Electric power carbon emission real-time monitoring system

Abstract

The invention relates to the technical field of environment monitoring, in particular to an electric power carbon emission real-time monitoring system, which comprises a data acquisition module, a data processing module, a data processing module, a data processing module, a data processing module, a data processing module, a data processing module and a data processing module, wherein the data processing module receives sensor data of a plurality of monitoring points of a power plant and comprises a carbon emission value, a timestamp and an environment parameter of each sensor; and sorting according to the time sequence characteristics of the data, and establishing a serialized data set. According to the method, data cleaning and time serialization processing are applied, data accuracy and availability are improved, data management is more adaptive and prospective by dynamically adjusting calculation of strategies such as kurtosis and skewness, geometric and topological analysis methods are introduced to improve understanding of data structure features, complex modes in data are effectively recognized, and the method is suitable for being popularized and applied. According to the method and the system, the dependency relationship of a hole, a cluster and an emission source and the application of an anomaly detection technology are applied, so that timely identification of abnormal behaviors is enhanced, potential equipment faults or operation errors are allowed to be found in an early stage, timely intervention is realized, and unnecessary environmental influence is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring, and in particular to a real-time monitoring system for carbon emissions from electricity. Background Art

[0002] The field of environmental monitoring technology involves the use of various technical means and systems to monitor, analyze and manage physical, chemical and biological parameters in natural and artificial environments. These technologies help detect changes in environmental quality, thereby preventing potential environmental risks and improving environmental management. Environmental monitoring systems can be widely used for air quality monitoring, water quality testing, soil condition monitoring, and real-time tracking of specific environmental impacts, such as industrial emissions and urban development. These systems typically include sensor networks, data collection units and analysis software to provide accurate real-time data to support environmental decision-making and compliance assessment.

[0003] Among them, the real-time monitoring system for power carbon emissions refers to a technical system specifically used to monitor and manage the carbon dioxide emissions generated by the power industry during the production process. Its main purpose is to help power companies or regulators track carbon emissions in real time to ensure compliance with environmental standards and facilitate participation in the carbon trading market. Through this monitoring, the environmental impact of the power industry can be more effectively evaluated and controlled, while providing data support for policy making and environmental protection. The implementation of this system is of great significance to promoting sustainable development and achieving carbon neutrality goals.

[0004] Although existing environmental monitoring technologies provide basic monitoring of real-time data, they are still lacking in in-depth analysis of multi-parameter interactions and identification of data structure complexity. Existing systems mainly focus on monitoring single data points and lack an in-depth understanding of the integrity and relevance of data. In addition, static data analysis methods cannot effectively adapt to the dynamic changes of environmental data, which limits the timeliness of monitoring responses. In terms of data integrity and anomaly detection, existing technologies fail to fully eliminate erroneous data or identify abnormal behavior in a timely manner, affecting data quality and the accuracy of decision-making. These technical limitations lead to incomplete or in-time assessments of environmental threats, affecting the formulation and implementation of environmental policies, and may fail to provide necessary support at critical moments, increasing environmental risks. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a real-time monitoring system for carbon emissions from electricity.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: The real-time monitoring system of electric power carbon emissions includes: The data acquisition module receives sensor data from multiple monitoring points in the power plant, including the carbon emission value, timestamp and environmental parameters of each sensor, performs data cleaning, removes abnormal values ​​and incomplete records, sorts the data according to its temporal characteristics, and establishes a serialized data set; The dynamic adjustment module extracts carbon emissions and environmental factors from the serialized data set, calculates the kurtosis and skewness of the data, dynamically adjusts the activation threshold, evaluates the degree of deviation of multiple data points from the established emission pattern, and generates adjusted data features in combination with time series analysis; The topological analysis module uses the adjusted data features and applies geometric and topological methods to analyze holes, connections and clusters between data, identify dependencies and spatial distributions between emission sources, extract topological features of the data, and generate a topological feature summary; The anomaly detection module compares the existing emission patterns based on the topological feature summary, identifies data points with abnormal behavior, distinguishes abnormal emissions caused by equipment failure or operational errors through cluster analysis, and generates abnormal emission identification results.

