Intelligent analysis method for carbon emission data

Abstract

The invention relates to the technical field of data preprocessing before machine learning and deep learning, in particular to a carbon emission data intelligent analysis method, which comprises the steps of S1, acquiring carbon emission data of each production link of an enterprise in real time; s2, preprocessing the carbon emission data to obtain preprocessed carbon emission data; s3, analyzing the preprocessed carbon emission data by using a K-Means clustering algorithm to obtain classification of carbon emission data types; s4, analyzing the carbon emission data of the divided types by using a fractal theory model to obtain carbon emission data trend prediction; s5, newly generated carbon emission related data are input into the carbon emission calculation model, model input is continuously updated, and data trend prediction fits the actual situation; and S6, analyzing carbon emission back main promotion factors fitting an actual trend prediction result, and planning an emission reduction target and measures. A decision maker can more accurately evaluate the emission reduction effect and cost of different scenes, thereby making a more scientific and reasonable emission reduction policy.

Description

Technical Field

[0001] The present invention relates to the technical field of data preprocessing before machine learning and deep learning, and particularly relates to an intelligent analysis method for carbon emission data. Background Art

[0002] With the increasingly serious problem of global warming, countries have successively formulated emission reduction targets and are committed to controlling greenhouse gas emissions to address various challenges brought about by climate change, such as rising sea levels and frequent extreme climate events. In this context, whether at the national level, corporate level, or various organizations, it is necessary to accurately grasp their own carbon emission situations and make reasonable predictions based on the research and analysis of the regularity of carbon emission data, which puts forward higher requirements for the intelligent analysis method of carbon emission data.

[0003] In today's society, data shows an explosive growth, and massive data resources have been accumulated in various fields, and the carbon emission field is no exception. Carbon emission-related data is continuously generated from multiple channels such as energy consumption records, production activity data, and transportation data. The maturity of big data technology, including the improvement of capabilities in data storage, management, and processing, makes it possible to collect and integrate large-scale carbon emission data.

[0004] Traditional manual statistics and simple analysis methods are difficult to handle massive and complex carbon emission data. Therefore, by means of intelligent analysis methods, it is possible to more efficiently and accurately mine valuable information in carbon emission data and provide strong support for formulating scientific and reasonable emission reduction strategies. Summary of the Invention

[0005] The present invention provides an intelligent analysis method for carbon emission data. By analyzing the regularity of carbon emission data, a prediction model is established to solve the defects in the prior art that carbon emission data cannot be reasonably predicted, thus making it difficult to effectively verify the rationality of the data and quickly make response strategies.

[0006] The present invention provides an intelligent analysis method for carbon emission data, including: S1: Real-time collect carbon emission data of each production link of an enterprise.

[0007] S2: Preprocess the carbon emission data to obtain preprocessed carbon emission data.

[0008] S3: Use the K-Means clustering algorithm to analyze the preprocessed carbon emission data to obtain the classification of carbon emission data types.

[0009] S4: Use the fractal theory model to analyze the classified carbon emission data to obtain the trend prediction of carbon emission data.

[0010] S5: Input the newly generated carbon emission-related data into the carbon emission calculation model, continuously update the model input, and make the data trend prediction fit the actual situation.

[0011] S6: Analyze the main driving factors behind the carbon emissions in the carbon emission trend prediction results that fit the actual situation, and plan emission reduction targets and measures.

[0012] According to an intelligent carbon emission data analysis method provided by the present invention, it includes: In step S1, the carbon emission data includes: energy consumption, production process parameters, non-carbon greenhouse gas emission data, production activity data, energy conversion efficiency data, equipment maintenance and operation data, and environmental parameter data.

[0013] According to an intelligent carbon emission data analysis method provided by the present invention, it includes: In step S2, the preprocessing of the carbon emission data includes: S21: Extract and construct valuable features in the carbon emission data.

[0014] S22: Use a standardization method to process the data to the same magnitude range.

[0015] S23: Identify and correct format errors and outliers in the data.

[0016] According to an intelligent carbon emission data analysis method provided by the present invention, it includes: According to an intelligent carbon emission data analysis method provided by the present invention, it includes: In step S3, the specific steps of using the K-Means clustering algorithm to analyze the preprocessed carbon emission data are as follows: S31: Use the elbow method to calculate and determine the squared clustering error for each K value, and set a thermal power unit as a unit.

