A method and system for excavating carbon reduction capacity of a transformer district

Abstract

This invention discloses a method and system for mining the carbon reduction capacity of distribution transformer areas. The method includes: acquiring time-series operational data of distribution transformer areas, calculating carbon-related characteristic indicators and constructing basic feature vectors; generating text-based feature vectors by combining distribution transformer area attribute data; constructing a distribution transformer area relationship graph, using a graph neural network to learn node embedding features, and calculating a collaborative carbon reduction potential score; after fusing multimodal features, identifying distribution transformer areas with similar electrical carbon behavior patterns through dynamic clustering; and finally calculating a comprehensive carbon reduction potential index, classifying potential levels, and formulating differentiated carbon reduction strategies. This invention, through multimodal feature fusion and dynamic clustering, achieves a multi-dimensional and accurate assessment of the carbon reduction capacity of distribution transformer areas, overcoming the problems of existing technologies having a single assessment dimension and neglecting the collaborative relationships between distribution transformer areas, and significantly improving the precision level and decision-making efficiency of low-carbon management of distribution networks.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of carbon emission reduction, and particularly relates to a transformer area carbon emission reduction capacity mining method and system. BACKGROUND

[0002] As a key link connecting the main grid and terminal users, the distribution network has widely connected distributed photovoltaic and energy storage clean energy at the transformer area level. The carbon emission characteristics and carbon emission reduction potential of the transformer area have become the focus of industry research.

[0003] However, the existing transformer area carbon emission evaluation and carbon emission reduction capacity mining method still has obvious deficiencies. First, the existing research and practice focus on a single dimension, such as only focusing on the proportion of clean energy or simply considering the load statistical characteristics, and lack a systematic and multi-dimensional evaluation index system. This leads to difficulty in comprehensively and scientifically describing the comprehensive electric carbon behavior of the transformer area, and cannot provide accurate basis for formulating differentiated carbon emission reduction strategies.

[0004] Secondly, the traditional method analyzes the load characteristics and carbon emission of the transformer area in a static and isolated manner, and does not fully consider the synergistic effect between transformer areas due to electrical connection, geographical proximity and other factors. The user types in the transformer area are diverse, and the load characteristics are complex. The traditional clustering method relies on subjective experience or a single feature, and it is difficult to effectively process multi-dimensional indexes and identify the inherent and representative behavior patterns between transformer area groups.

[0005] In addition, the carbon emission characteristics of the transformer area have significant time period and dynamic characteristics. The key behaviors such as load peak and photovoltaic output have great differences in different time periods in a day. The existing technology lacks effective means for dynamic time period analysis and visualization of the carbon image of the transformer area, and cannot intuitively show the carbon emission rules and carbon emission reduction capacity changes of the transformer area in different scenarios throughout the day, which limits the timeliness and accuracy of the regulation strategy.

[0006] Therefore, there is an urgent need in the art for a transformer area carbon emission reduction capacity evaluation method that can systematically integrate multi-source data, construct multi-dimensional features, mine transformer area synergistic relationships, and achieve dynamic and refined analysis, to overcome the above-mentioned deficiencies of the prior art. SUMMARY

[0007] The application provides a transformer area carbon emission reduction capacity mining method and system, which solves the technical problem that the prior art lacks effective means for dynamic time period analysis and visualization of the carbon image of the transformer area, and cannot intuitively show the carbon emission rules and carbon emission reduction capacity changes of the transformer area in different scenarios throughout the day, which limits the timeliness and accuracy of the regulation strategy.

[0008] In a first aspect, the application provides a transformer area carbon emission reduction capacity mining method, comprising:

[0009] Obtain time sequence operation data of a plurality of transformer areas, calculate a plurality of carbon-related feature indicators of each transformer area based on the time sequence operation data, and splice the plurality of carbon-related feature indicators to obtain a basic feature vector of each transformer area;

[0010] Calculate a structural auxiliary feature vector based on transformer area attribute data, splice the structural auxiliary feature vector with the basic feature vector, and perform normalization processing to obtain a standardized text type feature vector;

