Intelligent carbon footprint analysis method and system driven by industry data and medium

Abstract

The invention discloses an industry data driven carbon footprint intelligent analysis method and system and a medium, and relates to the technical field of carbon footprint analysis. The method comprises the following steps: collecting multi-source carbon emission data; constructing an industry knowledge base and accounting rules, performing feature extraction and structured index modeling based on the rules, and generating a carbon footprint reference model; creating an analysis task table, and performing prediction analysis on the reference model according to the task table to obtain a prediction parameter set; and performing visual analysis on the prediction parameter set, outputting a carbon footprint thermodynamic diagram, and performing attribution analysis and emission reduction decision-making based on the thermodynamic diagram so as to solve the problem that the accuracy and effectiveness of carbon emission prediction and emission reduction decision-making are insufficient due to the fact that an existing carbon footprint analysis method cannot effectively integrate multi-source data and cannot perform targeted analysis on different industries. And the effects of performing prediction and attribution analysis by adopting a machine learning algorithm and improving the intelligence and accuracy of carbon footprint analysis by combining multi-source data, an industry knowledge base and a carbon emission accounting rule are achieved.

Description

Technical Field

[0001] This invention relates to the field of carbon footprint analysis technology, specifically to industry data-driven intelligent carbon footprint analysis methods, systems, and media. Background Technology

[0002] With increasing global focus on climate change and carbon emission reduction, carbon footprint analysis has become a crucial tool for measuring corporate and industry carbon emissions, helping companies assess and reduce their emissions and thus promote green development. However, traditional carbon footprint analysis methods largely rely on manual data collection and static accounting models, resulting in relatively singular data sources, often limited to direct energy consumption data and failing to cover complex indirect emission links such as supply chains. Furthermore, these methods lack deep integration with industry-specific production processes and knowledge systems, leading to generalized and limited-precision analysis results. More critically, traditional methods lack sufficient intelligence, failing to deeply integrate and feature-minimize massive amounts of multi-source data, and lacking the ability to trace historical and real-time data. This results in often outdated carbon footprint conclusions, making it difficult to support precise and efficient emission reduction decisions targeting emission hotspots. Summary of the Invention

[0003] This application provides an industry data-driven intelligent carbon footprint analysis method, system, and medium, which solves the technical problem that existing carbon footprint analysis methods cannot effectively integrate multi-source data and conduct targeted analysis for different industries, resulting in insufficient accuracy and effectiveness of carbon emission prediction and emission reduction decisions.

[0004] In view of the above problems, this application provides an industry data-driven intelligent carbon footprint analysis method, system and medium.

[0005] The first aspect of this application provides an industry data-driven intelligent carbon footprint analysis method, the method comprising: collecting a multi-source carbon emission dataset of a target industry, the multi-source carbon emission dataset including production operation data, energy consumption data, supply chain data, and carbon emission factor data; constructing an industry knowledge base and carbon emission accounting rules, performing feature extraction and structured index modeling on the multi-source carbon emission dataset based on the industry knowledge base and carbon emission accounting rules, generating a carbon footprint benchmark model; creating a carbon footprint analysis task table, performing carbon emission prediction analysis on the carbon footprint benchmark model according to the carbon footprint analysis task table, obtaining a carbon footprint prediction task parameter set; performing visualization analysis on the carbon footprint prediction task parameter set, outputting a carbon footprint distribution heatmap, and performing carbon emission attribution analysis and emission reduction strategy decision-making based on the carbon footprint distribution heatmap.

[0006] The second aspect of this application provides an industry data-driven intelligent carbon footprint analysis system, comprising: a data acquisition module for acquiring multi-source carbon emission datasets of the target industry, including production operation data, energy consumption data, supply chain data, and carbon emission factor data; a data analysis module for constructing an industry knowledge base and carbon emission accounting rules, and performing feature extraction and structured index modeling on the multi-source carbon emission datasets based on the industry knowledge base and carbon emission accounting rules to generate a carbon footprint benchmark model; a carbon emission prediction module for creating a carbon footprint analysis task table, performing carbon emission prediction analysis on the carbon footprint benchmark model according to the carbon footprint analysis task table to obtain a carbon footprint prediction task parameter set; and an emission reduction decision module for performing visual analysis of the carbon footprint prediction task parameter set, outputting a carbon footprint distribution heatmap, and performing carbon emission attribution analysis and emission reduction strategy decision based on the carbon footprint distribution heatmap.

[0007] A third aspect of this application provides a computer-readable medium storing a computer program that, when executed by a processor, implements the industry data-driven intelligent carbon footprint analysis method provided in this application.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, a multi-source carbon emission dataset is collected from the target industry. This dataset includes production operation data, energy consumption data, supply chain data, and carbon emission factor data. Then, an industry knowledge base and carbon emission accounting rules are constructed. Based on these rules, feature extraction and structured index modeling are performed on the multi-source carbon emission dataset to generate a carbon footprint benchmark model. Next, a carbon footprint analysis task table is created. Carbon emission prediction analysis is performed on the benchmark model according to this table, yielding a carbon footprint prediction task parameter set. Finally, the carbon footprint prediction task parameter set is visualized, outputting a carbon footprint distribution heatmap. Based on this heatmap, carbon emission attribution analysis and emission reduction strategy decisions are made. This approach solves the technical problem that existing carbon footprint analysis methods cannot effectively integrate multi-source data and provide targeted analysis for different industries, leading to insufficient accuracy and effectiveness in carbon emission prediction and reduction decisions. It achieves the technical effect of improving the intelligence and accuracy of carbon footprint analysis by combining multi-source data, industry knowledge bases, and carbon emission accounting rules, and employing machine learning algorithms for prediction and attribution analysis. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A schematic diagram of the industry data-driven intelligent carbon footprint analysis method provided in the embodiments of this application.

