Water conservancy project construction quality digital acceptance management system based on BIM

Abstract

The invention relates to the technical field of water conservancy project quality management, in particular to a BIM-based water conservancy project construction quality digital acceptance management system, which is used for solving the problems that BIM and OWL ontology cannot be fused, semantic modeling and multi-source data association of quality control points and time state attributes are difficult to realize, and the construction quality of a water conservancy project is influenced in the prior art. BIM model information cannot be dynamically updated, intelligent construction suggestions cannot be generated, and the intelligence, precision and automation level of water conservancy project construction quality management and control is reduced. According to the method, BIM and OWL ontologies are fused through the water conservancy component quality intelligent control module, semantic modeling and multi-source data association of quality control points and time state attributes are achieved, and the intelligence, precision and automation level of water conservancy project construction quality control is remarkably improved by dynamically updating BIM model information and generating intelligent construction suggestions.

Description

Technical Field

[0001] This invention relates to the field of water conservancy project quality management technology, and more specifically, to a BIM-based digital acceptance management system for water conservancy project construction quality. Background Technology

[0002] Traditional water conservancy project acceptance has long been constrained by multiple technical bottlenecks. Fragmented data systems lead to scattered design, construction, and acceptance information. Relying on manual comparison of drawings and measured data is inefficient and prone to errors. Changes in construction quality are difficult to update the acceptance model in real time, causing discrepancies between the model and the actual situation, which affects the accuracy and reliability of the results.

[0003] Referring to patent application CN118657415A, a digital acceptance management platform for water conservancy engineering construction quality is disclosed. This invention analyzes the excavation accuracy, stability, and seepage drainage compliance of earthwork projects to obtain the quality compliance of earthwork projects and issues early warnings. It also analyzes the layout compliance and concrete pouring compliance to obtain the quality compliance of foundation projects and issues early warnings. This invention obtains the quality compliance of earthwork projects by analyzing the excavation accuracy, stability, and seepage drainage compliance of earthwork projects, avoiding the intervention of manual operation, reducing data errors, and thus improving the acceptance compliance of earthwork projects. It also obtains the quality compliance of foundation projects by analyzing the layout compliance and concrete pouring compliance. The intelligent analysis of data in foundation projects further improves the acceptance compliance of foundation projects. However, while the aforementioned reference patents achieve accurate quality assessment by intelligently analyzing construction parameters of earthwork, foundation, and reinforced concrete frames, reducing human error and improving project acceptance rates and construction efficiency, they cannot integrate BIM and OWL ontology. This makes it difficult to achieve semantic modeling of quality control points and temporal attributes, as well as multi-source data association. Furthermore, they cannot dynamically update BIM model information and generate intelligent construction suggestions, thus reducing the level of intelligence, precision, and automation in water conservancy project construction quality control. Simultaneously, they cannot construct quantifiable and interpretable acceptance strategy parameter adjustment models, making it difficult to dynamically optimize detection frequency, detection point layout, and judgment thresholds. This hinders the transformation of acceptance from static and fixed to dynamic and precise, reducing acceptance efficiency, relevance, and the level of project quality assurance.

[0004] To address these issues, we propose a BIM-based digital acceptance management system for water conservancy project construction quality. Summary of the Invention

[0005] The purpose of this invention is to provide a BIM-based digital acceptance management system for water conservancy engineering construction quality. This system addresses the problems of existing technologies that cannot integrate BIM and OWL ontology, making it difficult to achieve semantic modeling of quality control points and temporal attributes, as well as multi-source data association. It also fails to dynamically update BIM model information and generate intelligent construction suggestions, thus reducing the intelligence, precision, and automation levels of water conservancy engineering construction quality control. Furthermore, it cannot construct quantifiable and interpretable models for adjusting acceptance strategy parameters, making it difficult to dynamically optimize detection frequency, detection point layout, and judgment thresholds. This hinders the transformation of acceptance from static and fixed to dynamic and precise, reducing acceptance efficiency, relevance, and the level of engineering quality assurance.

[0006] The objective of this invention is achieved through the following technical solution: A BIM-based digital acceptance management system for water conservancy project construction quality includes: The data acquisition and processing module is used to collect multi-source sensing data from various automated monitoring points deployed during the construction of water conservancy projects, and to preprocess the collected multi-source sensing data. The intelligent quality control module for hydraulic components uses an OWL ontology built on a BIM model to express quality control points and temporal attributes. It integrates pre-processed multi-source perception data, uses graph neural networks and multi-task learning methods to predict the quality status of components, dynamically updates BIM model information, and generates construction optimization suggestions. The water conservancy project acceptance execution module is used to automatically generate acceptance tasks based on construction progress, component quality status and dynamic acceptance strategies. It uses a rule engine to compare perceived data with standards and specifications to generate and archive digital acceptance reports. The water conservancy quality traceability and control module constructs a quality event map based on BIM components, associates abnormal data, early warning information and acceptance results, executes the problem handling process and performs defect clustering analysis. The water conservancy acceptance strategy optimization module integrates construction progress, environmental conditions, material status, and historical acceptance data based on the BIM model to construct an acceptance strategy parameter adjustment model, dynamically generating inspection frequency, inspection point spatial layout scheme, and key indicator judgment thresholds.