[0011] Take each transformer area as a node, take the electrical connection relationship and geographical proximity relationship between transformer areas as edges, construct a transformer area relationship graph, assign the text type feature vector to the nodes in the transformer area relationship graph as an initial attribute, use a graph neural network to learn the transformer area relationship graph, and obtain an embedded feature vector of each node that has fused neighborhood information;

[0012] Calculate a collaborative carbon reduction potential score of each transformer area based on the embedded feature vector, and splice the collaborative carbon reduction potential score and the basic feature vector of the same transformer area to obtain a fusion feature vector;

[0013] Perform dynamic clustering on all transformer areas based on the fusion feature vector to obtain a plurality of transformer area groups with similar electrical carbon behavior patterns;

[0014] Calculate a carbon reduction potential comprehensive index of each transformer area based on the plurality of carbon-related feature indicators and the collaborative carbon reduction potential score, perform hierarchical division on all transformer areas according to the carbon reduction potential comprehensive index, and develop differentiated carbon reduction strategies for transformer areas of different levels.

[0015] In a second aspect, the present application provides a transformer area carbon reduction capacity mining system, comprising:

[0016] An acquisition module configured to obtain time sequence operation data of a plurality of transformer areas, calculate a plurality of carbon-related feature indicators of each transformer area based on the time sequence operation data, and splice the plurality of carbon-related feature indicators to obtain a basic feature vector of each transformer area;

[0017] A first splicing module configured to calculate a structural auxiliary feature vector based on transformer area attribute data, splice the structural auxiliary feature vector with the basic feature vector, and perform normalization processing to obtain a standardized text type feature vector;

[0018] A construction module configured to take each transformer area as a node, take the electrical connection relationship and geographical proximity relationship between transformer areas as edges, construct a transformer area relationship graph, assign the text type feature vector to the nodes in the transformer area relationship graph as an initial attribute, use a graph neural network to learn the transformer area relationship graph, and obtain an embedded feature vector of each node that has fused neighborhood information;

[0019] The second splicing module is configured to calculate a cooperative carbon reduction potential score of each substation based on the embedding feature vector, and splice the cooperative carbon reduction potential score and the basic feature vector of the same substation to obtain a fusion feature vector.

[0020] The clustering module is configured to perform dynamic clustering on all substations based on the fusion feature vector to obtain a plurality of substation groups with similar electric carbon behavior patterns.

[0021] The division module is configured to calculate a carbon reduction potential comprehensive index of each substation based on the plurality of carbon-related feature indicators and the cooperative carbon reduction potential score, perform grade division on all substations according to the carbon reduction potential comprehensive index, and formulate differentiated carbon reduction strategies for substations of different grades.

[0022] In a third aspect, an electronic device is provided, which includes at least one processor, and a memory connected to the at least one processor in communication, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the substation carbon reduction capacity mining method of any of the embodiments of the present application.

[0023] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and the program instructions are executed by a processor to enable the processor to perform the steps of the substation carbon reduction capacity mining method of any of the embodiments of the present application.

[0024] The substation carbon reduction capacity mining method and system of the present application are aimed at the problem of single evaluation index of substations, calculate a plurality of carbon-related feature indicators, systematically evaluate the carbon reduction capacity of substations from the aspects of power characteristics, load fluctuation, carbon emission intensity and clean energy utilization, realize modeling operation from static structure to dynamic behavior, introduce a graph neural network modeling method to map the structure of substation users, geographical position and bus connection system into a graph structure, capture cooperative carbon reduction behavior through feature aggregation between nodes, improve the correlation recognition ability of the system level, design a dynamic clustering mechanism based on dynamic time warping and policy gradient reinforcement learning, enhance the adaptability of the model to the sequence of substation power load and carbon emission, and realize real-time dynamic recognition of the behavior pattern of the substation. BRIEF DESCRIPTION OF DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0026] Figure 1 A flow chart of a transformer area carbon reduction capacity mining method provided by an embodiment of the present application is shown in FIG. 1.

[0027] Figure 2 A structural block diagram of a transformer area carbon reduction capacity mining system provided by an embodiment of the present application is shown in FIG. 2.

[0028] Figure 3 A structural schematic diagram of an electronic device provided by an embodiment of the present application is shown in FIG. 3. DETAILED DESCRIPTION

[0029] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in conjunction with the drawings of the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0030] Please refer to Figure 1 , which shows a flow chart of a transformer area carbon reduction capacity mining method of the present application.