[0007] As a further solution of the present invention, the step of obtaining the serialized data set is specifically: Receive sensor data from multiple monitoring points in the power plant, including carbon emissions values , Timestamp and environmental parameters , using the formula:

[0008] Calculating the adjusted carbon emissions of each sensor to generate an initial adjusted carbon emissions set; in, Indicates sensor The original carbon emission value, Represents environmental parameters, represents the adjusted carbon emissions, , is the adjustment factor, Indicates a timestamp; The initial adjusted carbon emissions set is cleaned to remove outliers and incomplete records using the formula:

[0009] Generate a cleaned dataset; in, represents the cleaned data set, is the time adjustment factor, Indicates a timestamp; The cleaned data set data is sorted and combined by timestamp to form a time series using the formula:

[0010] Generate a serialized dataset; in, represents a serialized dataset at time t, is the weighting coefficient of the data point.

[0011] As a further solution of the present invention, the step of obtaining the adjusted data features is specifically: Extract the carbon emissions at each time point from the serialized dataset and environmental factors , integrated into a feature dataset , using the formula:

[0012] Generate feature extraction results; in, represents carbon emissions, Represents environmental factors, , is the feature weighting coefficient; Statistical analysis was applied to the feature extraction results to calculate the kurtosis and skewness using the formula:

[0013] and

[0014] Generate statistical analysis results; in, represents the kurtosis, represents skewness, represents the average value, represents the standard deviation, represents the number of samples; Dynamically adjust the activation threshold according to the statistical analysis results , using the formula:

[0015] Increase Factor , integrating the composite effects of kurtosis and skewness on the threshold to generate an adjusted activation threshold; in, is the original activation threshold, is the adjusted activation threshold; Using the adjusted activation threshold and statistical data, calculate how far the data point deviates from the established emission pattern using the formula:

[0016] Increase the sensitivity of deviation and generate adjusted data features; in, Indicates the deviation, represents a given emission pattern, is the deviation adjustment factor.

[0017] As a further solution of the present invention, the steps of obtaining the holes, connections and clusters between the data are specifically as follows: Analyze the characteristics of the adjusted data, apply high-dimensional geometric mapping to identify the holes between data, and use the formula:

[0018] Calculate the holes and generate the hole identification results; in, represents the hole recognition result, is the adjustment parameter, Represents data feature points, represents the feature mean, is the modulation coefficient, Represents the number of feature points of the adjusted data; Based on the hole identification results, the Voronoi diagram is used to determine the data connection points using the formula:

[0019] Calculate connection points and generate connection point analysis results; in, Represents the results of the connection point analysis. Cited from adjusted data characteristics, is a small constant to prevent division by zero, is the adjustment coefficient, is the attenuation coefficient, Represents the number of feature points that determine the data connection points; Using the results of the tie point analysis, a density clustering algorithm was applied to identify clusters in the data using the formula:

[0020] Calculate clusters and generate cluster analysis results; in, represents the cluster analysis results, For each tie point in the tie point analysis results, is the clustering coefficient, is the minimum value among all connection points. Represents the number of data points for the density clustering algorithm.

[0021] As a further solution of the present invention, the step of obtaining the topological feature summary is specifically: From the analyzed cluster data, the topological analysis technique is used to extract the dependencies and spatial distribution between emission sources, using the formula:

[0022] Calculate dependency and spatial distribution, and generate dependency and spatial distribution analysis results; in, Represents the results of dependency and spatial distribution analysis, represents the characteristic value of the data point, represents the hole recognition result, To prevent small constants from dividing by zero, and is the adjustment factor, is the modulation coefficient, Represents the total number of data points involved in the calculation; Using the above dependency and spatial distribution analysis results, the persistent homology method is applied to refine the topological characteristics of the data, using the formula:

[0023] Calculate topological features and generate data topological feature results; in, Represents the topological characteristics of data. For a single value in the dependence and spatial distribution analysis results, is the adjustment factor, is the attenuation coefficient, is the modulation coefficient, m represents the total number of single analysis values ​​in the dependence and spatial distribution analysis results; Based on the data topology feature results, the formula is used:

[0024] Simplify data representation by composite exponents and quadratic terms to generate summaries of topological features; in, represents a summary of topological features, , is the adjustment factor.

