SF6 gas equivalent carbon emission analysis method, system and platform based on clustering analysis

Abstract

The invention provides an SF6 gas equivalent carbon emission analysis method, system and platform based on clustering analysis, which realize efficient cleaning and fusion processing of multi-source heterogeneous data generated in the use, recovery and purification process of SF6 gas, ensure the integrity, consistency and reliability of the data, and provide a high-quality data basis for subsequent accurate accounting of carbon emission. On the basis of the cleaned standardized data, an SF6 equivalent carbon emission accounting model is constructed, automatic calculation and analysis of the carbon emission are achieved, and the accuracy, timeliness and transparency of an accounting result are comprehensively improved; the SF6 gas distribution and the carbon emission are subjected to clustering identification and feature analysis by adopting a clustering analysis method according to multi-dimensional feature attributes such as regions, voltage grades and equipment models, emission hot spots and difference rules under different dimensions are identified, and a scientific basis is provided for formulating a grading and classifying management and control strategy.

Description

Technical Field

[0001] This invention relates to the field of greenhouse gas emission monitoring and environmental protection technology for power equipment, specifically a method, system, and platform for analyzing the equivalent carbon emissions of sulfur hexafluoride gas based on cluster analysis. Background Technology

[0002] In power systems, sulfur hexafluoride (SF6) is widely used in key equipment such as high-voltage switches, gas-insulated switchgear (GIS), and transformers due to its excellent insulation and arc-quenching properties, playing an irreplaceable role in ensuring the safe and stable operation of the power grid. However, its global warming potential (GWP) is extremely high, 25,200 times that of CO2, making it a significant source of carbon emissions for power grid companies. Currently, power grid companies face the following problems in managing SF6 gas: Data silos and heterogeneity issues: The entire chain of SF6 gas production, transportation, storage, use and recycling data is scattered across multiple independent systems (such as PMS production management system, ERP resource management system, laboratory LIMS system and various field operation records). The data formats and standards are inconsistent, forming information silos and making it difficult to conduct global correlation analysis.

[0003] The current carbon emission accounting system is crude, lacking real-time performance and accuracy. It relies heavily on periodic manual statistics and report compilation, which is not only inefficient but also prone to data errors due to human factors. The accounting process lacks transparency and traceability, failing to meet the MRV (Measurable, Reportable, Verifiable) carbon management requirements.

[0004] Lack of in-depth insights and intelligent decision support: Existing management methods mostly remain at the level of total statistics, failing to delve into emission hotspots, identify management differences, and summarize best practices from multiple dimensions such as region, voltage level, equipment type, and operation process. The formulation of management strategies lacks data-driven approaches, relying heavily on experience, and lacks specificity and foresight.

[0005] Weak regulation of the recovery and purification process: There is insufficient regulation of the key emission reduction link of SF6 gas recovery, purification and reuse, and there is a lack of accurate measurement and assessment methods for its recovery efficiency, purification effectiveness and process carbon emissions.

[0006] In the existing technology, there is no intelligent management system for SF6 carbon emissions that integrates automatic data collection, multi-source information fusion, intelligent analysis, visual monitoring and decision support. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide an intelligent management system for SF6 carbon emissions that integrates automatic data collection, multi-source information fusion, intelligent analysis, visual monitoring and decision support.

[0008] The present invention solves the above-mentioned technical problems through the following technical means.

[0009] The SF6 gas equivalent carbon emission analysis method based on cluster analysis includes the following steps: S01. Collect raw SF6 gas operation data from systems such as the Power Equipment Oil and Gas Medium Laboratory of State Grid Anhui Electric Power Research Institute, and process it through cleaning, standardization and multi-source fusion. S02. Based on the standardized data generated in step S01, calculate the multi-dimensional net loss of SF6 gas using a gas accounting model, and convert it into equivalent carbon emissions using the GWP value; S03. By combining SF6 gas equivalent carbon emission data from different regions, work categories, and voltage levels, a clustering model is constructed to analyze and uncover the common characteristics and differences in SF6 gas management across different regions and types of equipment, providing a scientific basis for formulating more precise management strategies.

