A carbon efficiency diagnosis method and system based on intelligent knowledge graph
By constructing a carbon efficiency diagnosis method based on intelligent knowledge graphs and using graph neural network models to capture the graph structure and temporal characteristics of carbon emission data, this method addresses the shortcomings of traditional methods in detecting carbon emission anomalies and analyzing their causes, and achieves accurate anomaly cause localization and optimization suggestions.
Patent Information
- Application Number
- CN202611321314.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-28
- Publication Date
- 2026-11-10
AI Technical Summary
Traditional carbon efficiency analysis methods struggle to capture complex nonlinear relationships and ignore the correlations between entities, resulting in limited capabilities for detecting and analyzing carbon emission anomalies and failing to provide accurate location of anomaly causes and optimization suggestions.
A carbon efficiency diagnosis method based on intelligent knowledge graph is constructed. Through a graph neural network model consisting of graph embedding layer, graph convolutional layer, time series processing layer and prediction layer, the graph structure features and time features of carbon emission data are captured. Combined with an anomaly cause database, the anomaly path is traced back along entity relationships to locate the cause of the anomaly.
It improves the accuracy of identifying the causes of abnormal carbon emissions, provides specific reasons for the anomalies, and enhances the interpretability and practicality of the diagnostic results.
Smart Images

Figure CN123070976A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of carbon efficiency diagnosis technology, specifically to a carbon efficiency diagnosis method and system based on intelligent knowledge graphs. Background Technology
[0002] With the increasing severity of global climate change, carbon emission control and management have become a key focus for governments and businesses worldwide. To achieve carbon neutrality, companies need to accurately monitor, predict, and optimize their carbon emissions. However, the complexity and diversity of carbon emission data make traditional carbon efficiency analysis methods inadequate for practical needs. Traditional methods primarily rely on statistical analysis of historical data or rule-based expert systems. However, these methods often treat carbon emission data as isolated time series or static data, ignoring the relationships and knowledge structures between entities. Statistical methods struggle to capture complex nonlinear relationships, while rule-based methods rely on human experience and are ill-suited to dynamically changing production environments. Traditional methods also have limited capabilities in detecting and analyzing the causes of carbon emission anomalies, making it difficult to provide accurate anomaly location and optimization suggestions.
[0003] Therefore, improving the accuracy of anomaly localization is a current research direction. Summary of the Invention
[0004] (a) Purpose of the invention The purpose of this invention is to provide a carbon efficiency diagnosis method and system based on intelligent knowledge graphs that can improve the accuracy of locating the causes of abnormal carbon emissions.
[0005] (II) Technical Solution To address the above problems, this invention provides a carbon efficiency diagnosis method based on intelligent knowledge graphs, comprising: Carbon emission entities are identified based on historical carbon emission data, and these carbon emission entities include primary nodes and secondary nodes. Based on the carbon emission entity, a carbon efficiency knowledge graph is constructed. The carbon efficiency knowledge graph includes first-level nodes and second-level nodes. The first-level nodes and second-level nodes are connected by edges, and the edges represent the relationship between the connected nodes. Based on the carbon efficiency knowledge graph, a carbon emission prediction model is constructed, which outputs a carbon emission prediction value based on the input production process data. Carbon efficiency diagnosis is performed based on the predicted carbon emissions.
[0006] In another aspect of the present invention, preferably, the carbon efficiency knowledge graph includes multiple vector paths, each vector path including a first-level node and multiple second-level nodes, and the first-level node and multiple second-level nodes are connected by multiple edges.
[0007] In another aspect of the present invention, preferably, the carbon emission prediction model is based on a graph neural network; The carbon emission prediction model includes a graph embedding layer, a graph convolutional layer, a time series processing layer, and a prediction layer; The graph embedding layer embeds entities and relationships in the carbon efficiency knowledge graph into a low-dimensional vector space; The graph convolutional layer updates the entity representation by aggregating information of neighboring nodes of entities in the graph embedding layer through graph convolution, thereby obtaining graph structure features in carbon emission data; The time series processing layer is used to extract time features from carbon emission data; The prediction layer is used to output carbon emission prediction values based on the graph structure features and the time features.
