution is relatively close, that is, it can be aligned through a certain mapping, and the rule of "demand data→scheme data" in the source domain can be migrated on the target domain.

[0147] The decision tree model is trained on the source domain to convergence. Since the mapping relationship between the demand data and the scheme data cannot be explicitly obtained, a data-driven method is used for learning in the source domain, and a decision tree model is used for training in consideration of compatibility with mixed features (continuous data and discrete data).

[0148] Further, a model evaluation index is used to evaluate the trained decision tree model to obtain a model evaluation result, specifically including:

[0149] A plurality of model evaluation indexes are used to evaluate the trained decision tree model respectively to obtain a plurality of evaluation results;

[0150] Based on the index weights corresponding to different model evaluation indexes, the plurality of evaluation results are fused to obtain a model comprehensive index;

[0151] The model comprehensive index is compared with a preset model training threshold to obtain the model evaluation result.

[0152] In the examples provided in the present application, mean squared error (MSE) and mean absolute error (MAE) are used as model evaluation indexes to evaluate the performance of the model, and a model comprehensive index is constructed. The model comprehensive index can simultaneously focus on the large errors existing in the model training process and the overall error between the predicted value and the true value.

[0153] In other embodiments, the model evaluation index can be one or more of mean squared error, root mean squared error (RMSE), mean absolute error, and R-squared (R² Score). The mean squared error is the average of the squares of the differences between the predicted value and the actual value, which is sensitive to large errors. The root mean squared error is the square root of the mean squared error, and the unit is consistent with the original data, which is convenient for interpretation. The mean absolute error is the average of the absolute values of the differences between the predicted value and the actual value, which is not sensitive to outliers. The R-squared score is used to measure the proportion of the variance of the data explained by the model, and the closer the value is to 1, the better the model fitting effect.

[0154] In one specific example, the four model evaluation indicators are fused to obtain a model comprehensive indicator, and for each model evaluation indicator, a corresponding indicator weight is set , and In different examples, normalization processing can be performed on different model evaluation indicators to scale different model evaluation indicators to the same scale and reduce the dimensional difference between model evaluation indicators.

[0155] The model comprehensive indicator is specifically represented as:

[0156] ;

[0157] wherein, is the model comprehensive indicator, is the weight of the mean absolute error, is the weight of the mean square error, MAE is the mean absolute error of the decision tree model, and MSE is the mean square error of the decision tree model.

[0158] The above , is the indicator weight, and the mean square error of the decision tree model and the mean absolute error of the decision tree model are evaluation results.

[0159] The model comprehensive indicator is compared with a preset model training threshold value, and when the model comprehensive indicator is less than the model training threshold value, it indicates that the error of the decision tree model is small, i.e., the model evaluation result is qualified, and the decision tree model training meets the standard, and the next step can be directly performed. When the model comprehensive indicator is greater than or equal to the model training threshold value, it indicates that the error of the decision tree model is large, i.e., the model evaluation result is unqualified, and the decision tree model training does not meet the standard, and needs to be further processed to retrain the decision tree model.

[0160] When the model evaluation result is unqualified, the scheme similarity threshold value is adjusted so that more historical planning schemes in the historical planning scheme library enter the model training data set to train the decision tree model. This process is repeated until the model evaluation result is qualified.

[0161] In other embodiments, the historical planning schemes with a scheme similarity reaching the scheme similarity threshold value that are screened out are divided into a training set, a validation set, and a test set according to a certain proportion. For example, they are divided into a training set, a validation set, and a test set according to a ratio of 6:2:2. When the number of historical planning schemes screened out is less than the scheme quantity threshold value, the scheme similarity threshold value is adjusted to reduce the screening requirement.

[0162] Further, referring to Figure 6 , the model training data set includes demand training data and scheme training data;

[0163] Based on the model evaluation result, the decision tree model is corrected and trained to obtain a target decision tree model, specifically including:

[0164] If the model evaluation result is qualified, the demand training data and the scheme training data are integrated to obtain integrated features;

[0165] Based on the integrated features, the trained decision tree model is domain adapted to obtain a mapping function containing the mapping relationship between the demand training data and the scheme training data;

[0166] The trained decision tree model is trained in combination with the mapping function to obtain a target decision tree model.

[0167] In the embodiments provided in the application, transfer component analysis (TCA) is used for domain adaptation. Through unsupervised domain adaptation, the target domain and the source domain are distributed more closely in the same feature subspace, and then the tree model is trained on the subspace to predict the target domain more accurately.

