Carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion

Abstract

The invention discloses a carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion, and relates to the technical field of carbon emission accounting, a data acquisition module is constructed, multi-source heterogeneous original carbon emission data is acquired, and standardized processing is performed to generate standardized carbon emission data; constructing a data fusion module, performing dimension extraction on the standardized carbon emission data to generate a carbon emission dimension set, and constructing a multi-dimensional data geometry; constructing a projection accounting module, performing projection calculation on the multidimensional data geometry through different projection operators, and outputting carbon emission accounting results corresponding to different accounting standards; constructing an intelligent decision support module, and generating a corresponding decision support scheme based on a carbon emission accounting result; constructing a visual interaction module which is used for visually displaying accounting and decision data; according to the invention, through the construction of the multi-dimensional data structure of the multi-source heterogeneous carbon emission data and the multi-dimensional projection calculation, the whole-process integrated management of the carbon emission accounting and the energy-saving decision is realized.

Description

Technical Field

[0001] This invention relates to the field of carbon emission accounting technology, specifically to a carbon emission accounting and energy-saving decision-making system that integrates multi-source heterogeneous data. Background Technology

[0002] Currently, existing carbon emission accounting mainly relies on several methods such as the emission factor method based on activity data, the material balance method, and continuous monitoring systems. Enterprises typically use standardized calculation forms or stand-alone software to manually or semi-automatically collect key activity data such as energy consumption and material input according to the requirements of specific accounting standards, and select standard emission factors for calculation. However, the existing technology has certain shortcomings. First, its data processing mode is repetitive and inefficient. The same set of raw data needs to be manually extracted, format converted and cleaned multiple times to adapt to different accounting standards, which consumes a lot of manpower and time. At the same time, since the accounting for each standard is an independently initiated and completed closed-loop process, the data base, calculation model and parameter selection between these processes are isolated from each other, resulting in a lack of internal consistency among the results of multiple accounting reports. The source of numerical differences is difficult to trace and reconcile. Managers cannot obtain a unified and reliable data foundation for decision-making. Furthermore, the accounting process is rigid and it is difficult to flexibly and quickly incorporate new data sources or newly released accounting standard requirements into the existing framework, resulting in a serious lack of scalability and adaptability of the system. Therefore, it is of great significance to develop a carbon emission accounting and energy-saving decision-making system that integrates multi-source heterogeneous data. Summary of the Invention

[0003] The purpose of this invention is to provide a carbon emission accounting and energy-saving decision-making system based on the fusion of multi-source heterogeneous data, so as to solve the problems in the background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion, comprising: Data acquisition module: used to collect raw carbon emission data from multiple sources and heterogeneous sources, and to standardize the raw carbon emission data to generate standardized carbon emission data; Data fusion module: Connected to the data acquisition module, it extracts dimensions from standardized carbon emission data to generate a carbon emission dimension set, and constructs a multidimensional data geometry based on the carbon emission dimension set; Projection accounting module: Connected to the data fusion module, it is used to perform projection calculations on multidimensional data geometry using different projection operators and output carbon emission accounting results corresponding to different accounting standards; Intelligent Decision Support Module: Connected to the projection accounting module, it generates corresponding decision support solutions based on carbon emission accounting results; Visualization and Interaction Module: Connected to the Intelligent Decision Support Module, it is used to visually display standardized carbon emission data, carbon emission accounting results, and decision support solutions.

[0005] In a preferred embodiment, the data acquisition module includes: The multi-source interface unit is used to collect raw carbon emission data from different data sources through database interfaces, industrial protocol interfaces, and file interfaces. The scene perception unit is used to identify the business scene information corresponding to the raw carbon emission data and to label the corresponding scene identifier; The data preprocessing unit is used to clean, transform, and standardize the raw carbon emission data to generate standardized carbon emission data.

[0006] In a preferred embodiment, the data fusion module includes: The dimension extraction unit is used to extract time dimension, spatial dimension, material attribute dimension, process dimension and responsibility dimension from standardized carbon emission data to form a carbon emission dimension set; The coordinate mapping unit establishes a multidimensional coordinate system based on the carbon emission dimension set, and assigns coordinates in the multidimensional coordinate system to each standardized carbon emission data point, generating coordinate information; The fusion body building unit is used to establish topological connections based on the coordinates of each data point and the business logic relationships between the data points, and generate a multi-dimensional data geometry.

[0007] In a preferred embodiment, the step of establishing a multidimensional coordinate system based on the carbon emission dimension set and assigning coordinates in the multidimensional coordinate system to each standardized carbon emission data point is as follows: Establish a timeline based on time points; Establish spatial axes based on spatial dimensions, using geographical coordinates or organizational hierarchies as scales; Establish a material property axis based on material type or energy type; Establish a process axis based on the process dimension, with production stages or technological links as the scale; Establish a responsibility axis based on the scope of emission responsibility or the supply chain level; A multi-dimensional coordinate system is constructed using the time axis, spatial axis, material property axis, process axis, and responsibility axis; The specific time information, spatial location information, material attribute information, process flow information, and responsibility attribution information corresponding to each standardized carbon emission data point are mapped to a multi-dimensional coordinate system, and coordinate information is generated.