[0008] In another aspect of the present invention, preferably, the carbon efficiency diagnosis based on the predicted carbon emissions includes: Collect production process data for each primary node; The production process data is input into the carbon emission prediction model, and the carbon emission prediction value for each vector path is output. The predicted carbon emissions are compared with the actual carbon emissions, and the deviation is calculated. If the deviation value exceeds the preset threshold, the carbon emissions of that vector path are determined to be abnormal.
[0009] In another aspect of the present invention, preferably, the method further includes: For vector paths identified as having abnormal carbon emissions, the abnormal emissions are located based on entities and relationships in the carbon efficiency knowledge graph; this includes: Starting from the output entity of the vector path and ending at the first-level node of the vector path, backtrack along the relationships in the knowledge graph to obtain the path of abnormal carbon emissions. Based on the aforementioned database of carbon emission anomalies and their causes, the causes of the anomalies are located.
[0010] In another aspect of the present invention, preferably, the graph embedding layer embeds entities and relations in the carbon efficiency knowledge graph into a low-dimensional vector space, comprising the following steps: Each entity and each relationship is mapped to a low-dimensional vector, and the low-dimensional vectors of the entity and relationship are obtained. The mapping is represented by the following formula: Among them, e i Represents an entity, where i represents the entity number, and h represents the entity ID. i Represents entity e i low-dimensional vectors, Represents the mapping function, rk Let k represent the relation, and w represent the relation number. k Representing relation r k a low-dimensional vector; The entity low-dimensional vector and the relation low-dimensional vector are combined into a triplet; The triplet is then optimized.
[0011] In another aspect of the present invention, preferably, the triplet is optimized using the following formula: Where U represents the optimization function, e i Represents the entity with number i, e j Represents entity j, r k Indicates a relation, (e i r k e j ) represents a triple, and G represents the geometry of all triples. h represents the Euclidean distance. j Represents entity e j A low-dimensional vector.
[0012] In another aspect of the invention, preferably, the update is performed using the following formula: in, This represents the updated entity representation. It is the l-th level entity e i The representation of, σ It is an activation function. It is the learnable weight matrix of the l-th layer. It is entity e i The set of neighbors, Entity e j The set of neighbors.
[0013] In another aspect of the present invention, preferably, locating the cause of the carbon emission anomaly based on the carbon emission anomaly path and anomaly cause database includes: The probability of each anomaly cause is calculated based on the contribution of each node in the path of the carbon emission anomaly and the weight of each node in the anomaly cause database. Based on the stated probability, the abnormal cause of the carbon emission anomaly can be identified.
[0014] In another aspect, preferably, a carbon efficiency diagnostic system based on a smart knowledge graph includes: Determination Module: Determines carbon emission entities based on historical carbon emission data, wherein the carbon emission entities include primary nodes and secondary nodes; First construction module: Based on the carbon emission entity, construct a carbon efficiency knowledge graph, which includes first-level nodes and second-level nodes. The first-level nodes and second-level nodes are connected by edges, and the edges represent the relationship between the connected nodes. The second construction module: Based on the carbon efficiency knowledge graph, a carbon emission prediction model is constructed, and the carbon emission prediction model outputs the carbon emission prediction value according to the input production process data; Diagnostic module: Performs carbon efficiency diagnosis based on the predicted carbon emissions.
[0015] (III) Beneficial Effects The above-described technical solution of the present invention has the following beneficial technical effects: This invention constructs a carbon efficiency knowledge graph, representing carbon emission-related entities and relationships in a graph structure. This captures complex correlation information, providing richer features for carbon emission prediction. Furthermore, the carbon efficiency knowledge graph allows for tracing the causal paths of anomalies along the relationships between entities, clarifying the root causes of anomalies, and improving the accuracy of locating the causes of carbon emission anomalies. Combined with an anomaly cause database, it can provide specific causes for each anomaly, enhancing the interpretability and practicality of the diagnostic results. Attached Figure Description
[0016] Figure 1 This is an overall flowchart of one embodiment of the present invention. Detailed Implementation
[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.
[0018] Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.
[0019] In the description of this invention, it should be noted that the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.