[0031] As shown in Figure 1 , the transformer area carbon reduction capacity mining method specifically includes the following steps:

[0032] In step S101, time-series operation data of a plurality of transformer areas is acquired, a plurality of carbon-related characteristic indexes of each transformer area is calculated based on the time-series operation data, and the plurality of carbon-related characteristic indexes are spliced to obtain a basic feature vector of each transformer area.

[0033] In this step, the plurality of carbon-related characteristic indexes include normalized daily power consumption, load peak-valley difference rate, reducible load proportion, time-shiftable load proportion, clean energy proportion, and carbon emission intensity.

[0034] The expression for calculating the normalized daily power consumption is as follows:

[0035]

[0036]

[0037] In the formula, is the normalized daily power consumption of the transformer area, is the daily power consumption of the transformer area i, is the average daily power consumption of the transformer area in the statistical range, is the total number of transformer areas, is the daily power consumption of the jth transformer area.

[0038] ​​The expression for calculating the load peak-valley difference rate is:

[0039] ;

[0040] In the formula, is the load peak-valley difference rate, is the peak load of the transformer area i, that is, the maximum power consumption, is the valley load of the transformer area i, that is, the minimum power consumption;

[0041] The expression for calculating the clean energy proportion is:

[0042] ;

[0043] In the formula, is the clean energy proportion, is the photovoltaic power generation power of the transformer area i at the time period t, is the total daily power consumption of the transformer area i, is the total number of time periods in a day;

[0044] The expression for calculating the carbon emission intensity is:

[0045] ;

[0046] ;

[0047] In the formula, is the carbon emission intensity, is the total carbon emission of the transformer area i, is the carbon emission factor, indicating the carbon emission corresponding to unit power consumption, is the load power of the transformer area i at the time period t.

[0048] It should be noted that the reducible load proportion represents the load capacity that can be directly reduced by the transformer area when participating in demand response, and is positively correlated with the proportion of industrial users; the time-shiftable load proportion reflects the time flexibility of load optimization scheduling, and is closely related to the proportion of commercial users. According to the transformer area type (residential, commercial, and industrial), different parameters are directly assigned to the reducible load proportion and the time-shiftable load proportion, respectively, to reflect the regulation and control capabilities of different types of transformer areas.

[0049] In step S102, a structural auxiliary feature vector is calculated based on transformer area attribute data, and the structural auxiliary feature vector is spliced with the basic feature vector and then normalized to obtain a standardized text type feature vector.

[0050] In this step, the structural auxiliary feature vector includes the proportion of residential user load, the proportion of industrial user load, the proportion of commercial user load, the charging pile penetration rate, and the distributed photovoltaic penetration rate;

[0051] wherein the expression for calculating the residential user load ratio is:

[0052] ;

[0053] wherein, is the residential user load ratio, is the number of residential users, is the rated power consumption of the kth residential user, is the total power consumption of the transformer area;

[0054] The expression for calculating the industrial user load ratio is:

[0055] ;

[0056] wherein, is the industrial user load ratio, is the number of industrial users, is the working load of the mth industrial user;

[0057] The expression for calculating the commercial user load ratio is:

[0058] ;

[0059] ;

[0060] wherein, is the commercial user load ratio, is the number of commercial users, is the peak load of the nth commercial user, is the total power consumption of all users, is the total power consumption of all distributed devices, is the power consumption of a single user, is the power consumption of a single distributed device;

[0061] The expression for calculating the charging pile penetration rate is:

[0062] ;

[0063] wherein, is the charging pile penetration rate, is the number of charging piles, is the rated charging power of the ith charging pile, is the rated capacity of the transformer distribution;

[0064] The expression for calculating the distributed photovoltaic penetration rate is:

[0065] ;

[0066] In the formula, For distributed photovoltaic penetration rate, For the number of photovoltaic nodes, Let be the installed capacity of the j-th photovoltaic node.

[0067] Step S103: Using each transformer substation as a node and the electrical connection relationship and geographical proximity relationship between transformer substations as edges, construct a transformer substation relationship graph, and assign the text-type feature vector as an initial attribute to the nodes in the transformer substation relationship graph. Use a graph neural network to learn the transformer substation relationship graph to obtain an embedded feature vector for each node that integrates neighborhood information.

