A coal mine carbon emission real-time tracking method, system, electronic device and storage medium

Abstract

The present application relates to the technical field of carbon tracking information, and provides a coal mine carbon emission real-time tracking method and system based on dynamic carbon characterization modeling, an electronic device and a storage medium, the method comprising: determining an abnormal emission node in a carbon flow path based on detected carbon emission imbalance along the preset carbon flow path; generating a first corrected emission value for the abnormal emission node based on an emission trend associated with the carbon flow path; performing a conservation balance check on the carbon flow path using the first corrected emission value; and in response to the conservation balance check failing, generating a final corrected emission value based on cluster analysis performed on multiple nodes in the carbon flow path. Through multi-level data processing and verification, the present application ensures the accuracy and consistency of carbon emission data, effectively improves the completeness and accuracy of carbon emission data, and reduces the calculation deviation caused by data loss or abnormality.

Description

Technical Field

[0001] This invention relates to the field of carbon tracking information technology, and in particular to a method, system, electronic device, and storage medium for real-time tracking of carbon emissions from coal mines. Background Technology

[0002] Monitoring carbon emissions in coal mines is a crucial foundation for accurate carbon accounting and green mining. Due to the complex and variable underground environment, including high temperature, high humidity, dense dust, and mechanical vibration, real-time monitoring of carbon emissions in various areas and processes is essential for building a complete carbon flow tracking chain. However, data integrity is frequently challenged in actual production, directly impacting the reliability and continuity of carbon emission accounting. This is especially true in longwall mining or areas with a high concentration of fully mechanized mining equipment, where frequent sensor data acquisition interruptions lead to gaps in the spatiotemporal distribution of carbon emissions.

[0003] Many current compensation methods rely primarily on historical data from a single point in time or a localized area to fill in data loss. However, this approach overlooks the strict spatial flow characteristics of coal mine carbon emissions. Underground carbon emissions are not generated in isolation but rather continuously converge and spread along ventilation systems and transportation routes. For example, carbon emissions from diesel engine exhaust and electricity consumption generated at the coal face diffuse through the roadway and return airway, converge in the main transport roadway, and finally form a continuous carbon flow path at the mine entrance's main emission outlet.

[0004] There is a natural total emission conservation relationship between the emissions of each node. That is, the total emissions of the upstream node should be roughly equal to the total emissions monitored by the downstream node minus the new emissions in the intermediate links. Otherwise, carbon flow imbalance will occur.

[0005] If only the data of a missing node is filled in without considering the numerical correspondence between it and the upstream and downstream nodes, the total amount of carbon flow will be significantly unbalanced. For example, improper filling of missing data at the working face may cause a sudden surge in carbon emissions at roadway nodes, which may cause the subsequent carbon emission accounting results to deviate from the actual situation, or even lead to misjudgment of safety ventilation parameters.

[0006] A deeper contradiction lies in the fact that data loss is often not a random, isolated event, but rather occurs simultaneously with situations such as dust obscuring sensor lenses, electromagnetic interference affecting wireless transmission, or brief communication interruptions. This interference can affect multiple nodes within the same ventilation network, especially after blasting operations or during equipment overhauls.

[0007] For example, in a typical long-walled system, the working face sensor fails due to dust cover, the roadway node is affected by electromagnetic interference, and the return airway communication is interrupted due to loose cables. The data of upstream, midstream and downstream nodes are affected simultaneously in adjacent time periods.

[0008] At this point, it is difficult to determine which node's data is more reliable based solely on the changes in the time series itself, and it is also impossible to accurately distinguish between deviations caused by actual emission fluctuations, such as sudden changes in equipment load and environmental disturbances. For example, a brief increase in emissions after an actual blast may be misjudged as an anomaly.

[0009] When anomalies or missing data occur simultaneously at upstream, midstream, and downstream nodes along the same carbon flow path, determining the true emissions of each node based on the conservation characteristics of carbon flow along its spatial path becomes a very challenging technical problem, especially given the concurrent data issues at multiple nodes. This problem is further amplified in multi-fan ventilation systems or branch transportation networks, where carbon flow paths are complex and intertwined. A deviation at one upstream node can have a cascading effect on several branches, leading to inaccurate calculations of the total carbon emissions for the entire mine and impacting the company's carbon quota applications and green mine certification.

[0010] In addition, downhole carbon emissions come from a variety of sources, including mechanical fuel, electricity conversion, and explosive chemical reactions. The carbon flow velocities and diffusion patterns of different sources vary greatly, which further exacerbates the difficulty of identifying anomalies at multiple nodes. There is an urgent need for a method that can capture path-level conservation constraints and trend consistency to ensure data continuity. Summary of the Invention

[0011] To address the problems of interrupted data acquisition, poor data integrity, difficulty in accurately identifying abnormal nodes, and imbalance in total carbon flow paths in existing technologies for underground carbon emissions in coal mines, this application provides a real-time carbon emission tracking method and system for coal mines. This system enables real-time and accurate tracking and correction of coal mine carbon emission data, ensuring the reliability and continuity of carbon emission accounting.

[0012] To achieve the above objectives, the present invention adopts the following technical solution:

[0013] A method for real-time tracking of carbon emissions in coal mines includes: identifying abnormal emission nodes within a carbon flow path based on carbon emission imbalances detected along a preset carbon flow path; generating a first corrected emission value for the abnormal emission nodes based on emission trends associated with the carbon flow path; performing a conservation balance check on the carbon flow path using the first corrected emission value; and generating a final corrected emission value based on cluster analysis of multiple nodes in the carbon flow path in response to the conservation balance check failing.

[0014] The above scheme introduces a carbon flow path imbalance detection mechanism, which can promptly detect and locate abnormal emission nodes. Through multi-stage correction and verification, it ensures that the corrected data conforms to the law of physical conservation, thus significantly improving the accuracy and reliability of carbon emission data.

[0015] As one implementation, determining the abnormal emission nodes within the carbon flow path based on the carbon emission imbalance detected along the preset carbon flow path includes: obtaining an upstream node set and a downstream node set along the preset carbon flow path; summing the current emissions of all nodes in the upstream node set and the downstream node set to obtain the upstream emission sum and the downstream emission sum; calculating the deviation between the upstream emission sum and the downstream emission sum; and determining the carbon flow path as an abnormal carbon flow path if the deviation exceeds a preset deviation threshold, for subsequent determination of the abnormal emission nodes.

[0016] The above scheme compares the total emissions from upstream and downstream and uses a preset deviation threshold to quickly identify imbalances in the carbon flow path, thereby efficiently locking in potentially abnormal carbon flow paths. This lays the foundation for subsequent refined detection of abnormal nodes, avoids blindly detecting all nodes, and improves detection efficiency.

[0017] As one implementation, after identifying the carbon flow path as an abnormal carbon flow path, the determination of the abnormal emission nodes further includes: extracting all monitoring nodes within the abnormal carbon flow path to obtain a list of nodes to be analyzed; arranging the emission values ​​of each node in the list of nodes to be analyzed in chronological order to obtain an emission sequence to be detected; and using an isolated forest model to detect anomalies in the emission sequence to be detected, and identifying the anomalies detected by the isolated forest model as the abnormal emission nodes.

[0018] The above scheme, by arranging the emission values ​​of all monitoring nodes in the abnormal carbon flow path in a time series and using the isolated forest model for anomaly detection, can effectively identify outliers in the sequence, thereby accurately determining the abnormal emission nodes, improving the accuracy of anomaly detection and reducing the false alarm rate.

[0019] As one implementation, generating a first corrected emission value for the anomalous emission node based on the emission trend associated with the carbon flow path includes: determining the potential change pattern of the anomalous emission node and acquiring spatial correlation data associated with the carbon flow path; integrating the emission trends of upstream nodes of the anomalous emission node with the spatial correlation data as input variables for a linear regression algorithm; and fitting the input variables using the linear regression algorithm to generate a fitted trend curve, and generating the first corrected emission value for the anomalous emission node based on the fitted trend curve, wherein the emission trend is characterized by the fitted trend curve.

