A method and system for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants

Abstract

This invention discloses a method and system for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants, relating to the field of wastewater treatment plant monitoring technology. The method includes analyzing the wastewater treatment plant's on-site layout, constructing a structural representation model containing process blocks and equipment nodes, and mapping this model to the actual layout; constructing a wastewater treatment plant knowledge graph, and labeling the process blocks and equipment nodes in the structural representation model based on knowledge explanations to form a first model mapping structure; performing content analysis on heterogeneous data, extracting key factors and mapping them to the structural representation model to form a second model mapping structure; adapting the second model mapping structure to the first model mapping structure to determine the information mapping logic of the heterogeneous data, and storing it according to the knowledge graph logic. This method achieves the integration and accurate mapping of heterogeneous data through structural models and knowledge graphs, solving the problems of data silos and insufficient correlation, improving the data management and operational efficiency of wastewater treatment plants, and is suitable for intelligent wastewater treatment scenarios.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment plant monitoring technology, and in particular to a method and system for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants. Background Technology

[0002] With the rapid development of the wastewater treatment industry and the continuous improvement of environmental protection requirements, the operation and management of wastewater treatment plants are moving towards intelligent and refined management. Wastewater treatment plants involve multiple data sources, including equipment operating status, process parameters, operation and maintenance records, material consumption, personnel management, standards and specifications, video surveillance data, and inspection records. This data exhibits significant heterogeneity, manifested in inconsistent data formats (e.g., structured data, semi-structured data, video streams), dispersed sources (e.g., sensors, monitoring equipment, manual records), and complex storage structures. This heterogeneity makes effective data integration difficult, creating data silos and severely limiting the improvement of wastewater treatment plant operating efficiency and the accuracy of decision support.

[0003] Traditional data management methods, such as relational databases or simple file storage systems, are insufficient to meet the integration and analysis needs of wastewater treatment plants for diverse and heterogeneous data. First, the physical layout and process flow of wastewater treatment plants have a high degree of spatial and logical correlation, requiring data analysis to correspond to the actual scenarios of process blocks, equipment nodes, and video surveillance and inspection records. For example, video surveillance data needs to be associated with specific equipment or process areas, and inspection record data needs to be matched with equipment operating status and time series; traditional methods lack effective mapping mechanisms. Second, massive amounts of heterogeneous data contain a large amount of redundant information. Accurately extracting key factors (such as abnormal operating states and inspection anomalies) and associating them with the operating scenario of the wastewater treatment plant is a key challenge for achieving data-driven optimization. Furthermore, the operating environment of wastewater treatment plants is highly dynamic (e.g., equipment updates, process adjustments, changes in monitoring points), and traditional data warehouses lack sufficient adaptability and intelligent analysis capabilities, making it difficult to meet real-time and dynamic management requirements. Summary of the Invention

[0004] The purpose of this invention is to provide a data warehouse construction method and system for intelligent planning of diverse and heterogeneous data from wastewater treatment plants.

[0005] This invention discloses a method for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants, including:

[0006] Step S100: Analyze the on-site layout of the wastewater treatment plant and construct a structural representation model of the wastewater treatment plant. The structural representation model of the wastewater treatment plant includes several process blocks, each process block includes several equipment nodes, and the relative positions between process blocks and between equipment nodes are mapped to the on-site layout of the wastewater treatment plant.

[0007] Step S200: A knowledge graph is constructed for the wastewater treatment plant. For each information mapping logic in the knowledge graph, a corresponding knowledge explanation is set. Based on the knowledge explanation, the process blocks and equipment nodes in the wastewater treatment plant structural representation model are marked to form the first model mapping structure.

[0008] Step S300: Perform content analysis on the heterogeneous data, identify the key factors in each heterogeneous data, analyze the key factors, determine the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model, and form the second model mapping structure.

[0009] Step S400: Based on the first model mapping structure adapted to the second model mapping structure of heterogeneous data, determine the information mapping logic corresponding to the heterogeneous data, and classify and store the heterogeneous data according to the information mapping logic of the knowledge graph.

[0010] In some embodiments disclosed in this invention, the method for constructing a knowledge graph for a wastewater treatment plant includes:

[0011] Step S201: Determine the knowledge keywords of the wastewater treatment plant, including process block keywords, equipment keywords, parameter keywords, time keywords, spatial keywords, operation and maintenance record keywords, material keywords, personnel keywords, and standard keywords;

[0012] Step S202: Construct the logical relationship between knowledge keywords, and based on the logical relationship, connect the corresponding knowledge keywords in sequence to obtain the graph framework of the knowledge graph;

[0013] Step S203: Based on the knowledge keywords and logical relationships mapped by the graph framework, the preset knowledge explanation content is categorized.

[0014] In some embodiments disclosed in this invention, the method for marking process blocks and equipment nodes in a wastewater treatment plant structural representation model to form a first model mapping structure includes:

[0015] Step S204: Perform keyword analysis on the knowledge explanation content, extract process block keywords and equipment node keywords, and mark the process blocks and equipment nodes in the wastewater treatment plant structural representation model based on the process block keywords and equipment node keywords.

[0016] Step S205: Perform event impact weight analysis on process blocks or equipment nodes in the knowledge explanation content. Based on the magnitude of the event impact weight, sort the process block keywords and equipment node keywords. Based on the sorting results, connect the marked process blocks and equipment nodes in the wastewater structure representation model with vector lines to obtain the first model mapping structure.

[0017] In some embodiments disclosed in this invention, a method for performing content analysis on heterogeneous data to determine key factors in each heterogeneous data set includes:

[0018] Step S301: A factor extraction template is set for each type of heterogeneous data. Several factor types are set on the factor extraction template, and a factor parameter segment sequence is set for each factor type.

[0019] Step S302: Mark the key factors of the heterogeneous data of the historical records and determine the factor parameters of each key factor to obtain several sets of historical key factor parameters.