[0007] In a preferred embodiment of the present invention, the process by which the intelligent quality control module for hydraulic components constructs an OWL ontology to express quality control points and temporal attributes based on a BIM model and integrates preprocessed multi-source sensing data includes: Extract the unique identifier, type, spatial location and material properties of structural units from the BIM model, create a construction structural unit class instance for each structural unit, create one or more quality control point class instances for each construction structural unit class instance, and create quality status instances for each quality control point class instance at different time points. Obtain preprocessed multi-source sensing data, extract the preprocessed multi-source sensing data related to each quality control point instance, and aggregate the data by time point; Create a graph structure with node types of construction structure unit class and quality control point class. Use the generated feature vector as the input feature of the quality control point class node, and use the output graph structure as the input of the graph neural network.

[0008] In a preferred embodiment of the present invention, the process by which the intelligent quality control module for hydraulic components predicts the quality status of components using graph neural networks and multi-task learning methods includes: Extract unique identifiers for all structural components from the BIM model, and create a graph node for each component, forming a node set V={v i}; Analyze the physical connections between components in the BIM model. If there is a physical connection between component A and component B, then at the corresponding node v i and v j Add an edge between them to form the edge set E={(v i ,v j )}; For each component c i Obtain historical data for all quality control points, and calculate the time-weighted average of these data as the feature vector f for this node. i ; The network has three layers. Each layer updates node features using a propagation rule. The initial node features are input and subjected to a three-layer graph convolution operation. The final output is the embedding representation z of each component. i ; The embedding representation of each node z i The input is fed into two independent fully connected layers, and the trained model is used to perform forward propagation on the nodes corresponding to each component to obtain the classification results and the predicted value of the quality degradation degree, respectively.

[0009] In a preferred embodiment of the present invention, the process by which the intelligent quality control module for hydraulic components dynamically updates BIM model information and generates construction optimization suggestions includes: Obtain the embedded representation z of each structural component i , embedding representation z i Input the classification header and output the state with the highest probability as the prediction result; The embedding representation z i Input the regression head, and output the value as the overall quality score of the component; The loss values ​​of the classification task and the regression task are added together by weight to obtain the joint loss function. An attribute hasPredictedStatus is created to write the predicted quality status into the ontology instance of the corresponding component. An attribute predictedAt is created to write the current system time in xsd:dateTime format. Write the updated OWL ontology data into the quality attribute field of the BIM model, and input three data items: the predicted quality status of the component, the construction stage number, and the predicted quality status of adjacent components. Matching is performed according to a preset rule table. Each component generates one instruction. If multiple rules match, the one with the highest priority is selected. Create the attribute hasConstructionAdvice, write the generated construction operation instructions into the corresponding component's ontology instance, insert a record into the scheduling system interface table, and commit the database transaction after the insertion operation is completed.

[0010] In a preferred embodiment of the present invention, the process by which the water conservancy project acceptance execution module automatically generates acceptance tasks based on construction progress, component quality status, and dynamic acceptance strategies includes: Read the current system time and retrieve the planned completion rate and actual completion rate corresponding to the current time from the construction plan data table; Subtract the planned completion percentage from the actual completion percentage to obtain the construction progress deviation value. Read the progress tolerance threshold, take the absolute value of the construction progress deviation value, and determine whether the construction progress status is normal by comparing the absolute value with the progress tolerance threshold. Read the unique identifier of the first structural component from the component list, and obtain all quality index test values ​​of the component from the quality monitoring system; The measured value of each quality indicator is multiplied by its corresponding weight value, and all the product results are added together to obtain the overall quality score of the component. The quality score pass benchmark value is read, and the overall quality score is compared with the quality score pass benchmark value to determine whether the quality status of the component meets the acceptance conditions. Check whether the construction progress status is normal and whether the component quality status meets the acceptance requirements. Create a new record and write it into the acceptance task table. Read the unique identifier of the next structural component from the component list. If there are still unprocessed components in the component list, repeat all operations starting from obtaining the detection value from the quality monitoring system.

[0011] In a preferred embodiment of the present invention, the process by which the water conservancy project acceptance execution module compares the perceived data with the standards and specifications using a rule engine and generates a digital acceptance report includes: The system obtains the component's unique identifier, a list of quality indicator test values, construction progress deviation values, and comprehensive quality score as input data. Based on the component's unique identifier, it queries the minimum and maximum allowable values ​​for each quality indicator from the standard database. The rules engine is invoked, and it outputs the judgment results for each quality indicator, creating a digital acceptance report data structure and serializing the report data structure into a structured data format.