[0025] As a further solution of the present invention, the step of obtaining the abnormal emission identification result is specifically as follows: The topological feature summary is compared with the existing emission pattern using the formula:

[0026] Calculate deviations and generate preliminary abnormal emission points; in, Indicates the initial abnormal emission point, is the weight factor, represents a summary of topological features, represents the emission value of the existing model, and is the tuning parameter, n5 represents the total number of data points for comparison between the topological feature summary and the existing emission pattern; Based on the preliminary abnormal emission points, cluster analysis was applied using the formula:

[0027] Identify the root causes of the anomaly and generate anomaly cause classification results; in, Indicates the classification result of abnormal cause, Cited from the preliminary abnormal emission point, is the standard deviation of the outliers, is a small constant, is the adjustment coefficient; Based on the above abnormal cause classification results, the formula is adopted:

[0028] By strengthening the weight of each classification, abnormal emission identification results are generated; in, Indicates the abnormal emission identification result, For each classification point in the abnormal cause classification result, n6 represents the number of classification points involved in the abnormal cause classification result.

[0029] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, through the comprehensive application of data cleaning and time series processing, the accuracy and availability of data are significantly improved, and the accuracy of carbon emissions monitoring is significantly improved. Dynamic adjustment strategies such as the calculation of kurtosis and skewness make data management more adaptive and forward-looking, and enhance the predictive value of data behavior. The introduction of geometric and topological analysis methods improves the understanding of data structure characteristics and effectively identifies complex patterns in the data, such as holes, clusters, and dependencies between emission sources. The application of anomaly detection technology enhances the timely identification of abnormal behavior, allowing early detection of potential equipment failures or operating errors, timely intervention, and reduction of unnecessary environmental impacts. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a system flow chart of the present invention; Figure 2 A flowchart of the steps for obtaining a serialized data set of the present invention; Figure 3 A flow chart of the steps for obtaining the adjusted data features of the present invention; Figure 4 A flow chart of steps for obtaining holes, connections and clusters between data of the present invention; Figure 5 A flowchart of the steps for obtaining a topological feature summary of the present invention; Figure 6 This is a flow chart of the steps for obtaining abnormal emission identification results of the present invention. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0032] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined. Embodiment 1

[0033] See also Figure 1 , the real-time monitoring system for electricity carbon emissions includes: The data acquisition module receives sensor data from multiple monitoring points in the power plant, including the carbon emission value, timestamp and environmental parameters of each sensor, performs data cleaning, removes abnormal values ​​and incomplete records, sorts the data according to its temporal characteristics, and establishes a serialized data set; The dynamic adjustment module extracts carbon emissions and environmental factors from the serialized data set, calculates the kurtosis and skewness of the data, dynamically adjusts the activation threshold, evaluates the degree of deviation of multiple data points from the established emission pattern, and generates adjusted data features in combination with time series analysis; The topological analysis module uses the adjusted data features and applies geometric and topological methods to analyze the holes, connections and clusters between data, identify the dependencies and spatial distribution between emission sources, extract the topological features of the data, and generate a summary of the topological features; The anomaly detection module compares the existing emission patterns based on the topological feature summary, identifies data points with abnormal behavior, distinguishes abnormal emissions caused by equipment failure or operational errors through cluster analysis, and generates abnormal emission identification results.

[0034] The serialized data set specifically includes carbon emission time series and environmental parameter time series. The adjusted data features specifically refer to the adjusted data kurtosis, data skewness, and dynamic activation threshold. The topological feature summary includes data hole features, connection features, and cluster features. The abnormal emission identification results specifically include equipment failure anomalies and operation error anomalies.