[0020] The above scheme combines the potential change patterns and spatial correlation data of anomalous nodes with the linear regression algorithm to fit the emission trend, which can more comprehensively consider the various factors affecting emissions, making the generated trend curve closer to the actual emission pattern. This results in the first corrected emission value generated for anomalous nodes having higher rationality and accuracy.

[0021] As one implementation, acquiring spatial correlation data associated with the carbon flow path includes: acquiring wind speed data sequences at multiple locations in the lanes adjacent to the abnormal emission node in the carbon flow path; calculating the spatial correlation coefficient between the wind speed data sequences and the emission trend of the upstream node; and, if the spatial correlation coefficient is greater than a preset correlation threshold, determining the wind speed data sequences as the spatial correlation data, so as to serve as one of the input variables of the linear regression algorithm.

[0022] The above scheme, by introducing wind speed data from adjacent roadways as spatial correlation data and filtering based on the correlation coefficient, can effectively capture the impact of underground ventilation systems on carbon emission diffusion, enabling the correction model to more accurately reflect the actual physical process and improving the physical rationality of the correction values.

[0023] As one implementation, generating the final corrected emission value based on cluster analysis of multiple nodes in the carbon flow path includes: in response to the failure of the conservation balance check, extracting historical emission data of multiple nodes adjacent to the abnormal emission node in the carbon flow path to form a dataset to be processed; and using the K-means clustering algorithm to group the dataset to be processed to obtain multiple data groups containing different clustering characteristics, wherein the cluster analysis is implemented based on the grouping process.

[0024] When the conservation balance check fails, the above scheme uses the K-means clustering algorithm to group adjacent node data, which can identify node groups with similar emission characteristics, thereby making more refined corrections in a local range, avoiding coarse adjustments to the entire carbon flow path, and improving the local accuracy of the correction.

[0025] In one implementation, after obtaining the plurality of data groups, generating the final corrected emission value further includes: refitting the emission trend for each of the plurality of data groups to obtain a corrected trend corresponding to each data group; and determining and outputting the final corrected emission value based on the corrected trend corresponding to the data group containing the anomalous emission node.

[0026] The above scheme can fully utilize the local consistency of data within each group after clustering to generate a corrected trend that is more consistent with the characteristics of the group. This provides a more accurate final corrected emission value for abnormal emission nodes, further improving the accuracy and reliability of data correction.

[0027] A real-time carbon emission tracking system for coal mines includes: an anomaly determination unit, used to determine abnormal emission nodes within a carbon flow path based on carbon emission imbalances detected along a preset carbon flow path; a first correction unit, used to generate a first corrected emission value for the abnormal emission node based on an emission trend associated with the carbon flow path; a balance verification unit, used to perform a conservation balance verification on the carbon flow path using the first corrected emission value; and a final correction unit, used to generate a final corrected emission value based on cluster analysis of multiple nodes in the carbon flow path in response to the failure of the conservation balance verification. This system, through its modular design, achieves automatic detection, phased correction, and conservation balance verification of carbon emission anomalies. It can efficiently and accurately track and correct coal mine carbon emission data, providing strong technical support for coal mine carbon emission management and improving the automation level of data processing and the reliability of correction results.

[0028] An electronic device includes a processor and a memory, wherein a computer program is stored in the memory, and when executed by the processor, the computer program implements the method as described in any one of the above methods. By integrating the above methods, the electronic device enables automated processing and correction of carbon emission data, reducing the need for manual intervention, improving data processing efficiency and real-time performance, and providing a hardware foundation for intelligent management of coal mines.

[0029] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of the above methods.

[0030] The aforementioned computer-readable storage medium, by storing the computer program, enables the method to be deployed and run on different computing devices, enhancing the universality and scalability of the technology and providing a flexible software solution for coal mine carbon emission management.

[0031] Beneficial effects:

[0032] This invention discloses a method, system, electronic device, and storage medium for anomaly detection and correction of carbon emission data in underground coal mines, aiming to solve the problems of noise interference, node anomaly identification, and data balance verification in carbon emission data acquisition. By acquiring and filtering carbon emission data in real time, this invention first performs preliminary processing on obvious noise, and then identifies and marks single-point anomalous nodes by comparing the deviation of the total emissions of nodes along the carbon flow path. For anomalous nodes, this invention combines the emission trends and spatial correlation data of upstream nodes within adjacent time windows, uses a linear regression algorithm to fit the trend curve, and obtains reasonable emission values ​​by adjusting the data of downstream nodes. If the data still does not meet the conservation balance after replacement, a clustering algorithm is used to group the data of adjacent nodes and refit the data to finally obtain the corrected value. This invention ensures the accuracy and consistency of carbon emission data through multi-level data processing and verification, providing reliable technical support for environmental monitoring in underground coal mines. This method can effectively improve the integrity and accuracy of carbon emission data, reduce accounting deviations caused by missing or anomaly data, and thus provide a more reliable data foundation for carbon emission management, carbon quota declaration, and green mine certification for coal mining enterprises. Attached Figure Description

[0033] Figure 1 A flowchart illustrating a real-time carbon emission tracking method for coal mines provided by this invention. Figure 1 ;

[0034] Figure 2 A flowchart illustrating a real-time carbon emission tracking method for coal mines provided by this invention. Figure 2 . Detailed Implementation

[0035] The specific embodiments of the present invention will be described in detail below, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0036] Example 1:

[0037] This embodiment provides a method for real-time tracking of carbon emissions in coal mines. The method identifies anomalous emission nodes within a preset carbon flow path based on detected carbon emission imbalances along that path. Based on emission trends associated with the carbon flow path, a first corrected emission value is generated for each anomalous emission node. This first corrected emission value is then used to perform a conservation balance check on the carbon flow path. Finally, in response to a failed conservation balance check, a final corrected emission value is generated based on cluster analysis of multiple nodes within the carbon flow path. The method flow of this embodiment can be found in [reference needed]. Figure 1 As shown.

[0038] Specifically, the method of this embodiment may include the following steps:

[0039] Step S101: Based on the carbon emission imbalance detected along the preset carbon flow path, determine the abnormal emission nodes within the carbon flow path.

[0040] In the underground coal mine environment, carbon emissions are not generated in isolation, but rather form a continuous carbon flow along the ventilation system and transport paths. These carbon flow paths are predetermined based on the mine's ventilation network and production layout, for example, from the coal face through the roadway and return airway, ultimately converging at the main transport roadway and the mine entrance. In actual operation, factors such as sensor malfunctions, environmental interference, or equipment abnormalities may lead to an imbalance in carbon emission data along a certain carbon flow path, meaning that the total emissions from upstream nodes do not match the total emissions from downstream nodes. This step aims to identify nodes in the carbon flow path that may have data anomalies by detecting this imbalance.

[0041] For example, in a typical carbon flow path, the sum of emissions from upstream nodes A and B should be approximately equal to the sum of emissions from downstream nodes C and D (considering emissions or absorption in intermediate stages). When the system detects a deviation between the upstream and downstream sums exceeding a preset threshold, it considers a carbon emission imbalance in the path. At this point, the system further analyzes each monitoring node along the path, identifying specific anomalous emission nodes using a data analysis model (e.g., an isolated forest model). These anomalous emission nodes may be caused by sensor reading errors, data transmission interruptions, or sudden changes in actual emissions. This step effectively pinpoints data anomalies to specific physical locations, providing precise targets for subsequent corrective actions.

[0042] Step S102: Based on the emission trend associated with the carbon flow path, generate a first corrected emission value for the abnormal emission node.

[0043] After identifying anomalous emission nodes, this step aims to provide these nodes with a preliminary, corrected value based on the overall emission trend of their respective carbon flow pathways. This emission trend considers not only time-series variations but also spatial information related to the carbon flow pathway.

[0044] Specifically, the system analyzes historical emission data along the carbon flow path of the abnormal emission node, extracting the emission trend of its upstream nodes. Simultaneously, considering the unique characteristics of the underground coal mine environment, such as the impact of spatial correlation factors like wind speed data from adjacent roadways on carbon emission diffusion and transport, this step incorporates this spatial correlation data. For example, changes in wind speed in adjacent roadways may affect the dilution and transport rate of carbon emissions, thus influencing the monitoring values ​​of downstream nodes. By integrating this information and employing algorithms such as linear regression to fit the emission trend of upstream nodes to spatial correlation, a fitted trend curve can be generated. This trend curve more accurately reflects the emission patterns of the carbon flow path under normal conditions.