[0020] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.
[0021] The invention will now be described in more detail with reference to the accompanying drawings. In the various drawings, the same elements are indicated by similar reference numerals. For clarity, the various parts in the drawings are not drawn to scale.
[0022] Example 1 A carbon efficiency diagnosis method based on intelligent knowledge graphs. Figure 1 An overall flowchart of one embodiment of the present invention is shown, as follows: Figure 1 As shown, it includes: Carbon emission entities are identified based on historical carbon emission data, which includes primary nodes and secondary nodes. Historical carbon emission data can be obtained from an enterprise's energy management system, environmental monitoring system, and production management system. Data types include carbon emission data, such as emissions of greenhouse gases like carbon dioxide and methane; production process data, such as output, energy consumption, and equipment operating status; and environmental data, such as temperature, humidity, and air pressure. In this embodiment, carbon emission entities can be extracted from text data using Natural Language Processing (NLP) technology, or from structured data using data association analysis, with attribute information added to each entity. The specific content of primary and secondary nodes is not limited here; primary nodes can represent entities at the macro level, and secondary nodes can represent entities at the micro level. In this embodiment, the primary node can be a factory or production entity, and the secondary nodes include workshops, energy types, production equipment, technological processes, and environmental indicators.
[0023] Based on the carbon emission entities, a carbon efficiency knowledge graph is constructed. This knowledge graph includes primary nodes and secondary nodes, connected by edges representing the relationships between the connected nodes. For example, factory A includes production equipment B, and production equipment B depends on process flow C. In this embodiment, the carbon efficiency knowledge graph can be constructed and managed using a graph database or a knowledge graph construction platform. The carbon efficiency knowledge graph includes multiple vector paths, each consisting of one primary node and multiple secondary nodes connected by multiple edges. The physical meaning of the vector path can be understood as the transmission path or influence chain of carbon emissions, describing the entire process from the source to the final manifestation of carbon emissions. For example, factory A is transported via transportation vehicle D, and transportation vehicle D consumes energy type C.
[0024] Based on the carbon efficiency knowledge graph, a carbon emission prediction model is constructed. The carbon emission prediction model outputs a carbon emission prediction value based on the input production process data. The specific content of the carbon emission prediction model is not limited here. In this embodiment, the carbon emission prediction model is based on a graph neural network. The carbon emission prediction model includes a graph embedding layer, a graph convolutional layer, a time series processing layer, and a prediction layer; The graph embedding layer embeds entities and relationships in the carbon efficiency knowledge graph into a low-dimensional vector space. The goal of the graph embedding layer is to map entities and relationships in the knowledge graph to a low-dimensional vector space. Through this mapping, the structural information (relationships between entities) in the knowledge graph can be represented as mathematical vector operations, thereby facilitating subsequent prediction. This embodiment does not limit the specific content of the graph embedding layer embedding entities and relationships in the carbon efficiency knowledge graph into the low-dimensional vector space; the following steps are included: Each entity and each relationship is mapped to a low-dimensional vector, and the low-dimensional vectors of the entity and relationship are obtained. The mapping is represented by the following formula: Among them, e i Represents an entity, where i represents the entity number, and h represents the entity ID. i Represents entity e i low-dimensional vectors, Represents the mapping function, r k Let k represent the relation, and w represent the relation number. k Representing relation r k a low-dimensional vector; The entity low-dimensional vector and the relation low-dimensional vector are combined into a triplet; (e i r k e j ) represents a triple, and represents entity e. i and entity e j There is a relationship r k For example, the ternary combination (factory A, emissions, carbon dioxide) means that factory A emits carbon dioxide; The triples are optimized. The goal of optimization is to optimize each triple (e... i r k e j ), entity e i vector h i Add relationship r k vector w k It should be close to entity e j vector h j .