[0035] See also Figure 2 , the specific steps for obtaining the serialized data set are: Receive sensor data from multiple monitoring points in the power plant, including carbon emissions values , Timestamp and environmental parameters , using the formula:

[0036] Calculating the adjusted carbon emissions of each sensor to generate an initial adjusted carbon emissions set; in, Indicates sensor The original carbon emission value, Represents environmental parameters, represents the adjusted carbon emissions, , is the adjustment factor, Indicates a timestamp; The initial adjusted carbon emissions set was cleaned to remove outliers and incomplete records using the formula:

[0037] Generate a cleaned dataset; in, represents the cleaned data set, is the time adjustment factor, Indicates a timestamp; Sort and combine the cleaned data set data by timestamp to form a time series using the formula:

[0038] Generate a serialized dataset; in, represents a serialized dataset at time t, is the weighting coefficient of the data point.

[0039] formula:

[0040] Parameter Description: : The raw carbon emission value of the sensor, in tons / hour.

[0041] : Environmental parameters, including temperature and humidity, are adjusted proportionally.

[0042] : Adjustment index of environmental parameters, assumed to be 0.1.

[0043] : Time decay coefficient, assumed to be 0.05.

[0044] : Timestamp from the start of monitoring, in hours.

[0045] Calculation process and examples: Assume that the carbon emission value of a sensor at the target time 50 tons / hour, environmental parameters is 30 (representing a certain environmental quality index), timestamp is 10 hours. Substitute into the formula to calculate the adjusted carbon emissions:

[0046] here Tons / hour, represents the actual carbon emissions after reference to environmental parameters and time decay.

[0047] formula:

[0048] Parameter Description: : Time adjustment coefficient, assumed to be 0.05.

[0049] : The adjusted carbon emissions calculated in step 1.

[0050] : Timestamp, in hours.

[0051] Calculation process and examples: Continue using the result from step 1 Tons / hour, timestamp Hour.

[0052]

[0053] because , which is an unreasonable result and suggests the need to re-evaluate Setting or checking In practical applications, and The setting should ensure that it does not result in a negative adjustment factor.

[0054] formula:

[0055] Parameter Description: : Weighting coefficient, referring to the importance of recent data.

[0056] Calculation process and examples: If the correction The calculation error is assuming that the last 6 hours arrive The data is as follows In order , In order .

[0057]

[0058] here Tons / hour, representing the average carbon emissions weighted by reference time.

[0059] See also Figure 3 , the steps to obtain the adjusted data features are as follows: Extract the carbon emissions at each time point from the serialized dataset and environmental factors , integrated into a feature dataset , using the formula:

[0060] Generate feature extraction results; in, represents carbon emissions, Represents environmental factors, , is the feature weighting coefficient; Apply statistical analysis to the feature extraction results to calculate the kurtosis and skewness using the formula:

[0061]

[0062]

[0063] Generate statistical analysis results; in, represents the kurtosis, represents skewness, represents the average value, represents the standard deviation, represents the number of samples; Dynamically adjust the activation threshold based on statistical analysis results , using the formula:

[0064] Increase Factor , integrating the composite effects of kurtosis and skewness on the threshold to generate an adjusted activation threshold; in, is the original activation threshold, is the adjusted activation threshold; Using the adjusted activation threshold and statistical data, calculate how far the data point deviates from the established emission pattern using the formula:

[0065] Increase the sensitivity of deviation and generate adjusted data features; in, Indicates the deviation, represents a given emission pattern, is the deviation adjustment factor.

[0066] formula:

[0067] Assumed parameters and calculation process: Assume that at a certain time point t, the observed carbon emissions 100ppm (parts per million), environmental factors It is 50 units.

[0068] Select weighting factor and ,The coefficients are selected based on previous studies or empirical data, so that the data adjustment can better reflect the sensitivity of the environment.

[0069] Calculate adjusted carbon emissions and environmental factors:

[0070]

[0071] Result analysis: Adjusted feature set It can better reflect the changes in sensitivity of the actual environment. The results show that the original data is adjusted to make it more expressive for the model.

[0072] formula:

[0073] and

[0074] Assumed parameters and calculation process: Assumptions The result of the previous step and other similar data points.

[0075] Assumptions ( The average value of the data), (standard deviation), (number of data points).

[0076] Calculation of kurtosis and skewness:

[0077]

[0078] Result analysis: The calculated kurtosis of 1.5 indicates that the data is flatter than a normal distribution, while the skewness of 0.3 indicates that the data is slightly skewed to the right. This helps to understand the shape of the data distribution.