[0008] In a preferred embodiment, the step of establishing topological connections based on the coordinates of each data point and the business logic relationships between the data points to generate a multidimensional data geometry is as follows: Based on neural networks, identify the business logic relationships between data points; The business logic relationships include the sequential relationship between data points in the production process, the conversion correspondence between data points in the energy and material conversion, the inclusion and subordination relationship between data points in the spatial distribution, and the continuous association relationship between data points in the time series. Based on business logic relationships, establish topological connections between related data points, where the topological connections are edges in vector form; By integrating all data points with coordinate information and their topological connections, a multidimensional data geometry is generated, which is based on a multidimensional coordinate system, with data points as nodes and topological connections as edges, forming a relational network structure.

[0009] In a preferred embodiment, the projection calculation module includes: The projection operator library stores multiple projection operators corresponding to different carbon emission accounting standards. The parameter configuration unit is used to configure the observation view parameters for the projection operator, wherein the observation view parameters include the observation position parameters, the observation direction parameters, and the projection accuracy parameters; The parallel projection engine is used to call multiple projection operators simultaneously, perform synchronous projection calculations on multidimensional data geometry according to the observation perspective parameters corresponding to each projection operator, and output multiple carbon emission calculation results corresponding to different accounting standards.

[0010] In a preferred embodiment, the intelligent decision support module includes: The result fusion unit receives multiple carbon emission accounting results, performs consistency verification and weighted fusion, and outputs a comprehensive accounting conclusion. The strategy matching unit matches applicable decision strategies from a pre-set decision strategy library based on the preset decision objectives. The decision strategies stored in the decision strategy library include compliance priority strategies, economic optimization strategies, and risk control strategies. The solution generation engine generates corresponding decision support solutions through multi-objective optimization calculations based on comprehensive accounting conclusions and corresponding decision-making strategies.

[0011] In a preferred embodiment, the visual interaction module includes: The data visualization unit is used to render standardized carbon emission data into charts and graphs, generating a visual view of carbon emission data. The calculation result comparison unit is used to display and analyze the differences between multiple carbon emission calculation results in parallel, and generate a multi-standard calculation comparison view. The decision-making scheme display unit is used to present the specific measures and expected effects in the decision support scheme in a structured way, and generate a decision-making scheme execution view.

[0012] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention unifies multi-source heterogeneous carbon emission data into a multi-dimensional data geometry and utilizes multi-angle projection technology to achieve data fusion and parallel accounting of multiple standards. This effectively solves the core technical problems of data silos and incompatible accounting standards in traditional methods. The system uses the multi-dimensional data geometry as the core data model. Through dimension extraction and coordinate mapping, it integrates the originally discrete and heterogeneous data points into a unified multi-dimensional coordinate system and establishes topological connections based on business logic, thereby forming a complete carbon emission data chain. By abstracting the accounting standards into different observation perspectives and projection operators, the system does not need to repeatedly process the original data during accounting. Instead, it uses a parallel projection engine to perform synchronous projection operations on the static and unified data geometry from multiple preset angles to generate multiple accounting results, which greatly improves the efficiency of data processing and accounting. It also ensures that all reports originate from the same data ontology and eliminates the risk of data inconsistency. 2. This invention integrates intelligent decision support with intuitive visualization interaction, achieving a closed loop from multi-dimensional carbon accounting to executable energy-saving decisions. The multi-standard accounting results output by the system are not simply listed, but are weighted and verified for consistency through the result fusion unit, generating a comprehensive accounting conclusion that reflects the commonalities and characteristics of different standards. The decision strategy library has preset strategies such as compliance priority, economic optimization, and risk control, which can be matched with specific business objectives or external instructions to ensure clear decision guidance. At the same time, the solution generation engine, based on this, uses multi-objective optimization calculation to generate quantitative decision support solutions that balance environmental benefits, economic benefits, and operational risks. Meanwhile, the data visualization view presents the original data situation, the multi-standard accounting comparison view clearly reveals the root causes of differences between different accounting methods, and the decision solution execution view transforms abstract strategies into specific measures and expected effect maps. This enables more scientific and accurate energy-saving and carbon-reduction decisions based on a panoramic data view and clear solution comparison, directly transforming data value into management action. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0014] Figure 1 This is a system flowchart of the present invention.