[0010] Furthermore, step S01 specifically includes: Data Acquisition: Obtain raw data on SF6 gas operations; the data should include at least the following key fields: city name, work category, equipment voltage level, total equipment charge volume, total actual recovered gas volume, total unrecovered gas volume, SF6 gas mass before purification, SF6 gas mass after purification, SF6 global warming potential, and SF6 equivalent CO2 emissions. Data cleaning and standardization: Perform integrity verification and logical consistency checks on the collected raw data, identify and handle missing values, outliers and duplicate records; standardize the units and formats of key numerical fields to ensure data quality and consistency; Data fusion and integration: The cleaned data is associated and integrated according to key fields such as city name, voltage level, and job category; a standardized dataset including equipment operation, gas recovery, and carbon emission potential information is constructed.

[0011] Furthermore, step S02 specifically includes: Multi-dimensional gas loss calculation: Taking each gas operation of the equipment as the basic unit, based on the "Total Equipment Gas Charge" and "Total Actual Recovered Gas" fields in the operation record, the net SF6 gas loss for that operation is calculated using the following formula:

[0012] Based on this, the loss of all operation records in the region is grouped, aggregated and summarized by city name, work category and equipment voltage level as key dimensions to form SF6 gas loss distribution table under different dimension combinations, accurately identifying high loss links and key equipment types. Accurate calculation of equivalent carbon emissions: Using the global warming potential standard value of SF6, the calculated SF6 losses for each group are converted into equivalent CO2 emissions; the conversion formula is as follows:

[0013] Output the carbon emission results dataset.

[0014] Furthermore, step S03 specifically includes: 3.1 Clustering Feature Engineering: (1) Using various cities as the basic analysis sample; (2) Select the following feature indicators to construct the feature vector: Carbon emission intensity: Equivalent CO2 emissions / number of devices; Gas recovery rate: (Actual total recovered gas volume / Total gas volume used) × 100%; Unit inflation emission: Equivalent CO2 emission / Total inflation volume; Maintenance operation emissions percentage: Maintenance category emissions / Total emissions × 100%; (3) The Z-score standardization method is used to process the feature data to eliminate the influence of the units and obtain the standardized feature matrix; 3.2 Clustering Model Construction: The K-means clustering algorithm was selected; the optimal number of clusters K was determined by combining the silhouette coefficient with the elbow rule; the standardized feature matrix was input into the model for training, and the cluster centers were iteratively optimized until convergence or the maximum number of iterations was reached. 3.3 Clustering Results Analysis: (1) Calculate the mean and standard deviation of each cluster feature index to generate a cluster feature profile: High-efficiency, low-carbon clusters: high recovery rate (>95%), low emission intensity; High-emission improved cluster: low recovery rate, high emissions per unit of inflated gas; Maintenance-sensitive clusters: high proportion of emissions from maintenance operations; (2) Use visualization tools to display the clustering results; 3.4 Differentiation Strategy Generation: Summarize best practices for high-efficiency, low-carbon clusters; and develop precise improvement measures for high-emission improvement clusters and maintenance-sensitive clusters.

[0015] This invention also provides a cluster analysis-based system for analyzing the equivalent carbon emissions of SF6 gas, comprising the following steps: Data acquisition and processing module: Collects raw SF6 gas operation data from systems such as the Power Equipment Oil and Gas Medium Laboratory of State Grid Anhui Electric Power Research Institute, and performs cleaning, standardization and multi-source fusion processing; Gas accounting module: Based on standardized data, it calculates the net SF6 gas loss from multiple dimensions using a gas accounting model, and converts the GWP value into equivalent carbon emissions; Clustering analysis module: By combining SF6 gas equivalent carbon emission data from different regions, work categories, and voltage levels, a clustering model is constructed to analyze and uncover the common characteristics and differences in SF6 gas management across different regions and types of equipment, providing a scientific basis for formulating more precise management strategies.

