Carbon neutralization layering optimization method and device
By collecting and constructing a hierarchical structure of historical carbon emission data, and combining random forests and deep learning models, the overfitting problem in carbon neutrality path optimization in existing technologies is solved, enabling accurate prediction and dynamic optimization of carbon emissions, and supporting the achievement of carbon neutrality goals.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-31
- Publication Date
- 2026-04-07
AI Technical Summary
Existing carbon neutrality path optimization methods mostly employ decision tree models, which can easily lead to overfitting when there are too many features. This reduces the model's generalization ability and the accuracy and stability of carbon emission prediction, making it difficult to meet the refined management requirements of carbon neutrality path optimization.
Historical carbon emission data is collected according to multiple predefined carbon neutrality levels to construct hierarchical data. Pre-trained random forest models and deep learning models are used for accurate prediction. Combined with carbon emission baselines, target optimization strategies are determined to achieve dynamic optimization and control.
It significantly improves the accuracy and stability of carbon emission prediction, provides dynamic optimization and precise control from macro to micro levels, and provides efficient technical support for achieving the goal of carbon neutrality.
Smart Images
Figure CN121809775A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of artificial intelligence technology, and in particular to a carbon neutrality hierarchical optimization method and apparatus. Background Technology
[0002] As global climate change becomes increasingly severe, achieving the goals of "carbon peaking" and "carbon neutrality" has become a global consensus. The core key to achieving this goal lies in the scientific and precise optimization of low-carbon transition pathways in key sectors such as energy, industry, transportation, and construction. This requires leveraging advanced computational models to analyze a vast array of multi-dimensional influencing factors (such as technology costs, energy structure, and economic growth) to predict carbon emission trajectories under different strategies, thereby identifying the path with the lowest cost, highest efficiency, and optimal feasibility.
[0003] Currently, most existing carbon neutrality pathway optimization methods employ decision tree models to predict carbon emissions and conduct preliminary assessments of pathway options. However, when the number of features is excessive, decision tree models are prone to overfitting to noise and details in the training data, leading to a decline in generalization ability. This results in a significant reduction in the accuracy and stability of the carbon emission predictions, making it difficult to meet the needs of refined management in carbon neutrality pathway optimization. Consequently, this fundamentally restricts the feasibility of achieving carbon neutrality with minimal socioeconomic costs or maximum overall benefits. Summary of the Invention
[0004] To address the aforementioned issues, this application provides a carbon neutrality stratified optimization method and apparatus, aiming to improve the accuracy of carbon emission prediction, achieve phased dynamic optimization of carbon emissions, and provide efficient technical support for achieving the carbon neutrality goal.
[0005] The embodiments of this application disclose the following technical solutions:
[0006] A first aspect of this application provides a carbon neutrality hierarchical optimization method, the method comprising:
[0007] Collect historical carbon emission data corresponding to each predefined carbon neutrality level; the historical carbon emission data includes at least one of the following categories: direct carbon emission data and indirect carbon emission data of the target object in a historical period, and carbon emission data of the entire product life cycle;
[0008] Data analysis is performed on multiple historical carbon emission data sets, and a hierarchical data structure is constructed; the hierarchical data structure is aggregated data organized hierarchically according to the dimensions of time, space, and emission source.
[0009] Using a pre-trained random forest model, the carbon emissions and carbon emission levels of the target object in the future time period and at each of the carbon neutrality levels are predicted based on the historical carbon emission data.
[0010] Using a pre-trained deep learning model, based on the historical carbon emission data, carbon emission amount, carbon emission level, and hierarchical structure data corresponding to each carbon neutrality level, the target carbon emission amount of the target object in the future period corresponding to each carbon neutrality level is predicted.
[0011] Based on the target carbon emissions and carbon emission baselines, a target optimization strategy matching the target carbon emissions is determined from multiple preset optimization strategies; the target optimization strategy is used to optimize and regulate the target carbon emissions of the target object in a future time period.
[0012] In an optional implementation, the step of analyzing multiple historical carbon emission data sets and constructing a hierarchical data structure includes:
[0013] Data cleaning and preprocessing are performed on multiple historical carbon emission data sets, and the cleaned historical carbon emission data is converted to obtain converted historical carbon emission data.
[0014] Statistical analysis is performed on the converted historical carbon emission data to determine the carbon emission data corresponding to each of the multiple first key factors; the multiple first key factors include: direct carbon emission related factors, indirect carbon emission related factors, and supply chain and consumption behavior related factors;
[0015] Using a pre-trained linear relationship model, the carbon emissions corresponding to each of the first key factors are predicted based on the carbon emission data corresponding to each of the first key factors; the carbon emissions corresponding to the first key factor are the predicted carbon emissions of the target object under the influence of the first key factor in the future period.
[0016] The hierarchical data is constructed based on the carbon emissions corresponding to each of the first key factors.
[0017] In an optional implementation, constructing the hierarchical data based on the carbon emissions corresponding to each of the first key factors includes:
[0018] Acquire multiple second key factor data; the multiple second key factor data include: energy factor data, economic factor data, population factor data, and technology factor data related to the carbon emissions of the target object in historical periods;
[0019] Using a pre-trained multiple regression model, the comprehensive carbon emissions are predicted based on the data of the multiple second key factors and the transformed historical carbon emission data; the comprehensive carbon emissions are the predicted total carbon emissions of the target object under the influence of multiple second key factors in the future period.
[0020] The hierarchical data is constructed based on the carbon emissions corresponding to each of the first key factors and the overall carbon emissions.
[0021] In an optional implementation, constructing the hierarchical data based on the carbon emissions corresponding to each of the first key factors and the overall carbon emissions includes:
[0022] The converted historical carbon emission data is input into a pre-trained autoregressive integral moving average model, which performs time-series analysis on the converted historical carbon emission data to predict the carbon emissions of the target object in future periods.
[0023] Data analysis is performed on the carbon emissions corresponding to each of the first key factors, the total carbon emissions, and the carbon emissions of the target object in the future period, as well as multiple historical carbon emission data, to construct the hierarchical data structure.
[0024] In an optional implementation, the predefined multiple carbon neutrality levels include: a direct carbon neutrality level, an indirect carbon neutrality level, and a product lifecycle carbon neutrality level;
[0025] The direct carbon neutrality hierarchy includes carbon emission data generated by emission sources directly controlled by the target object;
[0026] The indirect carbon neutrality hierarchy includes carbon emission data from the direct carbon neutrality hierarchy, as well as carbon emission data from external energy production sources related to the target object.
[0027] The product lifecycle carbon neutrality hierarchy includes carbon emission data from the indirect carbon neutrality hierarchy, as well as carbon emission data of the product generated by the target object throughout its lifecycle.