[0015] Figure 2 This is a logic block diagram of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1, please refer to Figure 1 and Figure 2 As shown in this embodiment, a carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion includes: Data acquisition module: used to collect raw carbon emission data from multiple sources and heterogeneous sources, and to standardize the raw carbon emission data to generate standardized carbon emission data; Data fusion module: Connected to the data acquisition module, it extracts dimensions from standardized carbon emission data to generate a carbon emission dimension set, and constructs a multidimensional data geometry based on the carbon emission dimension set; Projection accounting module: Connected to the data fusion module, it is used to perform projection calculations on multidimensional data geometry using different projection operators and output carbon emission accounting results corresponding to different accounting standards; Intelligent Decision Support Module: Connected to the projection accounting module, it generates corresponding decision support solutions based on carbon emission accounting results; Visualization and Interaction Module: Connected to the Intelligent Decision Support Module, it is used to visually display standardized carbon emission data, carbon emission accounting results, and decision support solutions. Furthermore, current carbon emission accounting mainly relies on several methods such as the emission factor method based on activity data, the material balance method, and continuous monitoring systems. Enterprises typically use standardized calculation forms or stand-alone software to manually or semi-automatically collect key activity data such as energy consumption and material input according to the requirements of specific accounting standards, and select standard emission factors for calculation. However, the existing technology has certain shortcomings. First, its data processing mode is repetitive and inefficient. The same set of raw data needs to be manually extracted, format converted and cleaned multiple times to adapt to different accounting standards, which consumes a lot of manpower and time. At the same time, since the accounting for each standard is an independently initiated and completed closed-loop process, the data base, calculation model and parameter selection between these processes are isolated from each other, resulting in a lack of internal consistency among the results of multiple accounting reports. The source of numerical differences is difficult to trace and reconcile. Managers cannot obtain a unified and reliable data foundation for decision-making. Furthermore, the accounting process is rigid and it is difficult to flexibly and quickly incorporate new data sources or newly released accounting standard requirements into the existing framework, resulting in a serious lack of scalability and adaptability of the system. This invention unifies multi-source heterogeneous carbon emission data into a multi-dimensional data geometry and utilizes multi-angle projection technology to achieve data fusion and parallel accounting of multiple standards. This effectively solves the core technical problems of data silos and incompatible accounting standards in traditional methods. The system uses the multi-dimensional data geometry as the core data model. Through dimension extraction and coordinate mapping, it integrates the originally discrete and heterogeneous data points into a unified multi-dimensional coordinate system and establishes topological connections based on business logic, thereby forming a complete carbon emission data chain. By abstracting the accounting standards into different observation perspectives and projection operators, the system does not need to repeatedly process the original data during accounting. Instead, it uses a parallel projection engine to perform synchronous projection operations on the static and unified data geometry from multiple preset angles to generate multiple accounting results, which greatly improves the efficiency of data processing and accounting and ensures that all reports originate from the same data ontology, eliminating the risk of data inconsistency. By integrating intelligent decision support with intuitive visualization interaction, a closed loop from multi-dimensional carbon accounting to executable energy-saving decisions is achieved. The multi-standard accounting results output by the system are not simply listed, but are weighted and verified for consistency through the result fusion unit to generate a comprehensive accounting conclusion that reflects the commonalities and characteristics of different standards. The decision strategy library has preset strategies such as compliance priority, economic optimization, and risk control, which can be matched with specific business objectives or external instructions to ensure clear decision-making guidance. At the same time, the solution generation engine uses multi-objective optimization calculation to generate quantitative decision support solutions that balance environmental benefits, economic benefits, and operational risks. The system also presents the original data situation based on data visualization views, and the multi-standard accounting comparison view clearly reveals the root causes of differences between different accounting methods. The decision solution execution view transforms abstract strategies into specific measures and expected effect maps, enabling more scientific and accurate energy-saving and carbon reduction decisions based on panoramic data views and clear solution comparisons, directly transforming data value into management action.