[0016] Furthermore, the specific execution process of the data acquisition and processing module in the above steps is as follows: Data Acquisition: Obtain raw data on SF6 gas operations; the data should include at least the following key fields: city name, work category, equipment voltage level, total equipment charge volume, total actual recovered gas volume, total unrecovered gas volume, SF6 gas mass before purification, SF6 gas mass after purification, SF6 global warming potential, and SF6 equivalent CO2 emissions. Data cleaning and standardization: Perform integrity verification and logical consistency checks on the collected raw data, identify and handle missing values, outliers and duplicate records; standardize the units and formats of key numerical fields to ensure data quality and consistency; Data fusion and integration: The cleaned data is associated and integrated according to key fields such as city name, voltage level, and job category; a standardized dataset including equipment operation, gas recovery, and carbon emission potential information is constructed.

[0017] Furthermore, the specific execution process of the gas accounting module is as follows: Multi-dimensional gas loss calculation: Taking each gas operation of the equipment as the basic unit, based on the "Total Equipment Gas Charge" and "Total Actual Recovered Gas" fields in the operation record, the net SF6 gas loss for that operation is calculated using the following formula:

[0018] Based on this, the loss of all operation records in the region is grouped, aggregated and summarized by city name, work category and equipment voltage level as key dimensions to form SF6 gas loss distribution table under different dimension combinations, accurately identifying high loss links and key equipment types. Accurate calculation of equivalent carbon emissions: Using the global warming potential standard value of SF6, the calculated SF6 losses for each group are converted into equivalent CO2 emissions; the conversion formula is as follows:

[0019] Output the carbon emission results dataset.

[0020] Furthermore, the clustering analysis module performs the following process: 3.1 Clustering Feature Engineering: (1) Using various cities as the basic analysis sample; (2) Select the following feature indicators to construct the feature vector: Carbon emission intensity: Equivalent CO2 emissions / number of devices; Gas recovery rate: (Actual total recovered gas volume / Total gas volume used) × 100%; Unit inflation emission: Equivalent CO2 emission / Total inflation volume; Maintenance operation emissions percentage: Maintenance category emissions / Total emissions × 100%; (3) The Z-score standardization method is used to process the feature data to eliminate the influence of the units and obtain the standardized feature matrix; 3.2 Clustering Model Construction: The K-means clustering algorithm was selected; the optimal number of clusters K was determined by combining the silhouette coefficient with the elbow rule; the standardized feature matrix was input into the model for training, and the cluster centers were iteratively optimized until convergence or the maximum number of iterations was reached. 3.3 Clustering Results Analysis: (1) Calculate the mean and standard deviation of each cluster feature index to generate a cluster feature profile: High-efficiency, low-carbon clusters: high recovery rate (>95%), low emission intensity; High-emission improved cluster: low recovery rate, high emissions per unit of inflated gas; Maintenance-sensitive clusters: high proportion of emissions from maintenance operations; (2) Use visualization tools to display the clustering results; 3.4 Differentiation Strategy Generation: Summarize best practices for high-efficiency, low-carbon clusters; and develop precise improvement measures for high-emission improvement clusters and maintenance-sensitive clusters.

[0021] The present invention also provides an SF6 gas equivalent carbon emission analysis platform based on cluster analysis, comprising a data layer, a carbon emission accounting layer, an intelligent cluster analysis layer, and an application layer; the data layer, carbon emission accounting layer, and intelligent cluster analysis layer are respectively used to execute steps S01, S02, and S03 in any one of the methods described in claims 1 to 4; the application layer is used to generate strategies and visualize the results based on the cluster analysis.

[0022] The advantages of this invention are: 1. Systematic integration and governance mechanism for multi-source heterogeneous data: Construct a complete process for the collection, cleaning, integration and standardization of multi-source SF6 gas business data, effectively integrate heterogeneous data sources from internal and external sources such as equipment operation and maintenance, gas recovery, and laboratory testing, and form a high-quality, standardized SF6 carbon emission thematic dataset through key field extraction, unit unification and logical verification, providing a reliable data foundation for full-process analysis.

[0023] 2. Multi-dimensional intelligent clustering and feature recognition model based on machine learning: Using unsupervised machine learning algorithms (such as K-Means clustering), combined with feature engineering and model optimization methods, the model automatically groups and profiles the SF6 emission behavior of different regions, voltage levels and operation types, objectively identifies management strengths and weaknesses, realizes the transformation from massive data to actionable insights, and supports the formulation of differentiated control strategies.