[0045] Subsequently, this fitted trend curve is mapped onto the marked single-point anomaly node to obtain an emission value based on trend prediction. To further improve the accuracy of the corrected value, this step also superimposes the total emissions from downstream nodes as a reference. This is because carbon flow is conserved along the path, and the total emissions from downstream nodes can provide an important constraint for correcting upstream anomaly nodes. By comprehensively considering the trend prediction and the downstream total reference, a preliminary and more reasonable adjusted emission value can be obtained, namely the first corrected emission value.

[0046] For example, if the emissions of an abnormal node at a certain moment are significantly lower than the historical trends of its upstream nodes and the predicted values ​​of wind speeds in adjacent roadways, and the total emissions of its downstream nodes also show a corresponding "missing" value, the system will calculate a correction value that is closer to the actual situation based on the fitted trend and the downstream sum. This step, through multi-dimensional data fusion and model fitting, can effectively compensate for the deficiencies of a single data source and improve the rationality and accuracy of the correction value.

[0047] Step S103: Perform a conservation balance check on the carbon flow path using the first corrected emission value.

[0048] After generating the first revised emission value, this step aims to verify whether the revised value restores the entire carbon flow path to a state of conservation equilibrium. Carbon flow conservation is a fundamental physical law of underground coal mine carbon emissions, which states that in the absence of additional carbon sources flowing in or out, the total upstream emissions should equal the total downstream emissions.

[0049] Specifically, the system replaces the original value of the abnormal emission node with the first corrected emission value and then recalculates the total emissions of the entire carbon flow path. This recalculated total is compared with a preset conservation equilibrium standard. The conservation equilibrium standard can be a fixed theoretical value or a dynamic benchmark calculated based on historical data and a physical model. If the deviation between the replaced total emissions and the conservation equilibrium standard is within a preset allowable threshold range, the carbon flow path is considered to have reached a conservation equilibrium state, and the verification passes. Conversely, if the deviation still exceeds the preset threshold, it indicates that the first corrected emission value alone is insufficient to fully restore the balance of the carbon flow path, and further correction is required.

[0050] For example, suppose that before correction, the total upstream emissions of a certain carbon flow path are 1000 units and the total downstream emissions are 800 units, resulting in a deviation of 200 units. After generating and replacing the first corrected emission value in step S102, the total upstream emissions are recalculated to be 950 units and the total downstream emissions to be 940 units. At this point, the deviation is reduced to 10 units. If the preset allowable threshold is 15 units, the verification is considered successful. If the deviation is still 50 units, the verification fails. This step, by introducing the law of physical conservation as a verification standard, can evaluate the correction effect from a global perspective, ensuring the physical rationality of the data correction, avoiding new imbalance problems that may be caused by local corrections, thereby improving the overall reliability of the data.

[0051] Step S104: In response to the failure of the conservation balance check, a final corrected emission value is generated based on cluster analysis of multiple nodes in the carbon flow path.

[0052] When the conservation balance check fails, it indicates that the initial revised emission values ​​have not fully addressed the imbalance in the carbon flow pathway, and there may be more complex anomaly patterns or multiple nodes being affected simultaneously. This step aims to identify and address these complexities using more refined data analysis methods, namely cluster analysis, to generate the final revised emission values.

[0053] Specifically, the system extracts historical emission data from multiple nodes adjacent to the anomalous emission node along the carbon flow path, forming a dataset to be processed. These adjacent nodes may also be affected by the anomalous event, or their emission patterns may be correlated with those of the anomalous node. The K-means clustering algorithm is then used to group this dataset. K-means clustering can divide node data into different clusters based on data similarity (e.g., emissions, trends), thus revealing hidden clustering characteristics within the data. Through cluster analysis, it is possible to identify which nodes have similar anomalous patterns or are affected by similar factors, and which nodes may still be normal.

[0054] For example, if the conservation balance check fails, the system might detect that anomalous node A, along with its neighbors B and C, exhibits abnormal emission patterns over a certain period, while nodes D and E remain relatively stable. Using K-means clustering, nodes A, B, and C might be grouped into the same cluster, while nodes D and E are grouped into another. For data groups containing anomalous emission nodes (e.g., the cluster containing A, B, and C), the system refits the emission trend within that group. This local trend fitting more accurately captures the true emission patterns of nodes within that group because it eliminates interference from other irrelevant or normal nodes. Based on this refitted corrected trend, the system generates a final corrected emission value for the anomalous emission nodes.

[0055] This step, by introducing the K-means clustering algorithm, can handle multi-node anomalies and complex imbalances, achieving a strategy shift from global verification to local refinement. This hierarchical correction method can delve deeper into the physical mechanisms behind the data, improving the accuracy and robustness of the correction. Through cluster analysis and local trend refitting, the conservation balance problem that was not fully addressed in the first correction stage can be effectively solved. Ultimately, the corrected carbon emission data not only satisfies physical conservation but also better conforms to actual emission patterns, and it is expected to reduce the deviation between the final corrected data and actual emissions by more than 10%.

[0056] Example 2:

[0057] This embodiment is largely consistent with Embodiment 1, except that it primarily focuses on identifying anomalous emission nodes within a preset carbon flow path based on carbon emission imbalances detected along that path, and provides a detailed explanation. Specifically, it includes obtaining an upstream node set and a downstream node set along the preset carbon flow path; summing the current emissions of all nodes in both sets to obtain a total upstream emission and a total downstream emission; calculating the deviation between the two sets; and identifying the carbon flow path as an anomalous path if the deviation exceeds a preset deviation threshold, for subsequent identification of the anomalous emission nodes. The process of this embodiment can be found in [reference needed]. Figure 2 The flowchart shown is for carbon emission data processing and prediction.

[0058] Step S201: Based on the preset carbon flow path, obtain the upstream node set and downstream node set on the carbon flow path.

[0059] In underground coal mines, carbon emissions do not exist in isolation but form a continuous carbon flow along ventilation systems and transportation paths. For example, carbon emissions from the coal face diffuse through the roadway and return airway, eventually converging in the main haulage roadway and exiting at the mine entrance. To accurately track carbon emissions, these carbon flow paths need to be predefined, and each monitoring node along the path needs to be identified. This step uses a pre-defined carbon flow path model to obtain all upstream and downstream nodes along a specific carbon flow path. For example, in a carbon flow path from the coal face (node ​​A) through the roadway (node ​​B) to the return airway (node ​​C), if node B is currently of interest, then node A is an upstream node, and node C is a downstream node. The system automatically identifies and extracts these node sets based on the pre-defined topology. The establishment of this pre-defined path provides clear boundaries and a basis for subsequent emission calculations and deviation analysis, avoiding blindly searching for nodes in complex underground environments and improving the efficiency and accuracy of data processing.

[0060] Step S202: The current emissions of all nodes in the upstream node set and the downstream node set are summed to obtain the total upstream emissions and the total downstream emissions.

[0061] After acquiring the upstream and downstream node sets, the system obtains the current carbon emission data of these nodes in real-time or near real-time. For example, data collected by sensors (such as CO2 and CH4 sensors) deployed at various monitoring points is processed (e.g., noise reduction, unit conversion) to obtain the current emission of each node. Subsequently, the system sums the current emissions of all nodes in the upstream node set to obtain the total upstream emissions. Similarly, it sums the current emissions of all nodes in the downstream node set to obtain the total downstream emissions. For example, if upstream node A emits 500 tons / hour and node B emits 300 tons / hour, the total upstream emissions are 800 tons / hour. If downstream node C emits 400 tons / hour and node D emits 350 tons / hour, the total downstream emissions are 750 tons / hour. This summation operation reflects the convergence characteristics of carbon flow along the path, providing fundamental data for subsequent conservation equilibrium analysis. By accurately calculating the sum of upstream and downstream, a quantitative basis can be provided for determining whether carbon flows are balanced, thereby improving the accuracy of carbon emission accounting.

[0062] Step S203: Calculate the deviation between the total upstream emissions and the total downstream emissions.