[0011] Figure 2 A schematic diagram of the structure of an industry data-driven intelligent carbon footprint analysis system provided in this application embodiment.

[0012] Figure labeling: Data acquisition module 11, data analysis module 12, carbon emission prediction module 13, emission reduction decision module 14. Detailed Implementation

[0013] This application addresses the technical problem that existing carbon footprint analysis methods cannot effectively integrate multi-source data and conduct targeted analysis on different industries, resulting in insufficient accuracy and effectiveness of carbon emission prediction and emission reduction decisions by providing industry data-driven intelligent carbon footprint analysis methods, systems, and media.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0015] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0016] Example 1, as Figure 1 As shown, this application provides an industry data-driven intelligent carbon footprint analysis method, wherein the method includes: Collect multi-source carbon emission datasets for the target industry, including production operation data, energy consumption data, supply chain data, and carbon emission factor data.

[0017] In one embodiment, a multi-source carbon emission dataset is first collected from the target industry. This dataset includes production operation data, energy consumption data, supply chain data, and carbon emission factor data. Production operation data primarily refers to various data reflecting the enterprise's production processes and equipment operation, such as the operating status of production equipment, production cycle, output, process flow, and production efficiency. This data helps analyze the carbon emissions that may be generated in different production processes and provides a foundation for subsequent carbon emission prediction and optimization. Energy consumption data involves the types and quantities of energy consumed by the enterprise during production, typically electricity, fuel, and steam. This data is crucial for estimating the carbon emissions generated by energy consumption. Supply chain data encompasses carbon emission-related data from raw material procurement, transportation, storage to distribution. Detailed analysis of carbon emission sources within the supply chain identifies high-carbon-emission segments and provides a basis for optimizing supply chain management and reducing carbon emissions. Carbon emission factor data refers to the ratio between carbon emissions and consumption for each energy source, material, or activity. These factors are typically established by national or international organizations based on different energy types or activity types and are used to calculate carbon emissions. Different production methods, energy types, and transportation modes correspond to different emission factors. Collecting carbon emission factor data allows for accurate calculation of carbon emissions at different stages and processes. Comprehensive collection of this multi-source data provides rich and accurate foundational data support for subsequent carbon emission analysis and forecasting, ensuring the reliability and timeliness of decision-making.

[0018] An industry knowledge base and carbon emission accounting rules are constructed. Based on the industry knowledge base and carbon emission accounting rules, feature extraction and structured index modeling are performed on the multi-source carbon emission dataset to generate a carbon footprint benchmark model.

[0019] In one embodiment, firstly, based on industry-related knowledge of the target industry and suitable carbon emission algorithms, an industry knowledge base and carbon emission accounting rules are constructed. The industry knowledge base is a collection of information related to the target industry, including technologies, processes, energy usage, production equipment, and best practices. This provides strong support for subsequent carbon emission analysis, ensuring that the data analysis process considers the industry's specific needs and characteristics. The carbon emission accounting rules are methods and standards applicable to the industry, developed based on the content of the industry knowledge base and combined with international standards and industry realities. This ensures the accuracy and consistency of carbon emission accounting. Then, based on the constructed industry knowledge base, features closely related to carbon emissions, such as energy consumption types, production process parameters, and supply chain links, are extracted from the collected multi-source carbon emission datasets. These features are then used to calculate the carbon emissions of each stage of the target industry using the carbon emission accounting rules, serving as the basis for subsequent analysis and decision-making. Finally, based on these characteristics and the corresponding carbon emissions, a structured indicator model is constructed to create a carbon footprint benchmark model. This carbon footprint benchmark model can provide enterprises with accurate carbon footprint benchmark values ​​and help them identify the main sources of carbon emissions and potential emission reduction space, providing a scientific basis for the subsequent formulation of effective carbon emission reduction strategies.

[0020] Furthermore, the construction of the industry knowledge base and carbon emission accounting rules includes: Collect industry-related knowledge datasets of the target industry, standardize the industry-related knowledge datasets to obtain usable industry knowledge datasets; perform knowledge classification and cascading analysis on the usable industry knowledge datasets to construct an industry knowledge base; select an appropriate carbon emission algorithm based on the completeness of industry data and the complexity of industry processes; and parse the calculation rules of the appropriate carbon emission algorithm based on the usable industry knowledge datasets to construct the carbon emission calculation rules.