[0017] S32: Randomly select the feature vectors of K units as the initial clustering centers, and select points with relatively dispersed and representative distributions as the initial centers.

[0018] S33: Calculate the distance from each unit to the K initial centers according to the number of clusters K and the initial clustering centers, assign each unit to the cluster where the nearest clustering center is located, recalculate the centroid of each cluster, and continuously iterate the process of assigning samples and updating the centroid.

[0019] S34: Determine whether the iteration process meets the maximum number of iterations. If so, stop the iteration and output the clustering result.

[0020] According to an intelligent carbon emission data analysis method provided by the present invention, it includes: In step S31, the formula for calculating the squared clustering error for each K value is:

[0021] where, ∑ xi∈Cj represents the summation over all sample points belonging to the j-th cluster, and ||x i − μ j || 2 is the square of the distance from the sample point x i to the centroid μ j of its belonging cluster C j .

[0022] According to an intelligent analysis method for carbon emission data provided by the present invention, it includes: In step S4, the specific steps for constructing a carbon emission calculation model are as follows: S41: Calculate the fractal dimension of the carbon emission data using the box dimension method.

[0023] S42: Analyze the complexity and internal laws of carbon emissions based on the fractal dimension and fractal features such as observed self-similarity.

[0024] S43: By comparing the fractal dimensions and their change trends in different regions or different time periods, find out the similarities and differences in carbon emission laws.

[0025] S44: Based on the analysis of fractal features and the quantification results of fractal dimensions, use an appropriate mathematical model to predict the future carbon emission development trends in different regions.

[0026] According to an intelligent analysis method for carbon emission data provided by the present invention, it includes: In step S41, the specific steps for calculating the fractal dimension of the carbon emission data using the box dimension method are as follows: S411: Set "boxes" of different sizes, where the "boxes" are square regions on a two-dimensional plane or time intervals in a time series.

[0027] S412: For each selected box scale, calculate the number of boxes required to cover all carbon emission data points.

[0028] S413: Construct coordinate points by taking the logarithm of the box side lengths at different box scales and the logarithm of the corresponding number of boxes required for coverage, and fit a straight line through linear regression. The slope of this straight line is the fractal dimension.

[0029] According to an intelligent analysis method for carbon emission data provided by the present invention, it includes: In step S412, the calculation formula for the number of boxes required to cover all carbon emission data points is expressed as:

[0030] where, r is the box length and T is the length of the entire time series.

[0031] An intelligent analysis method for carbon emission data provided by the present invention includes: in step S5, establishing a real-time data acquisition mechanism to enable newly generated carbon emission data to be timely connected to the analysis system and continuously update the model input, so that the analysis results always conform to the actual situation.

[0032] An intelligent analysis method for carbon emission data provided by the present invention includes: in step S6, deeply analyzing the results output by the model and using data visualization tools to display the analysis results in an intuitive chart form to formulate emission reduction strategies.

[0033] An intelligent analysis method for carbon emission data provided by the present invention analyzes the classified carbon emission data by using a fractal theory model, solving the problem of being unable to make response strategies in time according to the predicted carbon emissions in advance. Compared with the prior art, the present invention has the following advantages: 1. The present invention collects the carbon emission data of each production link of the enterprise in real time, including multi-dimensional data such as energy consumption data, production process parameters, carbon content of raw materials and fuels, and emissions of non-carbon greenhouse gases, which can comprehensively and accurately reflect the carbon emission status of the enterprise, lay a solid foundation for carbon emission accounting, enable the enterprise to timely understand its own carbon emission situation, provide strong data support for formulating emission reduction plans, and ensure that the data collection is comprehensive, accurate and reliable, enabling the enterprise to have a clear understanding of the carbon emission situation based on detailed data, which is conducive to the development of subsequent emission reduction related work.

[0034] 2. The present invention applies the K-Means clustering algorithm to the analysis of carbon emission data, thereby dividing the complex carbon emission data set into several simplified clusters, and each cluster represents a specific carbon emission pattern, which helps to reveal the carbon emission characteristics of different regions or production links, and at the same time can also discover potential carbon emission laws, and realizes the effective analysis of complex data by means of the algorithm, accurately excavates the laws behind the carbon emission data, facilitates the mastery of the carbon emission situation from different dimensions, and provides a basis for targeted emission reduction.