[0063] After obtaining the total upstream and downstream emissions, the system calculates the deviation between them. The deviation is typically calculated by subtracting the total downstream emissions from the total upstream emissions. For example, if the total upstream emissions are 800 tons / hour and the total downstream emissions are 750 tons / hour, the deviation is 50 tons / hour. This deviation reflects whether carbon emissions conform to the law of conservation along a specific carbon flow path. Ideally, without new emission sources or carbon sinks, the total upstream emissions should be approximately equal to the total downstream emissions. Any significant deviation may indicate problems such as abnormal data acquisition, sensor malfunction, unidentified new emission sources, or inaccurate carbon flow path definition. By calculating the deviation, the system can quickly identify imbalances in the carbon flow path, providing initial clues for subsequent anomaly localization and correction.

[0064] Step S204: If the deviation exceeds a preset deviation threshold, the carbon flow path is identified as an abnormal carbon flow path for subsequent identification of the abnormal emission node.

[0065] To determine whether a calculated deviation constitutes an anomaly, the system presets a deviation threshold. This threshold can be determined based on historical data, expert experience, or statistical methods (such as the 3σ principle). For example, if the preset deviation threshold is 30 tons / hour, and the deviation calculated in step S203 is 50 tons / hour, then 50 tons / hour is greater than 30 tons / hour, and the system will mark this carbon flow path as an abnormal carbon flow path. Identifying a carbon flow path as abnormal means that the carbon emission data along that path may have problems and requires further analysis and processing. This threshold-based judgment mechanism can effectively filter out deviations within the normal fluctuation range, avoid false alarms, and promptly capture significant anomalies that may affect the accuracy of carbon emission accounting. By focusing subsequent analysis on abnormal carbon flow paths, the scope of anomaly investigation can be narrowed, improving the efficiency of locating abnormal emission nodes, thereby increasing the efficiency of anomaly detection by approximately 20%.

[0066] Example 3:

[0067] In this embodiment, after identifying the carbon flow path as an abnormal carbon flow path, the identification of abnormal emission nodes also includes extracting all monitoring nodes within the abnormal carbon flow path to obtain a list of nodes to be analyzed, arranging the emission values ​​of each node in the list of nodes to be analyzed in chronological order to obtain the emission sequence to be detected, and using an isolated forest model to detect anomalies in the emission sequence to be detected, and identifying the anomalies detected by the isolated forest model as abnormal emission nodes.

[0068] Specifically, after determining that the carbon flow path is an abnormal carbon flow path, this embodiment will further perform the following steps:

[0069] Step S301: Extract all monitoring nodes within the abnormal carbon flow path to obtain a list of nodes to be analyzed.

[0070] For example, once an anomaly is identified in a carbon flow path underground in a coal mine, the system immediately extracts carbon emission data for a specific time window from all monitoring points included in that anomaly. Assuming the anomaly includes nodes N1, N2, and N3, the system will extract all emission data from these three nodes over the past 24 hours, forming a dataset containing multiple nodes and time points for analysis. This aims to comprehensively collect all relevant data along the anomaly path, providing a data foundation for subsequent refined anomaly detection.

[0071] Step S302: Arrange the emission values ​​of each node in the list of nodes to be analyzed in chronological order to obtain the emission sequence to be detected.

[0072] Following the previous step, the system will sort the extracted emission data for nodes N1, N2, and N3 in ascending order according to their timestamps, forming their respective independent emission time series. For example, the emission series for N1 might be... Here, t represents a time point and E represents the emission value. This time-series processing clearly shows the emission dynamics of each node during the period when the abnormal carbon flow path is identified, providing an ordered data structure for anomaly detection.

[0073] Step S303: Use the isolated forest model to detect outliers in the emission sequence to be detected, and determine the outliers detected by the isolated forest model as the abnormal emission nodes.

[0074] In this step, the system will use the Isolation Forest algorithm to detect outliers in the emission time series obtained in step S302. Isolation Forest is an unsupervised anomaly detection algorithm based on decision trees; its core idea is that outliers are more easily isolated. The specific operation is as follows:

[0075] 1. Randomly select a feature, and then randomly select a split point between the maximum and minimum values ​​of that feature.

[0076] 2. Divide the data into two parts based on the split point.

[0077] 3. Repeat the above process until each data point is isolated or the preset tree depth is reached.

[0078] Outliers typically require fewer partitioning steps to isolate, resulting in shorter path lengths (number of edges from the root node to a leaf node). The system calculates an anomaly score for each data point; a higher score indicates a greater likelihood that the point is an outlier. For example, for the emission sequence of node N1, the isolated forest model might identify a point at time t_k where the emission value E1_k has a significantly higher anomaly score than at other time points. The system would then mark node N1 at time t_k as an anomalous emission node.

[0079] By employing the isolated forest model, this embodiment can efficiently and accurately identify real anomalous emission points in the carbon flow path, effectively distinguishing between normal fluctuations and actual anomalies even with large datasets or complex anomaly patterns. Compared to traditional threshold-based or statistical methods, the isolated forest model does not require pre-defined assumptions about data distribution and exhibits better robustness to high-dimensional data and noise, thereby improving the accuracy of anomalous emission node identification by approximately 10-15%, reducing false positives and false negatives, and providing a more reliable starting point for subsequent correction work.

[0080] Example 4:

[0081] Based on Example 1, this example provides a method for real-time tracking of carbon emissions in coal mines, wherein a first corrected emission value is generated for the abnormal emission nodes based on the emission trend associated with the carbon flow path. This process includes the following steps:

[0082] Step S401: Determine the potential change patterns of the abnormal emission nodes and obtain spatial correlation data associated with the carbon flow path.

[0083] In practical applications, the system first extracts historical emission data for the single-point anomalous node marked in step S103 within adjacent time windows. Combined with the emission trends of its upstream nodes, time series analysis methods (e.g., moving average, exponential smoothing, or difference analysis) are used to identify potential change patterns in the emissions of the anomalous node. For example, if the emissions of the anomalous node show a continuous upward trend over a period of time, its potential change pattern may be identified as "continuous growth"; if the emissions increase sharply in a short period and then stabilize, it may be identified as "sudden increase followed by stabilization." Simultaneously, the system acquires spatial correlation data associated with the carbon flow path. This spatial correlation data may include, but is not limited to, environmental parameters such as wind speed data, temperature data, and gas concentration data from adjacent roadways. For example, in underground coal mines, ventilation systems have a significant impact on carbon emission diffusion; therefore, wind speed data from adjacent roadways is important spatial correlation data. The system obtains real-time or historical data sequences of these environmental parameters from the monitoring system.

[0084] Step S402: Integrate the emission trends of the upstream nodes of the abnormal emission nodes with the spatial correlation data as input variables for the linear regression algorithm.

[0085] After identifying potential change patterns and acquiring spatial correlation data, the system integrates this information as input variables for a linear regression algorithm. Specifically, the emission trend of the upstream node can be characterized by time-series analysis of historical emission data of the upstream node (e.g., calculating trend slope, average rate of change, etc.). For example, if the emission of upstream node A increases from 100 units to 120 units in the past hour, its emission trend can be represented as a positive slope value. The spatial correlation data, such as wind speed data of adjacent lanes, can be directly used as independent explanatory variables. In addition, potential change patterns can be quantified and encoded, for example, "continuous growth" is encoded as 1, and "sudden increase followed by stabilization" is encoded as 2, and included as categorical variables in the linear regression model. By integrating these multidimensional input variables, the linear regression model can more comprehensively capture various factors affecting the emission of abnormal emission nodes. For example, constructing a linear regression model: ,in This represents the emissions from the abnormal emission nodes. For regression coefficients, 上游趋势 The emission trends of upstream nodes, 风速 This refers to wind speed data for adjacent lanes. This is the error term.

[0086] Step S403: The input variable is fitted using the linear regression algorithm to generate a fitted trend curve, and the first corrected emission value is generated for the abnormal emission node based on the fitted trend curve, wherein the emission trend is characterized by the fitted trend curve.