[0021] Preferably, the first step is to collect various data and knowledge related to the target industry. This knowledge typically covers industry standards and regulations, production process information, equipment and technology information, and market and supply chain information. Standards and regulations usually include industry-specific carbon emission regulations, carbon reduction policies, and environmental standards. Production process information typically includes common production processes, raw material usage, and energy consumption within the industry. Equipment and technology information typically includes the types of production equipment used, technology updates, and energy efficiency improvement solutions. Market and supply chain information typically includes information on the industry's supply chain structure, raw material procurement channels, and logistics. These data can be sourced from publicly available industry reports, industry standards, and the company's own operational data. By collecting this knowledge, an industry-related knowledge dataset can be formed, providing a foundation for subsequent analysis. Subsequently, since different data sources within the industry may use different formats, units, dimensions, and representations, these data need to be processed uniformly. This includes removing outliers, removing redundant information, converting different units, and filling in missing values. The result is a unified and consistent usable industry knowledge dataset. Outlier removal can be performed using methods such as the 3σ principle and isolated forests; redundant information can be removed using hash deduplication and field comparison; unit conversion can be performed using unit conversion mapping tables; and missing values ​​can be filled using interpolation and mean imputation. Next, the usable industry knowledge dataset is categorized according to industry knowledge classification dimensions. Then, cascading analysis is used to link the classification results, establishing relationships and influences between different categories. This forms a dynamically updated industry knowledge base with industry-specific knowledge, serving as the basis for subsequent carbon emission accounting and prediction. Furthermore, a suitable carbon emission accounting algorithm is selected based on the completeness of industry data and the complexity of industry processes. For example, for industries with relatively complete data, data-driven machine learning models can be used; for industries with scarce or incomplete data, estimations based on physical or empirical models are required; if the industry's production processes are complex and variable, more complex models, such as multi-factor regression models or deep neural networks, are needed to better capture the correlations between various stages. Then, the selected suitable carbon emission algorithm is analyzed to obtain the parameters required for the algorithm. These parameters are then correlated with the data centrally recorded in the available industry knowledge dataset to form carbon emission calculation rules that meet the needs of the target industry. These carbon emission calculation rules typically include carbon emission factors for various production stages, energy consumption conversion coefficients, factors affecting production efficiency, and carbon emission factors for raw materials, providing a scientific and standardized basis for subsequent carbon emission data analysis, prediction, and optimization.

[0022] Furthermore, the step of performing knowledge classification and cascading analysis on the available industry knowledge dataset to construct an industry knowledge base includes: The available industry knowledge dataset is classified into multiple levels according to the industry knowledge classification dimensions to obtain a multi-level industry dimension knowledge set. Association analysis is performed on the knowledge in each dimension of the multi-level industry dimension knowledge set to obtain a multi-level industry knowledge relationship set. A graph cascade is constructed based on the multi-level industry knowledge relationship set to obtain an industry knowledge graph. Each knowledge node in the industry knowledge graph is encoded, standardized, and dynamically verified and updated to construct the industry knowledge base.

[0023] Optionally, when performing knowledge classification and cascading analysis on the available industry knowledge dataset, the industry knowledge classification dimensions are first set. These dimensions are determined based on the specific characteristics of the target industry, aiming to effectively categorize and organize industry-related knowledge. Common classification dimensions include production processes, energy consumption, equipment information, supply chains, emission factors, and laws and regulations. Setting these dimensions helps to better organize industry knowledge data, ensuring that each type of knowledge can be categorized and used in a targeted manner. Subsequently, based on the set industry knowledge classification dimensions, the available industry knowledge dataset undergoes multi-level knowledge classification. Initially, the dataset is roughly classified according to the main dimensions, such as production processes, energy consumption, and equipment information. Based on this initial classification, each category is further subdivided. For example, under production processes, it can be further divided into different production stages, such as raw material handling, processing, and assembly. Under energy consumption, it can be subdivided according to different energy types, such as electricity and natural gas. Furthermore, further subdivisions can be made based on actual needs, such as dividing each production stage into specific steps. After the segmentation is completed, the results are stored hierarchically, forming a structured, multi-level, multi-industry-dimensional knowledge set. Then, based on predefined rules, such as the use of blast furnace equipment in ironmaking processes, the consumption of coke by blast furnace equipment, and the generation of blast furnace gas, explicit correlation analysis is performed on the knowledge in each dimension of the multi-level industry-dimensional knowledge set to establish explicit relationships between them. Next, by analyzing the co-occurrence frequency of knowledge items in literature and data and comparing it with co-occurrence thresholds, implicit relationships between the knowledge in each dimension are determined. For example, sinter grade and fuel ratio in ironmaking processes appear simultaneously in numerous reports. By summarizing these explicit and implicit relationships, a multi-level industry knowledge relationship set is formed. Then, based on the relationships recorded in the multi-level industry knowledge relationship set, the multi-level industry-dimensional knowledge sets are cascaded in the knowledge graph. That is, each industry knowledge in the multi-level industry-dimensional knowledge set is treated as a node, and these nodes are connected hierarchically according to the hierarchical relationships within the multi-level industry-dimensional knowledge set. Furthermore, nodes with related relationships are connected according to the recorded relationships, constructing an industry knowledge graph that clearly displays the hierarchical structure and relationships between different knowledge elements within an industry, making industry knowledge more intuitive and easily accessible. Further, each knowledge node in the industry knowledge graph is assigned a unique coded identifier. These identifiers can be numbers, letters, or a mixed format to ensure that each node has a unique identifier, thereby avoiding duplication or conflict. The assigned industry knowledge graph is then encapsulated and stored in the industry knowledge base.As industry knowledge continues to develop and change, when new industry knowledge is introduced, it will be divided into corresponding knowledge nodes according to the above analysis method, and the above correlation analysis will be performed for verification. If a correlation is found with other knowledge nodes, the knowledge node will be cascaded with these knowledge nodes, and the industry knowledge graph stored in the industry knowledge base will be updated using the cascaded industry knowledge graph, thereby ensuring that the industry knowledge graph is consistent with the latest industry development.