[0024] 3. A visual monitoring platform that supports multi-dimensional drilling and real-time early warning: It provides a visual monitoring interface that integrates geographic information, equipment level and management dimensions, and supports dynamic display and multi-dimensional drilling analysis of key indicators such as SF6 emissions, recovery rate and unit intensity, thereby improving the intuitiveness of monitoring and the timeliness of decision-making.

[0025] 4. "Data-Accounting-Insight-Strategy" Closed-Loop Intelligent Management Mechanism: It connects the entire process from multi-source data access, cleaning and fusion, dynamic accounting, intelligent analysis to strategy generation, forming a data-driven continuous optimization closed loop, realizing end-to-end intelligent support for SF6 gas emission management, and significantly improving management efficiency and the scientific nature of emission reduction actions. Attached Figure Description

[0026] Figure 1 This is an analysis flowchart of the method in an embodiment of the present invention; Figure 2 This is a data acquisition architecture diagram in an embodiment of the present invention; Figure 3 This is a diagram illustrating the SF6 carbon emission calculation and analysis architecture in an embodiment of the present invention. Figure 4 This is a schematic diagram illustrating the working principle of the method in an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] This invention provides a method for monitoring and analyzing the equivalent carbon emissions of sulfur hexafluoride (SF6) based on power big data and cluster analysis. The following is in conjunction with... Figure 1 The implementation process of this method is described in detail, and the method includes the following steps: Step 01: Acquisition and Fusion Processing of Multi-Source SF6 Gas Data This step collects raw SF6 gas business data from multiple data sources, including the State Grid Anhui Electric Power Research Institute's Power Equipment Gas Medium Laboratory Information System. Through data cleaning, key field extraction, and multi-source data fusion, a standardized and reliable SF6 gas operation dataset is constructed to provide data support for subsequent accurate carbon emission accounting and multi-dimensional analysis.

[0029] Specifically, it includes the following sub-steps: Data Acquisition: Obtain raw SF6 gas operation data from data sources such as the Power Equipment Gas Medium Laboratory Information System via secure file transfer. The data should include at least the following key fields: city name, work category (e.g., gas replenishment, maintenance, decommissioning), equipment voltage level (e.g., 110kV, 220kV, 500kV), total equipment gas supply (kg), total actual recovered gas supply (kg), total unrecovered gas supply (kg), SF6 gas mass before purification (kg), SF6 gas mass after purification (kg), SF6 global warming potential (GWP), and SF6 equivalent CO2 emissions (tons). Example data is shown in the table below: Data cleaning and standardization: Perform integrity verification and logical consistency checks on the collected raw data, identify and handle missing values, outliers and duplicate records; standardize the units and formats of key numerical fields to ensure data quality and consistency.

[0030] Data fusion and integration: The cleaned data is associated and integrated according to key fields such as city name, voltage level, and job category; a standardized dataset including equipment operation, gas recovery and carbon emission potential information is constructed to support subsequent SF6 carbon emission accounting and cluster analysis.

[0031]

[0032] Step 02: Calculation of SF6 gas loss and equivalent carbon emissions This step, based on the standardized dataset generated in step 01, calculates the net loss of SF6 gas under different regions, work categories, and voltage levels by establishing a gas balance accounting model, and scientifically calculates the equivalent carbon dioxide emissions based on the internationally recognized Global Warming Potential (GWP) coefficient.

[0033] like Figure 2 As shown, the specific implementation process includes the following steps: Multi-dimensional gas loss calculation: Taking each gas operation of the equipment as the basic unit, based on the "Total Equipment Gas Charge" and "Total Actual Recovered Gas" fields in the operation record, the net SF6 gas loss for that operation is calculated using the following formula:

[0034] Based on this, according to business analysis needs, the loss of all operation records in the region is grouped, aggregated and summarized by key dimensions such as city name, work category (e.g., gas replenishment, recovery, dismantling, maintenance) and equipment voltage level (e.g., 110kV, 220kV, 500kV, etc.), forming SF6 gas loss distribution tables under different dimension combinations, accurately identifying high-loss links and key equipment types.