[0018] In one embodiment, the data acquisition module includes: The multi-source interface unit is used to collect raw carbon emission data from different data sources through database interfaces, industrial protocol interfaces, and file interfaces. The scene perception unit is used to identify the business scene information corresponding to the raw carbon emission data and to label the corresponding scene identifier; The data preprocessing unit is used to clean, transform, and standardize the raw carbon emission data to generate standardized carbon emission data. Furthermore, the multi-source interface unit establishes connections with the enterprise's relational database through standardized database connection interfaces such as JDBC or ODBC and executes predefined SQL queries to obtain structured data. Simultaneously, it establishes network connections with the field industrial control system through integrated industrial protocol communication libraries, such as dedicated clients supporting Modbus TCP, OPC UA, and Profibus protocols, and periodically reads real-time data points. In addition, it automatically identifies and parses various storage file formats such as CSV, Excel, JSON, and XML through an embedded file parsing engine to extract semi-structured or unstructured data, thus acquiring raw carbon emission data from multiple heterogeneous modeling data sources. The acquired raw carbon emission data is then input into the scene perception unit, which has a built-in scene classifier based on a rule engine and a lightweight machine learning model. This classifier analyzes the metadata features in the data, such as data point labels, time series patterns, and numerical characteristics. The system identifies the value range and source system identifier, and matches it with predefined scenario feature templates to automatically determine the specific business scenario to which the data belongs, such as continuous production, batch operation, equipment maintenance, or office energy consumption. Each data record is then labeled with a machine-readable scenario identifier. Finally, the raw data with scenario identifiers flows into the data preprocessing unit. This unit, based on data quality rules and standardization specifications, first performs data cleaning operations, such as using interpolation or scenario-based default value filling to handle missing values, and applying statistical thresholds or outlier detection algorithms to identify and correct anomalies. Secondly, it performs data transformation operations, including adjusting the consumption of various energy and materials according to... Pre-defined conversion factors are uniformly converted to standard units of measurement, and all time information is synchronized to the standard time zone and formatted as a unified timestamp. All data fields are mapped and reorganized according to a predefined standard data model to ensure consistency in data structure, type, and value range. This predefined standard data model is a structured data specification framework, defined starting with the abstraction of core entities and attributes. It first identifies the entity types that must be covered in the carbon emission field, including but not limited to emission source equipment, energy and material consumption activities, monitoring instruments, and organizational units. Each entity type corresponds to a standardized data structure template. The required and optional attribute fields for this type of entity are clearly defined. For example, for a boiler emission source entity, its required attributes may include the equipment unique identifier, the department code, the geographical coordinates of the installation location, and the fuel type code, while optional attributes may include the design thermal efficiency, the year of commissioning, etc. The specific definition of the attribute fields follows strict specifications. Each attribute must specify its data value type, such as string, number, boolean value, or date and time; the unit of measurement, such as kilogram, megawatt-hour, cubic meter; the value range or enumeration list, such as the fuel type enumeration values ​​including natural gas, coal, fuel oil; and the constraint condition of whether null values ​​are allowed.To ensure global consistency, the model mandates the use of a unified coding system to standardize the representation of key concepts such as equipment type, fuel type, process flow, and departmental classification. These codes typically reference or map from international, national, or industry standard code libraries.

[0019] In one embodiment, the data fusion module includes: The dimension extraction unit is used to extract time dimension, spatial dimension, material attribute dimension, process dimension and responsibility dimension from standardized carbon emission data to form a carbon emission dimension set; The coordinate mapping unit establishes a multidimensional coordinate system based on the carbon emission dimension set, and assigns coordinates in the multidimensional coordinate system to each standardized carbon emission data point, generating coordinate information; The fusion body building unit is used to establish topological connections based on the coordinates of each data point and the business logic relationships between data points, and generate a multi-dimensional data geometry. Furthermore, the dimension extraction unit parses each standardized carbon emission data point according to a predefined dimension mapping rule table to extract key dimension information. Specifically, time dimension extraction obtains UTC time values ​​accurate to the second by parsing the standard timestamp field in the data points; spatial dimension extraction converts the geographic location coordinates field in the data points or the equipment location master data associated with the equipment code into a unified spatial grid code; material attribute dimension extraction obtains standardized category identifiers by matching the material or energy type codes in the data points with the standard material classification code library; and process dimension extraction determines the stage in the production process by associating the source system identifier of the data points with the process flow diagram. The extraction of the carbon emission dimension is based on the emission source type and supply chain mapping rules, which automatically determine the scope of the emission, such as direct emissions, indirect emissions from purchased electricity, or indirect emissions from upstream and downstream of the supply chain. This constructs a carbon emission dimension set containing five categories of identifiers for each data point. After receiving the complete carbon emission dimension set, the coordinate mapping unit first establishes a time axis with absolute timestamps as continuous scale, a space axis with three-dimensional spatial coordinates or hierarchical spatial codes as discrete scale, a material attribute axis with standard material classification codes as discrete scale, a process axis with process flow node codes as discrete scale, and a responsibility axis with emission responsibility scope codes as discrete scale. The system uses five axes to form a multidimensional coordinate system. A normalized mapping algorithm is then used to convert the specific dimension identifier value of each data point into its corresponding coordinate value within this multidimensional coordinate system. For example, a timestamp is converted into a seconds offset relative to the origin, spatial grid codes are mapped to coordinates of a specific point in three-dimensional space, and various classification codes are mapped to specific ordinal numbers on their respective dimension axes using a lookup table. Ultimately, a unique set of multidimensional coordinate information is generated for each data point. The fusion body construction unit, based on the coordinate information of all data points, identifies potential logical relationships between data points through a business relationship rule engine. This engine, according to predefined production process sequence rules, identifies relationships between adjacent process stages. The system establishes topological connections between data points to represent flow direction, establishes topological connections between energy or material data points with input-output relationships to represent transformation based on the material-energy conservation rule, establishes topological connections between equipment and workshop data points with subordinate relationships to represent inclusion based on the spatial organization hierarchy rule, and establishes topological connections between continuously monitored data points of the same entity to represent evolution based on the time series continuity rule. Finally, the system integrates all data points as nodes and the various topological connections established between them as edges to form a multidimensional network structure with complete node coordinate information and edge connection relationships. This structure is the multidimensional data geometry that carries the intrinsic correlation of all carbon emission data.