[0012] In a preferred embodiment of the present invention, the process by which the water conservancy quality traceability and control module constructs a quality event map based on BIM components and associates abnormal data, early warning information, and acceptance results includes: Read the unique identifiers, types, spatial locations, and material properties of all structural components from the BIM model, and create a component node for each structural component; When the sensor detects that the temperature, stress or displacement data exceeds the preset threshold, an abnormal data record is generated. When the abnormal data meets the preset warning rules, corresponding operations are performed for processing. When the event is rectified, corresponding operations are performed for processing. When the rectification is completed and acceptance is carried out, the corresponding operations are performed to process the data, and all nodes and edges are written into the graph database to form a quality event graph.

[0013] In a preferred embodiment of the present invention, the process by which the water conservancy quality traceability and control module executes the problem handling procedure and performs defect clustering analysis includes: Read the current status of all structural components from the component status table, execute the corresponding status update process for each structural component, read all historical defect records from the quality event graph, and perform corresponding operations to process each defect record. Set the number of clusters K, use the K-means algorithm to iteratively calculate the standardized feature sample set, output the clustering results, and write the clustering results into the defect analysis result table.

[0014] In a preferred embodiment of the present invention, the process by which the water conservancy acceptance strategy optimization module integrates construction progress, environmental conditions, material status, and historical acceptance data based on a BIM model and constructs an acceptance strategy parameter adjustment model includes: Read the integrated component dataset from the BIM model, standardize the construction progress status, and generate standardized values ​​for the construction progress status. Temperature, humidity, and wind speed in environmental conditions are standardized to generate comprehensive environmental indicators. The material condition is structured to generate a comprehensive material condition score, and historical acceptance data is processed to generate historical acceptance indicators. The standardized value of construction progress status, comprehensive environmental index, comprehensive score of material status, and historical acceptance index are used as four input features. The historical acceptance frequency adjustment value is read from the acceptance execution record and used as the model output label. Define the acceptance strategy parameters to adjust the model structure, and use the least squares method to perform regression analysis on the input feature matrix and the output label vector; Write the model structure information into the model metadata table, and establish a relationship between the model parameters and the model structure information through the model number.

[0015] As a preferred embodiment of the present invention, the process by which the water conservancy acceptance strategy optimization module dynamically generates the detection frequency, the spatial layout scheme of detection points, and the judgment threshold of key indicators includes: The system acquires the detection task trigger signal, reads the current construction completion percentage from the progress management system, reads the baseline detection frequency from the system configuration table, and obtains the current detection frequency through a series of processes. Generate inspection frequency parameters, read the geometric center coordinates of the component to be inspected from the BIM model, read the risk level of the component to be inspected from the risk level table, and determine the number of inspection points based on the risk level; Based on the geometric center, a spatial arrangement scheme for the detection points is generated on the surface of the component in a symmetrical manner; The initial judgment thresholds of key indicators are read from the design specification library, and the material correction items are calculated. The environmental condition influence coefficient is read from the system configuration table, the current environmental conditions are read from the environmental monitoring system, the environmental correction item is calculated, and the initial judgment threshold, material correction item, and environmental correction item are added together to obtain the current key indicator judgment threshold. Generate key indicator judgment threshold parameters, generate detection task plan data package, and write the detection task plan data package into the detection task configuration table.

[0016] Compared with the prior art, the advantages of this invention are: (1) In this invention, by integrating BIM and OWL ontology through the intelligent quality control module for water conservancy components, semantic modeling of quality control points and temporal attributes and association of multi-source data are realized, a structured knowledge graph is constructed, and graph neural networks and multi-task learning are used to predict the quality status and degradation degree simultaneously by combining component topological relationships and temporal features. By dynamically updating BIM model information and generating intelligent construction suggestions, a closed loop of "perception-analysis-decision-feedback" is formed, which significantly improves the intelligence, precision and automation level of water conservancy project construction quality control. (2) In this invention, the construction progress, environment, materials and historical data are integrated through the water conservancy acceptance strategy optimization module to construct a quantifiable and interpretable acceptance strategy parameter adjustment model, realize the dynamic optimization of detection frequency, detection point layout and judgment threshold, improve the scientificity and adaptability of acceptance resource allocation through data-driven approach, enhance the response capability to complex working conditions, and the output optimization strategy effectively supports the intelligent decision-making of the water conservancy project acceptance execution module, promote the transformation of acceptance from static and fixed to dynamic and precise, and significantly improve the efficiency, pertinence and engineering quality assurance level of acceptance. Attached Figure Description