[0035] Accurate calculation of equivalent carbon emissions: Using the standard value of SF6 global warming potential (GWP) as determined by the Intergovernmental Panel on Climate Change (IPCC) assessment report, the SF6 losses calculated for each group are converted into equivalent CO2 emissions. The conversion formula is as follows:

[0036] The output provides a clear and standardized dataset of carbon emission results, which can serve as a reliable basis for subsequent emission reduction benefit assessment, difference analysis and low-carbon strategy optimization.

[0037] Step 03: Feature Mining and Differentiated Strategy Generation for SF6 Gas Management Based on Multi-Dimensional Cluster Analysis This step performs unsupervised machine learning cluster analysis on the multi-dimensional equivalent carbon emission data output from step 02, revealing the characteristic differences in SF6 gas management across different regions, voltage levels, and job categories, identifying best practice groups and groups needing improvement, and providing a basis for formulating differentiated low-carbon management strategies.

[0038] like Figure 3 As shown, the specific steps include the following: 3.1 Clustering Feature Engineering: (1) Using various cities as the basic analysis sample; (2) Select the following feature indicators to construct the feature vector: Carbon emission intensity: Equivalent CO2 emissions / Number of devices (tons / unit) Gas recovery rate: (Actual total recovered gas volume / Total gas volume used) × 100% Unit inflation emission: Equivalent CO2 emission / Total inflation volume (tons / kg) Maintenance operation emissions percentage: Maintenance category emissions / Total emissions × 100% (3) The Z-score standardization method is used to process the feature data to eliminate the influence of the units.

[0039] 3.2 Clustering Model Construction: The K-means clustering algorithm was selected; the optimal number of clusters K was determined by combining the silhouette coefficient with the elbow method; the standardized feature matrix was input into the model for training, and the cluster centers were iteratively optimized until convergence or the maximum number of iterations was reached.

[0040] 3.3 Clustering Results Analysis: (1) Calculate the mean and standard deviation of each cluster feature index to generate a cluster feature profile, for example: High-efficiency, low-carbon clusters: high recovery rate (>95%), low emission intensity; High-emission improved cluster: low recovery rate, high emissions per unit of inflated gas; Maintenance-sensitive clusters: high proportion of emissions from maintenance operations; (2) Use visualization tools such as scatter plots to display the clustering results and assist in interpreting the results.

[0041] 3.4 Differentiation Strategy Generation: Summarize best practices for the "high-efficiency and low-carbon cluster"; develop precise improvement measures for the "high-emission improvement cluster" and "maintenance-sensitive cluster," such as upgrading recycling devices, strengthening sealing inspections, and conducting specialized training; and provide data support for power grid investment decisions, equipment upgrades, and technology pilot projects based on clustering results.

[0042] This embodiment achieves accurate monitoring, multi-dimensional analysis, and management strategy optimization of SF6 gas equivalent carbon emissions through the above steps, thereby improving the intelligence and refinement of gas environment management in power grid companies.

[0043] like Figure 4 As shown, the working principle of the method in this embodiment is as follows: (1) System architecture and data flow principles: As shown in the architecture diagram, the system adopts a layered design, with its workflow beginning at the data layer. Multi-dimensional and multi-modal raw data on SF6 gas throughout its entire lifecycle (installation, maintenance, replenishment, decommissioning, and recovery / purification) is extracted from heterogeneous data sources such as the Power Production Management Information System (GIS), Equipment Asset Management System (PMS), and SF6 laboratories. This data is then processed to ultimately form a high-quality, time-series "SF6 Gas Wide Table," providing a consistent and reliable data foundation for upper-level analysis.

[0044] (2) Principle of carbon emission accounting engine: The accounting layer is the computational core of the system, and its principle is based on the law of conservation of mass and the IPCC international standardized algorithm. This engine takes the standard fact table output from the data layer as input, and its core algorithm first calculates the net emissions for each gas operation (…). ):

[0045] in, For the mass of the gas being filled, To recover gas mass, the engine then calls upon the GWP (Global Warming Potential), based on the value set in the latest IPCC assessment report (currently 25200), and efficiently completes batch calculations of equivalent carbon emissions using a parallel computing framework.