[0020] In one embodiment, the step of establishing a multidimensional coordinate system based on the carbon emission dimension set and assigning coordinates in the multidimensional coordinate system to each standardized carbon emission data point is as follows: Establish a timeline based on time points; Establish spatial axes based on spatial dimensions, using geographical coordinates or organizational hierarchies as scales; Establish a material property axis based on material type or energy type; Establish a process axis based on the process dimension, with production stages or technological links as the scale; Establish a responsibility axis based on the scope of emission responsibility or the supply chain level; A multi-dimensional coordinate system is constructed using the time axis, spatial axis, material property axis, process axis, and responsibility axis; The specific time information, spatial location information, material property information, process flow information and responsibility attribution information corresponding to each standardized carbon emission data point are mapped to a multi-dimensional coordinate system and coordinate information is generated. Furthermore, based on the carbon emission dimension set, coordinate axes with clear scales and mathematical definitions are established for each dimension. The time axis uses millisecond-level timestamps calculated from the Coordinated Universal Time epoch as continuous scales, ensuring that the time of all data points can be mapped to a unique value. The spatial axis selects one of two modes based on data characteristics: for data points with precise latitude and longitude information, a three-dimensional Cartesian coordinate system is used, converting latitude, longitude, and altitude into spatial coordinate values ​​with the Earth's center as the origin; for data points with only organizational hierarchy information, a discrete hierarchical coding space is used, encoding the hierarchical structure of the enterprise's factory area, workshops, and production lines as a series of discrete spatial nodes with inclusion relationships. The material attribute axis relies on a standardized material and energy classification coding library, mapping each material or energy type to a specific unique classification code in the library, and using this code as a discrete sequence number on that dimension axis. The process axis is based on the enterprise's predefined precise experimental flowchart, encoding each production stage or process link as a node identifier with a sequential order, and using this as the discrete scale of the process axis. The responsibility axis is based on the internationally accepted scope-one-scope-two-scope-three classification system and... By combining the enterprise's own supply chain management code, a unique responsibility level code is assigned to each possible emission responsibility attribution, forming a discrete scale of the responsibility axis. After defining the five coordinate axes, the system orthogonally combines these five axes to form a five-dimensional abstract mathematical space, i.e., a multi-dimensional coordinate system. The subsequent coordinate mapping process is a point-by-point lookup and calculation process. For each standardized carbon emission data point, the system reads its time information and converts it into a timestamp value as a coordinate component on the time axis, reads its spatial location information and calculates its coordinate components on the spatial axis according to a preset pattern, reads its material property information and determines its coordinate components on the material property axis by querying the standard classification code library, reads its process flow information and matches the flow chart nodes to determine its coordinate components on the process axis, reads its responsibility attribution information and matches the responsibility level code to determine its coordinate components on the responsibility axis. Finally, these five coordinate components are combined into a five-dimensional coordinate vector, which is the unique position identifier of the data point in the multi-dimensional coordinate system, i.e., the generated coordinate information. The set of coordinate information of all data points provides a precise mathematical basis for the subsequent construction of multi-dimensional data geometry.

[0021] In one embodiment, the step of establishing topological connections based on the coordinates of each data point and the business logic relationships between the data points to generate a multidimensional data geometry is as follows: Based on neural networks, identify the business logic relationships between data points; The business logic relationships include the sequential relationship between data points in the production process, the conversion correspondence between data points in the energy and material conversion, the inclusion and subordination relationship between data points in the spatial distribution, and the continuous association relationship between data points in the time series. Based on business logic relationships, establish topological connections between related data points, where the topological connections are edges in vector form; Integrate all data points with coordinate information and their topological connections to generate a multidimensional data geometry with a relational network structure based on a multidimensional coordinate system, with data points as nodes and topological connections as edges; Furthermore, the fusion building unit relies on a composite relationship recognition and graph construction engine specifically designed for multi-dimensional carbon data. Its core is a dual-channel graph neural network model pre-trained with domain-specific knowledge. One channel processes the multi-dimensional coordinate information of data points, capturing spatial and temporal proximity and sequence patterns through geometric operations between coordinate vectors, such as calculating Euclidean distance and direction. The other channel analyzes the standardized attributes of data points, such as energy type coding, material mass flow rate, and equipment status labels. The outputs of the two channels are fused in an intermediate layer, weighted by an attention mechanism, and then a fully connected classifier determines whether a specific type of business logic relationship exists between any two data points and the strength of that relationship. The system predefines four key business logic relationship types. The sequential relationship of the production process is mainly confirmed by judging the sequentiality of the process axis coordinates and combining it with the process knowledge base. The correspondence between energy and material conversion is determined by analyzing... The system infers the relationships between material property axes using transformation rules and mass-energy balance models. Spatial distribution is matched based on the hierarchical coding tree of spatial axis coordinates, while continuous relationships in time series are identified by analyzing the continuity of time axis coordinates and the covariance pattern of numerical changes. For each identified relationship, the system generates a topological connection edge in vector form with clear semantics. The data structure of this edge includes the relationship type encoding, the relationship confidence weight output by the neural network, the direction vector represented by the difference between the coordinate vectors of the source node and the target node, and optional association attributes such as transformation coefficients. These vector edges, along with data point nodes with precise coordinates, are integrated and managed in a dynamic graph database. This database uses a multidimensional coordinate system as the global reference system, data points as vertices, and vector edges as directed edges to construct and persistently store a property graph with complete geometric information and topological semantics. This structured property graph is the multidimensional data geometry.