[0035] 3. The present invention analyzes the carbon emission data by applying a fractal theory model, which provides a method for deeply analyzing the complexity and internal laws of carbon emissions. By calculating parameters such as the fractal dimension, the self-similarity and scale-free characteristics of the carbon emission data are revealed, and based on the prediction of the fractal theory model, the future development trend of carbon emissions can be predicted, deeply analyzing the complexity and internal laws of carbon emissions, accurately estimating the future carbon emission trend, helping the enterprise to plan emission reduction measures in advance to cope with the carbon emission challenge, and also providing a scientific reference for the government to formulate long-term emission reduction targets. It can also evaluate the effect of emission reduction strategies by comparing the prediction results under different emission reduction scenarios, helping to formulate the optimal plan, deeply analyzing the carbon emission characteristics and realizing the prediction function, providing scientific and effective support for multiple parties in aspects such as emission reduction planning, target setting and strategy evaluation, and improving the scientificity and forward-looking of the overall emission reduction work.

[0036] 4. The carbon emission data analysis method proposed by the present invention plays an important role in supporting decision-making. Among them, intuitive data visualization tools can display the carbon emission analysis results in the form of charts, enabling decision-makers to quickly understand and grasp the carbon emission situation, display data on carbon emission-related activities, production data trends, and reveal information such as the sources and composition of carbon emissions, facilitating decision-makers to accurately evaluate the emission reduction effects and costs of different scenarios, formulate scientific and reasonable emission reduction policies, and timely discover and correct problems to ensure the achievement of emission reduction targets. In addition, the data visualization tool can also monitor the implementation effects of emission reduction measures, timely discover and correct existing problems, ensure the achievement of emission reduction targets, and ensure the smooth progress of emission reduction work. Description of the Drawings

[0037] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0038] Figure 1 It is a step diagram of an intelligent carbon emission data analysis method provided by an embodiment of the present invention.

[0039] Figure 2 It is a step diagram of constructing a carbon emission calculation model provided by an embodiment of the present invention. Detailed Embodiments

[0040] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention with reference to the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present invention fall within the scope of protection of the present invention.

[0041] The following will be combined with Figure 1 - Figure 2 Describe an intelligent carbon emission data analysis method of the present invention.

[0042] As Figure 1 shown, an intelligent carbon emission data analysis method provided by an embodiment of the present invention includes: S1: Real-time collect carbon emission data of each production link of the enterprise, including energy consumption, production process parameters, non-carbon greenhouse gas emission data, production activity data, energy conversion efficiency data, equipment maintenance and operation data, and environmental parameter data. Through devices such as sensors and smart meters, real-time monitor the consumption of energy such as coal, coke, natural gas, and electricity, as well as key parameters in the production process, such as furnace temperature, pressure, and flow rate. These data can accurately reflect the usage amount and usage time of energy, providing basic data for carbon emission accounting. There may be losses during the energy conversion process. Understanding the energy conversion efficiency helps to more accurately calculate the actual amount of energy used and the corresponding carbon emissions. The maintenance and operation status of equipment also affect carbon emissions, and data such as equipment maintenance records and failure rates need to be collected.

[0043] S2: Preprocess the carbon emission data to obtain preprocessed carbon emission data.

[0044] S21: According to the objectives of carbon emission analysis, extract and construct valuable features. Convert the energy consumption data into standard coal equivalent according to different energy types, which is convenient for uniformly measuring the scale of carbon emissions. For the production link, summarize the data according to different dimensions such as different processes and workshops, and construct features such as carbon emissions per unit product and carbon emission intensity of a single production line. Also combine the time dimension to construct carbon emission change features by day, month, and year, etc.

[0045] S22: Since the magnitudes of the indicators in the collected data may vary greatly, adopt a suitable standardization method to process the data into the same magnitude range, avoiding deviation effects on the results due to data magnitude problems in the subsequent analysis model.

[0046] S23: Identify and correct format errors, outliers, etc. in the data to ensure the data quality for subsequent analysis.

[0047] S3: Use the K-Means clustering algorithm to analyze the preprocessed carbon emission data to obtain the current situation of carbon emissions.