[0028] In an optional implementation, the carbon emission baseline includes: a direct carbon emission baseline corresponding to the direct carbon neutrality level; the carbon neutrality stratification optimization method further includes:
[0029] Based on the historical carbon emission data corresponding to the direct carbon neutrality level, calculate the average direct carbon emission of the target object during the historical period;
[0030] The average value is used as the baseline for direct carbon emissions.
[0031] In an optional implementation, the carbon emission baseline includes: an indirect carbon emission baseline corresponding to the indirect carbon neutrality level; the carbon neutrality stratification optimization method further includes:
[0032] The energy consumption data and standard carbon emission conversion factor of the target object during the historical period are obtained; the standard carbon emission conversion factor is a benchmark conversion factor for converting energy consumption data into carbon emissions.
[0033] Based on the standard carbon emission conversion factor and the energy consumption data, the indirect carbon emissions are determined, and the indirect carbon emissions are used as the indirect carbon emission baseline.
[0034] In an optional implementation, the carbon emission baseline includes: a product lifecycle carbon emission baseline corresponding to the product lifecycle carbon neutrality level; the carbon neutrality stratification optimization method further includes:
[0035] A life cycle assessment is performed on the carbon emission data of the product throughout its entire life cycle to obtain the carbon emission amount of the product throughout its entire life cycle;
[0036] The carbon emissions throughout the product's entire life cycle are used as the baseline for the product's entire life cycle carbon emissions.
[0037] In an optional implementation, after determining a target optimization strategy matching each of the target carbon emissions from a plurality of preset optimization strategies based on each of the target carbon emissions and carbon emission baselines, the method further includes:
[0038] Based on the target optimization strategy, the hierarchical data, the historical carbon emission data, carbon emission amount, carbon emission level and the target carbon emission amount for each carbon neutrality level, a carbon neutrality optimization report is generated.
[0039] The carbon neutrality optimization report is stored in the storage space, and a corresponding shared access link is generated.
[0040] A second aspect of this application provides a carbon neutrality stratification optimization apparatus, the apparatus comprising:
[0041] The data acquisition module is used to collect historical carbon emission data corresponding to each carbon neutrality level according to multiple predefined carbon neutrality levels. The historical carbon emission data includes at least one of the following categories: direct carbon emission data and indirect carbon emission data of the target object in a historical period, and carbon emission data of the entire product life cycle.
[0042] The data analysis module is used to perform data analysis on multiple historical carbon emission data and construct hierarchical data; the hierarchical data is aggregated data organized hierarchically according to the dimensions of time, space and emission source.
[0043] The first prediction module is used to use a pre-trained random forest model to predict the carbon emissions and carbon emission levels of the target object in the future time period and at each of the carbon neutrality levels based on the historical carbon emission data.
[0044] The second prediction module is used to use a pre-trained deep learning model to predict the target carbon emissions of the target object in the future period corresponding to each of the carbon neutrality levels, based on the historical carbon emission data, carbon emission amount, carbon emission level and the hierarchical structure data corresponding to each of the carbon neutrality levels.
[0045] The optimization strategy determination module is used to determine a target optimization strategy that matches each of the target carbon emissions from a plurality of preset optimization strategies, based on each of the target carbon emissions and the carbon emission baseline; the target optimization strategy is used to optimize and regulate the target carbon emissions of the target object in a future time period.
[0046] Compared with the prior art, this application has the following beneficial effects:
[0047] In this technical solution, firstly, historical carbon emission data corresponding to each predefined carbon neutrality level is collected. This historical carbon emission data can cover the direct and indirect carbon emission data of the target object during historical periods, as well as the carbon emission data of the entire product life cycle, significantly expanding the breadth and completeness of data coverage and avoiding prediction bias caused by incomplete data, thus providing a clear data foundation for carbon neutrality path optimization. Secondly, data analysis is performed on multiple historical carbon emission data, and aggregated data (i.e., hierarchical structure data) is constructed according to the dimensions of time, space, and emission source. This data can be adapted to the subsequent hierarchical prediction and optimization process, providing a data foundation for accurately locating carbon emission issues in different dimensions.
[0048] Then, using a pre-trained random forest model, based on historical carbon emission data, the carbon emissions and levels corresponding to each carbon neutrality level for the target object in the future are accurately predicted, providing an accurate data foundation for subsequent carbon neutrality path optimization. Next, using a pre-trained deep learning model, based on historical carbon emission data, carbon emission amounts, carbon emission levels, and hierarchical structure data corresponding to each carbon neutrality level, the target carbon emissions corresponding to each carbon neutrality level in the future are accurately predicted. This achieves effective [prediction / prediction] by capturing the complex nonlinear relationships and deep temporal dependencies in the aforementioned data. Overcoming the overfitting problem that easily arises when the model has too many features, this method significantly improves the accuracy, stability, and generalization ability of carbon emission prediction. Finally, based on the accurate prediction results (i.e., the carbon emission amounts of each target) and the carbon emission baseline, a target optimization strategy matching each target carbon emission amount is determined from multiple preset optimization strategies. Through this target optimization strategy, the target carbon emission amount of the target object in the future period can be optimized and controlled, thereby realizing the dynamic optimization and precise control of carbon emissions at different stages and levels from macro-targets to micro-levels, providing efficient technical support for achieving the carbon neutrality goal. Attached Figure Description
[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0050] Figure 1 A flowchart of a carbon neutrality stratification optimization method provided in this application embodiment;
[0051] Figure 2 A flowchart illustrating the process of constructing hierarchical data is provided as an embodiment of this application;
[0052] Figure 3 A flowchart illustrating another hierarchical data construction process provided in this application embodiment;
[0053] Figure 4 A flowchart illustrating another hierarchical data construction process provided in this application embodiment;
[0054] Figure 5 This is a schematic diagram of a carbon neutralization stratification optimization device provided in an embodiment of this application. Detailed Implementation
[0056] As described earlier, current carbon neutrality path optimization methods mostly employ decision tree models to predict carbon emissions and conduct preliminary assessments of path options. However, when the number of features is too large, decision tree models are prone to overfitting to noise and details in the training data, leading to a decline in the model's generalization ability. This results in a significant reduction in the accuracy and stability of the carbon emissions predicted by the model, making it difficult to meet the needs of refined management in carbon neutrality path optimization. Consequently, it fundamentally restricts the feasibility of achieving carbon neutrality with minimal socioeconomic costs or maximum overall benefits.