[0022] In one embodiment, the projection calculation module includes: The projection operator library stores multiple projection operators corresponding to different carbon emission accounting standards. The parameter configuration unit is used to configure the observation view parameters for the projection operator, wherein the observation view parameters include the observation position parameters, the observation direction parameters, and the projection accuracy parameters; The parallel projection engine is used to call multiple projection operators simultaneously, perform synchronous projection calculations on multidimensional data geometry according to the observation perspective parameters corresponding to each projection operator, and output multiple carbon emission calculation results corresponding to different accounting standards. Furthermore, the projection operator library is essentially a versioned container of rules and algorithms. Projection operators include, but are not limited to, international standard projection operators, regional compliance projection operators, and industry-specific projection operators. Each projection operator is an independently encapsulated software module, embedding the core calculation logic and judgment rules corresponding to a specific carbon emission accounting standard. For example, operators for ISO standards will solidify their judgment algorithms regarding organizational and operational boundaries, as well as query interfaces for official emission factor databases. Operators for specific regional carbon markets encapsulate all calculation steps and data format requirements required by the monitoring reports and verification specifications recognized by that market. These operators are registered in the system's service directory in a pluggable manner. The parameter configuration unit provides a dynamic parameter setting interface, allowing users or external systems to configure a set of observation perspective parameters for the projection operator to be executed. The observation position parameter essentially defines the origin or observation point of the coordinate system used for projection calculations in multidimensional data geometry. For example, it can be set to focus only on the spatial coordinate range corresponding to a specific plant area. The observation direction parameter defines the projection direction. The annotation includes a subset of dimensions and their weights. For example, the directional parameter corresponding to an accounting standard focused on direct emissions might set the weight of range one in the responsibility dimension to 1 while setting the weights of other responsibility ranges to 0. The projection precision parameter controls the granularity of the calculation, such as determining whether the time dimension is aggregated by hour or by day to balance calculation speed and result granularity. The parallel projection engine is the core component for performing calculations. It first loads one or more specified projection operator instances from the projection operator library and injects them with the configured observation perspective parameters. Then, it creates multiple concurrent calculation threads or tasks. Each task independently holds a read-only snapshot of the multidimensional data geometry. Each task performs calculations on the geometry according to the logic built into its operator and the injected parameters. This calculation process can be mathematically regarded as a mapping from a high-dimensional space to a specific low-dimensional reporting plane, which is accomplished by traversing and aggregating node and edge data on relevant dimensions. The engine synchronously manages all these concurrent tasks and ensures the isolation and efficient use of computing resources. After all parallel calculation tasks are completed, the engine collects their respective outputs and formats them into a standardized carbon emission accounting result report.

[0023] In one embodiment, the intelligent decision support module includes: The result fusion unit receives multiple carbon emission accounting results, performs consistency verification and weighted fusion, and outputs a comprehensive accounting conclusion. The strategy matching unit matches applicable decision strategies from a pre-set decision strategy library based on the preset decision objectives. The decision strategies stored in the decision strategy library include compliance priority strategies, economic optimization strategies, and risk control strategies. The solution generation engine generates corresponding decision support solutions through multi-objective optimization calculations based on comprehensive accounting conclusions and corresponding decision-making strategies. Furthermore, the result fusion unit receives multiple carbon emission calculation results from the projection calculation module. First, it performs consistency verification, a process accomplished by a rule-based logical conflict detection engine. The engine compares the values ​​of different calculation results at the same emission source or within the same time interval, identifies significant differences exceeding a preset threshold, and traces these differences back to their source—the corresponding original data subsets in the multidimensional data geometry and the differences in the applied projection parameters—thus marking inconsistencies and generating a consistency report. Based on this verification, the unit employs a confidence-weighted fusion algorithm. This algorithm assigns a dynamic weight to each calculation result, determined by the authority level of the corresponding calculation standard, the quality score of the original data used, and the uncertainty assessment results during the projection calculation process. Then, a weighted average is calculated for all indicators in all calculation results, ultimately... The system outputs a comprehensive accounting conclusion that eliminates obvious conflicts and reflects multi-standard consensus and credibility assessment. The strategy matching unit maintains a structured decision strategy library, in which each strategy clearly defines its applicable decision objectives, constraints, and core optimization orientation. For example, the core of the compliance priority strategy is to ensure that all emission data meets the most stringent regulatory standards, the core of the economic optimization strategy is to maximize emission reduction benefits under cost constraints, and the core of the risk control strategy is to identify and prioritize emission links with high volatility or high regulatory risks. Based on the user's clearly defined decision objectives or the decision requirements automatically derived by analyzing the current business scenario and external policy environment, this unit uses a combination of rule matching and machine learning classification to match one or more of the most applicable decision strategies from the strategy library and passes their complete parameter set, including the objective function, constraints, and priority coefficients, to the downstream system. The solution generation engine, as the final execution component, receives the comprehensive calculation conclusions and selected decision strategy parameters. It first constructs a multi-objective optimization mathematical model. This model uses the comprehensive calculation conclusions as baseline data and the objective functions defined by the decision strategy parameters, such as minimizing total compliance costs, maximizing return on investment, or minimizing the risk of exceeding emission limits, as optimization objectives. It also uses practical conditions such as technical feasibility, budget limits, and time windows as constraints. The engine then calls its built-in optimization solver, which can use linear programming, integer programming, or heuristic algorithms such as genetic algorithms to solve the model, exploring and generating a series of candidate decision solutions that achieve different trade-offs among the objectives. Each candidate solution includes a specific list of measures, such as equipment modification items, operational parameter adjustments, and procurement recommendations, along with detailed expected emission reductions, cost estimates, and risk assessment data. Finally, the engine ranks all candidate solutions using multi-criteria decision analysis methods such as TOPSIS, recommends the optimal decision support solution, and generates a detailed report containing implementation paths and key performance indicators.