[0020] Step S303: Based on the consistency of key factor types as the classification condition, different historical key factor parameter groups are classified, and a concentrated feature analysis is performed on the historical key factor parameter groups in each category. Based on the analysis results, several factor parameter segment array templates are constructed. The factor parameter segment array template includes several content factor types, and each content factor type corresponds to several factor parameter segment sequences. The factor parameter segment sequence includes several factor parameter segments connected in order of size, and the factor parameter segments that need to be focused on are marked.

[0021] Step S304: Identify the content factors in the heterogeneous data, compare the content factors with different factor parameter segment array templates, determine the factor parameter segment array template that is fully mapped and adapted, and identify it as the reference factor parameter segment array template.

[0022] Step S305: Identify the factor types and the factor parameter segments of interest in the reference factor parameter segment array template as key factors of heterogeneous data.

[0023] In some embodiments disclosed in this invention, a method for performing centralized feature analysis on the historical key factor parameter group in each category includes:

[0024] Step S3031: Extract historical key factor parameters of the same category from the historical key factor parameter group, sort the historical key factor parameters in order, determine the clusters of historical key factor parameters gathered in the sequence, and calculate the average value in each historical key factor parameter cluster, which is recorded as the key factor parameter to be focused on.

[0025] Step S3032: Analyze the number of factor parameters and the span of factor parameters in the historical key factor parameter cluster, calculate the number density between the number of factor parameters and the span of factor parameters, and determine the extension segment length of the factor parameter segment corresponding to the historical key factor parameter cluster based on the preset density interval to which the number density belongs.

[0026] Step S3033: Randomly combine the factor parameter segments corresponding to different types of historical key factor parameter clusters to obtain several initial factor parameter segment array templates.

[0027] Step S3034: Map and compare the historical key factor parameter groups and the initial factor parameter segment array templates to determine the mapping matching performance. Based on the mapping matching performance, sort the initial factor parameter segment array templates and select the first few initial factor parameter segment array templates as the applied factor parameter segment array templates.

[0028] In some embodiments disclosed in this invention, the method for mapping and comparing historical key factor parameter groups and initial factor parameter segment array templates to determine the mapping fit performance includes:

[0029] Step S30341: Determine the segment center value of each type of factor parameter segment in the initial factor parameter segment array template, and calculate the factor parameter difference value between the corresponding historical key factor parameter and the segment center value in the historical key factor parameter group.

[0030] Step S30342: Determine the proportion of factor type matching between the initial factor parameter segment array template and the historical key factor parameter group, and combine the factor parameter difference value of each factor type to determine the mapping matching sub-parameter between the initial factor parameter segment array template and the single historical key factor parameter group.

[0031] Step S30343: Randomly select several mapping matching sub-parameters that are greater than or equal to preset values, and calculate their average value to obtain the mapping matching degree.

[0032] In some embodiments disclosed in this invention, the method for analyzing key factors and determining the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model includes:

[0033] Step S306: Establish a key factor ~ block and equipment mapping table. The key factor ~ block and equipment mapping table includes several key factor types, and each key factor type corresponds to a key factor parameter segment. Each key factor type and key factor parameter segment corresponds to several process blocks and equipment nodes, and there is an order relationship between the process blocks and equipment nodes.

[0034] Step S307: Substitute the determined key factors into the key factor ~ block and equipment mapping table to determine the process blocks and equipment nodes mapped in the wastewater structure representation model, and connect the process blocks and equipment nodes mapped in the wastewater structure representation model with vector lines based on the determined order relationship between the process blocks and equipment nodes.

[0035] In some embodiments disclosed in this invention, the method for determining the compatibility between the second model mapping structure and the first model mapping structure includes:

[0036] Step S401: Determine the envelope blocks of the first model mapping structure and the second model mapping structure on the wastewater treatment plant structural representation model, denoted as the first envelope block and the second envelope block, respectively. Analyze the cross volume between the first envelope block and the second envelope block, and determine the sum of the volumes of the first envelope block and the second envelope block, denoted as the total block volume. Calculate the volume ratio of the cross volume to the total block volume.

[0037] Step S402: Compare the line overlap ratio of the vector lines in the first model mapping structure and the second model mapping structure, and combine it with the volume ratio to determine the degree of fit between the first model mapping structure and the second model mapping structure.

[0038] In some embodiments disclosed in this invention, a multi-dimensional heterogeneous data warehouse construction system for wastewater treatment plants is also disclosed, characterized in that it includes:

[0039] The first module is used to analyze the on-site layout of the wastewater treatment plant and construct a structural representation model of the wastewater treatment plant. The structural representation model of the wastewater treatment plant includes several process blocks, each process block includes several equipment nodes, and the relative positions between process blocks and between equipment nodes are mapped to the on-site layout of the wastewater treatment plant.

[0040] The second module is used to construct a knowledge graph for the wastewater treatment plant. For each information mapping logic in the knowledge graph, a corresponding knowledge explanation is set. Based on the knowledge explanation, the process blocks and equipment nodes in the wastewater treatment plant structural representation model are marked to form the first model mapping structure.

[0041] The third module is used to perform content analysis on heterogeneous data, identify the key factors in each heterogeneous data, analyze the key factors, determine the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model, and form the second model mapping structure.

[0042] The fourth module is used to adapt the second model mapping structure based on heterogeneous data to the first model mapping structure, determine the information mapping logic corresponding to the heterogeneous data, and classify and store the heterogeneous data according to the information mapping logic of the knowledge graph.

[0043] This invention discloses a method and system for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants, relating to the field of wastewater treatment plant monitoring technology. The method includes analyzing the wastewater treatment plant's on-site layout, constructing a structural representation model containing process blocks and equipment nodes, and mapping this model to the actual layout; constructing a wastewater treatment plant knowledge graph, and labeling the process blocks and equipment nodes in the structural representation model based on knowledge explanations to form a first model mapping structure; performing content analysis on heterogeneous data, extracting key factors and mapping them to the structural representation model to form a second model mapping structure; adapting the second model mapping structure to the first model mapping structure to determine the information mapping logic of the heterogeneous data, and storing it according to the knowledge graph logic. This method achieves the integration and accurate mapping of heterogeneous data through structural models and knowledge graphs, solving the problems of data silos and insufficient correlation, improving the data management and operational efficiency of wastewater treatment plants, and is suitable for intelligent wastewater treatment scenarios.