[0168] In a specific implementation, the demand training data and the scheme training data are integrated to obtain integrated features , wherein , , is the number of source domain samples, is the number of target domain samples, and d is the feature dimension.

[0169] A mapping function is found to project to a new subspace , u is the spatial dimension of the new subspace, and the source domain mapping is obtained; the target domain mapping is: In the new subspace, maximum mean discrepancy (MMD) is used as a standard for measuring the minimum distribution difference between the source domain and the target domain, specifically represented as:

[0170] ;

[0171] ;

[0172] wherein is a function for measuring the distribution difference between the source domain and the target domain using MMD, is a mapping function for mapping data to a reproducing kernel Hilbert space (RKHS), To regenerate the norm in the Hilbert space.

[0173] Determine whether it is within the preset standard range, if it is within the preset standard range, the mapping function is obtained. The input , can be mapped to a new subspace through the mapping function . Conversely, if is not within the preset standard range, relearning is needed, that is, to find the mapping function again until the corresponding to the mapping function falls within the preset standard range.

[0174] After obtaining the mapping function, train the decision tree model in the new subspace corresponding to the mapping function. In a specific example, based on the mapping function , the mapped source domain features are obtained: , and the decision tree model is trained on to obtain the target decision tree model . In other embodiments, the performance of the target decision tree model can be evaluated by using the corresponding validation set.

[0175] Further, using the pre-acquired target decision tree model, combined with the to-be-planned system data, the target planning scheme is obtained, specifically including:

[0176] Based on the mapping function obtained in the target decision tree model, the to-be-planned system data is mapped to obtain the mapping feature;

[0177] The mapping feature is input into the decision tree model to obtain the target planning scheme.

[0178] Specifically, based on the mapping function, the mapped target domain feature , that is, the mapping feature, is obtained, wherein . The mapped target domain feature is input into the target decision tree model to obtain the target planning scheme .

[0179] Further, referring to Figure 7 , the pre-constructed scheme generation model is determined by the following steps:

[0180] A double-layer planning model is constructed, and an upper layer objective function and a lower layer objective function are set;

[0181] The iteration threshold is determined and the position and speed of the particle are initialized;

[0182] The double-layer planning model is iteratively solved by combining the upper layer objective function and the lower layer objective function using the particle swarm algorithm.

[0183] If the number of iterations reaches the iteration threshold, the solution of the bi-level programming model is completed, and a planning scheme is generated.

[0184] The upper objective function is used to limit the energy loss of each pipe network in the comprehensive energy system, and the lower objective function is used to limit the total economic cost of the comprehensive energy system in the life cycle, wherein the total economic cost includes the installation cost of each device and pipe network in the comprehensive energy system, the operation and maintenance cost of the comprehensive energy system, and the new energy subsidy income in the comprehensive energy system.

[0185] Further, the upper objective function is specifically represented as:

[0186] ;

[0187] The lower objective function is specifically represented as:

[0188] ;

[0189] Wherein, is the upper objective function, i is the pipe network number of the comprehensive energy system, is the resistance loss of the pipe network i, is the inductance and capacitance loss of the pipe network i, is the fixed loss of the pipe network i, is the variable loss of the pipe network i, is the other loss of the pipe network i, is the lower objective function, is the installation cost of the comprehensive energy system, is the operation and maintenance cost of the comprehensive energy system, is the new energy subsidy income of the comprehensive energy system.

[0190] In a specific example, a bi-level programming model is established in combination with actual engineering requirements. The upper level considers the minimum energy loss of each pipe network to determine the optimal location of the energy station and the pipe network. The lower level takes the minimum total economic cost of the comprehensive energy system in the life cycle as the target, involves the initial investment and installation of devices and pipe networks, the operation and maintenance cost, and the subsidy income of new energy generation, thereby determining the optimal installation capacity of each device of the regional energy station, and obtaining a target planning scheme.

[0191] In this example, the particle swarm algorithm is adopted, the fitness values of particles and populations are constructed into an evolution function, the solution optimization range of the upper objective function and the lower objective function is moved in the direction of global optimization. The evolution function is integrated into the inertia weight and the learning factor, which are two parameters, to dynamically adjust the values of the inertia weight and the learning factor in each iteration process. It is judged whether the iteration threshold is reached. If the requirement is reached, the target planning scheme is output; otherwise, the iteration is continued.

[0192] Specifically, the particle number and the iteration number threshold are determined, wherein the particle number is set according to actual engineering requirements, and the initial value of each particle is a randomly generated position and speed in the search space, and the historical optimal position and the global optimal position of each particle are recorded. The iteration number threshold is set according to actual requirements or experience.