[0057] The inventors have proposed a tiered optimization method for carbon neutrality. This method first collects historical carbon emission data corresponding to each predefined carbon neutrality level. This historical carbon emission data covers both direct and indirect carbon emission data of the target object within a historical period, as well as carbon emission data throughout the product's entire lifecycle. This significantly expands the breadth and completeness of the data coverage, avoiding prediction bias caused by incomplete data, thus providing a clear data foundation for carbon neutrality path optimization. Secondly, the method analyzes multiple historical carbon emission data sets and constructs aggregated data (i.e., hierarchical data) organized hierarchically according to time, space, and emission source dimensions. This hierarchical structure data is adaptable to subsequent tiered prediction and optimization processes, providing a data foundation for accurately identifying carbon emission issues at different levels.
[0058] Then, using a pre-trained random forest model, based on historical carbon emission data, the carbon emissions and levels corresponding to each carbon neutrality level for the target object in the future are accurately predicted, providing an accurate data foundation for subsequent carbon neutrality path optimization. Next, using a pre-trained deep learning model, based on historical carbon emission data, carbon emission amounts, carbon emission levels, and hierarchical structure data corresponding to each carbon neutrality level, the target carbon emissions corresponding to each carbon neutrality level in the future are accurately predicted. This achieves effective [prediction / prediction] by capturing the complex nonlinear relationships and deep temporal dependencies in the aforementioned data. Overcoming the overfitting problem that easily arises when the model has too many features, this method significantly improves the accuracy, stability, and generalization ability of carbon emission prediction. Finally, based on the accurate prediction results (i.e., the carbon emission amounts of each target) and the carbon emission baseline, a target optimization strategy matching each target carbon emission amount is determined from multiple preset optimization strategies. Through this target optimization strategy, the target carbon emission amount of the target object in the future period can be optimized and controlled, thereby realizing the dynamic optimization and precise control of carbon emissions at different stages and levels from macro-targets to micro-levels, providing efficient technical support for achieving the carbon neutrality goal.
[0059] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.
[0060] Keyword definition:
[0061] An Energy Management System (EMS) is a comprehensive solution designed to help businesses and organizations optimize energy use, reduce costs, and improve sustainability.
[0062] Life Cycle Assessment (LCA) is a tool used to assess the environmental impact of a product, process, or service throughout its entire life cycle, from raw material acquisition, production, transportation, use to final disposal.
[0063] Hash function: A mathematical function that transforms input (usually a string or block of data) into a fixed-size string (called a hash value or digest) using a specific algorithm.
[0064] Bootstrap sampling, also known as bootstrap sampling, is a nonparametric statistical method. The main idea of bootstrap sampling is to simulate the population distribution by drawing a large number of bootstrap samples with replacement from the original sample, thereby making statistical inferences. This method avoids the limitations of assumptions about the data distribution and can better handle nonparametric statistical problems.
[0065] BI tools: refers to Business Intelligence, which is a set of processes and tools for analyzing business data. It aims to help businesses transform data into actionable insights to make more informed decisions. Common BI tools include Tableau, Power BI, Qlik Sense, and Looker.
[0066] Method Implementation Examples
[0067] This application provides an embodiment of a carbon neutrality stratification optimization method. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although the flowchart shows a logical order, in some cases, the steps shown or described may be executed in a different order than that shown here.
[0068] See Figure 1 The figure is a flowchart of a carbon neutrality stratification optimization method provided in an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps:
[0069] Step S101: Collect historical carbon emission data corresponding to each predefined carbon neutrality level.
[0070] In one optional embodiment, a carbon neutrality hierarchical optimization system can serve as the execution subject of the carbon neutrality hierarchical optimization method of this application. For ease of description, the carbon neutrality hierarchical optimization system will be referred to simply as the system below.
[0071] In step S101, the historical carbon emission data includes at least one of the following categories: direct and indirect carbon emission data of the target object during a historical period, and carbon emission data throughout the product's entire life cycle.
[0072] In this application embodiment, the predefined multiple carbon neutrality levels include: direct carbon neutrality level C0, indirect carbon neutrality level C1, and product lifecycle carbon neutrality level C2; wherein, direct carbon neutrality level C0 includes carbon emission data generated by emission sources directly controlled by the target object (i.e., direct and easily measurable carbon emission sources), which may include industrial processes, transportation, power plants, stationary combustion sources, waste treatment, and agricultural activities; indirect carbon neutrality level C1 includes carbon emission data from the direct carbon neutrality level, as well as external energy production sources related to the target object (i.e., short-term carbon emission sources). Carbon emission data for emission sources that can achieve carbon neutrality within the period, including short-term carbon neutrality sources such as electricity consumption, heat consumption, transportation, industrial processes, agricultural activities, and waste disposal; and carbon emission data for the product lifecycle carbon neutrality level C2, including indirect carbon neutrality level carbon emission data and carbon emission data for the target product throughout its lifecycle. The carbon emission data throughout the lifecycle corresponds to longer-term carbon neutrality reserves or long-term carbon neutrality projects, which include the entire supply chain, consumption behavior, forest carbon sinks, and negative emission technologies.
[0073] The system can collect historical carbon emission data corresponding to each of the predefined carbon neutrality levels. This historical data covers direct and indirect carbon emission data of the target object over a historical period, as well as carbon emission data throughout the product's entire lifecycle. This ensures data completeness and multidimensionality, avoiding prediction bias caused by incomplete data, and providing a clear data foundation for optimizing the carbon neutrality path. Specifically, the system can obtain direct carbon emission data through real-time monitoring of direct carbon emission sources such as factory chimneys and vehicle exhaust using high-precision sensors (such as gas analyzers and flow meters). The system can collect indirect carbon emission data such as electricity consumption and natural gas usage through an energy management system. Furthermore, the system can collect carbon emission data from the supply chain and consumption behavior (i.e., carbon emission data throughout the product's entire lifecycle) using lifecycle assessment methods (such as LCA). This includes direct carbon emissions during the production process, indirect carbon emissions such as energy consumption, and carbon emissions from other stages such as the upstream and downstream of the supply chain and consumption behavior.
[0074] Optionally, the system uses a life cycle assessment approach to collect carbon emission data from the supply chain and consumption behavior, and the specific steps are as follows:
[0075] Step 1: Determine the reasons for conducting LCA and the expected results, and determine the functional units of the product and the scope of the LCA study.
[0076] Step 2: Collect all relevant process data for LCA and compile a list of all input and output data for all materials and energy involved in each process, while evaluating the accuracy and reliability of the input and output data.
[0077] Step 3: Classify the various environmental impact factors in the list, convert each type of environmental impact into a common unit of measurement, and compare the characteristic results with parameters to assess their relative importance.