[0068] In this step, the physical network of transformer substations is abstracted into a mathematical graph structure. Specifically, each transformer area is defined as a graph node. The total number of nodes equals the number of stations to be analyzed. It is a set of nodes, where each node corresponds to a transformer area. This is a set of edges representing the electrical connections or geographical proximity between transformer substations. Then, each node is assigned attributes derived from the obtained standardized text-based feature vectors. The edge relationships between nodes are then defined. At the same time, electrical connectivity and geographical proximity are taken into account. If transformer area i and transformer area j are connected to the same busbar and the geographical distance between them is less than a set threshold, then... Then an undirected edge is established between them. To quantify the tightness of connections between nodes, each edge is assigned a weight, which can be calculated using a distance-based Gaussian kernel function:

[0069] ;

[0070] In the formula, For the first The weight values ​​of the edges, The bandwidth parameter is used to control the weight decay rate. Let be the geographical distance between station i and station j. For indicator functions;

[0071] The final output is a weighted graph of the transformer substation relationship with nodes already assigned attributes. .

[0072] After obtaining the diagram of the relationship between the power stations Then, utilizing the message-passing mechanism of Graph Neural Networks (GNNs), the feature information of each node is propagated and aggregated among its neighboring nodes, thereby learning a new feature representation containing local topological context information. A Graph Convolutional Network (GCN) is used as the specific implementation model. For each node in the graph, its GCN... Layer Embedded Feature Vector Update using the following formula:

[0073] ;

[0074] In the formula, It is a non-linear activation function. Let be the degree of the node in region i, i.e., the number of connected edges. For the first The layer's trainable weight matrix is ​​used for feature transformation. For nodes The set of neighboring nodes, For the first The embedded feature vector of the layer, Let j be the node degree of the substation. For the first Embedded feature vectors of layer -1 Add a self-loop to node i so that node i retains its original features during feature aggregation, instead of relying solely on the features of its neighbors.

[0075] Step S104: Based on the embedded feature vector, calculate the synergistic carbon reduction potential score for each transformer area, and concatenate the synergistic carbon reduction potential score and the basic feature vector of the same transformer area to obtain a fused feature vector.

[0076] In this step, the expression for calculating the synergistic carbon reduction potential score for each distribution area is as follows:

[0077] ;

[0078] In the formula, Assess the collaborative carbon reduction potential of the i-th district. For the i-th node Final embedded feature vector The first in dimensional components, For the first The weighted coefficients of the dimensions reflect the degree to which different dimensions contribute to the synergistic potential. The carbon emission intensity of area i.

[0079] Step S105: Based on the fused feature vector, all transformer substations are dynamically clustered to obtain multiple substation substation groups with similar electric carbon behavior patterns.

[0080] In this step, the fused feature vectors of each transformer area at multiple time periods throughout the day are arranged in chronological order to construct a time-series feature sequence; the similarity distance between time-series feature sequences of different transformer areas is calculated using a dynamic time warping algorithm; based on the similarity distance, the K-shapes clustering algorithm is used to group all transformer areas to obtain multiple transformer area clusters.

[0081] Step S106: Based on the multiple carbon-related characteristic indicators and the synergistic carbon reduction potential score, calculate the comprehensive carbon reduction potential index for each transformer substation, classify all transformer substations according to the comprehensive carbon reduction potential index, and formulate differentiated carbon reduction strategies for transformer substations of different levels.

[0082] In this step, the expression for calculating the comprehensive carbon reduction potential index for each distribution area is as follows:

[0083] ;

[0084] In the formula, As a comprehensive index of carbon reduction potential, To normalize daily electricity consumption, Peak-to-valley difference rate In order to reduce the load ratio, For the proportion of time-shiftable loads, For the proportion of clean energy, For carbon emission intensity, Assess the collaborative carbon reduction potential of the i-th district. to Here are the weighting coefficients for the corresponding indicators, and , This is the k-th weight coefficient.