[0079] formula:

[0080] Assumed parameters and calculation process: Assuming the original activation threshold .

[0081] and .

[0082] Adjust the threshold:

[0083] Result analysis: The new activation threshold of 5.225 represents a fine-tuning of the original threshold based on the kurtosis and skewness of the data to better match the characteristics of the data.

[0084] formula:

[0085] Assumed parameters and calculation process: Assuming a given emission pattern , the above .

[0086] Calculate the deviation:

[0087] Result analysis: The deviance of 2.46 indicates the overall deviation of the data from the model, with lower values ​​indicating a better model fit.

[0088] See also Figure 4 ,The specific steps for obtaining holes, connections and clusters between data are: Analyze the characteristics of the adjusted data and apply high-dimensional geometric mapping to identify the holes between the data using the formula:

[0089] Calculate the holes and generate the hole identification results; in, represents the hole recognition result, is the adjustment parameter, Represents data feature points, represents the feature mean, is the modulation coefficient, Represents the number of feature points of the adjusted data; Based on the hole identification results, the Voronoi diagram is used to determine the data connection points using the formula:

[0090] Calculate connection points and generate connection point analysis results; in, Represents the results of the connection point analysis. Cited from adjusted data characteristics, is a small constant to prevent division by zero, is the adjustment coefficient, is the attenuation coefficient, Represents the number of feature points that determine the data connection points; Using the results of the join point analysis, a density clustering algorithm was applied to identify clusters in the data using the formula:

[0091] Calculate clusters and generate cluster analysis results; in, represents the cluster analysis results, For each tie point in the tie point analysis results, is the clustering coefficient, is the minimum value among all connection points. Represents the number of data points for the density clustering algorithm.

[0092] formula:

[0093] Detailed explanation and examples: Parameter definition: : The eigenvalue of the data point.

[0094] : The average of the eigenvalues ​​of all data points.

[0095] : Control the width of the Gaussian function, increase will make the function sharper.

[0096] : Modulate the amplitude of the sine wave to adjust the effect of the sine wave on the result.

[0097] Calculation process: Assume there are three data points with eigenvalues ,but .

[0098] choose and as a parameter.

[0099] Calculate the contribution to each data point:

[0100]

[0101]

[0102] Calculate the sum .

[0103] Result analysis: Represents the total influence of the data point after reference holes and tie point contributions.

[0104] formula:

[0105] Detailed explanation and examples: Parameter definition: : To prevent zero decimals, set it to a very small positive number, such as 0.01.

[0106] : Exponential coefficient for adjusting the weights between data points.

[0107] : Represents the attenuation coefficient, which is used to reduce the influence of data points far away from the center point.

[0108] Calculation process: Use the previous Values ​​and the same data points ,set up and 0.1, and .

[0109] For each data point calculate:

[0110]

[0111]

[0112] Calculate the sum .

[0113] Result analysis: Represents the normalized total influence obtained from the join point analysis.

[0114] formula:

[0115] Detailed explanation and examples: Parameter definition: : Clustering coefficient. Increasing the value will increase the compactness within the cluster.

[0116] Calculation process: Using the result of step 2 above, , assuming (Minimum connection point analysis result).

[0117] For each tie point calculate:

[0118]

[0119]

[0120] Calculate the sum .

[0121] Result analysis: It represents the total score of cluster analysis, reflecting the internal compactness of the cluster.

[0122] See also Figure 5 , the specific steps for obtaining the topological feature summary are: From the analyzed cluster data, the topological analysis technique is used to extract the dependencies and spatial distribution between emission sources, using the formula:

[0123] Calculate dependency and spatial distribution, and generate dependency and spatial distribution analysis results; in, Represents the results of dependency and spatial distribution analysis, represents the characteristic value of the data point, represents the hole recognition result, To prevent small constants from dividing by zero, and is the adjustment factor, is the modulation coefficient, Represents the total number of data points involved in the calculation; Using the results of dependency and spatial distribution analysis, the persistent homology method is applied to refine the topological characteristics of the data using the formula:

[0124] Calculate topological features and generate data topological feature results; in, Represents the topological characteristics of data. For a single value in the dependence and spatial distribution analysis results, is the adjustment factor, is the attenuation coefficient, is the modulation coefficient, m represents the total number of single analysis values ​​in the dependence and spatial distribution analysis results; Based on the data topology feature results, the formula is used:

[0125] Simplify data representation by composite exponents and quadratic terms to generate summaries of topological features; in, represents a summary of topological features, , is the adjustment factor.