[0044] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating the method steps for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant, as disclosed in an embodiment of the present invention. Detailed Implementation

[0046] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0047] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only for illustration and explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Those skilled in the art can make some non-essential improvements and adjustments based on the following content of the present invention. In the present invention, unless otherwise expressly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art.

[0048] Example:

[0049] The purpose of this invention is to provide a data warehouse construction method and system for intelligent planning of diverse and heterogeneous data from wastewater treatment plants.

[0050] This invention discloses a method for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants. (See reference...) Figure 1 ,include:

[0051] Step S100: Analyze the on-site layout of the wastewater treatment plant and construct a structural representation model of the wastewater treatment plant. The structural representation model of the wastewater treatment plant includes several process blocks, and each process block includes several equipment nodes. The relative positions between process blocks and between equipment nodes are mapped to the on-site layout of the wastewater treatment plant.

[0052] Step S100 involves an in-depth analysis of the wastewater treatment plant's on-site layout to construct a structural representation model. This model provides a spatialized and structured foundation for extracting key factors and mapping heterogeneous data, ensuring a precise correlation between the data and the actual physical layout and process flow. The operation and management of wastewater treatment plants are highly dependent on the characteristics of their physical layout and process flow. Key factors (such as equipment malfunctions, process parameter deviations, abnormal events in video surveillance, and fault descriptions in inspection records) are often directly related to specific process blocks or equipment nodes. Therefore, this step first conducts a systematic survey of the wastewater treatment plant's on-site layout, collecting information on the spatial distribution and functional division of each process block (such as pretreatment area, biological treatment area, and advanced treatment area), as well as the specific locations and connections of equipment nodes (such as pumps, valves, and aerators) within each process block. This information is transformed into a structural representation model using digital modeling technology, where the relative positions between process blocks and equipment nodes are mapped one-to-one with the actual on-site layout. For example, a biochemical treatment block may contain multiple equipment nodes (such as agitators and dissolved oxygen sensors), and the spatial relationships between them reflect functional dependencies in the process flow, such as the pipe connections between agitators and downstream filtration equipment. This mapping not only preserves the authenticity of the physical layout but also lays the foundation for the spatial association of key factors. For instance, abnormal events in video surveillance (such as pipe blockages) need to be located to specific process blocks or equipment nodes, while fault descriptions in inspection records may be related to the operating status of a particular piece of equipment. By constructing a structural representation model, the system can associate key factors (such as equipment malfunctions or process deviations) with physical locations, avoiding the analytical difficulties caused by the lack of spatial context in traditional data management. Furthermore, the model is dynamically scalable, adapting to actual changes in the wastewater treatment plant (such as the addition of new equipment or process adjustments), ensuring the continued effectiveness of the key factor mapping. For example, when a new aeration device is added to a biochemical treatment block, the model can update its node locations and associations, supporting the analysis of subsequent key factors (such as equipment malfunctions). This modeling approach effectively addresses the disconnect between key factors and physical layout, providing a unified structural framework for subsequent semantic labeling of knowledge graphs and accurate mapping of heterogeneous data, thereby enhancing the data-driven management capabilities of wastewater treatment plants.

[0053] In step S200, a knowledge graph is constructed for the wastewater treatment plant. For each information mapping logic in the knowledge graph, a corresponding knowledge explanation is set. Based on the knowledge explanation, the process blocks and equipment nodes in the wastewater treatment plant structural representation model are marked to form the first model mapping structure.

[0054] Step S200 constructs a knowledge graph of the wastewater treatment plant and semantically labels the structural representation model based on the knowledge explanation content, forming a first model mapping structure. This provides a logical foundation for the semantic association of key factors and the intelligent analysis of heterogeneous data. The operation of a wastewater treatment plant involves complex processes and equipment interactions. Key factors (such as equipment malfunctions, process parameter deviations, abnormal events in video surveillance, and fault descriptions in inspection records) need to accurately correspond to the functional semantics of the process flow and equipment. Knowledge graphs, as a graph-structured knowledge representation method, effectively integrate the operational knowledge of wastewater treatment plants by describing entities and their relationships through nodes and edges. First, the system identifies knowledge keywords related to the wastewater treatment plant, including process blocks, equipment, parameters, time, space, operation and maintenance records, video surveillance events, and inspection faults, and constructs the logical relationships between these keywords. For example, a process block may have an "inclusion" relationship with a specific piece of equipment, equipment may have a "monitoring" relationship with operating parameters, and abnormal events in video surveillance may have an "association" relationship with fault descriptions in inspection records. These logical relationships are structurally connected through a graph framework to form a knowledge graph, where each information mapping logic corresponds to specific knowledge explanations (such as the operating parameter range of equipment, the functional description of a process block, and the characteristics of abnormal events). Based on this, the system semantically labels the process blocks and equipment nodes in the structural representation model of step S100, mapping the semantic information of the knowledge graph onto the physical structure. For example, a biochemical treatment block may be labeled as "anaerobic treatment," and its contained equipment nodes may be labeled as "anaerobic reactors," and associated with operating parameters (such as dissolved oxygen concentration) and key factors (such as equipment malfunctions or fault images in video surveillance). This labeling process combines the physical structure with the semantic context, forming the first model mapping structure, enabling key factors (such as process parameter deviations or inspection fault descriptions) to be systematically analyzed based on the logical relationships of the knowledge graph. For example, the description of "insufficient aeration" in the inspection record can be associated with the operating parameters of the aeration equipment and abnormal images in the video surveillance through the knowledge graph, avoiding the problem of isolated information. Furthermore, the dynamic updating capability of the knowledge graph enables it to adapt to operational changes in wastewater treatment plants (such as the addition of monitoring points or process adjustments), ensuring the continuity and accuracy of key factor analysis. Through this step, the system lays the foundation for semantic management of key factors and the integration of heterogeneous data, significantly improving the level of intelligence in data analysis.