[0085] According to all TV districts Based on the statistical distribution characteristics, the percentile method was used to scientifically classify them into different levels of carbon reduction potential, for example: Distribution areas ranking in the top 25% are classified as having high carbon reduction potential, typically characterized by high load volatility and a high proportion of carbon reduction / time-shiftable capacity. Areas between the 25% and 60% percentiles are classified as having medium carbon reduction potential, and the remaining areas are classified as having low carbon reduction potential. For example, for high-potential areas, demand response projects should be prioritized, with precise peak-shaving control implemented during morning and evening peak load periods, supplemented by strong electricity price incentives. For medium-potential areas, the focus should be on optimizing internal energy management, promoting energy-saving equipment, and implementing smart electricity consumption strategies based on time-of-use pricing. For low-potential areas characterized by "low volatility and high penetration," the primary focus should be on maintaining their efficient and low-carbon operation, while encouraging their participation in green electricity trading to further tap their potential. This closed-loop management approach of "assessment-classification-policy implementation" using quantitative indices can significantly improve the efficiency and effectiveness of the overall low-carbon operation of the distribution network, strongly supporting the achievement of "dual carbon" goals.

[0086] In summary, the method in this application addresses the problem of a single evaluation indicator for transformer substations by calculating multiple carbon-related characteristic indicators. It systematically evaluates the carbon reduction capacity of transformer substations from aspects such as power characteristics, load fluctuations, carbon emission intensity, and clean energy utilization. It achieves modeling operations from static structure to dynamic behavior, introduces graph neural network modeling method, and uniformly maps the user structure, geographical location, and bus connection relationship of transformer substations into a graph structure. It captures collaborative carbon reduction behavior through feature aggregation between nodes, improves the correlation recognition capability of system hierarchy, and designs a dynamic clustering mechanism based on dynamic time warping and policy gradient reinforcement learning to enhance the model's adaptability to transformer substation power load and carbon emission sequences, and realizes real-time dynamic recognition of transformer substation behavior patterns.

[0087] Please see Figure 2 The diagram shows a structural block diagram of a system for tapping carbon reduction capacity in a transformer substation, as described in this application.

[0088] like Figure 2 As shown, the carbon reduction capacity mining system 200 for the transformer area includes an acquisition module 210, a first splicing module 220, a construction module 230, a second splicing module 240, a clustering module 250, and a division module 260.

[0089] The acquisition module 210 is configured to acquire time-series operational data of multiple transformer substations, calculate multiple carbon-related feature indicators for each substation based on the time-series operational data, and concatenate the multiple carbon-related feature indicators to obtain the basic feature vector of each substation. The first concatenation module 220 is configured to calculate a structural auxiliary feature vector based on the substation attribute data, and concatenate the structural auxiliary feature vector with the basic feature vector and then perform normalization processing to obtain a standardized text-type feature vector. The construction module 230 is configured to construct a substation relationship graph with each substation as a node and the electrical connection relationship and geographical proximity relationship between substations as edges, and assign the text-type feature vector as the initial attribute to the nodes in the substation relationship graph, and use a graph neural network to process the data. The system learns from the relationship graph of power distribution areas to obtain an embedded feature vector for each node that incorporates neighborhood information. A second splicing module 240 is configured to calculate the collaborative carbon reduction potential score for each power distribution area based on the embedded feature vector, and splice the collaborative carbon reduction potential score and the basic feature vector of the same power distribution area to obtain a fused feature vector. A clustering module 250 is configured to dynamically cluster all power distribution areas based on the fused feature vector to obtain multiple power distribution area groups with similar carbon emission behavior patterns. A partitioning module 260 is configured to calculate a comprehensive carbon reduction potential index for each power distribution area based on the multiple carbon-related feature indicators and the collaborative carbon reduction potential score, classify all power distribution areas according to the comprehensive carbon reduction potential index, and formulate differentiated carbon reduction strategies for power distribution areas of different levels.

[0090] It should be understood that Figure 2 The modules and references described in the document Figure 1 The steps described in the text correspond to those in the method described above. Therefore, the operations, features, and corresponding technical effects described above also apply to the method described in the text. Figure 2 The various modules in the document will not be described in detail here.

[0091] In other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the method for tapping the carbon reduction capacity of the transformer area in any of the above method embodiments.