[0126] formula:

[0127] Parameter Description: The feature value of the data point.

[0128] :Adjustment Index of influence.

[0129] : The value obtained from the void analysis reflects the degree of voids between data.

[0130] :Adjustment Index of influence.

[0131] : To avoid a small constant with a denominator of zero, set it to 0.01.

[0132] : Adjust the coefficient of the impact of the logarithmic term.

[0133] Assume the following example: set up , the characteristic value of the data point , the degree of void ,parameter .

[0134] Calculation process: calculate :

[0135] Calculate the denominator :

[0136] Compute the numerator of each term:

[0137] Compute the logarithmic term:

[0138] Comprehensive calculation :

[0139] result: Calculated It can be interpreted as a measure of the dependence and spatial distribution under the influence of reference data point importance and holes.

[0140] formula:

[0141] Parameter Description: : A single value in the results of dependence and spatial distribution analysis.

[0142] : Used to adjust The index of impact.

[0143] : Attenuation coefficient, used to adjust impact.

[0144] : Used to enhance small values The coefficient of influence.

[0145] Assume the following example: set up , the result value ,parameter .

[0146] Calculation process: Calculate each Adjustments:

[0147] Accumulation calculation : result: Represents the overall topological feature strength under a given attenuation and enhancement model.

[0148] formula:

[0149] Parameter Description: : Data topology characteristics.

[0150] :Used to adjust The influence of the linear part.

[0151] :Used to adjust The influence of nonlinear part.

[0152] Assume the following example: set up ,parameter .

[0153] Calculation process: calculate :

[0154] result: Represents the topological feature summary, a value between 0 and 1, which expresses the summary of the overall topological features of the data.

[0155] See also Figure 6 , the specific steps for obtaining abnormal emission identification results are: The topological feature summary is compared with the existing emission pattern using the formula:

[0156] Calculate deviations and generate preliminary abnormal emission points; in, Indicates the initial abnormal emission point, is the weight factor, represents a summary of topological features, represents the emission value of the existing model, and is the tuning parameter, n5 represents the total number of data points for comparison between the topological feature summary and the existing emission pattern; Based on the preliminary abnormal emission points, cluster analysis was applied using the formula:

[0157] Identify the root causes of the anomaly and generate anomaly cause classification results; in, Indicates the classification result of abnormal cause, Cited from the preliminary abnormal emission point, is the standard deviation of the outliers, is a small constant, is the adjustment coefficient; Based on the abnormal cause classification results, the formula is used:

[0158] By strengthening the weight of each classification, abnormal emission identification results are generated; in, Indicates the abnormal emission identification result, For each classification point in the abnormal cause classification result, n6 represents the number of classification points involved in the abnormal cause classification result.

[0159] formula:

[0160] Parameter explanation and source: : Topological feature summary The value of the data point.

[0161] : The current emission pattern The value of the data point.

[0162] : No. The weight factor of each data point is set according to its importance in the data set.

[0163] , : Adjust parameters to adjust the sensitivity and decay rate of the deviation.

[0164] Calculation process: For each data point , calculate its topological feature summary Compared with the current emission pattern The absolute value of the difference between .

[0165] Raise the absolute value of the difference to a power , and multiplied by an exponential decay factor to reduce the impact of large differences.

[0166] Multiply the result of each data point by the weight factor , and then sum to get the abnormal emission point index .

[0167] Calculation example: Assume there are 3 data points, given the following values:

[0168]

[0169]

[0170]

[0171]

[0172] The calculation process is:

[0173] result The value indicates that, based on the current parameters and weights, the data points as a whole have slightly abnormal emission behavior, because The value is close to 0, indicating that the deviation is not large.