[0055] Step S300: Perform content analysis on the heterogeneous data, identify the key factors in each heterogeneous data, analyze the key factors, determine the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model, and form the second model mapping structure.

[0056] Step S300 involves content analysis of heterogeneous data to extract key factors (such as equipment malfunctions, process parameter deviations, abnormal events in video surveillance, and fault descriptions in inspection records) and mapping them to a structural representation model, forming a second model mapping structure. This achieves precise correlation between key factors and physical layout and process flow. Wastewater treatment plants generate a wide range of heterogeneous data, including equipment sensor data, video surveillance streams, inspection records, and process parameters. These data differ significantly in format and content, and key factors are often hidden within massive amounts of information. This step first uses a factor extraction template to perform structured analysis on the heterogeneous data. The template defines the types of key factors (such as abnormal events, parameter deviations, and fault descriptions) and their parameter segment sequences (such as parameter ranges and event times). For example, for video surveillance data, a key factor might be an abnormal event (such as equipment malfunction images or pipe blockages); for inspection records, a key factor might be a fault description (such as "motor overheating") or inspection time. Subsequently, the system accurately extracts key factors through historical data analysis and the factor parameter segment array template, determining their mapping positions in the structural representation model. For example, a "pipe blockage" event in video surveillance might be mapped to a specific pipeline equipment node in the pretreatment block, while a "insufficient aeration" description in an inspection record might be mapped to an aerator node in the biochemical treatment block. This mapping process, through the association of key factors with process blocks and equipment nodes, forms a second model mapping structure. This not only preserves the original content of the data but also enhances the spatial and process context of the key factors. For instance, parameter data indicating abnormal equipment operation can be mapped to specific equipment, facilitating subsequent fault diagnosis; abnormal events in video surveillance can be cross-validated with fault descriptions in inspection records, improving the reliability of the analysis. Compared to traditional data analysis methods, this step, through template-based and historical data analysis, significantly improves the accuracy and efficiency of key factor extraction, solving the challenge of associating heterogeneous data with the process flow. Furthermore, this step supports dynamic updates; when new key factor types (such as anomalies captured by newly added monitoring equipment) appear, the system can adjust the templates and mapping rules to ensure the adaptability of the analysis. By forming a second model mapping structure, the system provides a reliable foundation for the systematic management of key factors and subsequent data classification and storage, laying the technical support for the intelligent operation optimization of wastewater treatment plants.

[0057] Step S400: Based on the first model mapping structure adapted to the second model mapping structure of heterogeneous data, determine the information mapping logic corresponding to the heterogeneous data, and classify and store the heterogeneous data according to the information mapping logic of the knowledge graph.

[0058] Step S400 determines the information mapping logic of key factors (such as equipment malfunctions, process parameter deviations, abnormal events in video surveillance, and fault descriptions in inspection records) by adapting the second model mapping structure to the first model mapping structure. It then classifies and stores the heterogeneous data according to the logic of the knowledge graph, constructing an efficient and semantic data warehouse. The heterogeneous data from the wastewater treatment plant has already formed a second model mapping structure through key factor mapping in step S300, but this data still needs to be integrated with the semantic logic of the knowledge graph to achieve systematic storage and intelligent analysis. This step first evaluates the fit between the second model mapping structure (based on heterogeneous data and key factors) and the first model mapping structure (based on knowledge graph labels), calculating the fit by analyzing the overlap volume of their envelope blocks and the proportion of overlapping vector lines. For example, an "equipment malfunction" event in video surveillance might be mapped to the "fault diagnosis" logic related to that equipment in the knowledge graph, while a "parameter exceeding limits" description in the inspection record might be mapped to the "process anomaly" logic. Based on this adaptation, the system determines the information mapping logic corresponding to key factors and classifies and stores heterogeneous data according to the structure of a knowledge graph. For example, abnormal equipment operation data is stored in the "Equipment Status" category, abnormal events from video surveillance are stored in the "Monitoring Events" category, and fault descriptions from inspection records are stored in the "Maintenance Records" category. This categorized storage not only improves data retrieval efficiency but also achieves a systematic organization of key factors through the semantic relationships of the knowledge graph, enabling subsequent analysis to quickly locate relevant process blocks or equipment nodes. For example, when querying abnormal events in a certain process block, the system can simultaneously retrieve relevant video surveillance data and inspection records, improving the comprehensiveness of the analysis. In addition, this step supports the scalability of the data warehouse through a dynamic adaptation mechanism. When the wastewater treatment plant introduces new data types (such as data from newly added monitoring equipment) or key factors (such as new failure modes), the system can update the mapping logic and storage structure to ensure continuous applicability. Compared to traditional data warehouses, this step addresses the complexity and lack of intelligence in the classification and storage of heterogeneous data through semantic integration of key factors and model adaptation. It provides an efficient and unified data management platform for the operation optimization of wastewater treatment plants, significantly improving decision support capabilities based on key factors.

[0059] In some embodiments disclosed in this invention, the method for constructing a knowledge graph for a wastewater treatment plant includes:

[0060] Step S201: Determine the knowledge keywords of the wastewater treatment plant, including process block keywords, equipment keywords, parameter keywords, time keywords, spatial keywords, operation and maintenance record keywords, material keywords, personnel keywords, and standard keywords.

[0061] Step S202: Construct the logical relationships between knowledge keywords, and based on the logical relationships, connect the corresponding knowledge keywords in sequence to obtain the graph framework of the knowledge graph.

[0062] Step S203: Based on the knowledge keywords and logical relationships mapped by the graph framework, the preset knowledge explanation content is categorized.

[0063] In some embodiments disclosed in this invention, the method for marking process blocks and equipment nodes in a wastewater treatment plant structural representation model to form a first model mapping structure includes:

[0064] Step S204: Perform keyword analysis on the knowledge explanation content, extract process block keywords and equipment node keywords, and mark the process blocks and equipment nodes in the wastewater treatment plant structural representation model based on the process block keywords and equipment node keywords.