[0078] Step 4: Summarize the results of the inventory analysis and impact assessment to form a comprehensive report, assess the uncertainty and sensitivity of the LCA results, identify key influencing factors, and propose improvement suggestions and optimization schemes based on the LCA results.
[0079] Step S102: Perform data analysis on multiple historical carbon emission data and construct hierarchical data structure.
[0080] In step S102, the hierarchical data is aggregated data organized hierarchically according to the dimensions of time, space and emission source.
[0081] In this embodiment, the system can perform data analysis on multiple historical carbon emission data and construct aggregated data (i.e., hierarchical data) organized hierarchically according to the dimensions of time, space and emission source. This can be adapted to subsequent hierarchical prediction and optimization processes, providing a data foundation for accurately locating carbon emission issues in different dimensions.
[0082] Specifically, see Figure 2 The figure is a flowchart of a hierarchical data construction process provided in an embodiment of this application. The process includes the following steps:
[0083] Step S1021: Perform data cleaning and preprocessing on multiple historical carbon emission data, and convert the units of the cleaned historical carbon emission data to obtain the converted historical carbon emission data.
[0084] In this embodiment, data cleaning preprocessing may include identifying and removing unreasonable data values, filling in missing data, and deleting duplicate data records. The system can use box plots to identify and remove unreasonable data values from multiple historical carbon emission data sets, and then identify and fill in missing values through data visualization or statistical methods. Specifically, when processing missing values, if the number of missing values is small and randomly distributed, the system can delete the relevant records; otherwise, it can fill in the missing data using interpolation. When deleting duplicate data records, the system can identify duplicates by checking the unique identifier of each data record and use a hash function to calculate the hash value of each data record to identify identical data records. Duplicate data records are then deleted, retaining only one record. If duplicate records have differences, the system can merge the data as needed.
[0085] Furthermore, the system can perform unit conversion on the cleaned historical carbon emission data to unify data from different units into a standard unit, resulting in converted historical carbon emission data. For example, energy consumption can be converted from kilowatt-hours (kWh) to megajoules (MJ), or internationally accepted standard units can be used.
[0086] It should be noted that when the system converts the cleaned historical carbon emission data to units, it must ensure that the timestamps of all data are consistent to facilitate time series analysis. For periodic data (such as annual reports), it must ensure that the data covers the same time range.
[0087] In one feasible implementation, after the system converts the cleaned historical carbon emission data to different units, it can also normalize the converted data, scaling it to a preset range (such as 0 to 1) to facilitate subsequent comparison and analysis. Normalization may include minimum-maximum normalization and Z-score normalization, wherein minimum-maximum normalization is specifically as shown in formula (1), and Z-score normalization is specifically as shown in formula (2).
[0088] (1)
[0089] (2)
[0090] Where x is the numerical value in the converted data; The values are the minimum-maximum normalized values; max(x) is the maximum value in the transformed data, and min(x) is the minimum value in the transformed data. σ is the normalized Z-score; μ is the mean of the transformed data; and σ is the standard deviation of the transformed data.
[0091] Optionally, after performing normalization calculations, the system can check the consistency within the data to ensure there are no contradictions between different parts of the data, and compare it with known benchmark data or industry standards to verify the rationality of the data.
[0092] Step S1022: Perform statistical analysis on the converted historical carbon emission data to determine the carbon emission data corresponding to each of the multiple primary key factors.
[0093] In step S1022, several primary key factors include: direct carbon emission-related factors (such as carbon emissions from industrial processes, transportation, power plants, stationary combustion sources, waste treatment, and agricultural activities), indirect carbon emission-related factors (such as carbon emissions from electricity consumption, heat consumption, transportation, industrial processes, agricultural activities, and waste treatment), and supply chain and consumption behavior-related factors (such as carbon emissions from the entire supply chain, consumption behavior, forest carbon sinks, and negative emission technologies).
[0094] In this embodiment of the application, statistical analysis may include the calculation of basic statistics such as mean, median, and standard deviation; the system can calculate basic statistics such as mean, median, and standard deviation, draw time series graphs, and determine the corresponding carbon emission data based on the above basic statistics, time series graphs, and multiple preset first key factors.
[0095] Optionally, the formula for calculating the average value (Mean) is as follows:
[0096]
[0097] Where, x i It is the i-th data point; n is the total number of data points.
[0098] Optionally, the median is the middle value when a set of data is arranged in ascending order. If the number of data points n is odd, the median is the (n+1) / 2th data point; if the number of data points n is even, the median is the average of the (n / 2)th and (n / 2+1)th data points.
[0099] Alternatively, the formula for the standard deviation (SD) can be as follows:
[0100]
[0101] Where, x i It is the i-th data point; It is the average of all data; n is the total number of data points.
[0102] Step S1023: Using a pre-trained linear relationship model, predict the carbon emissions corresponding to each first key factor based on the carbon emission data corresponding to each first key factor.
[0103] In step S1023, the carbon emissions corresponding to the first key factor are the predicted carbon emissions of the target object under the influence of the first key factor in the future period.
[0104] In this embodiment, the pre-trained linear relationship model can be a simple linear regression model. The system can use the simple linear regression model to predict the carbon emissions corresponding to each first key factor based on the carbon emission data corresponding to each first key factor. The specific formula for the simple linear regression model is shown below:
[0105]
[0106] Where Y represents the carbon emissions output by the simple linear regression model; X represents the first key factor. For the intercept term; These are the regression coefficients; This is the error term.
[0107] Alternatively, the system can use the least squares method to fit the model to estimate the parameters. and And calculate R 2 The model's fit is evaluated using the mean squared error (ESM) and root mean squared error (RESM).
[0108] It should be noted that a simple linear regression model can clearly describe the linear relationship between energy consumption and carbon emissions, thereby accurately predicting carbon emission trends.
[0109] Step S1024: Construct hierarchical data based on the carbon emissions corresponding to each primary key factor.
[0110] In this embodiment, the system can construct hierarchical data based on the carbon emissions corresponding to each first key factor and the predicted total carbon emissions under the influence of multiple second key factors (such as economic development, industrial structure, energy consumption structure, technological level, and climate factors). Specifically, see [link to relevant documentation]. Figure 3 The figure is a flowchart of another hierarchical data construction process provided in an embodiment of this application, which includes the following steps:
[0111] Step S10241: Obtain data on multiple second key factors.
[0112] In step S10241, the multiple second key factor data include: energy factor data, economic factor data, population factor data, and technological factor data related to the target object's carbon emissions over historical periods. In addition, the multiple second key factor data also include: climate factor data.