[0048] The K-Means clustering algorithm is based on distance measurement. Its core idea is to divide the samples in the dataset into K different clusters, so that the sum of the distances from the data points within each cluster to the center (centroid) of the cluster is the smallest. As mentioned before, it is applicable to situations where there is a general understanding of the number of categories to be clustered in advance, and the data is relatively regular in the feature space, showing a spherical or near-spherical distribution. When analyzing the carbon emission levels of multiple regions, if it can be roughly judged, based on past experience or some preliminary investigations, that the regions can be divided into high, medium, and low categories according to carbon emissions, then the K-Means clustering is a feasible choice.

[0049] In a specific embodiment, assume that a thermal power unit is taken as a unit, and there are carbon emission-related data of 50 units, including characteristic data such as carbon emissions per unit area, the proportion of coal in the energy structure, and the proportion of industrial output value in the regional GDP. Set K = 3, which means that these 50 units are to be clustered into 3 categories, namely high, medium, and low carbon emission levels. When the algorithm starts, first randomly select the feature vectors of K units as the initial clustering centers, and then calculate the distance from each of the remaining units to these 3 initial centers. The calculation formula is:

[0050] where x and y are the feature vectors of two units, and n is the number of features.

[0051] Assign each unit to the cluster where the nearest clustering center is located according to the distance. Recalculate the centroid of each cluster, that is, the average value of the feature vectors of all samples within the cluster, and repeat the above processes of assignment and centroid update, and continuously iterate until the samples within the cluster no longer change or reach the stopping conditions such as the preset number of iterations, and finally obtain the carbon emission level category to which each unit belongs.

[0052] S31: Use the elbow method to calculate and determine the sum of squared errors of clustering under each K value to find the optimal number of clusters K.

[0053] Try different K values through the elbow method and calculate the sum of squared errors of clustering (SSE) under each K value. As the K value increases, the number of samples in each cluster decreases, and the sample points are divided more finely, so the SSE usually gradually decreases. However, when the K value increases to a certain extent, the decrease amplitude of the SSE will tend to level off because increasing the number of clusters at this time no longer significantly improves the clustering effect. The elbow method is to find the turning point of the curve from a sharp decline to a gentle trend according to the trend of the SSE changing with the K value. The K value corresponding to this turning point is usually considered the optimal number of clusters. For a given data set , where xi represents the i-th sample point. Assume that after K-Means clustering, the data set is divided into K clusters, denoted as C 1 , C 2 , …, C K , and each cluster C j (j = 1, 2, …, K) has its centroid μ j , then the calculation formula for the sum of squared errors of clustering within the cluster is:

[0054] where ∑ xi∈Cj represents the sum over all sample points belonging to the j-th cluster. ||x i - μ j || 2Denote the sample point as \(x\). i to the centroid \(\mu\) of its belonging cluster \(C\). j The square of the distance is j .

[0055] Assume that the sample point \(x\). i and the centroid \(\mu\). j are both \(d\)-dimensional vectors, and its calculation formula is:

[0056]

[0057] In the formula, represents the value of the \(l\)-th feature of the \(i\)-th sample point \(x\). i represents the value of the \(l\)-th feature of the centroid \(\mu\) of the \(j\)-th clustering cluster. \(d\) is the dimension of the data. j

[0058] Plot the relationship curve between the value of \(K\) and SSE. The value of \(K\) corresponding to the turning point where the curve drops sharply and then flattens out is usually the more appropriate number of clusters.

[0059] S32: Randomly select the feature vectors of \(K\) units as the initial clustering centers, and select points with relatively dispersed and representative distributions as the initial centers.

[0060] The selection of the initial clustering centers is also crucial. Random selection is a simple method, but it may lead to different results in different runs. According to the characteristics of the data, select points with relatively dispersed and representative distributions in the data as the initial centers, or use methods such as taking the average result of multiple random selections to improve stability.

[0061] S33: Calculate the distances from each unit to the \(K\) initial centers according to the number of clusters \(K\) and the initial clustering centers, assign each unit to the cluster where the nearest clustering center is located, and recalculate the centroid of each cluster, continuously iterating the process of assigning samples and updating the centroid.