[0092] In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows:

[0093] The time-series operation data of multiple transformer substations are acquired, and multiple carbon-related feature indicators of each transformer substation are calculated based on the time-series operation data. The multiple carbon-related feature indicators are then concatenated to obtain the basic feature vector of each transformer substation.

[0094] Calculate structural auxiliary feature vectors based on the attribute data of the transformer area, and then normalize the structural auxiliary feature vectors after concatenating them with the basic feature vectors to obtain standardized text-type feature vectors.

[0095] With each transformer substation as a node and the electrical connection and geographical proximity between substations as edges, a transformer substation relationship graph is constructed. The text-type feature vector is assigned as an initial attribute to the nodes in the transformer substation relationship graph. The graph neural network is used to learn the transformer substation relationship graph to obtain an embedded feature vector for each node that incorporates neighborhood information.

[0096] Based on the embedded feature vector, the collaborative carbon reduction potential score of each transformer area is calculated, and the collaborative carbon reduction potential score of the same transformer area and the basic feature vector are concatenated to obtain the fused feature vector.

[0097] Based on the fused feature vector, all transformer substations are dynamically clustered to obtain multiple substation substation groups with similar electric carbon behavior patterns.

[0098] Based on the multiple carbon-related characteristic indicators and the synergistic carbon reduction potential score, a comprehensive carbon reduction potential index is calculated for each transformer substation. All transformer substations are classified according to the comprehensive carbon reduction potential index, and differentiated carbon reduction strategies are formulated for transformer substations of different levels.

[0099] Computer-readable storage media may include a stored program area and a stored data area, wherein the stored program area may store an operating system and an application program required for at least one function; the stored data area may store data created based on the use of the carbon reduction capacity mining system, etc. Furthermore, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include memory remotely configured relative to a processor, which can be connected to the carbon reduction capacity mining system via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0100] Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 3 As shown, the device includes a processor 310 and a memory 320. The electronic device may also include an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 can be connected via a bus or other means. Figure 3Taking a bus connection as an example, the memory 320 is the computer-readable storage medium described above. The processor 310 executes various server functions and data processing by running non-volatile software programs, instructions, and modules stored in the memory 320, thereby implementing the above-described method embodiment for exploring the carbon reduction capacity of the distribution area. The input device 330 can receive input digital or character information and generate key signal inputs related to user settings and function control of the distribution area carbon reduction capacity exploration system. The output device 340 may include a display screen or other display device.

[0101] The aforementioned electronic device can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0102] In one implementation, the above-described electronic device is applied to a transformer substation carbon reduction capacity exploration system for a client application, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to:

[0103] The time-series operation data of multiple transformer substations are acquired, and multiple carbon-related feature indicators of each transformer substation are calculated based on the time-series operation data. The multiple carbon-related feature indicators are then concatenated to obtain the basic feature vector of each transformer substation.

[0104] Calculate structural auxiliary feature vectors based on the attribute data of the transformer area, and then normalize the structural auxiliary feature vectors after concatenating them with the basic feature vectors to obtain standardized text-type feature vectors.

[0105] With each transformer substation as a node and the electrical connection and geographical proximity between substations as edges, a transformer substation relationship graph is constructed. The text-type feature vector is assigned as an initial attribute to the nodes in the transformer substation relationship graph. The graph neural network is used to learn the transformer substation relationship graph to obtain an embedded feature vector for each node that incorporates neighborhood information.

[0106] Based on the embedded feature vector, the collaborative carbon reduction potential score of each transformer area is calculated, and the collaborative carbon reduction potential score of the same transformer area and the basic feature vector are concatenated to obtain the fused feature vector.

[0107] Based on the fused feature vector, all transformer substations are dynamically clustered to obtain multiple substation substation groups with similar electric carbon behavior patterns.

[0108] Based on the multiple carbon-related characteristic indicators and the synergistic carbon reduction potential score, a comprehensive carbon reduction potential index is calculated for each transformer substation. All transformer substations are classified according to the comprehensive carbon reduction potential index, and differentiated carbon reduction strategies are formulated for transformer substations of different levels.