[0113] In this embodiment, the system can collect historical carbon emission data and data on possible influencing factors, perform data cleaning, and then convert data from different units into a consistent unit to ensure data consistency and comparability. Through correlation coefficient matrix or scatter plot, the system can preliminarily determine the relationship between various influencing factors and carbon emissions, and obtain the aforementioned data on multiple second key factors.
[0114] Step S10242: Using a pre-trained multivariate regression model, predict the overall carbon emissions based on multiple second key factor data and converted historical carbon emission data.
[0115] In step S10242, the comprehensive carbon emissions are the predicted total carbon emissions of the target object under the influence of multiple second key factors in the future time period.
[0116] In this embodiment, the pre-trained multiple regression model can select multiple influencing factors (such as energy factors, economic factors, population factors, technological factors, and climate factors), and its model formula is as follows:
[0117]
[0118] Where Y is the total carbon emissions output by the multiple regression model; Each corresponds to data from multiple influencing factors; For the intercept term; These are the regression coefficients; This is the error term.
[0119] The system utilizes the pre-trained multiple regression model to predict the overall carbon emissions based on multiple secondary key factor data and transformed historical carbon emission data. This allows for a more comprehensive description of the relationship between multiple influencing factors and carbon emissions, thereby improving the accuracy of the model in predicting the overall carbon emissions.
[0120] Step S10243: Construct hierarchical data based on the carbon emissions and total carbon emissions corresponding to each primary key factor.
[0121] In this embodiment, the system can construct hierarchical data based on the carbon emissions corresponding to each primary key factor and the overall carbon emissions, as well as the carbon emissions of the target object in future time periods. Specifically, see... Figure 4 The figure is a flowchart of another hierarchical data construction process provided in an embodiment of this application, which includes the following steps:
[0122] Step S102431: Input the converted historical carbon emission data into the pre-trained autoregressive integral moving average model. The autoregressive integral moving average model performs time series analysis on the converted historical carbon emission data to predict the carbon emissions of the target object in the future period.
[0123] In this embodiment of the application, the pre-trained autoregressive integral moving average model can process time series data, predict future carbon emission trends, and take into account the autocorrelation of the data. The specific formula of the model is as follows:
[0124]
[0125] Where p is the order of the autoregressive term, d is the interpolation number, and q is the order of the moving average term.
[0126] The system uses an autoregressive integral moving average model to perform time-series analysis on the converted historical carbon emission data, which can accurately predict future carbon emission trends and thus obtain the accurate carbon emission amount of the target object in the future period.
[0127] Optionally, the system can use the ADF test or other methods to check the stationarity of the time series, while plotting the autocorrelation function and partial autocorrelation function, determining appropriate p and q values, and trying different combinations of p, d, and q through a grid search method to determine the optimal parameters of the model.
[0128] Step S102432 involves performing data analysis on the carbon emissions corresponding to each primary key factor, the total carbon emissions, the carbon emissions of the target object in the future period, and multiple historical carbon emission data to construct a hierarchical data structure.
[0129] In this embodiment of the application, the system can perform data analysis on the carbon emissions corresponding to each first key factor, the total carbon emissions, the carbon emissions of the target object in the future period, and multiple historical carbon emission data, divide the above data into several clusters, identify carbon emission sources with similar characteristics, and gradually merge similar data clusters to obtain hierarchical data.
[0130] Step S103: Using a pre-trained random forest model, predict the carbon emissions and carbon emission levels of the target object in the future time period and at each carbon neutrality level based on historical carbon emission data.
[0131] In this embodiment, carbon emission levels may include high carbon emission levels, medium carbon emission levels, low carbon emission levels, and negative carbon neutrality levels. The system can utilize a pre-trained random forest model to accurately predict the carbon emissions and carbon emission levels corresponding to each carbon neutrality level for the target object in future time periods based on historical carbon emission data, providing an accurate data foundation for subsequent carbon neutrality path optimization. Optionally, the system can also utilize the pre-trained random forest model to perform anomaly detection on the historical carbon emission data, identifying anomalies in the carbon emission data, including abnormal data values, abnormal data trends, abnormal data seasonality, and abnormal data correlations.
[0132] Optionally, in this embodiment, the system can obtain a training dataset including carbon emission data corresponding to C0, C1, and C2 levels within a historical time period, and preprocess missing and outlier values in the training dataset, dividing the processed training dataset into a training set and a test set. Then, the system can generate multiple training subsets through Bootstrap sampling, and randomly select some features for node splitting of each decision tree. The classification task is often based on the Gini coefficient or information gain, while the regression task is based on minimizing the mean squared error, thereby constructing multiple weak learners. Finally, the prediction results of all decision trees are aggregated through ensemble learning to obtain the final output, and the model performance is comprehensively evaluated using metrics such as accuracy, confusion matrix, mean squared error, and R². The classification task uses a voting method, and the regression task uses an averaging method.
[0133] Step S104: Using a pre-trained deep learning model, based on the historical carbon emission data, carbon emission amount, carbon emission level and hierarchical structure data corresponding to each carbon neutrality level, predict the target carbon emission amount of the target object in the future period corresponding to each carbon neutrality level.
[0134] In this embodiment, the deep learning model can be any one of the following: Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), or Long Short-Term Memory (LSTM), and the choice can be made according to the actual situation. The system can utilize a pre-trained deep learning model to accurately predict the target carbon emissions of the target object at each carbon neutrality level in the future, based on historical carbon emission data, carbon emission amounts, carbon emission levels, and hierarchical structure data corresponding to each carbon neutrality level, as well as feature data (such as energy-related data, economic activity data, population data, industrial structure data, technological level data, geographical environment data, and transportation data). This effectively overcomes the overfitting problem that easily arises when the model has too many features by capturing the complex nonlinear relationships and deep temporal dependencies in the aforementioned data, significantly improving the accuracy, stability, and generalization ability of carbon emission prediction.
[0135] Step S105: Based on each target carbon emission and carbon emission baseline, determine the target optimization strategy that matches each target carbon emission from multiple preset optimization strategies.
[0136] In step S105, the target optimization strategy is used to optimize and regulate the target carbon emissions of the target object in the future time period.
[0137] In this embodiment, the system can determine a target optimization strategy that matches each target carbon emission based on accurate prediction results (i.e., various target carbon emissions) and a carbon emission baseline, from multiple preset optimization strategies. This target optimization strategy can optimize and control the target carbon emissions of the target object in future time periods, thereby achieving dynamic optimization and precise control of carbon emissions at different stages and levels, from macro-level targets to micro-levels, providing efficient technical support for achieving carbon neutrality goals. For example, the target optimization strategy matching each target carbon emission could be as follows:
[0138] (1) For the CO level: reduce emissions of carbon dioxide and other greenhouse gases from direct combustion. For example, use low-NOx burners to reduce NOx generation while reducing carbon dioxide emissions, and use biomass fuels to replace fossil fuels.