[0024] Furthermore, the generated carbon footprint benchmark model includes: The multi-source carbon emission dataset is cleaned, time-aligned, and standardized to obtain a standard carbon emission dataset. Key features are extracted from the standard carbon emission dataset based on the industry knowledge base to obtain a carbon emission key feature set. Carbon emission accounting is performed on the carbon emission key feature set according to the carbon emission accounting rules to determine the carbon emission accounting dataset. Structured index modeling is performed based on the carbon emission key feature set and the carbon emission accounting dataset to generate a carbon footprint benchmark model.

[0025] Preferably, after collecting the multi-source carbon emission dataset, to ensure the accuracy, consistency, and usability of the data, the dataset undergoes data cleaning, time alignment, and standardization to obtain a standard carbon emission dataset. Data cleaning includes removing outliers, removing redundancy, and imputing missing values, using the same methods described above. Time alignment can be achieved by resampling all data to a uniform time granularity. Standardization includes converting different units and normalization. The unit conversion method is the same as described above, and normalization can be performed using min-max standardization or Z-score standardization to ensure that all data are compared on the same scale. Subsequently, based on the key features marked in the industry knowledge base—that is, the parameters required for subsequent carbon emission accounting rule calculations—the specific data for each key feature is extracted from the standard carbon emission dataset to form a carbon emission key feature set. This set is then input into the corresponding adaptive carbon emission algorithm in the carbon emission accounting rule to calculate the carbon emissions at each stage. These carbon emission amounts are stored to form a carbon emission accounting dataset. Subsequently, the obtained key carbon emission feature set and carbon emission accounting dataset are structured and modeled according to the graded carbon emission indicator system to form a target industry carbon footprint model. Then, industry benchmark values ​​are introduced into this target industry carbon footprint model to construct a carbon footprint benchmark model. This carbon footprint benchmark model can be used to assess the impact of different processes, production processes, equipment configurations, etc. on carbon emissions, providing enterprises with a scientific basis for carbon emission analysis and optimization.

[0026] Furthermore, the step of generating a carbon footprint benchmark model by modeling structured indicators based on the key carbon emission feature set and the carbon emission accounting dataset includes: Based on the industry knowledge base, a tiered carbon emission indicator system is designed; the indicators at each level of the tiered carbon emission indicator system are sequentially associated and distributed with the carbon emission key feature set and the carbon emission accounting dataset to obtain a tiered carbon emission indicator dataset; structured indicator modeling is performed based on the tiered carbon emission indicator dataset to construct a target industry carbon footprint model; industry benchmark values ​​are introduced to perform benchmark comparison and data-driven updates on the target industry carbon footprint model to generate the carbon footprint benchmark model.

[0027] Optionally, based on an industry knowledge base, different carbon emission dimensions are first established. Common tiered dimensions include energy consumption, production processes, equipment efficiency, and supply chain management. Then, using these tiered dimensions as a benchmark, indicators at different levels are determined to form a tiered carbon emission indicator system. For example, primary indicators can be more macro-level indicators, such as total energy consumption and total carbon emissions; secondary indicators can be more detailed, such as energy consumption categorized by production stage and carbon emissions categorized by equipment type; and tertiary indicators can be even more detailed, such as carbon emissions per piece of equipment or emissions per mode of transportation. Subsequently, the indicators at each level of the tiered carbon emission indicator system are associated with relevant data in the carbon emission key feature set and carbon emission accounting dataset, and then distributed to each indicator level. For example, based on the carbon emission characteristics of different production stages, relevant data are assigned to corresponding primary, secondary, or tertiary indicators, thereby distributing carbon emissions from different energy types to different energy consumption indicators, forming a tiered carbon emission indicator dataset. This dataset contains carbon emission indicators categorized according to different levels, providing complete data support for subsequent modeling. Subsequently, based on the key carbon emission characteristics and indicators recorded in the tiered carbon emission indicator dataset, a target industry carbon footprint model is constructed using structured indicator modeling. Specifically, carbon emission indicators are mapped to corresponding indicators in the tiered carbon emission indicator system, and key carbon emission characteristics are linked to the tiered carbon emission indicator system through these indicators, forming the required target industry carbon footprint model. Then, to ensure the accuracy and industry applicability of the target industry carbon footprint model, an industry benchmark value is introduced to construct a carbon footprint benchmark model. This benchmark value is a carbon emission reference value calculated based on industry standards, best practices, or national / international standards. By comparing subsequent predictions with the benchmark value, it is possible to check whether the company's carbon emission level meets industry standards or low-carbon goals. Furthermore, whenever there are knowledge updates in the target industry, new key characteristics and corresponding carbon emission indicators are introduced to update the carbon footprint model data-driven, improving its accuracy and providing scientific decision support for companies to help achieve low-carbon development goals.

[0028] Create a carbon footprint analysis task table, and perform carbon emission prediction analysis on the carbon footprint benchmark model according to the carbon footprint analysis task table to obtain the carbon footprint prediction task parameter set.

[0029] In one embodiment, during the intelligent carbon footprint analysis process, a carbon footprint analysis task table is first created. This task table covers all the key tasks required for carbon emission analysis and is established based on actual business needs, typically including analysis objectives. Then, based on this task table, the required key features are parsed from the carbon footprint benchmark model, and the data corresponding to these key features are input into an adapted deep neural network algorithm for carbon emission prediction analysis. This predicts future carbon emission levels, forming a carbon footprint prediction task parameter set. This parameter set includes the analysis objectives specified in the task table, such as the predicted total carbon emissions and the carbon emission proportions at each stage, providing data support for subsequent decision-making, helping enterprises understand the current carbon emission status, and formulating reasonable carbon reduction plans.