[0065] Step S205: Perform event impact weight analysis on process blocks or equipment nodes in the knowledge explanation content. Based on the magnitude of the event impact weight, sort the process block keywords and equipment node keywords. Based on the sorting results, connect the marked process blocks and equipment nodes in the wastewater structure representation model with vector lines to obtain the first model mapping structure.

[0066] Step S205 performs event impact weight analysis on process blocks or equipment nodes in the knowledge explanation content, adjusts the priority of keywords based on weight ranking, and connects the marked process blocks and equipment nodes with vector lines to form a first model mapping structure. This provides dynamic and structured support for the systematic association of key factors (such as equipment malfunctions, process parameter deviations, abnormal events in video surveillance, and fault descriptions in inspection records) and the intelligent integration of heterogeneous data. In the operation of a wastewater treatment plant, different process blocks and equipment nodes have varying degrees of influence on key factors. For example, the impact of "aerators" on "dissolved oxygen concentration deviation" may be much greater than that of other equipment, while the "pretreatment" block has a more significant impact on the "pipeline blockage" event. This step first quantifies the event impact in the knowledge explanation content, assessing the correlation strength between each process block and equipment node and events related to key factors. For example, by analyzing historical data and knowledge explanation content, the system may determine that the weight of "aerators" and "dissolved oxygen concentration abnormalities" events is higher than that of "pumps," because the former directly affects the biochemical treatment effect. Similarly, a "pipeline blockage event in video surveillance" might have a higher weight in the "preprocessing" block, while a "motor overheating fault in inspection records" might be related to the weight of a specific equipment node (such as "motor"). Based on these weights, the system sorts the process block keywords and equipment node keywords, placing high-priority keywords (such as "aerator" or "preprocessing" closely related to key factors) at the top. Subsequently, based on the sorting results, the system connects the marked process blocks and equipment nodes with vector lines in the structural representation model, generating the first model mapping structure. These vector lines not only represent physical or process connections (such as pipeline connections) but also reflect the priority and logical relationships of event impacts. For example, a "biochemical treatment" block might be connected to "aerator" via a high-weight vector line and further connected to the "dissolved oxygen concentration" parameter node, forming a mapping path reflecting the key factor of "dissolved oxygen concentration deviation." This structured mapping allows key factors to be systematically represented and analyzed in the model. For example, when analyzing "pipeline blockage," the system can quickly locate the "preprocessing" block and its associated video surveillance and inspection data through vector lines. Compared to traditional data association methods, this step, through weight analysis and vector line connections, solves the problem of difficulty in quantifying and associating the influence of key factors. Furthermore, this step supports dynamic adjustment; when the operating scenario of the wastewater treatment plant changes (such as the addition of new equipment or failure modes), the system can re-analyze the event impact weights and update the mapping structure, ensuring the accuracy and adaptability of key factor analysis. The formation of the first model mapping structure not only provides a semantic and priority foundation for the subsequent adaptation and classification storage of heterogeneous data, but also significantly improves the intelligence level of the wastewater treatment plant's data warehouse, providing strong technical support for operational optimization and decision support.

[0067] In some embodiments disclosed in this invention, a method for performing content analysis on heterogeneous data to determine key factors in each heterogeneous data set includes:

[0068] Step S301: For each type of heterogeneous data, a factor extraction template is set. The factor extraction template has several factor types, and each factor type has a factor parameter segment sequence.

[0069] Step S301 establishes factor extraction templates for each type of heterogeneous data, defining factor types and their parameter segment sequences. This provides a structured framework for the accurate extraction of key factors (such as equipment malfunctions, process parameter deviations, abnormal events in video surveillance, and fault descriptions in inspection records), ensuring that heterogeneous data can be systematically analyzed and associated with the operational scenarios of wastewater treatment plants. Wastewater treatment plants generate diverse heterogeneous data from various sources, including equipment sensor data, video surveillance streams, inspection records, and process parameters. These data vary significantly in format and content, and key factors are often hidden within complex information. To effectively extract these key factors, this step first designs dedicated factor extraction templates based on the type of heterogeneous data (such as structured sensor data, semi-structured inspection records, and unstructured video surveillance data). Each template contains several factor types, such as "equipment status" (corresponding to equipment malfunctions), "process parameters" (corresponding to parameter deviations), "abnormal events" (corresponding to anomalies in video surveillance), and "fault descriptions" (corresponding to inspection records). Each factor type further defines a factor parameter segment sequence, i.e., the quantification range or characteristic description of the key factor. For example, for the "process parameter" type, the parameter segment sequence might include the normal range (e.g., 2-4 mg / L) and abnormal range (e.g., <1 mg / L) of dissolved oxygen concentration; for the "abnormal event" type, the parameter segment might include features of abnormal images in video surveillance (e.g., "pipeline blockage" or "equipment vibration"). These templates are designed based on the operating characteristics of the wastewater treatment plant, historical data analysis, and industry standards to ensure coverage of the diversity of key factors. For example, a "motor overheating" fault description in an inspection record might be defined as a specific parameter segment of the "fault description" factor type, directly related to abnormal equipment operation. By setting factor extraction templates, the system provides a standardized reference framework for extracting key factors, avoiding the extraction difficulties caused by data heterogeneity in traditional data analysis. Furthermore, the templates are scalable and can be updated based on new data types in the wastewater treatment plant (e.g., data from newly added monitoring equipment) or key factors (e.g., new fault modes). This structured, template-based approach not only improves the efficiency and accuracy of key factor extraction but also lays the foundation for subsequent historical data analysis and factor classification, ensuring that key factors can be accurately associated with process flows and equipment nodes, thus providing crucial support for the intelligent construction of wastewater treatment plant data warehouses.

[0070] Step S302: Mark the key factors of the heterogeneous data of the historical records and determine the factor parameters of each key factor to obtain several sets of historical key factor parameters.