[0139] (2) For C1 level: Reduce indirect carbon emissions, especially those related to energy consumption. For example, adopt efficient air conditioning and heating systems, and utilize smart grid computing to achieve effective management and scheduling of distributed energy resources.
[0140] (3) For C2 level: Further reduce carbon emissions through long-term projects and technological advancements, and ultimately achieve net-zero emissions. For example, optimize logistics and transportation methods to reduce carbon emissions during transportation, and carry out large-scale afforestation projects to increase carbon sinks and absorb carbon dioxide from the atmosphere.
[0141] Optionally, the carbon emission baseline includes: a direct carbon emission baseline corresponding to the direct carbon neutrality level, an indirect carbon emission baseline corresponding to the indirect carbon neutrality level, and a product lifecycle carbon emission baseline corresponding to the product lifecycle carbon neutrality level; wherein, the direct carbon emission baseline can be calculated through the following steps:
[0142] Step S11: Based on the historical carbon emission data corresponding to the direct carbon neutrality level, calculate the average direct carbon emissions of the target object during the historical period.
[0143] Step S12: Use the average value as a baseline for direct carbon emissions.
[0144] In this embodiment of the application, the indirect carbon emission baseline can be calculated through the following steps:
[0145] Step S21: Obtain the energy consumption data and standard carbon emission conversion factor of the target object in the historical period; the standard carbon emission conversion factor is the benchmark conversion factor for converting energy consumption data into carbon emissions.
[0146] Step S22: Based on the standard carbon emission conversion factor and energy consumption data, determine the indirect carbon emissions and use the indirect carbon emissions as the baseline for indirect carbon emissions.
[0147] In this embodiment of the application, the carbon emission baseline for the entire product lifecycle can be calculated through the following steps:
[0148] Step S31: Perform a life cycle assessment on the carbon emission data of the product throughout its entire life cycle to obtain the carbon emission amount of the product throughout its entire life cycle.
[0149] Step S32: Use the carbon emissions of the product's entire life cycle as the baseline for the product's entire life cycle carbon emissions.
[0150] It should be noted that steps S11 to S12, S21 to S22, and S31 to S32 are not shown in the figure.
[0151] It should be noted that by setting carbon emission baselines for each level, the emission reduction effect can be accurately assessed, providing a reference standard for carbon emissions. By comparing the actual carbon emissions with the baseline, the emission reduction effectiveness of each level over a certain period can be intuitively understood.
[0152] In one feasible implementation, after determining the target optimization strategy that matches each target carbon emission based on each target carbon emission and carbon emission baseline from multiple preset optimization strategies, the system can generate a carbon neutrality optimization report based on the target optimization strategy, hierarchical structure data, historical carbon emission data corresponding to each carbon neutrality level, carbon emission amount, carbon emission level, and target carbon emission amount; then the system can store the carbon neutrality optimization report in the storage space and generate the corresponding shared access link.
[0153] In this embodiment, the system can periodically use BI tools to generate carbon emission reports (i.e., carbon neutrality optimization reports) based on historical carbon emission data, carbon emission amounts, carbon emission levels and target carbon emission amounts, anomalies in carbon emission data, target optimization strategies and hierarchical structure data corresponding to each carbon neutrality level, according to customer needs. The system can then store the carbon emission reports in storage space (such as the cloud) and provide API interfaces to enable third-party systems to obtain and use the carbon emission reports by generating corresponding shared access links. At the same time, blockchain technology is used to ensure the transparency and immutability of the carbon emission reports.
[0154] In another feasible implementation, after implementing a target optimization strategy matched to each target carbon emission level, the system can periodically monitor carbon emissions from direct emission sources (i.e., C0 level), indirect emission sources (i.e., C1 level), and carbon sink projects (i.e., C2 level), and assess the carbon emission situation to obtain monitoring results, thereby ensuring the effectiveness of emission reduction measures. The system can then make the following adjustments based on the monitoring results:
[0155] (1) For the C0 level, the system can adjust emission reduction measures according to the monitoring results to ensure that the emission reduction measures achieve the expected results.
[0156] (2) For the C1 level, the system can optimize the energy structure and energy-saving measures to improve the overall energy efficiency.
[0157] (3) For the C2 level, the system can adjust supply chain management and consumer behavior strategies to continuously improve carbon neutrality.
[0158] The carbon neutrality hierarchical optimization method provided in this application achieves multiple predefined carbon neutrality levels, significantly expanding the breadth and completeness of data coverage and avoiding prediction bias caused by incomplete data, thus providing a clear data foundation for carbon neutrality path optimization. By constructing aggregated data (i.e., hierarchical structured data) organized hierarchically according to time, space, and emission source dimensions, it can adapt to subsequent hierarchical prediction and optimization processes, providing a data foundation for accurately locating carbon emission issues at different dimensions. It achieves accurate prediction of the carbon emission amount and carbon emission level of the target object at each carbon neutrality level in the future time period based on various historical carbon emission data using a pre-trained random forest model, providing an accurate data foundation for subsequent carbon neutrality path optimization. It also achieves accurate prediction of the carbon emission amount and carbon emission level of the target object at each carbon neutrality level in the future time period using a pre-trained deep... The degree-learning model captures complex nonlinear relationships and deep temporal dependencies in historical carbon emission data, carbon emission amounts, carbon emission levels, and hierarchical structure data corresponding to each carbon neutrality level. This effectively overcomes the overfitting problem that easily occurs when there are too many features, significantly improving the accuracy, stability, and generalization ability of carbon emission prediction. Based on accurate prediction results (i.e., each target carbon emission amount) and carbon emission baselines, a target optimization strategy matching each target carbon emission amount is determined from multiple preset optimization strategies. Through this target optimization strategy, the target carbon emission amount of the target object in the future period can be optimized and controlled, thereby realizing dynamic optimization and precise control of carbon emissions at different stages and levels from macro targets to micro levels, providing efficient technical support for achieving the carbon neutrality goal.
[0159] Device Examples
[0160] This application provides a carbon neutralization stratification optimization device, wherein... Figure 5 This is a schematic diagram of the structure of a carbon neutralization stratification optimization device provided in an embodiment of this application, as shown below. Figure 5 As shown, the device includes: a data acquisition module 11, a data analysis module 12, a first prediction module 13, a second prediction module 14, and an optimization strategy determination module 15. From Figure 5 You can see the connections between several modules.