[0193] The inertia weight is set, which plays a role in balancing the global exploration and local development capabilities in the particle swarm algorithm. The inertia weight determines the influence of the speed at the previous moment on the next movement, and by adjusting the inertia weight, the convergence speed and global search ability of the algorithm can be controlled. A larger inertia weight is conducive to global exploration and increases the diversity of the population, while a smaller inertia weight can improve the local mining ability of the algorithm and speed up the convergence speed.

[0194] The learning factor and the social learning factor are set, the learning factor represents the degree of movement of the particle to its historical optimal position, and the social learning factor represents the degree of movement of the particle to the global optimal position. The learning factor is used to adjust the weight of the particle updating the speed according to individual experience, and determines the degree of attention of the particle to the individual optimal solution in the search process. A larger learning factor can increase the global search ability of the particle, but may cause the particle to oscillate in the search space; a smaller learning factor can increase the local search ability of the particle, but may lead to falling into a local optimal solution. The social learning factor is used to adjust the weight of the particle updating the speed according to group experience, and determines the degree of attention of the particle to the group optimal solution in the search process. A larger social learning factor can increase the global search ability of the particle and help the particle better utilize group information, but may also cause the particle to oscillate in the search space; a smaller social learning factor can increase the local search ability of the particle, but may lead to falling into a local optimal solution. According to different optimization scenarios, the values of the learning factor and the social learning factor can be manually adjusted at different stages of the particle swarm algorithm to balance the global search and local search capabilities; the parameters can also be adjusted according to the convergence of the particle swarm algorithm, for example, a larger learning factor and a larger social learning factor are used in the early stage of the particle swarm algorithm to speed up global search, and a smaller learning factor and a smaller social learning factor are used in the later stage of the particle swarm algorithm for fine search.

[0195] The application provides a planning method of a comprehensive energy system, which filters historical planning schemes in a historical planning scheme library by fusing two different similarity indexes to obtain historical planning schemes with higher similarity, combines a decision tree model and TCA technology to obtain a target decision tree model, and finally inputs the mapped features of the target domain into the target decision tree model to obtain a target planning scheme. In this way, when a historical planning scheme with similar engineering conditions is encountered, the target planning scheme can be obtained by modifying the corresponding historical planning scheme without re-planning. If no historical planning scheme with similar engineering conditions is encountered, the target planning scheme can be solved by using a scheme generation model obtained by solving a double-layer planning model. The planning method of the comprehensive energy system avoids information loss or precision reduction caused by a single processing method, thereby ensuring the comprehensiveness of data processing, enhancing the adaptability of the planning method to actual demands, coping with complex comprehensive energy system design problems, and maintaining efficient solving performance in different scenarios.

[0196] With reference to Figure 8 The application embodiment provides a planning device of a comprehensive energy system, which comprises:

[0197] A data acquisition module 201 is configured to acquire to-be-planned system data, wherein the to-be-planned system data comprises to-be-planned discrete data and to-be-planned continuous data.

[0198] A similarity analysis module 202 is configured to combine a historical planning scheme library and use different similarity indexes to respectively analyze the scheme similarity between the to-be-planned discrete data and the to-be-planned continuous data and historical planning schemes in the historical planning scheme library.

[0199] A similarity judgment module 203 is configured to judge the scheme similarity based on a preset scheme similarity threshold to obtain a similarity result.

[0200] A scheme determination module 204 is configured to determine a scheme according to a scheme determination strategy corresponding to the similarity result, perform scheme planning in combination with the to-be-planned system data, and give a target planning scheme.

[0201] In addition, the application also provides a configuration system for a comprehensive energy system, which comprises a memory, a processor, and a computer program stored in the memory, and the computer program is executed by the processor to execute instructions according to the above method.

[0202] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described modules can refer to the corresponding process in the foregoing method embodiments, which will not be described herein.

[0203] In particular, according to the embodiments of the present application, the above reference flowchart Figure 1The described processes can be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program carried on a machine-readable medium, the computer program comprising program code for executing the methods illustrated in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication section, and / or installed from a removable medium. When the computer program is executed by a central processing unit, the above-described functions defined in the apparatus of the present application are executed.

[0204] It should be noted that the computer-readable medium described above in the present disclosure can be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. The computer-readable storage medium, for example, can be, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or any suitable combination thereof. More specific examples of the computer-readable storage medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program used or used in conjunction with an instruction execution system, apparatus or device. In the present disclosure, the computer-readable signal medium can include a data signal carried in a baseband or as a part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, which can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained in the computer-readable medium can be transmitted by any suitable medium, including but not limited to a wire, a cable, a RF (radio frequency) or the like, or any suitable combination thereof.