[0046] The principle of this process is to transform the amount of gas handled into measurable, reportable, and verifiable (MRV) environmental impact indicators through a deterministic and internationally recognized mathematical model, thereby ensuring the scientific validity, audit traceability, and international comparability of the accounting results.

[0047] (3) Principle of intelligent clustering analysis engine: The analysis layer is the core of this invention for achieving intelligent insights, and its working principle is based on unsupervised machine learning. This engine receives a multi-dimensional carbon emission result set output from the accounting layer, and first performs feature engineering. The mathematical principle behind this is to construct a set of indicator feature vectors that can quantitatively evaluate management effectiveness.

[0048] After Z-score normalization, the feature vectors are fed into the K-Means clustering algorithm. The mathematical principle of this algorithm is to iteratively optimize and find K cluster centers that minimize the sum of squared Euclidean distances from all data points to their respective cluster centers (i.e., minimizing intra-cluster variance). The optimal number of clusters K is determined by a combination of the silhouette coefficient (which measures intra-cluster compactness and inter-cluster separation) and the elbow rule (which plots the inflection point of the sum of squared errors as K changes). Through clustering, the system can objectively classify managed entities into typical groups such as "high-efficiency low-carbon clusters" and "high-emission improvement clusters" without prior labels, thus achieving data-driven management pattern discovery.

[0049] (4) Principles of strategy generation and visual feedback: The system's ultimate working principle is reflected in the formation of a closed-loop management system. The application-layer visualization platform (such as a BI dashboard) presents the analysis results intuitively, allowing users to drill down, slice, and rotate data from multiple dimensions such as region, time, voltage level, and equipment type, enabling rapid location of emission hotspots.

[0050] Based on the clustering results, a feature profile is created for each cluster. By comparing the differences in the mean features between the "best practice cluster" and the "clusters to be improved," diagnostic conclusions and improvement suggestions are generated. These strategies are fed back to the operations and maintenance management end to guide on-site operations, thereby generating new data and forming a continuously optimized and self-evolving intelligent management closed loop.

[0051] This embodiment can achieve the following effects: 1. Comprehensively improve data governance and accounting accuracy. By establishing a standardized multi-source data cleaning and fusion process, the core pain points of scattered, heterogeneous, and inconsistent quality of SF6 gas business data have been fundamentally solved, ensuring the reliability of data at the source of carbon emission accounting.

[0052] Based on the gas balance model and the automated accounting mechanism of the IPCC international standard, the error caused by human intervention is significantly reduced, making the carbon emission calculation process traceable and verifiable, and the calculation results have a high degree of scientific rigor and authority, meeting the stringent requirements of carbon accounting, carbon disclosure and carbon trading.

[0053] 2. Achieve refined monitoring and analysis with multi-dimensional insights. This method breaks through the limitations of traditional methods that only focus on total emissions statistics, and can perform penetrating analysis of emissions from multiple key dimensions such as geographical region, voltage level, work category, and equipment type.

[0054] Through cluster analysis, typical groups such as "high efficiency and low carbon" and "high emission sensitive" are automatically and objectively identified, and weak links in management and high emission hotspots are accurately located. This has achieved a major shift from "macro total quantity control" to "micro precise policy implementation", greatly improving the pertinence and effectiveness of management measures.

[0055] 3. Enhance environmental compliance capabilities and green development image This technical solution helps companies quickly and accurately establish SF6 carbon emission inventories and reports that comply with international standards and national policies, significantly improving the efficiency and compliance of environmental information disclosure and effectively responding to increasingly stringent environmental regulations.

[0056] By showcasing advanced carbon emission monitoring and management capabilities, companies can actively cultivate a responsible and transparent green brand image, enhance public and stakeholder trust, and create favorable conditions for sustainable development in the context of "dual carbon".

[0057] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. 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. Such 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 analyzing the equivalent carbon emissions of SF6 gas based on cluster analysis, characterized in that, Includes the following steps: S01. Collect raw SF6 gas operation data from systems such as the Power Equipment Oil and Gas Medium Laboratory of State Grid Anhui Electric Power Research Institute, and process it through cleaning, standardization and multi-source fusion. S02. Based on the standardized data generated in step S01, calculate the multi-dimensional net loss of SF6 gas using a gas accounting model, and convert it into equivalent carbon emissions using the GWP value; S03. By combining SF6 gas equivalent carbon emission data from different regions, work categories, and voltage levels, a clustering model is constructed to analyze and uncover the common characteristics and differences in SF6 gas management across different regions and types of equipment, providing a scientific basis for formulating more precise management strategies.