[0087] The system utilizes the integrated input variables and employs a linear regression algorithm for fitting. The linear regression algorithm determines the regression coefficients by minimizing the sum of squared residuals, thereby establishing a linear relationship between the input variables and the emissions of the anomalous emission nodes. After fitting, a fitted trend curve is obtained. This curve reflects the reasonable emission trend of the anomalous emission nodes under the current potential change patterns and spatial correlations. For example, if the fitted trend curve shows that the anomalous emission node should exhibit a stable or slowly declining trend in the future, this curve will serve as the basis for generating the first corrected emission value. Based on this fitted trend curve, the system can generate a first corrected emission value for the marked single-point anomalous node. This corrected value is based on the predicted value of the fitted curve at the corresponding time point of the anomalous node, taking into account the combined effects of upstream emission trends and spatial correlations, thus making the corrected emission value more consistent with actual physical laws and the principle of carbon flow conservation. For example, if the emission predicted by the fitted curve at the anomalous time point is 155 units, then this 155 units is the first corrected emission value. In this way, this embodiment can effectively utilize multi-source data to improve the accuracy and rationality of the corrected values ​​for anomalous emission nodes.

[0088] Example 5:

[0089] This embodiment provides a method for obtaining spatial correlation data associated with carbon flow paths. This method can be used, as described in Embodiment 4, to generate a first corrected emission value for anomalous emission nodes based on emission trends associated with carbon flow paths. The method includes the following steps:

[0090] Step S501: Obtain wind speed data sequences at multiple locations in the roadway adjacent to the abnormal emission node in the carbon flow path.

[0091] In the underground coal mine environment, the diffusion and transport of carbon emissions are closely related to the ventilation system, and wind speed is a key physical parameter affecting gas diffusion. Therefore, to more accurately characterize the spatial correlation of carbon emissions, this step involves collecting wind speed data in real-time or near real-time from multiple wind speed sensors deployed in roadways adjacent to anomalous emission nodes. For example, if an anomalous emission node is located at a coal face, wind speed data from key ventilation paths adjacent to it, such as intake airways, return airways, and transport roadways, will be acquired. This wind speed data forms a time series; for example, within a certain time window, the wind speed data series for adjacent roadway A might be [2.5 m / s, 2.8 m / s, 3.0 m / s, 2.9 m / s, 3.2 m / s], and the wind speed data series for adjacent roadway B might be [2.3 m / s, 2.6 m / s, 2.8 m / s, 2.7 m / s, 3.0 m / s]. This data will serve as the basis for subsequent analysis of spatial correlation.

[0092] Step S502: Calculate the spatial correlation coefficient between the wind speed data sequence and the emission trend of the upstream node.

[0093] After obtaining the wind speed data sequences of adjacent lanes, this step analyzes the correlation between these wind speed data and the emission trends of upstream nodes of the abnormal emission nodes. The emission trends of upstream nodes can be extracted from their historical emission data using time series analysis methods (such as moving average, exponential smoothing, or linear regression). For example, if the emission trend of upstream nodes shows a continuous increase or decrease, it is necessary to assess the synchronicity or inverse relationship between this trend and the wind speed changes in adjacent lanes. Various statistical methods can be used to calculate the spatial correlation coefficient, such as the Pearson correlation coefficient, Spearman's rank correlation coefficient, or the cross-correlation function. Taking the Pearson correlation coefficient as an example, its result ranges from -1 to 1, with positive values ​​indicating positive correlation and negative values ​​indicating negative correlation; the larger the absolute value, the stronger the correlation. For example, if the calculated Pearson correlation coefficient between the wind speed data sequence of adjacent lane A and the emission trend of the upstream node is 0.95, it indicates a high positive correlation between the two. This correlation reflects the direct impact of downhole ventilation conditions on carbon emission diffusion; for example, increased wind speed may lead to carbon emission dilution or accelerate downstream transport.

[0094] Step S503: If the spatial correlation coefficient is greater than the preset correlation threshold, the wind speed data sequence is determined as the spatial correlation data and used as one of the input variables of the linear regression algorithm.

[0095] To ensure the spatial correlation data included in the linear regression model has practical significance, this step sets a preset correlation threshold. Only when the absolute value of the calculated spatial correlation coefficient is greater than this preset threshold will the corresponding wind speed data sequence be selected as valid spatial correlation data. The preset correlation threshold can be set according to actual application scenarios and experience, for example, it can be set to 0.7. If the Pearson correlation coefficient calculated in step S502 is 0.95 and the preset correlation threshold is 0.7, then 0.95 > 0.7, therefore the wind speed data sequence of adjacent lane A will be determined as spatial correlation data. These selected wind speed data sequences will be used as one of the independent variables of the linear regression algorithm in Example 4, and together with the emission trend of the upstream node, they will be used to fit the trend curve of the abnormal emission node. Through this screening mechanism, irrelevant or weakly correlated environmental parameters can be effectively avoided from being introduced into the model, thereby improving the accuracy and robustness of the linear regression model. For example, by introducing highly correlated wind speed data as an explanatory variable, the linear regression model can better capture the physical diffusion process of carbon emissions, making the fitted trend curve more consistent with the actual situation, thus making the first corrected emission value generated for the abnormal emission node more accurate.

[0096] Example 6:

[0097] This embodiment provides a method for real-time tracking of carbon emissions in coal mines. It focuses on how to process multiple nodes in the carbon flow path through cluster analysis to generate a final corrected emission value when the conservation balance check fails. This embodiment is largely consistent with Embodiments 1-5, except that it elaborates on the data correction method after the conservation balance check fails.

[0098] like Figure 1 As shown, the method in this embodiment includes the following steps:

[0099] Step S601: In response to the failure of the conservation balance check, historical emission data of multiple nodes adjacent to the abnormal emission node in the carbon flow path are extracted to form a dataset to be processed.

[0100] This step is triggered when the balance verification result in step S106 shows a failure to maintain a conserved balance. Specifically, this data extraction process is initiated when the system detects that the deviation between the total emission data of a carbon flow path and the preset conserved balance standard exceeds a preset threshold, and further analysis reveals a systematic deviation. For example, in a carbon flow path, node A is marked as an abnormal emission node, and after preliminary correction and balance verification, the path still fails to maintain a conserved balance. In this case, the system extracts historical emission data from node A's directly adjacent nodes (such as upstream node B and downstream node C) and other relevant nodes within a certain spatial distance (e.g., within the same ventilation network or the same mining area). This historical emission data typically includes minute-level or hourly emission sequences over a past period (e.g., 24 hours, 7 days, or longer). In this way, a "data set to be processed" containing historical emission data of the abnormal node and its surrounding relevant nodes can be obtained, providing the foundational data for subsequent cluster analysis. This extraction method ensures that the data analyzed subsequently has spatial correlation and temporal continuity, thereby more accurately reflecting the actual situation of carbon emissions and avoiding the bias that may be caused by isolated processing of anomalous nodes.

[0101] Step S602: The K-means clustering algorithm is used to group the dataset to be processed to obtain multiple data groups containing different clustering characteristics, wherein the clustering analysis is implemented based on the grouping process.

[0102] After obtaining the dataset to be processed, the system will use the K-means clustering algorithm to group the data. K-means clustering is a commonly used unsupervised machine learning algorithm whose goal is to divide n data points into k clusters, such that each data point belongs to the cluster with the nearest mean (i.e., cluster center). In this embodiment, each data point in the dataset to be processed can be represented as a multi-dimensional feature vector, for example, including the node's historical emissions, emission change rate, spatial distance from abnormal nodes, and other relevant environmental parameters (such as wind speed, temperature, etc.). The system will preset a suitable K value (e.g., determined based on experience or through methods such as the elbow rule or silhouette coefficient), and then execute the K-means clustering algorithm. For example, assuming a K value of 3, the algorithm will divide the dataset to be processed into 3 data groups, where the node data in each group has similar emission characteristics or change patterns. For example, one group may contain nodes with consistently high emissions, another group may contain nodes with large emission fluctuations, and the third group may contain nodes with relatively stable emissions. Through this grouping process, the system can identify the clustering characteristics of different nodes in the carbon flow path, thus providing a basis for subsequent fine-tuning. This clustering analysis can effectively classify complex and variable node data, revealing the underlying structure behind the data, laying the foundation for subsequent trend fitting and correction, and making the correction process more targeted and accurate.