[0205] The above computer-readable medium can be included in the above electronic device; or can exist separately without being assembled into the electronic device.

[0206] Computer program code for carrying out operations of the present disclosure can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0207] The computer program instructions can also be loaded onto a computer or other programmable information processing apparatus to cause a series of operations to be performed on the computer or other programmable information processing apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable information processing apparatus implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0208] The units described in the embodiments of the present disclosure can be implemented by software, or by hardware, or by a combination of software and hardware. In some cases, the names of the units do not constitute a limitation on the units themselves.

[0209] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those of skill in the art once they have the benefit of the present disclosure. Therefore, the appended claims are intended to encompass within their scope all possible variations and modifications of the preferred embodiments. It will be apparent to those of ordinary skill in the art that various changes and modifications can be made to the application without departing from the spirit or scope thereof. It is intended, therefore, in the appended claims to cover all such changes and modifications that fall within the true scope of the application. Accordingly, while the preferred embodiments of the application have been described above, it will be appreciated that those skilled in the art will be able to make modifications and alterations to this description.

Claims

1. A method for planning an integrated energy system, characterized in that, The method comprises the following steps: acquiring to-be-planned system data, wherein the to-be-planned system data comprises to-be-planned discrete data and to-be-planned continuous data; combining a historical planning scheme library, using different similarity indexes, and respectively analyzing the scheme similarity between the to-be-planned discrete data and the historical planning scheme in the historical planning scheme library and the to-be-planned continuous data and the historical planning scheme in the historical planning scheme library; based on a preset scheme similarity threshold, judging the scheme similarity to obtain a similarity result; if the similarity result corresponding to at least one historical planning scheme in the historical planning scheme library is similar, sorting the historical planning schemes in the historical planning scheme library based on the scheme similarity, and giving a to-be-determined historical scheme according to the sorting result; based on a mapping function obtained from a target decision tree model, mapping the to-be-planned system data to obtain mapping features; inputting the mapping features into the decision tree model to obtain a target planning scheme, wherein the target decision tree model is obtained by analyzing and correcting the to-be-determined historical scheme; the target decision tree model is determined by the following steps: filtering the historical planning schemes in the historical planning scheme library based on the similarity results of the historical planning schemes to obtain a model training data set, wherein the model training data set comprises demand training data and scheme training data; initializing the decision tree model and setting the model parameters in the decision tree model; training the decision tree model using the model training data set until convergence; evaluating the trained decision tree model using a model evaluation index to obtain a model evaluation result; if the model evaluation result is qualified, integrating the demand training data and the scheme training data to obtain integrated features; based on the integrated features, performing domain adaptation on the trained decision tree model to obtain a mapping function containing the mapping relationship between the demand training data and the scheme training data; and training the trained decision tree model based on the mapping function to obtain the target decision tree model; if the similarity results of the historical planning schemes in the historical planning scheme library are all dissimilar, combining the to-be-planned system data and a pre-constructed scheme generation model to give a target planning scheme.

2. The method for planning an integrated energy system of claim 1, wherein, Combining the historical planning scheme library, using different similarity indexes, and respectively analyzing the scheme similarity between the to-be-planned discrete data and the historical planning scheme in the historical planning scheme library and the to-be-planned continuous data and the historical planning scheme in the historical planning scheme library, specifically comprising: using a first similarity index to analyze the discrete similarity between the to-be-planned discrete data and the discrete data in the historical planning scheme; using a second similarity index to analyze the continuous similarity between the to-be-planned continuous data and the continuous data in the historical planning scheme; combining a preset discrete similarity weight and a preset continuous similarity weight to fuse the discrete similarity and the continuous similarity to obtain the scheme similarity.

3. The method for planning an integrated energy system of claim 2, wherein, The first similarity index is the Jaccard coefficient, and the second similarity index is the Euclidean distance.

4. The method for planning an integrated energy system of claim 2, wherein, Combining the preset discrete similarity weight and the preset continuous similarity weight to fuse the discrete similarity and the continuous similarity to obtain the scheme similarity, specifically comprising: respectively normalizing the discrete similarity and the continuous similarity to obtain standard discrete similarity and standard continuous similarity; The standard discrete similarity and the standard continuous similarity are weighted and summed to obtain the scheme similarity by combining the discrete similarity weight and the continuous similarity weight.