[0030] Furthermore, the obtained carbon footprint prediction task parameter set includes: Carbon emission features are extracted from the carbon footprint baseline model according to the carbon footprint analysis task list to obtain the carbon footprint analysis task feature set; a deep neural network algorithm list is selected based on the target characteristic information of each analysis task in the carbon footprint analysis task list; the deep neural network algorithm list is used to perform carbon emission supervised training and prediction analysis on the carbon footprint analysis task feature set to obtain the carbon footprint prediction task parameter set.

[0031] Preferably, the set analysis targets, such as carbon emissions at a certain stage or in a certain process, are obtained from the carbon footprint analysis task table. These targets are then input into the carbon footprint benchmark model. The benchmark model compares the received analysis targets with its internal indicators at various levels and extracts the corresponding key carbon emission features based on the comparison results, forming a carbon footprint analysis task feature set. Subsequently, based on the carbon footprint analysis task table, the target characteristic information for each analysis task is extracted. This target characteristic information includes the required accuracy for prediction and the task requirements. Based on the accuracy requirements, all models that meet the required accuracy can be selected from a pre-set model library. This model library contains various deep neural networks, such as recurrent neural networks, long short-term memory networks, convolutional neural networks, graph neural networks, and multilayer perceptrons. These models are then further filtered according to the task requirements to form a list of deep neural network algorithms. For example, when the task target is long-term carbon emission trend prediction, recurrent neural networks and long short-term memory networks can be selected; when the task target is overall prediction integrating multidimensional features, multilayer perceptrons or integrated deep learning frameworks can be selected. Next, the feature set of the carbon footprint analysis task is input into a selected list of deep neural network algorithms. Predictive analysis is then performed based on the deep neural networks included in this list. These deep neural networks have all been pre-trained. Taking a Long Short-Term Memory (LSTM) network as an example, the corresponding training data is input into the LSM network for forward propagation to obtain predicted task parameters, such as the predicted total carbon emissions. Then, the mean squared error is used to calculate the loss between the predicted and actual task parameters. Backpropagation is then performed based on the calculated loss value, and the Adam optimizer is used to adjust the weights and biases of the LSM network until the maximum number of iterations is reached or the loss converges. Through the analysis of these deep neural networks, corresponding carbon footprint prediction task parameters can be generated for each analysis task. By summarizing these carbon footprint prediction task parameters, a carbon footprint prediction task parameter set is formed, providing basic data for subsequent carbon emission visualization and attribution analysis, and providing reliable predictive basis for enterprises to formulate targeted carbon reduction strategies.

[0032] The parameter set of the carbon footprint prediction task is visualized and analyzed to output a carbon footprint distribution heat map. Based on the carbon footprint distribution heat map, carbon emission attribution analysis and emission reduction strategy decision-making are carried out.

[0033] In one embodiment, after obtaining the carbon footprint prediction task parameter set, the task parameters in the set, such as total carbon emissions and carbon emissions at each process stage, are spatialized and hierarchically mapped. Then, a heatmap is generated to present different emission levels using color depth or gradient distribution, visually reflecting the carbon emission distribution across different time periods, processes, and regions, highlighting anomalies or high-emission hotspots. Next, by comparing carbon emissions at different spatial locations and time points in the carbon footprint distribution heatmap, the regions and times leading to high emissions are identified, and attribution parameters for carbon emission anomalies are calculated. Finally, emission reduction strategies are analyzed based on these attribution parameters to generate targeted emission reduction pathways. These pathways are then used to control the main sources and high-emission stages of carbon emissions, thereby achieving more efficient carbon emission management and optimization.

[0034] Furthermore, the carbon emission attribution analysis and emission reduction strategy decision-making based on the carbon footprint distribution heatmap includes: Anomalies in the carbon footprint distribution heatmap are identified spatially and temporally to obtain a set of anomalous carbon emission regions and a set of anomalous carbon emission time trends. Carbon emission attribution analysis is performed based on the set of anomalous carbon emission regions and the set of anomalous carbon emission time trends to determine the attribution parameters for anomalous carbon emissions. Emission reduction strategies are analyzed based on the attribution parameters for anomalous carbon emissions to obtain carbon emission reduction path schemes, and carbon emission reduction decisions are made based on the carbon emission reduction path schemes.

[0035] Preferably, after obtaining the carbon footprint distribution heatmap, the carbon emission data displayed in the heatmap is first subjected to anomaly spatial and temporal identification. Specifically, using the carbon emission information of each region in the heatmap, statistical analysis is performed on different spatial units (such as different production workshops, factories, or areas) and different time periods (such as hours, days, and months). Threshold detection or standard deviation analysis is used to identify spatial regions and time periods with emissions significantly higher than the average level, thus obtaining a set of anomalous carbon emission regions and a set of anomalous carbon emission time trends. These sets mark carbon emission hotspots and anomalous fluctuations, providing a target range for subsequent analysis. Subsequently, based on the obtained set of anomalous carbon emission regions, data is traced for each spatial unit within the set to obtain corresponding production process parameters, equipment operating load, equipment efficiency, energy consumption, and other data. These data are then identified using the Z-score method or box plot method, filtering out data with significant deviations and marking them as anomalous data. Finally, the Pearson correlation coefficient is used to calculate the correlation between each anomalous data point and carbon emissions, and highly correlated anomalous data are selected as candidate anomalous data. When multiple candidate anomaly data exist, partial correlation coefficients are used to eliminate interference from other variables. Only candidate anomaly data with partial correlation coefficients still above a threshold are identified as spatial carbon emission anomaly attribution parameters. Furthermore, based on the anomalous carbon emission time trend set, data source tracing is performed for each time period within the anomalous time trend set, and carbon emission attribution analysis is conducted using the same method described above to determine the temporal carbon emission anomaly attribution parameters. By intersecting the spatial and temporal carbon emission anomaly attribution parameters, a carbon emission anomaly attribution parameter is obtained. This parameter encompasses data that are anomalous in both spatial and temporal dimensions, representing the key reasons truly causing carbon emission anomalies. Then, based on the determined carbon emission anomaly attribution parameters, fuzzy matching is performed using an industry carbon emission reduction strategy library to find suitable carbon emission reduction pathways. Executing these pathways enables carbon emission reduction decisions, providing targeted emission reduction guidance to enterprises and ensuring that carbon emission anomalies can be detected, analyzed, and corrected in a timely manner, thereby effectively reducing their carbon footprint.