[0024] In one embodiment, the visual interaction module includes: The data visualization unit is used to render standardized carbon emission data into charts and graphs, generating a visual view of carbon emission data. The calculation result comparison unit is used to display and analyze the differences between multiple carbon emission calculation results in parallel, and generate a multi-standard calculation comparison view. The decision-making scheme display unit is used to present the specific measures and expected effects in the decision support scheme in a structured way, and generate a decision-making scheme execution view; Furthermore, the data visualization unit is built on a high-performance WebGL graphics library and visualization programming framework. Its core is a configurable rendering pipeline. This pipeline first receives a standardized carbon emission data stream from the data acquisition module, and then automatically matches and calls the corresponding visualization templates according to the data attributes. For time-series data such as energy consumption trends, line charts and area charts are used for rendering, and the time axis can be zoomed and panned. For spatial distribution data such as emission intensity of various factories, heat maps or hierarchical statistical maps are used for overlay display, and can be linked to geographic information system base maps. For categorical comparison data such as the proportion of different energy types, pie charts or stacked bar charts are used for presentation, and color coding and legend interaction are supported. All graphic elements are rendered by vector to ensure clarity at multiple resolutions, and a data binding mechanism is used to achieve real-time synchronous updates between the view and the underlying data. Finally, a carbon emission data visualization view that integrates multiple charts and supports linked filtering and drill-down operations is generated. The comparison unit for accounting results employs a multi-view collaborative rendering and difference detection algorithm. This unit obtains multiple carbon emission accounting result datasets from the projection accounting module and renders each dataset independently in a side-by-side visual panel. Each panel uses consistent visual encoding, such as the same coordinate axis scales, color mapping, and legends, to ensure comparability. The unit integrates a difference calculation engine, which compares the values ​​of corresponding data points in different panels item by item, calculates the absolute and relative differences, and uses highlighting techniques such as color overlay or difference arrow annotations to visually present significant differences on each panel. At the same time, the unit provides a central synchronization controller, allowing users to uniformly filter the time range, adjust the aggregation level, or switch key indicators for multiple views, thereby ensuring a consistent perspective for comparative analysis and ultimately generating a multi-standard accounting comparison view that clearly reveals the similarities and differences between different accounting standards. The design of the decision-making solution display unit focuses on transforming structured decision-making solutions into an understandable and traceable execution blueprint. This unit receives decision support solutions from the intelligent decision support module, which typically include a list of measures, a timeline, resource allocation, and expected performance indicators. The unit first uses an interactive Gantt chart component based on HTML5 Canvas to render the implementation timeline and dependencies of the measures. Each measure is represented as a timeline and can be categorized by color according to the responsible department. Simultaneously, the unit utilizes Sankey diagrams or node link diagrams to visualize the flow relationships between measures, resource inputs, and expected emission reduction effects. Key performance indicators such as emission reductions, cost savings, and return on investment are embedded around the view in the form of information cards or dashboards. Furthermore, the view supports interactive exploration; users can click on any measure bar to view its detailed description, prerequisites, and risk warnings in the sidebar. Adjusting the start and end dates of a timeline dynamically updates the calculated values ​​of the associated performance indicators. Ultimately, this multi-layered, interactive, and structured presentation generates an intuitive view of the decision-making solution execution, guiding users to understand and follow up on the implementation of the solution.