2. The method for analyzing equivalent carbon emissions of SF6 gas based on cluster analysis according to claim 1, characterized in that, The specific steps of S01 are as follows: Data Acquisition: Obtain raw data on SF6 gas operations; the data should include at least the following key fields: city name, work category, equipment voltage level, total equipment charge volume, total actual recovered gas volume, total unrecovered gas volume, SF6 gas mass before purification, SF6 gas mass after purification, SF6 global warming potential, and SF6 equivalent CO2 emissions. Data cleaning and standardization: Perform integrity verification and logical consistency checks on the collected raw data, identify and handle missing values, outliers and duplicate records; standardize the units and formats of key numerical fields to ensure data quality and consistency; Data fusion and integration: The cleaned data is associated and integrated according to key fields such as city name, voltage level, and job category; a standardized dataset including equipment operation, gas recovery, and carbon emission potential information is constructed.

3. The method for analyzing equivalent carbon emissions of SF6 gas based on cluster analysis according to claim 1, characterized in that, Step S02 specifically involves: Multi-dimensional gas loss calculation: Taking each gas operation of the equipment as the basic unit, based on the "Total Equipment Gas Charge" and "Total Actual Recovered Gas" fields in the operation record, the net SF6 gas loss for that operation is calculated using the following formula: Based on this, the loss of all operation records in the region is grouped, aggregated and summarized by city name, work category and equipment voltage level as key dimensions to form SF6 gas loss distribution table under different dimension combinations, accurately identifying high loss links and key equipment types. Accurate calculation of equivalent carbon emissions: Using the global warming potential standard value of SF6, the calculated SF6 losses for each group are converted into equivalent CO2 emissions; the conversion formula is as follows: Output the carbon emission results dataset.

4. The method for analyzing equivalent carbon emissions of SF6 gas based on cluster analysis according to any one of claims 1 to 3, characterized in that, Step S03 specifically involves: 3.1 Clustering Feature Engineering: (1) Using various cities as the basic analysis sample; (2) Select the following feature indicators to construct the feature vector: Carbon emission intensity: Equivalent CO2 emissions / number of devices; Gas recovery rate: (Actual total recovered gas volume / Total gas volume used) × 100%; Unit inflation emission: Equivalent CO2 emission / Total inflation volume; Maintenance operation emissions percentage: Maintenance category emissions / Total emissions × 100%; (3) The Z-score standardization method is used to process the feature data to eliminate the influence of the units and obtain the standardized feature matrix; 3.2 Clustering Model Construction: The K-means clustering algorithm was selected; the optimal number of clusters K was determined by combining the silhouette coefficient with the elbow rule; the standardized feature matrix was input into the model for training, and the cluster centers were iteratively optimized until convergence or the maximum number of iterations was reached. 3.3 Clustering Results Analysis: (1) Calculate the mean and standard deviation of each cluster feature index to generate a cluster feature profile: High-efficiency, low-carbon clusters: high recovery rate (>95%), low emission intensity; High-emission improved cluster: low recovery rate, high emissions per unit of inflated gas; Maintenance-sensitive clusters: high proportion of emissions from maintenance operations; (2) Use visualization tools to display the clustering results; 3.4 Differentiation Strategy Generation: Summarize best practices for high-efficiency, low-carbon clusters; and develop precise improvement measures for high-emission improvement clusters and maintenance-sensitive clusters.

5. A cluster analysis-based SF6 gas equivalent carbon emission analysis system, characterized in that, Includes the following steps: Data acquisition and processing module: Collects raw SF6 gas operation data from systems such as the Power Equipment Oil and Gas Medium Laboratory of State Grid Anhui Electric Power Research Institute, and performs cleaning, standardization and multi-source fusion processing; Gas accounting module: Based on standardized data, it calculates the net SF6 gas loss from multiple dimensions using a gas accounting model, and converts the GWP value into equivalent carbon emissions; Clustering analysis module: By combining SF6 gas equivalent carbon emission data from different regions, work categories, and voltage levels, a clustering model is constructed to analyze and uncover the common characteristics and differences in SF6 gas management across different regions and types of equipment, providing a scientific basis for formulating more precise management strategies.