[0109] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for tapping the carbon reduction capacity of a transformer substation, characterized in that, include: The time-series operation data of multiple transformer substations are acquired, and multiple carbon-related feature indicators of each transformer substation are calculated based on the time-series operation data. The multiple carbon-related feature indicators are then concatenated to obtain the basic feature vector of each transformer substation. Calculate structural auxiliary feature vectors based on the attribute data of the transformer area, and then normalize the structural auxiliary feature vectors after concatenating them with the basic feature vectors to obtain standardized text-type feature vectors. With each transformer substation as a node and the electrical connection and geographical proximity between substations as edges, a transformer substation relationship graph is constructed. The text-type feature vector is assigned as an initial attribute to the nodes in the transformer substation relationship graph. The graph neural network is used to learn the transformer substation relationship graph to obtain an embedded feature vector for each node that incorporates neighborhood information. Based on the embedded feature vector, a synergistic carbon reduction potential score for each transformer substation is calculated. The synergistic carbon reduction potential score and the basic feature vector for the same substation are then concatenated to obtain a fused feature vector. The expression for calculating the synergistic carbon reduction potential score for each substation is as follows: ， In the formula, Assess the collaborative carbon reduction potential of the i-th district. For the i-th node Final embedded feature vector The first in dimensional components, For the first The weighted coefficients of the dimensions reflect the degree to which different dimensions contribute to the synergistic potential. The carbon emission intensity of area i; The expression for the embedded feature vector is: ， In the formula, It is a non-linear activation function. Let be the degree of the node in region i, i.e., the number of connected edges. For the first The layer's trainable weight matrix is ​​used for feature transformation. For nodes The set of neighboring nodes, For the first The embedded feature vector of the layer, Let j be the node degree of the substation. For the first Embedded feature vectors of layer -1 Add a self-loop to node i so that node i retains its original features during feature aggregation, instead of relying solely on the features of its neighbors. Based on the fused feature vector, all transformer substations are dynamically clustered to obtain multiple substation substation groups with similar electric carbon behavior patterns. Based on the multiple carbon-related characteristic indicators and the synergistic carbon reduction potential score, a comprehensive carbon reduction potential index is calculated for each transformer substation. All transformer substations are classified according to the comprehensive carbon reduction potential index, and differentiated carbon reduction strategies are formulated for transformer substations of different levels.

2. The method for tapping the carbon reduction capacity of a transformer substation according to claim 1, characterized in that, The various carbon-related characteristic indicators include normalized daily electricity consumption, peak-valley load difference rate, proportion of load that can be reduced, proportion of load that can be shifted over time, proportion of clean energy, and carbon emission intensity. The expression for calculating the normalized daily electricity consumption is as follows: ， ， In the formula, This refers to the normalized daily electricity consumption of the power distribution area. The daily electricity consumption of transformer area i. This refers to the average daily electricity consumption of the transformer substations within the statistical scope. The total number of stations. Let j be the daily electricity consumption of the j-th transformer substation. The expression for calculating the load peak-valley difference rate is as follows: ， In the formula, The peak-to-valley load difference rate The peak load, or maximum power consumption, of transformer area i. The valley load, i.e., the minimum power consumption, is for transformer area i. The expression for calculating the proportion of clean energy is: ， In the formula, For the proportion of clean energy, Let i be the photovoltaic power generation capacity of the transformer substation during time period t. This represents the total daily electricity consumption of transformer area i. The total number of time periods in a day; The expression for calculating the carbon emission intensity is: ， ， In the formula, For carbon emission intensity, The total carbon emissions of area i. The carbon emission factor represents the amount of carbon emissions per unit of electricity consumed. Let be the load power of transformer area i during time period t.

3. The method for tapping the carbon reduction capacity of a transformer substation according to claim 1, characterized in that, The structural auxiliary feature vector includes the load share of residential users, the load share of industrial users, the load share of commercial users, the penetration rate of charging piles, and the penetration rate of distributed photovoltaic power. The expression for calculating the load ratio of residential users is as follows: ， In the formula, The proportion of residential users' load. For the number of residential users, Let k be the rated power consumption of the kth residential user. This represents the total power consumption of the transformer substation. The expression for calculating the load ratio of the industrial users is as follows: ， In the formula, The proportion of industrial user load, For the number of industrial users, Let m be the workload of the m-th industrial user; The expression for calculating the load percentage of the business users is as follows: ， ， In the formula, The percentage of load for business users, For the number of business users, For the peak load of the nth business user, The total power consumption of all users. This represents the total power consumption of all distributed devices. For the power consumption of a single user, The power consumption of a single distributed device; The expression for calculating the penetration rate of the charging pile is: ， In the formula, For charging pile penetration rate, The number of charging stations. Let i be the rated charging power of the i-th charging pile. Rated capacity of the distribution transformer in the area; The expression for calculating the distributed photovoltaic penetration rate is as follows: ， In the formula, For distributed photovoltaic penetration rate, For the number of photovoltaic nodes, Let be the installed capacity of the j-th photovoltaic node.