[0062] After determining \(K\) and the selection method of the initial clustering centers, according to the steps of iteratively assigning samples to clusters and updating cluster centers introduced above, use the K-Means algorithm in the Scikit-learn library in Python to implement clustering, passing in the corresponding parameters until the stopping condition is met.

[0063] S34: Determine whether the iteration process meets the maximum number of iterations. If so, stop the iteration and output the clustering result. The maximum number of iterations is set to 100 times, or the assignment of samples within the cluster no longer changes, etc. Finally, obtain the clustering cluster to which each sample belongs, and complete the clustering process.

[0064] Such as Figure 2 ​​As shown, S4: Use the fractal theory model to analyze the carbon emission data of the divided types to obtain the trend prediction of the carbon emission data.

[0065] The fractal theory was proposed by the mathematician Benoît Mandelbrot, aiming to describe geometric shapes or structures with irregular, fragmented, and self-similar characteristics. The core characteristics of fractals are self-similarity and scale invariance. Self-similarity means that no matter at what scale the object is observed, the local form is similar to the overall form, as if it is a reduced copy of the overall structure. The classic Koch snowflake is constructed by continuously performing specific subdivisions and additions on each side of an equilateral triangle, and it is found that a small segment of the local curve has a similar complex shape to the outline of the entire snowflake. Scale invariance indicates that there is no characteristic scale for fractal objects, and their forms and structures do not change their properties with the change of the observation scale, unlike ordinary geometric figures that have obvious different characteristic performances at different magnification multiples. Fractals are quantitatively described by the fractal dimension, which is different from the dimension concept in traditional Euclidean geometry. The fractal dimension can be a fractional value, used to measure the complexity of fractal objects, the ability to fill space, etc. Common methods for calculating the fractal dimension include the box dimension method, Hausdorff dimension, etc. When actually applied to carbon emission analysis, an appropriate calculation method will be selected according to the data characteristics.

[0066] S41: Calculate the fractal dimension of the carbon emission data using the box dimension method. The general application steps of the commonly used box dimension method in carbon emission analysis are as follows: S411: Set "boxes" of different sizes. The "boxes" are square regions in the two-dimensional plane or time intervals in the time series, etc., depending on whether the fractal being analyzed is in the spatial or time dimension. The side lengths of these boxes will be successively reduced in a certain proportion.

[0067] S412: For each selected box scale, calculate the number of boxes required to cover all carbon emission data points. The formula is expressed as:

[0068] In the formula, r is the box length, and T is the length of the entire time series.

[0069] S413: Take the logarithm of the box side lengths at different box scales and construct coordinate points with the logarithm of the corresponding number of boxes required for coverage. Then, fit a straight line through methods such as linear regression. The slope of this straight line is approximately equal to the fractal dimension to be obtained.

[0070] S42: Based on the calculated fractal dimension and fractal features such as observed self-similarity, deeply analyze the complexity degree and internal laws of carbon emissions. Generally speaking, the higher the fractal dimension, the more complex the structure of carbon emission data, and the greater the degree of its change affected by the interaction of multiple factors. For example, a relatively high fractal dimension of carbon emissions in a region may indicate a diverse industrial structure, complex and interacting types of energy use within the region, and being simultaneously affected by a combination of multiple policies and market factors, resulting in a complex fluctuation and irregularity in the changing trend of carbon emissions.

[0071] S43: By comparing the fractal dimensions and their changing trends in different regions or different time periods, identify the similarities and differences in the laws of carbon emissions. For example, if it is found that the fractal dimension of a newly developing region in a unit is gradually increasing, it indicates that as the region is constructed and developed, the carbon emission situation is becoming more and more complex, and more refined carbon management measures may be required. While the fractal dimension of a traditional old industrial area has decreased, it may mean that through measures such as industrial upgrading and energy structure optimization, the complexity of carbon emissions has decreased, and the emission reduction effect is significant.

[0072] S44: Based on the analysis of fractal features and the quantification results of fractal dimensions, use appropriate mathematical models to predict the future development trends of carbon emissions in different regions. Since fractals reflect an inherent regularity, even in complex changes, it is possible to speculate on the possible future trends of carbon emissions based on this regularity.