[0071] Step S303: Based on the consistency of key factor types as the classification condition, different historical key factor parameter groups are classified, and a centralized feature analysis is performed on the historical key factor parameter groups in each category. Based on the analysis results, several factor parameter segment array templates are constructed. The factor parameter segment array template includes several content factor types, and each content factor type corresponds to several factor parameter segment sequences. The factor parameter segment sequence includes several factor parameter segments connected in order of size, and the factor parameter segments that need to be focused on are marked.

[0072] Step S303 categorizes historical key factor parameter groups based on the consistency of key factor types, performs centralized feature analysis on each category, and constructs a factor parameter segment array template. This provides a dynamic and standardized reference framework for the accurate identification of key factors (such as equipment operation anomalies, process parameter deviations, abnormal events in video surveillance, and fault descriptions in inspection records) and the systematic mapping of heterogeneous data. Wastewater treatment plant operation data is complex and variable, and the characteristics of key factors may differ in different scenarios. For example, the threshold for "process parameter deviation" varies in different process blocks, and the characteristics of "abnormal events in video surveillance" vary depending on the equipment type. This step first classifies the historical key factor parameter groups generated in step S302 based on the key factor type (such as "equipment status," "abnormal events," and "fault descriptions"). For example, the "dissolved oxygen concentration deviation" parameter group is classified into the "process parameters" category, the "pipeline blockage image" parameter group is classified into the "abnormal events" category, and the "motor overheating" parameter group is classified into the "fault description" category. Subsequently, the system performs a concentrated feature analysis on the parameter groups within each category, extracting the statistical characteristics and distribution patterns of key factors. For example, by analyzing the parameter groups in the "process parameters" category, the system may identify common abnormal ranges in dissolved oxygen concentration (such as <1 mg / L or >6 mg / L); by analyzing the parameter groups in the "abnormal events" category, the system may identify typical features of pipeline blockage images (such as image grayscale changes). Based on these analysis results, the system constructs factor parameter segment array templates. Each template contains several content factor types, and each type corresponds to a set of factor parameter segment sequences (such as numerical ranges or feature descriptions in sequential order), and marks the segments of interest. For example, a "process parameters" template may contain segment sequences of the "dissolved oxygen concentration" type (0-1 mg / L, 1-4 mg / L, 4-6 mg / L), where 0-1 mg / L is marked as a segment of interest, indicating a high-risk deviation; an "abnormal events" template may contain image feature sequences of the "pipeline blockage" type, where specific grayscale ranges are marked as areas of interest. These templates not only summarize the characteristic patterns of key factors but also provide standardized comparison criteria for factor extraction from new data. Compared to traditional data analysis methods, this step solves the problems of scattered and difficult-to-standardize key factor features through classification and feature analysis. Furthermore, the templates support dynamic optimization; when the operating environment of the wastewater treatment plant changes (such as the addition of new equipment or failure modes), the system can update the templates based on new historical data, ensuring the accuracy of key factor identification. These factor parameter segment array templates provide core support for subsequent factor comparison and key factor identification, promoting the intelligent construction of the data warehouse and data-driven decision-making for wastewater treatment plant operation optimization.

[0073] Step S304: Identify the content factors in the heterogeneous data, compare the content factors with different factor parameter segment array templates, determine the factor parameter segment array template that is fully mapped and adapted, and identify it as the reference factor parameter segment array template.

[0074] Step S305: Identify the factor types and the factor parameter segments of interest in the reference factor parameter segment array template as key factors of heterogeneous data.

[0075] In some embodiments disclosed in this invention, a method for performing centralized feature analysis on the historical key factor parameter group in each category includes:

[0076] Step S3031: Extract historical key factor parameters of the same category from the historical key factor parameter group, sort the historical key factor parameters in order, determine the clusters of historical key factor parameters gathered in the sequence, and calculate the average value in each historical key factor parameter cluster, which is recorded as the key factor parameter to be focused on.

[0077] Step S3032: Analyze the number of factor parameters and the span of factor parameters in the historical key factor parameter cluster, calculate the quantity density between the number of factor parameters and the span of factor parameters, and determine the extension segment length of the factor parameter segment corresponding to the historical key factor parameter cluster based on the preset density interval to which the quantity density belongs.

[0078] Step S3033: Randomly combine the factor parameter segments corresponding to different types of historical key factor parameter clusters to obtain several initial factor parameter segment array templates.

[0079] Step S3034: Map and compare the historical key factor parameter groups and the initial factor parameter segment array templates to determine the mapping matching performance. Based on the mapping matching performance, sort the initial factor parameter segment array templates and select the first few initial factor parameter segment array templates as the applied factor parameter segment array templates.

[0080] In some embodiments disclosed in this invention, the method for mapping and comparing historical key factor parameter groups and initial factor parameter segment array templates to determine the mapping fit performance includes:

[0081] Step S30341: Determine the segment center value of each type of factor parameter segment in the initial factor parameter segment array template, and calculate the factor parameter difference value between the corresponding historical key factor parameter and the segment center value in the historical key factor parameter group.

[0082] Step S30342: Determine the proportion of factor type matching between the initial factor parameter segment array template and the historical key factor parameter group, and combine the factor parameter difference value of each factor type to determine the mapping matching sub-parameter between the initial factor parameter segment array template and the single historical key factor parameter group.

[0083] Step S30343: Randomly select several mapping matching sub-parameters that are greater than or equal to preset values, and calculate their average value to obtain the mapping matching degree.

[0084] .

[0085] Where W represents the degree of mapping occurrence. This is the preset maximum factor parameter difference value. This represents the difference between the initial factor parameter segment array template and the i-th factor parameter in the historical key factor parameter group. This is a valid function for judging the difference value of factor parameters. If the difference value of the i-th factor parameter is less than or equal to the value of the corresponding factor parameter segment... ,but Output 1 otherwise output 0. This represents the number of factor types that match the initial factor parameter segment array template and the historical key factor parameter groups. The number of factor types in the initial factor parameter segment array template. L represents the number of factor types in the historical key factor parameter group, L is the factor type matching ratio influence adjustment coefficient, and b is the factor type matching ratio influence adjustment constant.