4. The method for tapping the carbon reduction capacity of a transformer substation according to claim 1, characterized in that, The dynamic clustering of all transformer substations based on the fused feature vectors to obtain multiple substation groups with similar electrical carbon behavior patterns includes: The fused feature vectors of each transformer area at multiple time periods throughout the day are arranged in chronological order to construct a time-series feature sequence; The similarity distance between time-series feature sequences of different stations is calculated using a dynamic time warping algorithm. Based on the similarity distance, the K-shapes clustering algorithm is used to group all transformer substations, resulting in multiple substation clusters.

5. The method for tapping the carbon reduction capacity of a transformer substation according to claim 1, characterized in that, The expression for calculating the comprehensive carbon reduction potential index for each distribution area is as follows: ， In the formula, As a comprehensive index of carbon reduction potential, To normalize daily electricity consumption, Peak-to-valley difference rate In order to reduce the load ratio, For the proportion of time-shiftable loads, For the proportion of clean energy, For carbon emission intensity, Assess the collaborative carbon reduction potential of the i-th district. to Here are the weighting coefficients for the corresponding indicators, and , This is the k-th weight coefficient.

6. A system for tapping into the carbon reduction capacity of a transformer substation, characterized in that, include: The acquisition module is configured to acquire time-series operation data of multiple transformer substations, calculate multiple carbon-related feature indicators for each transformer substation based on the time-series operation data, and concatenate the multiple carbon-related feature indicators to obtain the basic feature vector of each transformer substation. The first splicing module is configured to calculate a structural auxiliary feature vector based on the transformer area attribute data, and then splice the structural auxiliary feature vector with the basic feature vector and perform normalization processing to obtain a standardized text-type feature vector. The module is configured to construct a transformer area relationship graph with each transformer area as a node and the electrical connection relationship and geographical proximity relationship between transformer areas as edges. The module assigns the text-type feature vector as an initial attribute to the nodes in the transformer area relationship graph and uses a graph neural network to learn the transformer area relationship graph to obtain an embedded feature vector for each node that incorporates neighborhood information. The second splicing module is configured to calculate the synergistic carbon reduction potential score for each transformer substation based on the embedded feature vector, and splice the synergistic carbon reduction potential score of the same transformer substation with the basic feature vector to obtain a fused feature vector. The expression for calculating the synergistic carbon reduction potential score for each transformer substation is as follows: ， In the formula, Assess the collaborative carbon reduction potential of the i-th district. For the i-th node Final embedded feature vector The first in dimensional components, For the first The weighted coefficients of the dimensions reflect the degree to which different dimensions contribute to the synergistic potential. The carbon emission intensity of area i; The expression for the embedded feature vector is: ， In the formula, It is a non-linear activation function. Let be the degree of the node in region i, i.e., the number of connected edges. For the first The layer's trainable weight matrix is ​​used for feature transformation. For nodes The set of neighboring nodes, For the first The embedded feature vector of the layer, Let j be the node degree of the substation. For the first Embedded feature vectors of layer -1 Add a self-loop to node i so that node i retains its original features during feature aggregation, instead of relying solely on the features of its neighbors. The clustering module is configured to dynamically cluster all transformer substations based on the fused feature vectors to obtain multiple transformer substation groups with similar electric carbon behavior patterns. The classification module is configured to calculate a comprehensive carbon reduction potential index for each transformer substation based on the multiple carbon-related characteristic indicators and the synergistic carbon reduction potential score, classify all transformer substations according to the comprehensive carbon reduction potential index, and formulate differentiated carbon reduction strategies for transformer substations of different levels.

7. An electronic device, characterized in that, include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method described in any one of claims 1 to 5.

Images