[0073] In a specific embodiment, according to the fractal features of carbon emissions in a region over a past period of time and the stability of the fractal dimension, predict whether the carbon emissions in the region will continue to show a similar fluctuating increase or tend to be stable under the existing development model in the next few years. Key emission units formulate a reasonable regional energy layout plan based on these prediction results. For example, for regions with high carbon emission complexity and a growing trend, increase the layout and construction of clean energy supply facilities, and guide industries to transform towards low-carbon. For regions with relatively stable carbon emissions and a low fractal dimension, further consolidate the existing emission reduction achievements, optimize the energy distribution efficiency, etc., so as to achieve a scientific and reasonable carbon emission reduction plan, and predict the carbon emissions of enterprises in advance by predicting factors such as future production plans and market demands.

[0074] S5: Establish a real-time data collection mechanism so that newly generated carbon emission-related data can be timely accessed into the analysis system, in order to continuously update the model input and make the analysis results always conform to the actual situation. As the production processes of enterprises change, the energy structure is adjusted, etc., according to the new data feedback and the deviation between the analysis results and the actual situation, re-train and optimize the model in a timely manner to ensure the accuracy and effectiveness of the intelligent analysis method, and continuously serve carbon emission management.

[0075] S6: Conduct an in-depth analysis of the results output by the model to identify the main driving factors behind them. Use professional data visualization tools or professional business intelligence software to present the analysis results in an intuitive chart form, facilitating quick understanding and grasp of the carbon emission situation by decision-makers, providing strong support for formulating emission reduction strategies, etc., helping enterprises reasonably plan emission reduction targets and measures, and proactively address challenges brought about by policies such as the carbon border adjustment mechanism.

[0076] The present invention provides an intelligent analysis method for carbon emission data. By analyzing the laws of carbon emission data and establishing a prediction model, it aims to solve the deficiencies in the prior art, where it is impossible to reasonably predict carbon emission data, making it difficult to effectively verify the rationality of the data and quickly formulate response strategies. The beneficial effects achieved are as follows: By comparing the predicted data with the actual carbon emission data, the accuracy and rationality of the actual data can be verified. This helps to identify potential errors in data recording, statistics, or reporting, thereby improving the overall quality of carbon emission data. The reasonable predicted data can be used as a benchmark for comparative analysis with the actual carbon emission data to identify data anomalies or deviations, providing a basis for further investigation and correction. Based on accurate predicted data, the government and enterprises can formulate emission reduction targets and plans in advance to ensure effective response measures are taken before carbon emissions reach the peak or exceed the limit. The predicted data can also help enterprises evaluate the effects and costs of different emission reduction scenarios, thereby selecting the optimal emission reduction strategy, reducing emission reduction costs, and improving emission reduction efficiency.

[0077] By predicting the elemental carbon content of coal, enterprises can more precisely control the coal combustion process, ensure the full combustion of coal, and reduce unburned carbon losses. This helps to improve coal combustion efficiency, reduce coal consumption costs, and reduce pollutant emissions caused by incomplete combustion. The elemental carbon content prediction model can serve as an important tool for enterprise carbon management, helping enterprises to monitor and manage carbon emissions during the coal combustion process in real time. By continuously optimizing the elemental carbon content prediction model of coal, enterprises can further improve their carbon management capabilities and achieve more precise carbon emission control and emission reduction targets.

[0078] Energy consumption prediction during the power generation and heat supply process can help enterprises better understand their energy usage situation, thereby adopting more effective energy-saving measures. By optimizing energy distribution and scheduling, enterprises can improve energy utilization efficiency, reduce energy costs, and reduce carbon emissions caused by energy waste. Accurate predicted data can help enterprises make advance energy reserve and scheduling plans to ensure stable power supply and heat supply during peak energy demand periods. This helps to improve the service quality and customer satisfaction of enterprises, and also helps to reduce economic losses and social impacts caused by energy shortages or interruptions.

[0079] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., located in one place or distributed to multiple network units. Select some or all of the modules according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative work.