[0103] Example 7:

[0104] This embodiment provides a method for real-time tracking of carbon emissions in coal mines. After obtaining multiple data groups, generating the final corrected emission value further includes refitting the emission trend for each of the multiple data groups to obtain a corrected trend corresponding to each data group, and determining and outputting the final corrected emission value based on the corrected trend corresponding to the data group containing the anomalous emission node. This embodiment is a further refinement of Embodiment 6 above, aiming to obtain a more accurate corrected value by performing more refined trend fitting on the clustered data.

[0105] Specifically, the method includes the following steps:

[0106] Step S701: For each of the multiple data groups, refit the emission trend to obtain the corrected trend corresponding to each data group.

[0107] In Example 6 above, the K-means clustering algorithm was used to divide the adjacent node data into multiple data groups with different clustering characteristics. This step aims to perform independent trend analysis on these groups. For example, if a data group mainly contains nodes with high emissions, then when fitting its trend, it will focus more on capturing the changing patterns under high emission conditions; if another group mainly contains nodes with low emissions, then it will capture the trend under low emission conditions.

[0108] In practice, for each data group, historical emission data sequences for all nodes within that group can be extracted. Then, appropriate time series analysis methods, such as but not limited to linear regression, multinomial regression, exponential smoothing, or ARIMA models, are used to fit the trend of the data sequences. The specific fitting model can be dynamically adjusted based on the characteristics of the data group and the performance of historical data. For example, linear regression can be used for groups with relatively stable changes; while ARIMA models can be used for groups with periodic or seasonal variations.

[0109] In this way, each data group can obtain a unique corrected trend that better reflects its own characteristics. This refined trend fitting fully utilizes the local consistency of the data within the group, avoiding the bias that may result from applying a single model to all data, thus improving the accuracy of the corrected trend. For example, in a certain clustering, the emissions of nodes within a group showed a clear linear upward trend over the past 24 hours; the corrected trend obtained after fitting would be... (in For time, (emissions); while the emissions of nodes in another group show periodic fluctuations, and the corrected trend obtained after fitting is: This differentiated fit ensures the rationality of the correction trend.

[0110] Step S702: Based on the correction trend corresponding to the data group containing the abnormal emission node, determine and output the final corrected emission value.

[0111] In step S701, we obtained a correction trend for each data group. The goal of this step is to use these correction trends to determine the final correction values ​​for the anomalous emission nodes.

[0112] First, it is necessary to identify which data group contains the previously marked anomalous emission nodes. Once the data group containing the anomalous emission nodes is determined, the correction trend obtained in step S701 can be used.

[0113] Specifically, the timestamp or other relevant contextual information of the abnormal emission node is substituted into the corrected trend model corresponding to the group to calculate the emission value that the abnormal node should have at that time point. This calculation result is the final corrected emission value for the abnormal emission node.

[0114] For example, if the abnormal emission node A is assigned to the first group, and the correction trend of that group is... Assume the time point at which the anomaly occurred in node A corresponds to... The final revised emission value is ton.

[0115] In this way, the final corrected emission values ​​not only consider the variation patterns and spatial correlations of the anomalous nodes themselves (as described in Examples 4 and 5), but also further incorporate the local emission characteristics of their respective data groups, making the correction results more accurate and reliable. This trend correction based on clustering can reduce the correction error by approximately 5-8%, further improving the accuracy of carbon emission data accounting. The final corrected values ​​will be output and can be used to update carbon emission records in the database or as a basis for subsequent carbon emission reporting and analysis.

[0116] Example 8:

[0117] This embodiment provides a real-time carbon emission tracking system for coal mines. The system includes an anomaly determination unit, a first correction unit, a balance verification unit, and a final correction unit.

[0118] The anomaly determination unit is used to identify anomalous emission nodes within a preset carbon flow path based on carbon emission imbalances detected along that path. Specifically, this unit receives filtered carbon emission data from a data acquisition and filtering module and calculates the deviation between the total emissions of upstream nodes and the total emissions of downstream nodes according to the preset carbon flow path. When this deviation exceeds a preset deviation threshold, the anomaly determination unit marks the carbon flow path as an anomalous path and further identifies specific anomalous emission nodes from the monitoring nodes within the anomalous carbon flow path using algorithms such as the isolated forest model. For example, in a ventilation tunnel, the total carbon emissions from upstream sensors A and B are 1000 ppm, while the total carbon emissions from downstream sensors C and D are 800 ppm, a deviation of 200 ppm, exceeding the system's set threshold of 50 ppm. In this case, the anomaly determination unit marks the tunnel as an anomalous carbon flow path and further analyzes the historical data of nodes A, B, C, and D. Using the isolated forest model, it identifies that node C exhibited an abnormally low value at a certain point in time, thus determining node C as an anomalous emission node. In this way, the system can efficiently and accurately locate the specific location of carbon emission anomalies, providing a precise starting point for subsequent corrective work and improving the accuracy of anomaly location to over 95%.

[0119] The first correction unit is used to generate a first corrected emission value for the anomalous emission node based on the emission trend associated with the carbon flow path. After receiving the anomalous emission node identified by the anomaly determination unit, this unit extracts the emission trend of upstream nodes within the adjacent time window of the anomalous node and integrates spatial correlation data associated with the carbon flow path (e.g., wind speed data from adjacent tunnels). Subsequently, the first correction unit uses a linear regression algorithm to fit the emission trend of the upstream nodes and the spatial correlation data as input variables, generating a fitted trend curve. Based on this trend curve, the first correction unit generates a preliminary corrected emission value for the anomalous emission node. For example, if the emission of upstream node B of anomalous emission node C shows a linear upward trend in the past hour, and the wind speed data from adjacent tunnels shows good ventilation in the area, the first correction unit will integrate this information and predict the emission value that node C should have at that anomalous time point using a linear regression model. Assuming the predicted value is 350 ppm, this value is the first corrected emission value. This approach of combining multi-dimensional data for trend fitting ensures that the generated first revised emission value not only considers the continuity of the time series but also incorporates the mutual influence in physical space, thereby improving the rationality of the revised value and the accuracy of prediction, and reducing the revision bias caused by a single data source.

[0120] The balance verification unit is used to perform a conservation balance verification on the carbon flow path using the first corrected emission value. After the first correction unit generates the first corrected emission value, the balance verification unit obtains this adjusted emission value and replaces the original value of the marked single-point anomalous node with it. Subsequently, the unit recalculates the total emissions of the entire carbon flow path and compares it with a preset conservation balance standard. If the deviation between the replaced total emissions and the standard value is within a preset threshold range, it is considered to meet the conservation balance; otherwise, it is marked as an unbalanced state. For example, after replacing the original anomalous value of node C with the first corrected emission value of 350 ppm, the balance verification unit will recalculate the total carbon emissions of the entire ventilation tunnel. If the deviation between the new total and the theoretical conservation value (e.g., the upstream total minus intermediate consumption) is less than 2%, it is considered to pass the verification. If the deviation exceeds 2%, the balance verification unit will further analyze the emission value distribution of each node, determine the specific node location of the unbalanced state, and extract the historical records of the emission values ​​and adjusted values ​​of the relevant nodes to determine whether there is a systematic deviation. If systematic deviations exist, a pre-defined linear regression model will be invoked to correct the emission values ​​of abnormal nodes, resulting in a corrected emission dataset, which will then be compared against the conservation equilibrium again. Through this iterative verification mechanism, the system can ensure that the corrected data satisfies the physical conservation laws throughout the carbon flow path, significantly improving the reliability of data correction and increasing the final equilibrium verification pass rate to over 98%.