[0174] formula:

[0175] Parameter explanation and source: : No. The calculation results of the preliminary abnormal emission points for the data points.

[0176] : The standard deviation of the outliers can be obtained by The standard deviation of the values ​​was calculated.

[0177] : Small constant, set to 0.01 or less.

[0178] : Adjustment coefficient, used to adjust the sensitivity of abnormal emissions.

[0179] Calculation process: Calculate each abnormal emission point Normalized values ​​relative to the standard deviation.

[0180] Raise the normalized value to a power And multiply it by the cosine function to highlight those abnormal points with large deviations and periodic changes.

[0181] Average the calculation results of all data points to obtain the abnormal cause classification results .

[0182] Calculation example: Assume there are 3 results for abnormal emission points:

[0183]

[0184]

[0185]

[0186] The calculation process is:

[0187] result The value indicates that, based on the current parameters, the abnormal cause classification result shows an obvious abnormality because its Higher value.

[0188] formula:

[0189] Parameter explanation and source: The abnormal cause classification results data points.

[0190] Strengthen the influence of data points with higher anomaly classification scores by Normalization and logarithmic transformation are performed to enhance the nonlinear characteristics of the data.

[0191] Calculation process: First, the classification score Sum them up and get a sum for normalization.

[0192] Score each category Normalization is performed and then the square root of the logarithm is calculated to enhance the contribution of those categories with high scores.

[0193] Add the processing results of the data points to obtain the abnormal emission identification results .

[0194] Calculation example: Assume that the abnormal cause classification results are as follows:

[0195] The calculation process is:

[0196] result The value indicates the abnormal emission identification effect, and a high value indicates that the corresponding data point has significant abnormal behavior.

[0197] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. Real-time monitoring system for electricity carbon emissions, characterized by: The system comprises: The data acquisition module receives sensor data from multiple monitoring points in the power plant, including the carbon emission value, timestamp and environmental parameters of each sensor, performs data cleaning, removes abnormal values ​​and incomplete records, sorts the data according to its temporal characteristics, and establishes a serialized data set; The dynamic adjustment module extracts carbon emissions and environmental factors from the serialized data set, calculates the kurtosis and skewness of the data, dynamically adjusts the activation threshold, evaluates the degree of deviation of multiple data points from the established emission pattern, and generates adjusted data features in combination with time series analysis; The topological analysis module uses the adjusted data features and applies geometric and topological methods to analyze holes, connections and clusters between data, identify dependencies and spatial distributions between emission sources, extract topological features of the data, and generate a topological feature summary; The anomaly detection module compares the existing emission patterns based on the topological feature summary, identifies data points with abnormal behavior, distinguishes abnormal emissions caused by equipment failure or operational errors through cluster analysis, and generates abnormal emission identification results.

2. The real-time monitoring system for carbon emissions from electricity according to claim 1 is characterized in that: The steps for obtaining the serialized data set are specifically as follows: Receive sensor data from multiple monitoring points in the power plant, including carbon emissions values , Timestamp and environmental parameters , using the formula: ； Calculating the adjusted carbon emissions of each sensor to generate an initial adjusted carbon emissions set; in, Indicates sensor The original carbon emission value, Represents environmental parameters, represents the adjusted carbon emissions, , is the adjustment factor, Indicates a timestamp; The initial adjusted carbon emissions set is cleaned to remove outliers and incomplete records using the formula: ； Generate a cleaned dataset; in, represents the cleaned data set, is the time adjustment factor, Indicates a timestamp; The cleaned data set data is sorted and combined by timestamp to form a time series using the formula: ； Generate a serialized dataset; in, represents a serialized dataset at time t, is the weighting coefficient of the data point.