[0025] In this embodiment, the triplet is optimized using the following formula: Where U represents the optimization function, e i Represents the entity with number i, e j Represents entity j, r k Indicates a relation, (e i r ke j ) represents a triple, and G represents the geometry of all triples. h represents the Euclidean distance. j Represents entity e j The low-dimensional vector. By optimizing the objective function L, for each triplet (e i r k e j ), vector h i +w k As close as possible to vector h j This indicates that the model is correctly modeled in the vector space. If the distance is large, it means the model needs to adjust its vector representation. For example, in a knowledge graph, there is a triple (Factory A, Emissions, Carbon Dioxide), and the vector of entity Factory A is h. 工厂A The vector representing the relationship between emissions is w. 排放 The vector of the solid carbon dioxide is h. 二氧化碳 The optimization target is h 工厂A +w 排放 Approximately equal to h 二氧化碳 That is, the value of the above formula should be as small as possible.
[0026] The graph convolutional layer updates the entity representation by aggregating information from the neighbor nodes of entities in the graph embedding layer through graph convolution, thereby obtaining graph structure features in the carbon emission data. The graph convolutional layer is the core part of the graph neural network, capable of capturing local structural information between entities. Graph convolution can be implemented using GCN or GAT. If higher-order neighbor information needs to be captured, multiple layers of graph convolution can be stacked. In this embodiment, the update is performed using the following formula: in, This represents the updated entity representation. It is the l-th level entity e i The representation of, σ It is an activation function. It is the learnable weight matrix of the l-th layer. It is entity e i The set of neighbors, Entity e j The set of neighbors is aggregated. The aggregation method is to perform a weighted summation of the features of neighboring nodes. The weights are determined by the normalization coefficients and the learnable weight matrix. The normalization coefficients are used to balance the influence of node degree and avoid nodes with high degree having too much influence on the result. The activation function σ introduces nonlinearity, enabling the model to learn more complex feature representations.
[0027] The time series processing layer is used to extract time features from carbon emission data; in this embodiment, a Long Short-Term Memory (LSTM) network is used to model the time series.
[0028] The prediction layer is used to output predicted carbon emissions based on the graph structure features and the time features. Assume the output of the time series processing layer is h. T The output of the graph convolutional layer is h. i (L) , where L is the number of graph convolutional layers, which fuses time series features and graph structure features, and uses fully connected layers (FC) to output predicted carbon emissions; Among them, z i W is the fused feature vector. p It is a learnable weight matrix, b p It is a bias term. This is a predicted value for carbon emissions.
[0029] Furthermore, in this embodiment, historical carbon emission data is used to train the carbon emission prediction model. The training process includes: Initialize the parameters of the graph embedding layer, graph convolutional layer, time series processing layer, and prediction layer; Initialize the graph embedding layer using pre-trained knowledge graph embeddings; The entities and relations in the carbon efficiency knowledge graph are embedded into a low-dimensional vector space, and the embedded representation h of the entities and relations is output. i and w k ; By aggregating information from neighboring nodes through graph convolution, the entity representation is updated, and the graph structure feature h is output. i (L) ; LSTM is used to process time series data, extract time features, and output the time feature h. t ; By fusing graph structure features and time features, a carbon emission prediction is output.
[0030] Using mean squared error as the loss function: in, Represents the loss function. y represents the predicted carbon emissions. i N represents historical carbon emissions, and N is the sample size.
[0031] Carbon efficiency diagnosis based on the predicted carbon emissions includes: Collect production process data for each primary node; this can include raw material usage, equipment operating status, production time, etc. The production process data is input into the carbon emission prediction model, and the carbon emission prediction value for each vector path is output. The predicted carbon emissions are compared with the actual carbon emissions to calculate the deviation value. The deviation value is calculated as |actual carbon emissions - predicted carbon emissions|, obtained from monitoring equipment or systems. A preset threshold is set based on historical data or industry standards to determine whether there is an anomaly.
[0032] If the deviation value exceeds the preset threshold, the carbon emissions of that vector path are determined to be abnormal.