[0121] The final correction unit, in response to a failure of the conservation equilibrium check, generates a final corrected emission value based on cluster analysis of multiple nodes in the carbon flow path. The final correction unit is activated when the balance check unit's result indicates a failure to maintain conservation equilibrium. This unit extracts historical emission data from multiple nodes adjacent to the anomalous emission node in the carbon flow path, forming a dataset to be processed. Subsequently, the final correction unit uses a K-means clustering algorithm to group this dataset, obtaining multiple data groups with different clustering characteristics. For each data group, the final correction unit refits the emission trend to obtain a corrected trend corresponding to each data group. Finally, based on the corrected trends corresponding to the data groups containing the anomalous emission node, the final corrected emission value is determined and output. For example, if the balance check fails, the final correction unit collects emission data from anomalous node C and its adjacent nodes (such as B, D, and E) over a past period. Through K-means clustering, this data may be divided into two groups: one group consists of nodes with normal fluctuations, and the other group consists of nodes with similar anomalous characteristics to anomalous node C. The final correction unit refits the trend for the anomalous group containing node C. For example, if it finds that the overall emissions of this group are too low, it will further adjust the emission value of node C based on the new fitting trend to obtain a more accurate final corrected emission value. This clustering refitting method can handle complex anomalies in local areas more precisely, avoiding coarse adjustments to the entire system. Therefore, even under multi-node anomalies or complex disturbances, it can still generate highly accurate correction values, improving the accuracy of data correction to 99%.

[0122] Example 9:

[0123] This embodiment provides an electronic device, which includes a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements a method for real-time tracking of carbon emissions from coal mines.

[0124] The electronic device may be, but is not limited to, a server, personal computer, tablet computer, smartphone, embedded device, or any other hardware device with computing and storage capabilities. The processor may be a central processing unit (CPU), graphics processing unit (GPU), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or any other computing unit capable of executing instructions. The memory may be random access memory (RAM), read-only memory (ROM), flash memory, hard disk drive, or any other medium capable of storing data.

[0125] When the computer program is executed by the processor, the method includes:

[0126] First, based on the carbon emission imbalances detected along a preset carbon flow path, abnormal emission nodes within the carbon flow path are identified. Specifically, the processor can read preset carbon flow path information from memory and obtain real-time carbon emission data for each node along the path. By comparing the total emissions from upstream nodes with the total emissions from downstream nodes, the deviation between the two is calculated. If this deviation exceeds a preset deviation threshold, the carbon flow path is marked as abnormal. Subsequently, the processor further analyzes all monitoring nodes within the abnormal carbon flow path, for example, by using an isolated forest model to detect outliers in the time series of emission values ​​for each node, thereby accurately identifying the abnormal emission nodes. For example, when a significant imbalance occurs in carbon emission data within a ventilation duct, the electronic equipment can quickly pinpoint which sensor node's data is abnormal. For instance, in a main return air duct, the total emissions from the three upstream monitoring points (M1, M2, M3) are 1500 ppm, while the total emissions from the two downstream monitoring points (M4, M5) are 1200 ppm, a deviation of 300 ppm, exceeding a preset threshold of 100 ppm. In this case, the electronic equipment will mark the return air duct as an abnormal carbon flow path. Furthermore, by analyzing the emission data sequences of nodes M1-M5 over the past 24 hours, the electronic equipment uses an isolated forest model to identify a sudden increase in emissions at node M2 ​​within a certain time period, thus identifying it as an abnormal emission node. This process refines the granularity of anomaly detection from macroscopic paths to microscopic nodes, improving the accuracy of anomaly localization and providing a precise target for subsequent correction work.

[0127] Secondly, based on the emission trends associated with the carbon flow path, a first corrected emission value is generated for the abnormal emission node. The processor extracts the emission trends of upstream nodes within the adjacent time window of the abnormal emission node and combines them with spatial correlation data associated with the carbon flow path (e.g., wind speed data of adjacent lanes). A linear regression algorithm is used to fit these input variables to generate a fitted trend curve. Based on this trend curve, the processor can generate a preliminary corrected emission value for the abnormal emission node. For example, for the aforementioned abnormal emission node M2, the electronic device analyzes the emission trend of its upstream node M1 and obtains the wind speed data of the lane where M2 is located and its adjacent lanes. Assuming that the emission trend of M1 is steadily increasing, and the wind speed of the lane where M2 is located is positively correlated with the emission amount of M1, the processor uses this data as input and fits the reasonable emission trend of M2 within the abnormal time period using a linear regression model. For example, the fitting result shows that the reasonable emission value of M2 in this time period should be 550 ppm, rather than the original 800 ppm. This step, by comprehensively considering time trends and spatial correlations, makes the generated correction values ​​more consistent with actual physical laws, avoids the deviations that may be caused by simple interpolation, and improves the rationality of the correction values.

[0128] Next, the carbon flow path is checked for balance using the first corrected emission value. The processor replaces the original value of the abnormal emission node with the first corrected emission value and then recalculates the total emission of the entire carbon flow path. By comparing the replaced total with a preset balance standard, it is determined whether the replacement meets the balance condition, thus obtaining the balance check result. For example, after the electronic device replaces the emission value of node M2 ​​from 800 ppm to 550 ppm, it recalculates the total emission of the three upstream monitoring points (M1, M2, M3) in the main return air roadway and compares it with the total emission of the two downstream monitoring points (M4, M5). If the deviation between the upstream total and the downstream total is within a preset allowable range (e.g., less than 50 ppm), it is considered to meet the balance requirement. This check process ensures that the data correction not only considers the rationality of local nodes but also takes into account the physical conservation characteristics of the entire carbon flow path, avoiding the problem of "pressing down one gourd and another floats up," and improving the global consistency of data correction.

[0129] Finally, in response to the failure of the conservation balance check, a final corrected emission value is generated based on cluster analysis of multiple nodes in the carbon flow path. If the balance check result shows that it does not conform to the conservation balance, the processor will extract historical emission data of multiple nodes adjacent to the abnormal emission node in the carbon flow path to form a dataset to be processed. The processor uses the K-means clustering algorithm to group the dataset to be processed to obtain multiple data groups containing different clustering characteristics. For the data group containing the abnormal emission node, the processor will refit the emission trend to obtain the corrected trend corresponding to the data group, and determine and output the final corrected emission value based on this. For example, if the above conservation balance check fails, it indicates that correcting only M2 is insufficient to restore the balance of the entire carbon flow path. At this time, the electronic device will extract historical emission data of adjacent nodes such as M1, M2, M3, M4, and M5, and use the K-means clustering algorithm to divide these node data into several groups. For example, M1, M2, and M3 are grouped into one group, and M4 and M5 are grouped into another group. For the group M1, M2, and M3, the electronic equipment refits their emission trends and generates a more accurate final corrected emission value for M2 based on the new trend, such as 580 ppm. This multi-stage correction and verification mechanism ensures that, even under complex and abnormal conditions, highly accurate and physically reasonable carbon emission data can be obtained through iterative and refined processing.

[0130] By executing the above method through the electronic device, real-time and accurate tracking and correction of coal mine carbon emission data can be achieved, ensuring the reliability and continuity of carbon emission accounting. The electronic device, through automated processing, reduces the need for manual intervention, improves data processing efficiency and real-time performance, and provides a hardware foundation for intelligent coal mine management.

[0131] Example 10:

[0132] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements the method described in any of the method embodiments of this application.

[0133] The computer-readable storage medium may be any medium capable of storing computer programs, such as, but not limited to:

[0134] 1. Non-transient storage media: including read-only memory (ROM), random access memory (RAM), electrically erasable programmable read-only memory (EEPROM), flash memory or other solid-state storage technologies, CD-ROM, DVD or other optical disc storage, disk storage or other magnetic storage devices, or any other physical medium that can be used to carry or store program code, which may be in the form of instructions or data structures.

[0135] 2. Transient storage media: such as electrical signals, optical signals, or electromagnetic signals, which can be in the form of instructions or data structures.

[0136] When the computer program is executed by a processor, the processor can invoke instructions stored on the computer-readable storage medium to perform the steps in any method embodiment of this application. For example, when the computer program is executed by a processor, the processor can perform the following operations:

[0137] First, based on the carbon emission imbalance detected along a preset carbon flow path, abnormal emission nodes within the carbon flow path are identified. Specifically, the processor can obtain the upstream and downstream node sets along the preset carbon flow path, and sum the current emissions of all nodes in each set to obtain the upstream and downstream total emissions. Then, the processor calculates the deviation between the upstream and downstream total emissions. If this deviation exceeds a preset deviation threshold, the carbon flow path is identified as an abnormal carbon flow path, which is then used to further identify abnormal emission nodes. For example, in a coal mine ventilation system, the processor monitors the carbon emissions of the main duct (upstream) and branch ducts (downstream) in real time. If the difference between the emissions of the main duct and the branch duct exceeds 5%, the ventilation path is marked as abnormal, and the specific monitoring points along the path are further analyzed.