[0036] Furthermore, the proposed carbon emission reduction pathway includes: An industry carbon emission reduction strategy library is constructed. Based on the industry carbon emission reduction strategy library, fuzzy matching is performed on the carbon emission anomaly attribution parameters to determine the target carbon emission reduction strategy. Based on the target carbon emission reduction strategy, the strategy scheme is analyzed and the feasibility is optimized on the carbon emission anomaly attribution parameters to obtain the carbon emission reduction path scheme.

[0037] Optionally, after determining the carbon emission anomaly attribution parameters, a pre-built industry carbon emission reduction strategy library is connected. This library contains emission reduction measures for different industries, production stages, and types of carbon emission anomalies, such as energy optimization, equipment modification, process improvement, clean energy substitution, and logistics optimization. Each emission reduction measure is categorized according to the carbon emission stage, emission reduction type, and applicable conditions for subsequent calculation and matching. Then, the obtained carbon emission anomaly attribution parameters are fuzzily matched with the applicable conditions of each emission reduction measure in the industry carbon emission reduction strategy library. Cosine similarity is used to filter emission reduction measures from the strategy library that match the carbon emission anomaly attribution parameters—that is, emission reduction measures that meet the similarity threshold—forming the target carbon emission reduction strategy. Next, the cost-effectiveness of each target carbon reduction strategy is evaluated, the ratio between the cost required to implement the target carbon reduction strategy and its emission reduction effect is calculated, carbon reduction strategies that can achieve efficient emission reduction at a lower cost are screened out, and the technological level and resource conditions required for these carbon reduction strategies are evaluated. Those strategies that can be implemented quickly under existing conditions are given priority, thereby forming an executable emission reduction path plan, ensuring that emission reduction measures are scientific and feasible, and providing reliable technical support for enterprises to achieve low-carbon development.

[0038] In summary, the embodiments of this application have at least the following technical effects: First, a multi-source carbon emission dataset is collected from the target industry. This dataset includes production operation data, energy consumption data, supply chain data, and carbon emission factor data. Then, an industry knowledge base and carbon emission accounting rules are constructed. Based on these rules, feature extraction and structured index modeling are performed on the multi-source carbon emission dataset to generate a carbon footprint benchmark model. Next, a carbon footprint analysis task table is created. Carbon emission prediction analysis is performed on the benchmark model according to this table, yielding a carbon footprint prediction task parameter set. Finally, the carbon footprint prediction task parameter set is visualized, outputting a carbon footprint distribution heatmap. Based on this heatmap, carbon emission attribution analysis and emission reduction strategy decisions are made. This approach solves the technical problem that existing carbon footprint analysis methods cannot effectively integrate multi-source data and provide targeted analysis for different industries, leading to insufficient accuracy and effectiveness in carbon emission prediction and reduction decisions. It achieves the technical effect of improving the intelligence and accuracy of carbon footprint analysis by combining multi-source data, industry knowledge bases, and carbon emission accounting rules, and employing machine learning algorithms for prediction and attribution analysis.

[0039] Example 2, based on the same inventive concept as the industry data-driven intelligent carbon footprint analysis method in the foregoing examples, such as... Figure 2As shown, this application provides an industry data-driven intelligent carbon footprint analysis system, which includes: a data acquisition module 11: collecting multi-source carbon emission datasets of the target industry, including production operation data, energy consumption data, supply chain data, and carbon emission factor data; a data analysis module 12: constructing an industry knowledge base and carbon emission accounting rules, and performing feature extraction and structured index modeling on the multi-source carbon emission datasets based on the industry knowledge base and carbon emission accounting rules to generate a carbon footprint benchmark model; a carbon emission prediction module 13: creating a carbon footprint analysis task table, performing carbon emission prediction analysis on the carbon footprint benchmark model according to the carbon footprint analysis task table to obtain a carbon footprint prediction task parameter set; and an emission reduction decision module 14: performing visual analysis on the carbon footprint prediction task parameter set, outputting a carbon footprint distribution heat map, and performing carbon emission attribution analysis and emission reduction strategy decision based on the carbon footprint distribution heat map.

[0040] Furthermore, the data analysis module 12 is used to perform the following methods: Collect industry-related knowledge datasets of the target industry, standardize the industry-related knowledge datasets to obtain usable industry knowledge datasets; perform knowledge classification and cascading analysis on the usable industry knowledge datasets to construct an industry knowledge base; select an appropriate carbon emission algorithm based on the completeness of industry data and the complexity of industry processes; and parse the calculation rules of the appropriate carbon emission algorithm based on the usable industry knowledge datasets to construct the carbon emission calculation rules.