5. The method for planning an integrated energy system of claim 4, wherein, The scheme similarity is specifically represented as: ； wherein S ij is the similarity of the scheme corresponding to the system data i to be planned and the historical planning scheme j in the historical planning scheme library, J i,j is the standard discrete similarity of the discrete data C i to be planned in the system data i to be planned and the discrete data D j in the historical planning scheme j, is the standard continuous similarity of the continuous data a i to be planned in the system data i to be planned and the continuous data b j in the historical planning scheme j, w1 is the continuous similarity weight, and w2 is the discrete similarity weight.

6. The method for planning an integrated energy system of claim 1, wherein, The model evaluation index is at least one of a mean square error, a root mean square error, a mean absolute error, and an R2 score.

7. The method for planning an integrated energy system of claim 1, wherein, The pre-constructed scheme generation model is determined by the following steps: A bi-level programming model is constructed, and an upper layer objective function and a lower layer objective function are set; An iteration number threshold is determined, and the position and the speed of a particle are initialized; The bi-level programming model is iteratively solved by combining the upper layer objective function and the lower layer objective function and using a particle swarm algorithm; If the iteration number reaches the iteration number threshold, the bi-level programming model is solved, and the scheme generation model is obtained.

8. The integrated energy system planning method of claim 7, wherein, The upper layer objective function is used to limit the energy loss of each pipe network in the comprehensive energy system, and the lower layer objective function is used to limit the total economic cost of the comprehensive energy system in the life cycle, wherein the total economic cost includes the installation cost of each device and pipe network in the comprehensive energy system, the operation and maintenance cost of the comprehensive energy system, and the new energy subsidy income in the comprehensive energy system.

9. The method for planning an integrated energy system of claim 8, wherein, The upper layer objective function is specifically represented as: ; The lower layer objective function is specifically represented as: ; wherein, f upper is the upper layer objective function, i is the pipe network number of the integrated energy system, R i is the resistance loss of pipe network i, CL i is the inductance and capacitance loss of pipe network i, FD i is the fixed loss of pipe network i, VB i is the variable loss of pipe network i, OL i is the other loss of pipe network i, f lower is the lower layer objective function, Inst is the installation cost of the integrated energy system, OM is the operation and maintenance cost of the integrated energy system, and Grant is the new energy subsidy income of the integrated energy system. 10.A planning device of an integrated energy system, characterized by comprising: The planning method of the comprehensive energy system is adopted, and the planning method comprises the following steps: A data acquisition module is configured to acquire system data to be planned, wherein the system data to be planned comprises discrete data to be planned and continuous data to be planned; A similarity analysis module is configured to combine a historical planning scheme library, and analyze the scheme similarity between the discrete data to be planned and the continuous data to be planned and the historical planning schemes in the historical planning scheme library by using different similarity indexes; A similarity judgment module is configured to judge the scheme similarity based on a preset scheme similarity threshold to obtain a similarity result. The scheme determining module is configured to: if the similar results corresponding to at least one historical planning scheme in the historical planning scheme library are similar, sort the historical planning schemes in the historical planning scheme library based on the scheme similarity, and give a to-be-determined historical scheme according to the sorting result; map the to-be-planned system data based on the mapping function obtained from the target decision tree model to obtain mapping features; and input the mapping features into the decision tree model to obtain a target planning scheme, wherein the pre-acquired target decision tree model is obtained by analyzing and correcting the to-be-determined historical scheme; and the pre-acquired target decision tree model is determined by the following steps: filtering the historical planning schemes in the historical planning scheme library based on the similar results of the historical planning schemes to obtain a model training data set, wherein the model training data set includes demand training data and scheme training data; initializing the decision tree model and setting model parameters in the decision tree model; training the decision tree model using the model training data set until convergence; evaluating the trained decision tree model using a model evaluation index to obtain a model evaluation result; if the model evaluation result is qualified, integrating the demand training data and the scheme training data to obtain integrated features; performing domain adaptation on the trained decision tree model based on the integrated features to obtain a mapping function containing the mapping relationship between the demand training data and the scheme training data; training the trained decision tree model in combination with the mapping function to obtain the target decision tree model; and if the similar results of the historical planning schemes in the historical planning scheme library are all dissimilar, giving a target planning scheme in combination with the to-be-planned system data and a pre-constructed scheme generation model.

11. A configuration system for an integrated energy system, comprising a memory, a processor, and a computer program stored on the memory, wherein, The computer program is run by the processor, and the instructions of the method according to any one of claims 1-9 are executed.

Images