3. The real-time monitoring system for carbon emissions from electricity according to claim 2 is characterized in that: The steps for obtaining the adjusted data features are specifically as follows: Extract the carbon emissions at each time point from the serialized dataset and environmental factors , integrated into a feature dataset , using the formula: ； Generate feature extraction results; in, represents carbon emissions, Represents environmental factors, , is the feature weighting coefficient; Statistical analysis was applied to the feature extraction results to calculate the kurtosis and skewness using the formula: ； and ； Generate statistical analysis results; in, represents the kurtosis, represents the skewness, represents the average value, represents the standard deviation, represents the number of samples; Dynamically adjust the activation threshold according to the statistical analysis results , using the formula: ； Increase Factor , integrating the composite effects of kurtosis and skewness on the threshold to generate an adjusted activation threshold; in, is the original activation threshold, is the adjusted activation threshold; Using the adjusted activation threshold and statistical data, calculate how far the data point deviates from the established emission pattern using the formula: ； Increase the sensitivity of deviation and generate adjusted data features; in, Indicates the deviation, represents a given emission pattern, is the deviation adjustment factor.

4. The real-time monitoring system for carbon emissions from electricity according to claim 3 is characterized in that: The steps for obtaining the holes, connections and clusters between the data are specifically as follows: Analyze the characteristics of the adjusted data, apply high-dimensional geometric mapping to identify the holes between data, and use the formula: ； Calculate the holes and generate the hole identification results; in, represents the hole recognition result, is the adjustment parameter, Represents data feature points, represents the feature mean, is the modulation coefficient, Represents the number of feature points of the adjusted data; Based on the hole identification results, the Voronoi diagram is used to determine the data connection points using the formula: ； Calculate connection points and generate connection point analysis results; in, Represents the results of the connection point analysis. Cited from adjusted data characteristics, is a small constant to prevent division by zero, is the adjustment coefficient, is the attenuation coefficient, Represents the number of feature points that determine the data connection points; Using the results of the tie point analysis, a density clustering algorithm was applied to identify clusters in the data using the formula: ； Calculate clusters and generate cluster analysis results; in, represents the cluster analysis results, For each tie point in the tie point analysis results, is the clustering coefficient, is the minimum value among all connection points. Represents the number of data points for the density clustering algorithm.

5. The real-time monitoring system for carbon emissions from electricity according to claim 4 is characterized in that: The steps for obtaining the topological feature summary are specifically as follows: From the analyzed cluster data, the topological analysis technique is used to extract the dependencies and spatial distribution between emission sources, using the formula: ； Calculate dependency and spatial distribution, and generate dependency and spatial distribution analysis results; in, Represents the results of dependency and spatial distribution analysis, represents the characteristic value of the data point, represents the hole recognition result, To prevent small constants from dividing by zero, and is the adjustment factor, is the modulation coefficient, Represents the total number of data points involved in the calculation; Using the above dependency and spatial distribution analysis results, the persistent homology method is applied to refine the topological characteristics of the data, using the formula: ； Calculate topological features and generate data topological feature results; in, Represents the topological characteristics of the data. For a single value in the dependence and spatial distribution analysis results, is the adjustment factor, is the attenuation coefficient, is the modulation coefficient, m represents the total number of single analysis values ​​in the dependence and spatial distribution analysis results; Based on the data topology feature results, the formula is used: ； Simplify data representation by composite exponents and quadratic terms to generate summaries of topological features; in, represents a summary of topological features, , is the adjustment factor.

6. The real-time monitoring system for carbon emissions from electricity according to claim 5 is characterized in that: The steps for obtaining the abnormal emission identification result are specifically as follows: The topological feature summary is compared with the existing emission pattern using the formula: ； Calculate deviations and generate preliminary abnormal emission points; in, Indicates the initial abnormal emission point, is the weight factor, represents a summary of topological features, represents the emission value of the existing model, and is the tuning parameter, n5 represents the total number of data points for comparison between the topological feature summary and the existing emission pattern; Based on the preliminary abnormal emission points, cluster analysis was applied using the formula: ； Identify the root causes of the anomaly and generate anomaly cause classification results; in, Indicates the classification result of abnormal cause, Cited from the preliminary abnormal emission point, is the standard deviation of the outliers, is a small constant, is the adjustment coefficient; Based on the above abnormal cause classification results, the formula is adopted: ； By strengthening the weight of each classification, abnormal emission identification results are generated; in, Indicates the abnormal emission identification result, For each classification point in the abnormal cause classification result, n6 represents the number of classification points involved in the abnormal cause classification result.

Images