[0041] Furthermore, the data analysis module 12 is used to perform the following methods: The available industry knowledge dataset is classified into multiple levels according to the industry knowledge classification dimensions to obtain a multi-level industry dimension knowledge set. Association analysis is performed on the knowledge in each dimension of the multi-level industry dimension knowledge set to obtain a multi-level industry knowledge relationship set. A graph cascade is constructed based on the multi-level industry knowledge relationship set to obtain an industry knowledge graph. Each knowledge node in the industry knowledge graph is encoded, standardized, and dynamically verified and updated to construct the industry knowledge base.

[0042] Furthermore, the data analysis module 12 is used to perform the following methods: The multi-source carbon emission dataset is cleaned, time-aligned, and standardized to obtain a standard carbon emission dataset. Key features are extracted from the standard carbon emission dataset based on the industry knowledge base to obtain a carbon emission key feature set. Carbon emission accounting is performed on the carbon emission key feature set according to the carbon emission accounting rules to determine the carbon emission accounting dataset. Structured index modeling is performed based on the carbon emission key feature set and the carbon emission accounting dataset to generate a carbon footprint benchmark model.

[0043] Furthermore, the data analysis module 12 is used to perform the following methods: Based on the industry knowledge base, a tiered carbon emission indicator system is designed; the indicators at each level of the tiered carbon emission indicator system are sequentially associated and distributed with the carbon emission key feature set and the carbon emission accounting dataset to obtain a tiered carbon emission indicator dataset; structured indicator modeling is performed based on the tiered carbon emission indicator dataset to construct a target industry carbon footprint model; industry benchmark values ​​are introduced to perform benchmark comparison and data-driven updates on the target industry carbon footprint model to generate the carbon footprint benchmark model.

[0044] Furthermore, the carbon emission prediction module 13 is used to perform the following method: Carbon emission features are extracted from the carbon footprint baseline model according to the carbon footprint analysis task list to obtain the carbon footprint analysis task feature set; a deep neural network algorithm list is selected based on the target characteristic information of each analysis task in the carbon footprint analysis task list; the deep neural network algorithm list is used to perform carbon emission supervised training and prediction analysis on the carbon footprint analysis task feature set to obtain the carbon footprint prediction task parameter set.

[0045] Furthermore, the emission reduction decision module 14 is used to execute the following method: Anomalies in the carbon footprint distribution heatmap are identified spatially and temporally to obtain a set of anomalous carbon emission regions and a set of anomalous carbon emission time trends. Carbon emission attribution analysis is performed based on the set of anomalous carbon emission regions and the set of anomalous carbon emission time trends to determine the attribution parameters for anomalous carbon emissions. Emission reduction strategies are analyzed based on the attribution parameters for anomalous carbon emissions to obtain carbon emission reduction path schemes, and carbon emission reduction decisions are made based on the carbon emission reduction path schemes.

[0046] Furthermore, the emission reduction decision module 14 is used to execute the following method: An industry carbon emission reduction strategy library is constructed. Based on the industry carbon emission reduction strategy library, fuzzy matching is performed on the carbon emission anomaly attribution parameters to determine the target carbon emission reduction strategy. Based on the target carbon emission reduction strategy, the strategy scheme is analyzed and the feasibility is optimized on the carbon emission anomaly attribution parameters to obtain the carbon emission reduction path scheme.

[0047] Example 3: Based on the same inventive concept as the industry data-driven intelligent carbon footprint analysis method in the aforementioned examples, this application provides a medium storing a computer program. When the processor executes the computer program, it performs the following steps: A multi-parameter collaborative monitoring field is constructed by deploying multi-modal sensing devices in the underground space of a coal mine, wherein the multi-parameter collaborative monitoring field consists of N multi-parameter coverage monitoring fields; N geological hazard identification nodes are locally deployed in the N multi-parameter coverage monitoring fields, wherein the N geological hazard identification nodes establish bidirectional real-time communication with the early warning and control platform; the N geological hazard identification nodes collaboratively operate to perform risk identification on the N local multi-modal data returned from the N multi-parameter coverage monitoring fields, and output N geological hazard prediction results; the early warning and control platform receives and, based on the spatial correlation topology of the N multi-parameter coverage monitoring fields, performs disaster evolution prediction on the N geological hazard prediction results, and outputs real-time geological carbon footprint analysis.

[0048] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0049] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0050] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. An industry data-driven intelligent carbon footprint analysis method, characterized in that, The method includes: Collect multi-source carbon emission datasets for the target industry, including production operation data, energy consumption data, supply chain data, and carbon emission factor data; Construct an industry knowledge base and carbon emission accounting rules, and perform feature extraction and structured index modeling on the multi-source carbon emission dataset based on the industry knowledge base and carbon emission accounting rules to generate a carbon footprint benchmark model; Create a carbon footprint analysis task table, and perform carbon emission prediction analysis on the carbon footprint benchmark model according to the carbon footprint analysis task table to obtain a carbon footprint prediction task parameter set. The parameter set of the carbon footprint prediction task is visualized and analyzed to output a carbon footprint distribution heat map. Based on the carbon footprint distribution heat map, carbon emission attribution analysis and emission reduction strategy decision-making are carried out.