[0086] In some embodiments disclosed in this invention, the method for analyzing key factors and determining the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model includes:

[0087] Step S306: Establish a key factor ~ block and equipment mapping table. The key factor ~ block and equipment mapping table includes several key factor types, and each key factor type corresponds to a key factor parameter segment. Each key factor type and key factor parameter segment corresponds to several process blocks and equipment nodes, and there is an order relationship between the process blocks and equipment nodes.

[0088] Step S307: Substitute the determined key factors into the key factor ~ block and equipment mapping table to determine the process blocks and equipment nodes mapped in the wastewater structure representation model, and connect the process blocks and equipment nodes mapped in the wastewater structure representation model with vector lines based on the determined order relationship between the process blocks and equipment nodes.

[0089] In some embodiments disclosed in this invention, the method for determining the compatibility between the second model mapping structure and the first model mapping structure includes:

[0090] Step S401: Determine the envelope blocks of the first model mapping structure and the second model mapping structure on the wastewater treatment plant structural representation model, denoted as the first envelope block and the second envelope block, respectively. Analyze the cross volume between the first envelope block and the second envelope block, and determine the sum of the volumes of the first envelope block and the second envelope block, denoted as the total block volume. Calculate the volume ratio of the cross volume to the total block volume.

[0091] Step S402: Compare the line overlap ratio of the vector lines in the first model mapping structure and the second model mapping structure, and combine it with the volume ratio to determine the degree of fit between the first model mapping structure and the second model mapping structure.

[0092] .

[0093] Where S represents the degree of fit. The intersection volume of the first and second envelope blocks. The volume of the first envelope block. The volume of the second envelope block. The number of vector lines in the first model mapping structure. The number of vector lines in the second model mapping structure. The number of overlapping vector lines. The weights are affected by the cross volume. The weights are determined by the overlap of vector lines.

[0094] In some embodiments disclosed in this invention, a multi-dimensional heterogeneous data warehouse construction system for wastewater treatment plants is also disclosed, characterized in that it includes:

[0095] The first module is used to analyze the on-site layout of the wastewater treatment plant and construct a structural representation model of the wastewater treatment plant. The structural representation model of the wastewater treatment plant includes several process blocks, each process block includes several equipment nodes, and the relative positions between process blocks and between equipment nodes are mapped to the on-site layout of the wastewater treatment plant.

[0096] The second module is used to construct a knowledge graph for the wastewater treatment plant. For each information mapping logic in the knowledge graph, a corresponding knowledge explanation is set. Based on the knowledge explanation, the process blocks and equipment nodes in the wastewater treatment plant structural representation model are marked to form the first model mapping structure.

[0097] The third module is used to perform content analysis on heterogeneous data, identify the key factors in each heterogeneous data, analyze the key factors, determine the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model, and form the second model mapping structure.

[0098] The fourth module is used to adapt the second model mapping structure based on heterogeneous data to the first model mapping structure, determine the information mapping logic corresponding to the heterogeneous data, and classify and store the heterogeneous data according to the information mapping logic of the knowledge graph.

[0099] This invention discloses a method and system for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants, relating to the field of wastewater treatment plant monitoring technology. The method includes analyzing the wastewater treatment plant's on-site layout, constructing a structural representation model containing process blocks and equipment nodes, and mapping this model to the actual layout; constructing a wastewater treatment plant knowledge graph, and labeling the process blocks and equipment nodes in the structural representation model based on knowledge explanations to form a first model mapping structure; performing content analysis on heterogeneous data, extracting key factors and mapping them to the structural representation model to form a second model mapping structure; adapting the second model mapping structure to the first model mapping structure to determine the information mapping logic of the heterogeneous data, and storing it according to the knowledge graph logic. This method achieves the integration and accurate mapping of heterogeneous data through structural models and knowledge graphs, solving the problems of data silos and insufficient correlation, improving the data management and operational efficiency of wastewater treatment plants, and is suitable for intelligent wastewater treatment scenarios.

[0100] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant, characterized in that, include: Step S100: Analyze the on-site layout of the wastewater treatment plant and construct a structural representation model of the wastewater treatment plant. The structural representation model of the wastewater treatment plant includes several process blocks, each process block includes several equipment nodes, and the relative positions between process blocks and between equipment nodes are mapped to the on-site layout of the wastewater treatment plant. Step S200: A knowledge graph is constructed for the wastewater treatment plant. For each information mapping logic in the knowledge graph, a corresponding knowledge explanation is set. Based on the knowledge explanation, the process blocks and equipment nodes in the wastewater treatment plant structural representation model are marked to form the first model mapping structure. Step S300: Perform content analysis on the heterogeneous data, identify the key factors in each heterogeneous data, analyze the key factors, determine the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model, and form the second model mapping structure. Step S400: Based on the first model mapping structure adapted to the second model mapping structure of heterogeneous data, determine the information mapping logic corresponding to the heterogeneous data, and classify and store the heterogeneous data according to the information mapping logic of the knowledge graph. Methods for constructing knowledge graphs for wastewater treatment plants include: Step S201: Determine the knowledge keywords of the wastewater treatment plant, including process block keywords, equipment keywords, parameter keywords, time keywords, spatial keywords, operation and maintenance record keywords, material keywords, personnel keywords, and standard keywords; Step S202: Construct the logical relationship between knowledge keywords, and based on the logical relationship, connect the corresponding knowledge keywords in sequence to obtain the graph framework of the knowledge graph; Step S203: Based on the knowledge keywords and logical relationships mapped by the graph framework, classify the preset knowledge explanation content; Methods for marking process blocks and equipment nodes in the structural representation model of a wastewater treatment plant to form the first model mapping structure include: Step S204: Perform keyword analysis on the knowledge explanation content, extract process block keywords and equipment node keywords, and mark the process blocks and equipment nodes in the wastewater treatment plant structural representation model based on the process block keywords and equipment node keywords. Step S205: Perform event impact weight analysis on process blocks or equipment nodes in the knowledge explanation content. Based on the magnitude of the event impact weight, sort the process block keywords and equipment node keywords. Based on the sorting results, connect the marked process blocks and equipment nodes in the wastewater structure representation model with vector lines to obtain the first model mapping structure.