[0161] The acquisition module 11 is used to acquire historical carbon emission data corresponding to each carbon neutrality level according to multiple predefined carbon neutrality levels. The historical carbon emission data includes at least one of the following categories: direct carbon emission data and indirect carbon emission data of the target object in the historical period, and carbon emission data of the entire product life cycle.
[0162] Data analysis module 12 is used to perform data analysis on multiple historical carbon emission data and construct hierarchical data; the hierarchical data is aggregated data organized in layers according to the dimensions of time, space and emission source;
[0163] The first prediction module 13 is used to predict the carbon emissions and carbon emission levels of the target object in the future period and at each carbon neutrality level based on the pre-trained random forest model and various historical carbon emission data.
[0164] The second prediction module 14 is used to use a pre-trained deep learning model to predict the target carbon emissions of the target object in the future period and at each carbon neutrality level, based on the historical carbon emission data, carbon emission amount, carbon emission level and hierarchical structure data corresponding to each carbon neutrality level.
[0165] The optimization strategy determination module 15 is used to determine the target optimization strategy that matches each target carbon emission based on each target carbon emission and carbon emission baseline from multiple preset optimization strategies; the target optimization strategy is used to optimize and regulate the target carbon emission of the target object in the future period.
[0166] Optionally, the data analysis module specifically includes: a processing unit, a statistical analysis unit, a first prediction unit, and a first construction unit.
[0167] The processing unit is used to perform data cleaning and preprocessing on multiple historical carbon emission data, and to convert the cleaned historical carbon emission data to obtain the converted historical carbon emission data.
[0168] The statistical analysis unit is used to perform statistical analysis on the converted historical carbon emission data to determine the carbon emission data corresponding to each of the multiple primary key factors. These primary key factors include: direct carbon emission related factors, indirect carbon emission related factors, and supply chain and consumption behavior related factors.
[0169] The first prediction unit is used to predict the carbon emissions corresponding to each first key factor based on the carbon emission data corresponding to each first key factor using a pre-trained linear relationship model. The carbon emissions corresponding to the first key factor are the predicted carbon emissions of the target object under the influence of the first key factor in the future period.
[0170] The first building unit is used to construct hierarchical data based on the carbon emissions corresponding to each primary key factor.
[0171] Optionally, the first building unit specifically includes: an acquisition unit, a second prediction unit, and a second building unit.
[0172] The acquisition unit is used to acquire multiple second key factor data, which include: energy factor data, economic factor data, population factor data, and technology factor data related to the target object's carbon emissions during historical periods.
[0173] The second prediction unit is used to predict the total carbon emissions based on multiple second key factor data and transformed historical carbon emission data using a pre-trained multiple regression model. The total carbon emissions are the predicted total carbon emissions of the target object under the influence of multiple second key factors in the future period.
[0174] The second building unit is used to construct hierarchical data based on the carbon emissions corresponding to each primary key factor and the overall carbon emissions.
[0175] Optionally, the second building unit specifically includes a timing analysis unit and a third building unit.
[0176] The time series analysis unit is used to input the converted historical carbon emission data into a pre-trained autoregressive integral moving average model. The autoregressive integral moving average model performs time series analysis on the converted historical carbon emission data to predict the carbon emissions of the target object in the future period.
[0177] The third building unit is used to perform data analysis on the carbon emissions corresponding to each primary key factor, the total carbon emissions, the carbon emissions of the target object in the future period, and multiple historical carbon emission data to build a hierarchical data structure.
[0178] Optionally, the predefined carbon neutrality levels include: direct carbon neutrality level, indirect carbon neutrality level, and product lifecycle carbon neutrality level;
[0179] The direct carbon neutrality tier includes carbon emission data from emission sources directly controlled by the target entity;
[0180] The indirect carbon neutrality hierarchy includes carbon emission data from the direct carbon neutrality hierarchy, as well as carbon emission data from external energy production related to the target.
[0181] The product lifecycle carbon neutrality hierarchy includes carbon emission data for the indirect carbon neutrality hierarchy, as well as carbon emission data for the products generated by the target object throughout their lifecycle.
[0182] Optionally, the carbon emission baseline includes: a direct carbon emission baseline corresponding to the direct carbon neutrality level;
[0183] The carbon neutrality stratification optimization device also includes: a first calculation module and a first determination module.
[0184] The first calculation module is used to calculate the average direct carbon emissions of the target object over a historical period based on the historical carbon emission data corresponding to the direct carbon neutrality level.
[0185] The first determining module is used to use the average value as a direct carbon emission baseline.
[0186] Optionally, the carbon emission baseline includes: an indirect carbon emission baseline corresponding to the indirect carbon neutrality level;
[0187] The carbon neutrality stratification optimization device also includes: a data acquisition module and a second determination module.
[0188] The data acquisition module is used to acquire the target object’s energy consumption data and standard carbon emission conversion factor over a historical period; the standard carbon emission conversion factor is the benchmark conversion factor for converting energy consumption data into carbon emissions.
[0189] The second determination module is used to determine the amount of indirect carbon emissions based on the standard carbon emission conversion coefficient and energy consumption data, and to use the amount of indirect carbon emissions as the baseline for indirect carbon emissions.
[0190] Optionally, the carbon emission baseline includes: a product lifecycle carbon emission baseline corresponding to the product lifecycle carbon neutrality level;
[0191] The carbon neutrality stratification optimization device also includes an evaluation module and a third determination module.
[0192] The assessment module is used to conduct a life cycle assessment of the carbon emission data throughout the product's entire life cycle, and to obtain the carbon emission amount throughout the product's entire life cycle.
[0193] The third determining module is used to use the carbon emissions of the entire product life cycle as the baseline for the carbon emissions of the entire product life cycle.
[0194] Optionally, the carbon neutrality stratification optimization device also includes a generation module and a storage module.
[0195] The generation module is used to determine the target optimization strategy that matches each target carbon emission based on each target carbon emission and carbon emission baseline from multiple preset optimization strategies, and then generate a carbon neutrality optimization report based on the target optimization strategy, hierarchical data, historical carbon emission data corresponding to each carbon neutrality level, carbon emission, carbon emission level and target carbon emission.
[0196] The storage module is used to store the carbon neutrality optimization report to the storage space and generate the corresponding shared access link.
[0197] It should be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiments. The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separate. The components indicated as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment solution according to actual needs. Those skilled in the art can understand and implement this without creative effort.