2. The industry data-driven intelligent carbon footprint analysis method as described in claim 1, characterized in that, The construction of the industry knowledge base and carbon emission accounting rules includes: Collect industry-related knowledge datasets of the target industry, and standardize the industry-related knowledge datasets to obtain usable industry knowledge datasets. The available industry knowledge dataset is classified and cascaded to construct an industry knowledge base; Select an appropriate carbon emission algorithm based on the completeness of industry data and the complexity of industry processes; Based on the available industry knowledge dataset, the carbon emission calculation rules of the adapted carbon emission algorithm are parsed to construct the carbon emission calculation rules.

3. The industry data-driven intelligent carbon footprint analysis method as described in claim 2, characterized in that, The step of performing knowledge classification and cascading analysis on the available industry knowledge dataset to construct an industry knowledge base includes: Set industry knowledge classification dimensions, and perform multi-level knowledge classification on the available industry knowledge dataset according to the industry knowledge classification dimensions to obtain a multi-level industry dimension knowledge set. A correlation analysis is performed on the knowledge in each dimension of the multi-level industry knowledge set to obtain a multi-level industry knowledge relationship set; Based on the multi-level industry knowledge relationship set, a graph cascade is constructed from the multi-level industry dimension knowledge set to obtain an industry knowledge graph. The industry knowledge graph is encoded, standardized, and dynamically verified and updated to construct the industry knowledge base.

4. The industry data-driven intelligent carbon footprint analysis method as described in claim 1, characterized in that, The generated carbon footprint benchmark model includes: The multi-source carbon emission dataset is cleaned, time-aligned, and standardized to obtain a standard carbon emission dataset. Based on the industry knowledge base, key features are extracted from the standard carbon emission dataset to obtain a set of key carbon emission features. Carbon emission accounting is performed on the carbon emission key feature set according to the carbon emission accounting rules to determine the carbon emission accounting dataset. Based on the key feature set of carbon emissions and the carbon emission accounting dataset, structured index modeling is performed to generate a carbon footprint benchmark model.

5. The industry data-driven intelligent carbon footprint analysis method as described in claim 4, characterized in that, The process of building a structured index model based on the key carbon emission feature set and the carbon emission accounting dataset to generate a carbon footprint benchmark model includes: Based on the industry knowledge base, a tiered carbon emission index system was designed. According to the hierarchical carbon emission indicator system, each level of the indicator is sequentially associated and distributed with the carbon emission key feature set and the carbon emission accounting dataset to obtain the hierarchical carbon emission indicator dataset. Based on the aforementioned graded carbon emission index dataset, structured index modeling is performed to construct a carbon footprint model for the target industry. An industry benchmark value is introduced to perform benchmark comparison and data-driven updates on the target industry's carbon footprint model, thereby generating the carbon footprint benchmark model.

6. The industry data-driven intelligent carbon footprint analysis method as described in claim 1, characterized in that, The obtained carbon footprint prediction task parameter set includes: Carbon emission features are extracted from the carbon footprint baseline model according to the carbon footprint analysis task table to obtain the carbon footprint analysis task feature set. Based on the target characteristic information of each analysis task in the carbon footprint analysis task table, select the deep neural network algorithm list; The deep neural network algorithm list is used to perform carbon emission supervision training and prediction analysis on the feature set of the carbon footprint analysis task, respectively, to obtain the parameter set of the carbon footprint prediction task.

7. The industry data-driven intelligent carbon footprint analysis method as described in claim 1, characterized in that, The carbon emission attribution analysis and emission reduction strategy decision-making based on the carbon footprint distribution heatmap includes: Anomalies in the carbon footprint distribution heatmap are identified spatially and temporally to obtain a set of anomalous carbon emission regions and a set of anomalous carbon emission temporal trends. Carbon emission attribution analysis is performed based on the set of abnormal carbon emission regions and the set of abnormal carbon emission time trends to determine the attribution parameters for abnormal carbon emissions. The carbon emission anomaly attribution parameters are analyzed to obtain carbon emission reduction path schemes, and carbon emission reduction decisions are made based on these schemes.

8. The industry data-driven intelligent carbon footprint analysis method as described in claim 7, characterized in that, The proposed carbon emission reduction pathway includes: Construct an industry carbon emission reduction strategy library, and perform fuzzy matching on the carbon emission anomaly attribution parameters based on the industry carbon emission reduction strategy library to determine the target carbon emission reduction strategy. Based on the target carbon emission reduction strategy, the carbon emission anomaly attribution parameters are analyzed for strategy schemes and their feasibility is optimized to obtain the carbon emission reduction path scheme.

9. An industry data-driven intelligent carbon footprint analysis system, characterized in that: The system is used to implement the industry data-driven intelligent carbon footprint analysis method according to any one of claims 1-8, the system comprising: Data acquisition module: Collects multi-source carbon emission datasets from the target industry, including production operation data, energy consumption data, supply chain data, and carbon emission factor data; Data analysis module: Constructs an industry knowledge base and carbon emission accounting rules, performs feature extraction and structured index modeling on the multi-source carbon emission dataset based on the industry knowledge base and carbon emission accounting rules, and generates a carbon footprint benchmark model; Carbon emission prediction module: Creates a carbon footprint analysis task table, performs carbon emission prediction analysis on the carbon footprint benchmark model according to the carbon footprint analysis task table, and obtains a carbon footprint prediction task parameter set; Emission reduction decision module: Performs visual analysis on the parameter set of the carbon footprint prediction task, outputs a carbon footprint distribution heat map, and performs carbon emission attribution analysis and emission reduction strategy decision based on the carbon footprint distribution heat map.

10. A computer-readable medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the industry data-driven intelligent carbon footprint analysis method as described in any one of claims 1-8.

Images