[0138] Secondly, based on the emission trends associated with the carbon flow path, a first corrected emission value is generated for the anomalous emission node. The processor can first determine the potential change pattern of the anomalous emission node, for example, by analyzing its historical emission data to identify whether it is a sudden increase, a sudden decrease, or a periodic fluctuation. Simultaneously, the processor acquires spatial correlation data associated with the carbon flow path, such as wind speed data from adjacent lanes. Subsequently, the processor uses the emission trends and spatial correlation data of the upstream nodes of the anomalous emission node as input variables, and uses a linear regression algorithm to fit the data, generating a fitted trend curve. Based on this fitted trend curve, the processor generates a first corrected emission value for the anomalous emission node. For example, if the anomalous node shows a sudden increase pattern, and the wind speed in adjacent lanes suddenly increases, the processor will combine the emission trends and wind speed data of the upstream nodes, and use a linear regression model to predict the emission value of the anomalous node under normal conditions, as the first corrected value.

[0139] Next, the carbon flow path is checked for balance using the first corrected emission value. The processor replaces the original value of the abnormal emission node with the generated first corrected emission value and recalculates the total emissions of the entire carbon flow path. Then, the processor compares the recalculated total with a preset balance standard to determine whether the replacement conforms to the balance. For example, the processor substitutes the corrected emission value into the carbon flow path, recalculates the total emissions, and compares it with the theoretical total emissions. If the deviation is within ±3%, it is considered to conform to the balance.

[0140] Finally, in response to the failure of the conservation balance check, a final corrected emission value is generated based on cluster analysis of multiple nodes in the carbon flow path. If the conservation balance check result shows that it does not conform to the conservation balance, the processor extracts historical emission data from multiple nodes adjacent to the anomalous emission node in the carbon flow path, forming a dataset to be processed. Then, the processor uses the K-means clustering algorithm to group this dataset, obtaining multiple data groups with different clustering characteristics. For the data group containing the anomalous emission node, the processor refits the emission trend to obtain a corrected trend, and determines and outputs the final corrected emission value based on this corrected trend. For example, if the first correction is still unbalanced, the processor collects historical data of the anomalous node and its five nearest neighbors, divides these nodes into 2-3 groups using K-means clustering, and refits the trend for the data in the group containing the anomalous node to generate a more accurate final corrected value.

[0141] Through the execution of the aforementioned computer program, this embodiment enables real-time tracking, anomaly detection, and correction of coal mine carbon emission data, ensuring the accuracy and consistency of carbon emission data and providing reliable technical support for underground environmental monitoring in coal mines. This technical solution effectively improves the completeness and accuracy of carbon emission data, reduces accounting deviations caused by missing or abnormal data, and thus provides a more reliable data foundation for coal mining enterprises' carbon emission management, carbon quota declaration, and green mine certification.

[0142] The specific embodiments provided in this application are intended to further illustrate the principles and implementation methods of this application, and are not intended to limit the scope of protection of this application. Those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of this application. Therefore, the scope of protection of this application should be determined by the appended claims. This application, by introducing dynamic carbon characterization modeling, achieves real-time tracking and correction of coal mine carbon emission data, effectively solving the shortcomings of traditional methods in data integrity, anomaly identification, and conservation balance verification. This technical solution not only improves the accuracy and reliability of carbon emission accounting, providing a solid data foundation for coal mining enterprises' carbon emission management, carbon quota declaration, and green mine certification, but also ensures data accuracy and consistency through multi-level data processing and verification mechanisms, providing reliable technical support for underground environmental monitoring in coal mines. The implementation of this application will significantly improve the management level of coal mine carbon emission data and promote the green and sustainable development of the coal mining industry.

[0143] The above-disclosed embodiments are merely a few specific examples of the present invention. However, the embodiments of the present invention are not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A method for real-time tracking of carbon emissions from coal mines, characterized in that, include: Based on the carbon emission imbalance detected along the preset carbon flow path, abnormal emission nodes within the carbon flow path are identified. Based on the emission trend associated with the carbon flow path, a first corrected emission value is generated for the abnormal emission node; The carbon flow path is checked for conservation balance using the first corrected emission value. as well as, In response to the failure of the conservation balance check, a final corrected emission value is generated based on cluster analysis of multiple nodes in the carbon flow path.

2. The method according to claim 1, characterized in that, The step of determining abnormal emission nodes within the carbon flow path based on carbon emission imbalances detected along a preset carbon flow path includes: Based on the preset carbon flow path, obtain the upstream node set and downstream node set on the carbon flow path; The current emissions of all nodes in the upstream node set and the downstream node set are summed to obtain the total upstream emissions and the total downstream emissions. Calculate the deviation between the total upstream emissions and the total downstream emissions; and, If the deviation exceeds a preset deviation threshold, the carbon flow path is identified as an abnormal carbon flow path for subsequent identification of the abnormal emission node.

3. The method according to claim 2, characterized in that, After identifying the carbon flow path as an anomalous carbon flow path, the determination of the anomalous emission node further includes: Extract all monitoring nodes within the abnormal carbon flow path to obtain a list of nodes to be analyzed; The emission values ​​of each node in the list of nodes to be analyzed are arranged in chronological order to obtain the emission sequence to be detected; and, An isolated forest model is used to detect outliers in the emission sequence to be detected, and the outliers detected by the isolated forest model are identified as the abnormal emission nodes.

4. The method according to claim 1, characterized in that, The process of generating a first corrected emission value for the anomalous emission node based on the emission trend associated with the carbon flow path includes: Identify the potential change patterns of the abnormal emission nodes and obtain spatial correlation data associated with the carbon flow path; The emission trends of upstream nodes of the anomalous emission nodes are integrated with the spatial correlation data, serving as input variables for a linear regression algorithm; and, The input variables are fitted using the linear regression algorithm to generate a fitted trend curve, and a first corrected emission value is generated for the abnormal emission node based on the fitted trend curve, wherein the emission trend is characterized by the fitted trend curve.

5. The method according to claim 4, characterized in that, The acquisition of spatial correlation data associated with the carbon flow path includes: Obtain wind speed data sequences at multiple locations in the roadway adjacent to the abnormal emission node in the carbon flow path; Calculate the spatial correlation coefficient between the wind speed data sequence and the emission trend of the upstream node; and, If the spatial correlation coefficient is greater than a preset correlation threshold, the wind speed data sequence is determined as the spatial correlation data and used as one of the input variables of the linear regression algorithm.

6. The method according to claim 1, characterized in that, The generation of the final corrected emission values ​​based on cluster analysis of multiple nodes in the carbon flow path includes: In response to the failure of the conservation balance check, historical emission data of multiple nodes adjacent to the abnormal emission node in the carbon flow path are extracted to form a dataset to be processed; and, The dataset to be processed is grouped using the K-means clustering algorithm to obtain multiple data groups containing different clustering characteristics, wherein the clustering analysis is implemented based on the grouping process.

7. The method according to claim 6, characterized in that, After obtaining the multiple data groups, generating the final corrected emission values ​​further includes: For each of the plurality of data groups, the emission trend is refitted to obtain a corrected trend corresponding to each data group; and, Based on the correction trend corresponding to the data group containing the anomalous emission nodes, the final corrected emission value is determined and output.

8. A real-time carbon emission tracking system for coal mines, characterized in that, include: An anomaly determination unit is used to determine the abnormal emission nodes within the carbon flow path based on the carbon emission imbalance detected along the preset carbon flow path. The first correction unit is used to generate a first corrected emission value for the abnormal emission node based on the emission trend associated with the carbon flow path. A balance verification unit is used to perform a conservation balance verification on the carbon flow path using the first corrected emission value. as well as, The final correction unit is used to generate a final corrected emission value based on cluster analysis of multiple nodes in the carbon flow path in response to the failure of the conservation balance check.

9. An electronic device comprising a processor and a memory, the memory storing a computer program which, when executed by the processor, implements the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images