[0025] The above description is merely a 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 emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion, characterized in that, Data acquisition module: used to collect raw carbon emission data from multiple sources and heterogeneous sources, and to standardize the raw carbon emission data to generate standardized carbon emission data; Data fusion module: Connected to the data acquisition module, it extracts dimensions from standardized carbon emission data to generate a carbon emission dimension set, and constructs a multidimensional data geometry based on the carbon emission dimension set; Projection accounting module: Connected to the data fusion module, it is used to perform projection calculations on multidimensional data geometry using different projection operators and output carbon emission accounting results corresponding to different accounting standards; Intelligent Decision Support Module: Connected to the projection accounting module, it generates corresponding decision support solutions based on carbon emission accounting results; Visualization and Interaction Module: Connected to the Intelligent Decision Support Module, it is used to visually display standardized carbon emission data, carbon emission accounting results, and decision support solutions.

2. The carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The data acquisition module includes: The multi-source interface unit is used to collect raw carbon emission data from different data sources through database interfaces, industrial protocol interfaces, and file interfaces. The scene perception unit is used to identify the business scene information corresponding to the raw carbon emission data and to label the corresponding scene identifier; The data preprocessing unit is used to clean, transform, and standardize the raw carbon emission data to generate standardized carbon emission data.

3. The carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The data fusion module includes: The dimension extraction unit is used to extract time dimension, spatial dimension, material attribute dimension, process dimension and responsibility dimension from standardized carbon emission data to form a carbon emission dimension set; The coordinate mapping unit establishes a multidimensional coordinate system based on the carbon emission dimension set, and assigns coordinates in the multidimensional coordinate system to each standardized carbon emission data point, generating coordinate information; The fusion body building unit is used to establish topological connections based on the coordinates of each data point and the business logic relationships between the data points, and generate a multi-dimensional data geometry.

4. The carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion according to claim 3, characterized in that, The steps for establishing a multidimensional coordinate system based on the carbon emission dimension set and assigning coordinates in the multidimensional coordinate system to each standardized carbon emission data point are as follows: Establish a timeline based on time points; Establish spatial axes based on spatial dimensions, using geographical coordinates or organizational hierarchies as scales; Establish a material property axis based on material type or energy type; Establish a process axis based on the process dimension, with production stages or technological links as the scale; Establish a responsibility axis based on the scope of emission responsibility or the supply chain level; A multi-dimensional coordinate system is constructed using the time axis, spatial axis, material property axis, process axis, and responsibility axis; The specific time information, spatial location information, material attribute information, process flow information, and responsibility attribution information corresponding to each standardized carbon emission data point are mapped to a multi-dimensional coordinate system, and coordinate information is generated.

5. The carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion according to claim 3, characterized in that, The steps for establishing topological connections based on the coordinates of each data point and the business logic relationships between data points to generate a multidimensional data geometry are as follows: Based on neural networks, identify the business logic relationships between data points; The business logic relationships include the sequential relationship between data points in the production process, the conversion correspondence between data points in the energy and material conversion, the inclusion and subordination relationship between data points in the spatial distribution, and the continuous association relationship between data points in the time series. Based on business logic relationships, establish topological connections between related data points, where the topological connections are edges in vector form; By integrating all data points with coordinate information and their topological connections, a multidimensional data geometry is generated, which is based on a multidimensional coordinate system, with data points as nodes and topological connections as edges, forming a relational network structure.

6. The carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The projection calculation module includes: The projection operator library stores multiple projection operators corresponding to different carbon emission accounting standards. The parameter configuration unit is used to configure the observation view parameters for the projection operator, wherein the observation view parameters include the observation position parameters, the observation direction parameters, and the projection accuracy parameters; The parallel projection engine is used to call multiple projection operators simultaneously, perform synchronous projection calculations on multidimensional data geometry according to the observation perspective parameters corresponding to each projection operator, and output multiple carbon emission calculation results corresponding to different accounting standards.

7. The carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The intelligent decision support module includes: The result fusion unit receives multiple carbon emission accounting results, performs consistency verification and weighted fusion, and outputs a comprehensive accounting conclusion. The strategy matching unit matches applicable decision strategies from a pre-set decision strategy library based on the preset decision objectives. The decision strategies stored in the decision strategy library include compliance priority strategies, economic optimization strategies, and risk control strategies. The solution generation engine generates corresponding decision support solutions through multi-objective optimization calculations based on comprehensive accounting conclusions and corresponding decision-making strategies.

8. The carbon emission accounting and energy-saving decision-making system based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The visual interaction module includes: The data visualization unit is used to render standardized carbon emission data into charts and graphs, generating a visual view of carbon emission data. The calculation result comparison unit is used to display and analyze the differences between multiple carbon emission calculation results in parallel, and generate a multi-standard calculation comparison view. The decision-making scheme display unit is used to present the specific measures and expected effects in the decision support scheme in a structured way, and generate a decision-making scheme execution view.

Images