[0033] Furthermore, the method also includes: For vector paths identified as having abnormal carbon emissions, the abnormal emissions are located based on entities and relationships in the carbon efficiency knowledge graph; this includes: Starting from the output entity of the vector path and ending at the first-level node of the vector path, the path of carbon emission anomalies is obtained by tracing back along the relationships in the knowledge graph. The output entity is the final manifestation node of the carbon emission anomaly, such as abnormal product carbon emission data or environmental monitoring data. Following the relationships in the knowledge graph, a path is formed by tracing back from the output entity to the first-level node, such as abnormal product carbon emissions → production line B → energy type C (coal) → factory A. Based on the aforementioned carbon emission anomaly path and anomaly cause database, the causes of the carbon emission anomalies are located. The anomaly cause database contains the following information: anomaly type, such as equipment failure, abnormal process parameters, changes in environmental conditions, etc.; anomaly cause description, such as equipment B failure leading to increased emissions; and the probability or weight of the anomaly cause. Furthermore, locating the causes of carbon emission anomalies includes: The probability of each anomaly cause is calculated based on the contribution of each node in the path of the carbon emission anomaly and the weight of each node in the anomaly cause database. Based on the aforementioned probabilities, the cause of the carbon emission anomaly can be identified, with the one with the highest probability being the cause of the carbon emission anomaly.
[0034] Furthermore, the probability of each abnormal cause is calculated using the following formula: Among them, S ij C represents the contribution of secondary node i in the path of carbon emission anomalies to the cause j of the anomaly. i R represents the contribution of second-level node i in the path of carbon emission anomalies. The contribution can be obtained from the carbon emissions and influencing factors of the second-level node. ij Let T represent the degree of association between the i-th node and the j-th anomaly cause. The degree of association can be obtained from a knowledge graph. j W represents the total contribution of the anomaly cause j to the path, n represents the total number of nodes in the path, and W represents the total number of nodes in the path. jP represents the weight of the abnormal cause j, m represents the total number of abnormal causes, and P represents the weight of the abnormal cause j. j This represents the probability of the abnormal cause j.
[0035] This invention constructs a carbon efficiency knowledge graph, representing carbon emission-related entities and relationships in a graph structure. This captures complex correlation information, providing richer features for carbon emission prediction. Furthermore, the carbon efficiency knowledge graph allows for tracing the causal paths of anomalies along the relationships between entities, clarifying the root causes of anomalies, and improving the accuracy of locating the causes of carbon emission anomalies. Combined with an anomaly cause database, it can provide specific causes for each anomaly, enhancing the interpretability and practicality of the diagnostic results.
[0036] Example 2 A carbon efficiency diagnostic system based on intelligent knowledge graphs includes: First acquisition module: Acquire historical carbon emission data; Extraction module: Extracts carbon emission entities from the historical carbon emission data, wherein the carbon emission entities include primary nodes and secondary nodes; First construction module: Based on the carbon emission entities, construct a carbon efficiency knowledge graph, wherein the first-level nodes and second-level nodes are connected by edges, and the edges represent the relationships between entities; The second construction module: Based on the entities and relationships in the carbon efficiency knowledge graph, a carbon emission prediction model is constructed. The carbon emission prediction model outputs carbon emission prediction values based on the input production process data. Diagnostic module: Based on the predicted carbon emissions, perform carbon efficiency diagnosis and identify carbon emission anomalies; The second acquisition module: Based on the entities and relationships in the carbon efficiency knowledge graph, it acquires the abnormal causes of carbon emission anomalies.
[0037] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.
[0038] The present invention has been described above with reference to embodiments thereof. However, these embodiments are merely illustrative and not intended to limit the scope of the invention. The scope of the invention is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of the invention, and all such substitutions and modifications should fall within the scope of the invention.
[0039] Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions, and modifications can be made to the embodiments of the present invention without departing from the spirit and scope of the invention.
[0040] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.
Claims
1. A carbon efficiency diagnosis method based on intelligent knowledge graphs, characterized in that, include: Carbon emission entities are identified based on historical carbon emission data, and these carbon emission entities include primary nodes and secondary nodes. Based on the carbon emission entity, a carbon efficiency knowledge graph is constructed. The carbon efficiency knowledge graph includes first-level nodes and second-level nodes. The first-level nodes and second-level nodes are connected by edges, and the edges represent the relationship between the connected nodes. Based on the carbon efficiency knowledge graph, a carbon emission prediction model is constructed, which outputs a carbon emission prediction value based on the input production process data. Carbon efficiency diagnosis is performed based on the predicted carbon emissions.
2. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 1, characterized in that, The carbon efficiency knowledge graph includes multiple vector paths, each of which includes a first-level node and multiple second-level nodes, and the first-level node and multiple second-level nodes are connected by multiple edges.
3. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 1, characterized in that, The carbon emission prediction model is based on a graph neural network. The carbon emission prediction model includes a graph embedding layer, a graph convolutional layer, a time series processing layer, and a prediction layer; The graph embedding layer embeds entities and relationships in the carbon efficiency knowledge graph into a low-dimensional vector space; The graph convolutional layer updates the entity representation by aggregating information of neighboring nodes of entities in the graph embedding layer through graph convolution, thereby obtaining graph structure features in carbon emission data; The time series processing layer is used to extract time features from carbon emission data; The prediction layer is used to output carbon emission prediction values based on the graph structure features and the time features.
4. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 2, characterized in that, The carbon efficiency diagnosis based on the predicted carbon emissions includes: Collect production process data for each primary node; The production process data is input into the carbon emission prediction model, and the carbon emission prediction value for each vector path is output. The predicted carbon emissions are compared with the actual carbon emissions, and the deviation is calculated. If the deviation value exceeds the preset threshold, the carbon emissions of that vector path are determined to be abnormal.
5. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 4, characterized in that, The method further includes: For vector paths identified as having abnormal carbon emissions, the abnormal emissions are located based on entities and relationships in the carbon efficiency knowledge graph; this includes: Starting from the output entity of the vector path and ending at the first-level node of the vector path, backtrack along the relationships in the knowledge graph to obtain the path of abnormal carbon emissions. Based on the aforementioned database of carbon emission anomalies and their causes, the causes of the anomalies are located.
6. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 3, characterized in that, The graph embedding layer embeds entities and relationships in the carbon efficiency knowledge graph into a low-dimensional vector space, including the following steps: Each entity and each relationship is mapped to a low-dimensional vector, and the low-dimensional vectors of the entity and relationship are obtained. The mapping is represented by the following formula: Among them, e i Represents an entity, where i represents the entity number, and h represents the entity ID. i Represents entity e i low-dimensional vectors Represents the mapping function, r k Let k represent the relation, and w represent the relation number. k Representing relation r k a low-dimensional vector; The entity low-dimensional vector and the relation low-dimensional vector are combined into a triplet; The triplet is then optimized.
7. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 6, characterized in that, The triplet is optimized using the following formula: ; Where U represents the optimization function, e i Represents the entity with number i, e j Represents entity j, r k Indicates a relation, (e i r k e j ) represents a triple, and G represents the geometry of all triples. h represents the Euclidean distance. j Represents entity e j A low-dimensional vector.
8. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 6, characterized in that, The update is performed using the following formula: ; in, This represents the updated entity representation. It is the l-th level entity e i The representation of, σ It is an activation function. It is the learnable weight matrix of the l-th layer. It is entity e i The set of neighbors, Entity e j The set of neighbors.
9. The carbon efficiency diagnosis method based on intelligent knowledge graph according to claim 5, characterized in that, Based on the aforementioned database of carbon emission anomalies and their causes, the causes of the carbon emission anomalies are located, including: The probability of each anomaly cause is calculated based on the contribution of each node in the path of the carbon emission anomaly and the weight of each node in the anomaly cause database. Based on the stated probability, the abnormal cause of the carbon emission anomaly can be identified.
10. A carbon efficiency diagnostic system based on intelligent knowledge graphs, characterized in that, include: Determination Module: Determines carbon emission entities based on historical carbon emission data, wherein the carbon emission entities include primary nodes and secondary nodes; First construction module: Based on the carbon emission entity, construct a carbon efficiency knowledge graph, which includes first-level nodes and second-level nodes. The first-level nodes and second-level nodes are connected by edges, and the edges represent the relationship between the connected nodes. The second construction module: Based on the carbon efficiency knowledge graph, a carbon emission prediction model is constructed, and the carbon emission prediction model outputs the carbon emission prediction value according to the input production process data; Diagnostic module: Performs carbon efficiency diagnosis based on the predicted carbon emissions.