[0080] Through the description of the above embodiments, those skilled in the art clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on such an understanding, the above technical solution, in essence, or the part that contributes to the prior art is embodied in the form of a software product. This computer software product is stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiments.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A carbon emission data intelligent analysis method, characterized in that: include: S1: Real-time collection of carbon emission data from all production links of thermal power companies; S2: preprocessing the carbon emission data to obtain preprocessed carbon emission data; S3: Analyze the pre-processed carbon emission data using a K-Means clustering algorithm to obtain a classification of carbon emission data types; S4: using a fractal model to analyze the carbon emission data divided into the carbon emission data types to obtain a trend forecast of the carbon emission data; S5: Input the newly generated carbon emission related data into the carbon emission calculation model, continuously update the model input, and obtain the prediction results that fit the actual trend; S6: Analyze the main driving factors behind the carbon emissions forecast results that are in line with actual trends, and plan emission reduction targets and measures.

2. According to claim 1, a carbon emission data intelligent analysis method is characterized in that: In step S1, the carbon emission data includes: energy consumption, production process parameters, raw material fuel test data, non-carbon greenhouse gas emission data, production activity data, energy conversion efficiency data, equipment maintenance and operation data, and environmental parameter data.

3. The method for intelligent analysis of carbon emission data according to claim 1, characterized in that: In step S2, preprocessing the carbon emission data includes: S21: extracting and constructing valuable features from the carbon emission data; S22: Use standardized methods to process data to the same magnitude range; S23: Identify and correct format errors and outliers in the data.

4. The method for intelligent analysis of carbon emission data according to claim 1, characterized in that: In step S3, the specific steps of using the K-Means clustering algorithm to analyze the pre-processed carbon emission data are: S31: Use the elbow method to calculate and determine the sum of squares of clustering errors under each K value, assuming that one thermal power unit is a unit; S32: Randomly select the feature vectors of K units as the initial cluster centers, and select dispersed and representative points as the initial centers; S33: Calculate the distance between each unit and the K initial centers according to the number of clusters K and the initial cluster centers, assign each unit to the cluster with the nearest cluster center, recalculate the centroid of each cluster, and continuously iterate the process of assigning samples and updating the centroid; S34: Determine whether the iterative sample allocation process meets the maximum number of iterations, if yes, stop the iteration and output the clustering result.

5. A carbon emission data intelligent analysis method according to claim 4, characterized in that: In step S31, the clustering error square calculation formula under each K value is: In the formula, ∑ xi∈Cj To sum all sample points belonging to the jth cluster, |||x i −μ j ∣∣ 2 is the sample point x i To its own cluster C j The center of mass μ j The square of the distance.

6. The carbon emission data intelligent analysis method according to claim 1 is characterized in that: In step S4, the specific steps of constructing the carbon emission calculation model are: S41: Calculate the fractal dimension of carbon emission data using the box dimension method; S42: Analyze the complexity and internal laws of carbon emissions based on the fractal dimension and the observed self-similar fractal characteristics; S43: By comparing the fractal dimensions and their changing trends in different units or time periods, find out the similarities and differences in carbon emission patterns; S44: Based on the analysis of fractal characteristics and the quantification of fractal dimensions, mathematical models are used to predict the future carbon emission trends in different regions.

7. A carbon emission data intelligent analysis method according to claim 6, characterized in that: In step S41, the specific steps of using the box dimension method to calculate the fractal dimension of carbon emission data are: S411: Setting "boxes" of different sizes, where a "box" is a square area on a two-dimensional plane or a time interval in a time series; S412: For each selected box size, calculate the number of boxes required to cover all carbon emission data points; S413: The logarithm of the box side lengths at different box scales and the logarithm of the corresponding number of boxes required for coverage are constructed into coordinate points, and a straight line is fitted through linear regression, and the slope of the straight line is the fractal dimension.

8. A carbon emission data intelligent analysis method according to claim 7, characterized in that: In step S412, the calculation formula for the number of boxes required to cover all carbon emission data points is expressed as: Where r is the box length and T is the length of the entire time series.

9. The carbon emission data intelligent analysis method according to claim 1, characterized in that: In step S5, a real-time data collection mechanism is established to allow newly generated carbon emission data to be instantly connected to the analysis system, and the model input is continuously updated to ensure that the analysis results always fit the actual situation.

10. The method for intelligent analysis of carbon emission data according to claim 1, characterized in that: In step S6, the results of the model output are deeply analyzed, and the analysis results are displayed in the form of intuitive charts using data visualization tools to assist in supporting the formulation of emission reduction strategies.

Images