2. The method for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant according to claim 1, characterized in that, Methods for performing content analysis on heterogeneous data to identify key factors in each heterogeneous dataset include: Step S301: A factor extraction template is set for each type of heterogeneous data. Several factor types are set on the factor extraction template, and a factor parameter segment sequence is set for each factor type. Step S302: Mark the key factors of the heterogeneous data of the historical records and determine the factor parameters of each key factor to obtain several sets of historical key factor parameters. Step S303: Based on the consistency of key factor types as the classification condition, different historical key factor parameter groups are classified, and a concentrated feature analysis is performed on the historical key factor parameter groups in each category. Based on the analysis results, several factor parameter segment array templates are constructed. The factor parameter segment array template includes several content factor types, and each content factor type corresponds to several factor parameter segment sequences. The factor parameter segment sequence includes several factor parameter segments connected in order of size, and the factor parameter segments that need to be focused on are marked. Step S304: Identify the content factors in the heterogeneous data, compare the content factors with different factor parameter segment array templates, determine the factor parameter segment array template that is fully mapped and adapted, and identify it as the reference factor parameter segment array template. Step S305: Identify the factor types and the factor parameter segments of interest in the reference factor parameter segment array template as key factors of heterogeneous data.

3. The method for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant according to claim 2, characterized in that, Methods for performing clustered feature analysis on historical key factor parameter groups in each category include: Step S3031: Extract historical key factor parameters of the same category from the historical key factor parameter group, sort the historical key factor parameters in order, determine the clusters of historical key factor parameters gathered in the sequence, and calculate the average value in each historical key factor parameter cluster, which is recorded as the key factor parameter to be focused on. Step S3032: Analyze the number of factor parameters and the span of factor parameters in the historical key factor parameter cluster, calculate the number density between the number of factor parameters and the span of factor parameters, and determine the extension segment length of the factor parameter segment corresponding to the historical key factor parameter cluster based on the preset density interval to which the number density belongs. Step S3033: Randomly combine the factor parameter segments corresponding to different types of historical key factor parameter clusters to obtain several initial factor parameter segment array templates. Step S3034: Map and compare the historical key factor parameter groups and the initial factor parameter segment array templates to determine the mapping matching performance. Based on the mapping matching performance, sort the initial factor parameter segment array templates and select the first few initial factor parameter segment array templates as the applied factor parameter segment array templates.

4. The method for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant according to claim 3, characterized in that, The method for mapping and comparing historical key factor parameter sets with initial factor parameter segment array templates to determine the mapping fit performance includes: Step S30341: Determine the segment center value of each type of factor parameter segment in the initial factor parameter segment array template, and calculate the factor parameter difference value between the corresponding historical key factor parameter and the segment center value in the historical key factor parameter group. Step S30342: Determine the proportion of factor type matching between the initial factor parameter segment array template and the historical key factor parameter group, and combine the factor parameter difference value of each factor type to determine the mapping matching sub-parameter between the initial factor parameter segment array template and the single historical key factor parameter group. Step S30343: Randomly select several mapping matching sub-parameters that are greater than or equal to preset values, and calculate their average value to obtain the mapping matching degree.

5. The method for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant according to claim 1, characterized in that, Methods for analyzing key factors to determine the process blocks and equipment nodes mapped in the structural representation model of a wastewater treatment plant include: Step S306: Establish a key factor ~ block and equipment mapping table. The key factor ~ block and equipment mapping table includes several key factor types, and each key factor type corresponds to a key factor parameter segment. Each key factor type and key factor parameter segment corresponds to several process blocks and equipment nodes, and there is an order relationship between the process blocks and equipment nodes. Step S307: Substitute the determined key factors into the key factor ~ block and equipment mapping table to determine the process blocks and equipment nodes mapped in the wastewater structure representation model, and connect the process blocks and equipment nodes mapped in the wastewater structure representation model with vector lines based on the determined order relationship between the process blocks and equipment nodes.

6. The method for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant according to claim 1, characterized in that, Methods for determining the compatibility between the second model mapping structure and the first model mapping structure include: Step S401: Determine the envelope blocks of the first model mapping structure and the second model mapping structure on the wastewater treatment plant structural representation model, denoted as the first envelope block and the second envelope block, respectively. Analyze the cross volume between the first envelope block and the second envelope block, and determine the sum of the volumes of the first envelope block and the second envelope block, denoted as the total block volume. Calculate the volume ratio of the cross volume to the total block volume. Step S402: Compare the line overlap ratio of the vector lines in the first model mapping structure and the second model mapping structure, and combine it with the volume ratio to determine the degree of fit between the first model mapping structure and the second model mapping structure.

7. A system for constructing a multi-dimensional heterogeneous data warehouse for wastewater treatment plants, characterized in that, The method for constructing a multi-dimensional heterogeneous data warehouse for a wastewater treatment plant according to any one of claims 1-6 includes: The first module is used to analyze the on-site layout of the wastewater treatment plant and construct a structural representation model of the wastewater treatment plant. The structural representation model of the wastewater treatment plant includes several process blocks, each process block includes several equipment nodes, and the relative positions between process blocks and between equipment nodes are mapped to the on-site layout of the wastewater treatment plant. The second module is used to construct a knowledge graph for the wastewater treatment plant. For each information mapping logic in the knowledge graph, a corresponding knowledge explanation is set. Based on the knowledge explanation, the process blocks and equipment nodes in the wastewater treatment plant structural representation model are marked to form the first model mapping structure. The third module is used to perform content analysis on heterogeneous data, identify the key factors in each heterogeneous data, analyze the key factors, determine the process blocks and equipment nodes mapped in the wastewater treatment plant structural representation model, and form the second model mapping structure. The fourth module is used to adapt the second model mapping structure based on heterogeneous data to the first model mapping structure, determine the information mapping logic corresponding to the heterogeneous data, and classify and store the heterogeneous data according to the information mapping logic of the knowledge graph.

Images