6. The SF6 gas equivalent carbon emission analysis system based on cluster analysis according to claim 5, characterized in that, The specific execution process of the data acquisition and processing module in the aforementioned steps is as follows: Data Acquisition: Obtain raw data on SF6 gas operations; the data should include at least the following key fields: city name, work category, equipment voltage level, total equipment charge volume, total actual recovered gas volume, total unrecovered gas volume, SF6 gas mass before purification, SF6 gas mass after purification, SF6 global warming potential, and SF6 equivalent CO2 emissions. Data cleaning and standardization: Perform integrity verification and logical consistency checks on the collected raw data, identify and handle missing values, outliers and duplicate records; standardize the units and formats of key numerical fields to ensure data quality and consistency; Data fusion and integration: The cleaned data is associated and integrated according to key fields such as city name, voltage level, and job category; a standardized dataset including equipment operation, gas recovery, and carbon emission potential information is constructed.

7. The SF6 gas equivalent carbon emission analysis system based on cluster analysis according to claim 5, characterized in that, The specific execution process of the gas accounting module is as follows: Multi-dimensional gas loss calculation: Taking each gas operation of the equipment as the basic unit, based on the "Total Equipment Gas Charge" and "Total Actual Recovered Gas" fields in the operation record, the net SF6 gas loss for that operation is calculated using the following formula: Based on this, the loss of all operation records in the region is grouped, aggregated and summarized by city name, work category and equipment voltage level as key dimensions to form SF6 gas loss distribution table under different dimension combinations, accurately identifying high loss links and key equipment types. Accurate calculation of equivalent carbon emissions: Using the global warming potential standard value of SF6, the calculated SF6 losses for each group are converted into equivalent CO2 emissions; the conversion formula is as follows: Output the carbon emission results dataset.

8. The SF6 gas equivalent carbon emission analysis system based on cluster analysis according to any one of claims 5 to 7, characterized in that, The cluster analysis module performs the following process: 3.1 Clustering Feature Engineering: (1) Using various cities as the basic analysis sample; (2) Select the following feature indicators to construct the feature vector: Carbon emission intensity: Equivalent CO2 emissions / number of devices; Gas recovery rate: (Actual total recovered gas volume / Total gas volume used) × 100%; Unit inflation emission: Equivalent CO2 emission / Total inflation volume; Maintenance operation emissions percentage: Maintenance category emissions / Total emissions × 100%; (3) The Z-score standardization method is used to process the feature data to eliminate the influence of the units and obtain the standardized feature matrix; 3.2 Clustering Model Construction: The K-means clustering algorithm was selected; the optimal number of clusters K was determined by combining the silhouette coefficient with the elbow rule; the standardized feature matrix was input into the model for training, and the cluster centers were iteratively optimized until convergence or the maximum number of iterations was reached. 3.3 Clustering Results Analysis: (1) Calculate the mean and standard deviation of each cluster feature index to generate a cluster feature profile: High-efficiency, low-carbon clusters: high recovery rate (>95%), low emission intensity; High-emission improved cluster: low recovery rate, high emissions per unit of inflated gas; Maintenance-sensitive clusters: high proportion of emissions from maintenance operations; (2) Use visualization tools to display the clustering results; 3.4 Differentiation Strategy Generation: Summarize best practices for high-efficiency, low-carbon clusters; and develop precise improvement measures for high-emission improvement clusters and maintenance-sensitive clusters.

9. A cluster analysis-based platform for analyzing the equivalent carbon emissions of SF6 gas, characterized in that, It includes a data layer, a carbon emission accounting layer, an intelligent clustering analysis layer, and an application layer; the data layer, carbon emission accounting layer, and intelligent clustering analysis layer are respectively used to execute steps S01, S02, and S03 in any one of the methods described in claims 1 to 4; the application layer is used to generate strategies and visualize them based on the clustering analysis results.

Images