[0198] The above is merely one specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A carbon neutrality stratification optimization method, characterized in that, include: Collect historical carbon emission data corresponding to each predefined carbon neutrality level; the historical carbon emission data includes at least one of the following categories: direct carbon emission data and indirect carbon emission data of the target object in a historical period, and carbon emission data of the entire product life cycle; Data analysis is performed on multiple historical carbon emission data sets, and a hierarchical data structure is constructed; the hierarchical data structure is aggregated data organized hierarchically according to the dimensions of time, space, and emission source. Using a pre-trained random forest model, the carbon emissions and carbon emission levels of the target object in the future time period and at each of the carbon neutrality levels are predicted based on the historical carbon emission data. Using a pre-trained deep learning model, based on the historical carbon emission data, carbon emission amount, carbon emission level, and hierarchical structure data corresponding to each carbon neutrality level, the target carbon emission amount of the target object in the future period corresponding to each carbon neutrality level is predicted. Based on the target carbon emissions and carbon emission baselines, a target optimization strategy that matches the target carbon emissions is determined from a plurality of preset optimization strategies. The target optimization strategy is used to optimize and regulate the target carbon emissions of the target object in the future time period.
2. The method according to claim 1, characterized in that, The process of analyzing multiple historical carbon emission data sets and constructing a hierarchical data structure includes: Data cleaning and preprocessing are performed on multiple historical carbon emission data sets, and the cleaned historical carbon emission data is converted to obtain converted historical carbon emission data. Statistical analysis is performed on the converted historical carbon emission data to determine the carbon emission data corresponding to each of the multiple first key factors; the multiple first key factors include: direct carbon emission related factors, indirect carbon emission related factors, and supply chain and consumption behavior related factors; Using a pre-trained linear relationship model, the carbon emissions corresponding to each of the first key factors are predicted based on the carbon emission data corresponding to each of the first key factors; the carbon emissions corresponding to the first key factor are the predicted carbon emissions of the target object under the influence of the first key factor in the future period. The hierarchical data is constructed based on the carbon emissions corresponding to each of the first key factors.
3. The method according to claim 2, characterized in that, The construction of the hierarchical data based on the carbon emissions corresponding to each of the first key factors includes: Acquire multiple second key factor data; the multiple second key factor data include: energy factor data, economic factor data, population factor data, and technology factor data related to the carbon emissions of the target object in historical periods; Using a pre-trained multiple regression model, the comprehensive carbon emissions are predicted based on the data of the multiple second key factors and the transformed historical carbon emission data; the comprehensive carbon emissions are the predicted total carbon emissions of the target object under the influence of multiple second key factors in the future period. The hierarchical data is constructed based on the carbon emissions corresponding to each of the first key factors and the overall carbon emissions.
4. The method according to claim 3, characterized in that, The construction of the hierarchical data based on the carbon emissions corresponding to each of the first key factors and the overall carbon emissions includes: The converted historical carbon emission data is input into a pre-trained autoregressive integral moving average model, which performs time-series analysis on the converted historical carbon emission data to predict the carbon emissions of the target object in future periods. Data analysis is performed on the carbon emissions corresponding to each of the first key factors, the total carbon emissions, and the carbon emissions of the target object in the future period, as well as multiple historical carbon emission data, to construct the hierarchical data structure.
5. The method according to claim 1, characterized in that, The predefined carbon neutrality levels include: direct carbon neutrality level, indirect carbon neutrality level, and product lifecycle carbon neutrality level; The direct carbon neutrality hierarchy includes carbon emission data generated by emission sources directly controlled by the target object; The indirect carbon neutrality hierarchy includes carbon emission data from the direct carbon neutrality hierarchy, as well as carbon emission data from external energy production sources related to the target object. The product lifecycle carbon neutrality hierarchy includes carbon emission data from the indirect carbon neutrality hierarchy, as well as carbon emission data of the product generated by the target object throughout its lifecycle.
6. The method according to claim 5, characterized in that, The carbon emission baseline includes: a direct carbon emission baseline corresponding to the direct carbon neutrality level; the method further includes: Based on the historical carbon emission data corresponding to the direct carbon neutrality level, calculate the average direct carbon emission of the target object during the historical period; The average value is used as the baseline for direct carbon emissions.
7. The method according to claim 5, characterized in that, The carbon emission baseline includes: an indirect carbon emission baseline corresponding to the indirect carbon neutrality level; the method further includes: The energy consumption data and standard carbon emission conversion factor of the target object during the historical period are obtained; the standard carbon emission conversion factor is a benchmark conversion factor for converting energy consumption data into carbon emissions. Based on the standard carbon emission conversion factor and the energy consumption data, the indirect carbon emissions are determined, and the indirect carbon emissions are used as the indirect carbon emission baseline.
8. The method according to claim 5, characterized in that, The carbon emission baseline includes: a product lifecycle carbon emission baseline corresponding to the product's lifecycle carbon neutrality level; the method further includes: A life cycle assessment is performed on the carbon emission data of the product throughout its entire life cycle to obtain the carbon emission amount of the product throughout its entire life cycle; The carbon emissions throughout the product's entire life cycle are used as the baseline for the product's entire life cycle carbon emissions.
9. The method according to claim 1, characterized in that, After determining a target optimization strategy that matches each of the target carbon emissions from a plurality of preset optimization strategies based on each of the target carbon emissions and carbon emission baselines, the method further includes: Based on the target optimization strategy, the hierarchical data, the historical carbon emission data, carbon emission amount, carbon emission level and the target carbon emission amount for each carbon neutrality level, a carbon neutrality optimization report is generated. The carbon neutrality optimization report is stored in the storage space, and a corresponding shared access link is generated.
10. A carbon neutralization stratification optimization device, characterized in that, include: The data acquisition module is used to collect historical carbon emission data corresponding to each predefined carbon neutrality level. The historical carbon emission data includes at least one of the following categories: direct and indirect carbon emission data of the target object during a historical period, and carbon emission data throughout the product's entire life cycle; The data analysis module is used to perform data analysis on multiple historical carbon emission data and construct hierarchical data; the hierarchical data is aggregated data organized hierarchically according to the dimensions of time, space and emission source. The first prediction module is used to use a pre-trained random forest model to predict the carbon emissions and carbon emission levels of the target object in the future time period and at each of the carbon neutrality levels based on the historical carbon emission data. The second prediction module is used to use a pre-trained deep learning model to predict the target carbon emissions of the target object in the future period corresponding to each of the carbon neutrality levels, based on the historical carbon emission data, carbon emission amount, carbon emission level and the hierarchical structure data corresponding to each of the carbon neutrality levels. The optimization strategy determination module is used to determine, based on each of the target carbon emissions and carbon emission baselines, a target optimization strategy that matches each of the target carbon emissions from a plurality of preset optimization strategies. The target optimization strategy is used to optimize and regulate the target carbon emissions of the target object in the future time period.