[0017] Figure 1 This is a system block diagram of Embodiment 1 of the present invention; Figure 2 This is a system block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] Example 1: As Figure 1 As shown, the BIM-based digital acceptance management system for water conservancy engineering construction quality proposed in this invention includes: The data acquisition and processing module is used to collect multi-source sensing data from various automated monitoring points deployed during the construction of water conservancy projects. The multi-source sensing data includes the horizontal displacement of the cofferdam horizontal displacement monitoring point, the vertical displacement of the earth-rock cofferdam vertical displacement monitoring point, the wind speed of the high-altitude operation wind speed monitoring point, the internal temperature of the concrete temperature control monitoring point, the strain of the concrete component strain monitoring point, the water level elevation of the groundwater observation well, and the real-time compressive strength of the concrete strength maturity monitoring point. The module preprocesses the collected multi-source sensing data, including data cleaning, data calibration, time alignment, and data format standardization. The data acquisition and processing module automatically collects and integrates multi-source monitoring data such as displacement, wind speed, temperature, and strain, and performs preprocessing such as cleaning, calibration, time alignment, and standardization, significantly improving data quality and real-time performance. This enables comprehensive perception of construction status and rapid early warning of anomalies. Its standardized and modular design supports intelligent analysis and system expansion, reduces labor costs and errors, ensures data traceability, and provides solid support for the safety, quality, and intelligent management of water conservancy engineering construction.

[0020] The intelligent quality control module for hydraulic components uses an OWL ontology built on a BIM model to express quality control points and temporal attributes. It integrates pre-processed multi-source perception data, uses graph neural networks and multi-task learning methods to predict the quality status of components, dynamically updates BIM model information, and generates construction optimization suggestions. The process of constructing an OWL ontology to express quality control points and temporal attributes based on a BIM model and fusing preprocessed multi-source sensing data in the intelligent quality control module for hydraulic components includes: Extract the unique identifier, type, spatial location, and material properties of structural units from the BIM model, and create a construction structural unit class instance for each structural unit; Create a construction structural unit class to describe the components in the project, and create a quality control point class to describe the quality monitoring location or detection index attached to the component; Create a component that includes a quality control point attribute, set its target to the construction structural unit class, and set its value range to the quality control point class; Create a quality status attribute, set its target to the quality control point class, and set its value range to qualified, unqualified, and require re-inspection; Create a status time attribute, set its target to quality status, and its data type to date and time format; Create a previous state property, set its target to the quality state instance, and its value range to the quality state instance; Create a next state property, set its target object to the quality state instance, and its value range to the quality state instance; Create one or more quality control point class instances for each construction structural unit class instance. For each quality control point class instance, use the component to contain quality control point attributes and associate it with its corresponding construction structural unit class instance. For each quality control point class instance, create a quality status instance at different time points. Each quality status instance is associated with a quality status value and a status time value. Multiple quality status instances of the same quality control point class instance are arranged in ascending order of status time value. Use the previous state attribute to connect the next state to the previous state, and use the next state attribute to connect the previous state to the next state. The first quality state instance does not have the previous state attribute set, and the last quality state instance does not have the next state attribute set. Obtain preprocessed multi-source sensing data, extract the preprocessed multi-source sensing data related to each quality control point instance, aggregate the data by time point, and combine the temperature, humidity, stress, strain, moisture content, material batch number, and construction process code at the same time point into a feature vector; The vector contains eight elements. The first to fifth elements are temperature, humidity, stress, strain, and moisture content. The sixth element is the material batch number, the seventh element is the construction process code, and the eighth element is the timestamp code. Missing values ​​are filled with 0. The feature vector is associated with the corresponding quality status instance. Create a graph structure with node types of construction structure unit class and quality control point class. Add nodes: each instance of construction structure unit class and each instance of quality control point class corresponds to one graph node. Add edge: For each component that contains a quality control point attribute, add an edge from the construction structure unit class node to the quality control point class node; For each previous state attribute, add an edge from the previous quality state instance to the current quality state instance; For each next state attribute, add an edge from the current mass state instance to the next mass state instance; The generated feature vectors are used as input features for the quality control point class nodes, and the output graph structure is used as input for the graph neural network. The process by which the intelligent quality control module for hydraulic components predicts the quality status of components using graph neural networks and multi-task learning methods includes: Extract unique identifiers for all structural components from the BIM model, and create a graph node for each component, forming a node set V={v i}, where each node v i Corresponding to a component c i ; Analyze the physical connections between components in the BIM model. If there is a physical connection between component A and component B, then at the corresponding node v i and v j Add an edge between them to form the edge set E={(v i ,v j )}; For each component c i Obtain historical data for all quality control points, and calculate the time-weighted average of these data as the feature vector f for this node. i The time weight is calculated using a decay factor, which is equal to 0.95. Construct a graph G=(V,E) using a set of nodes V and a set of edges E, and then add self-loops to the adjacency matrix A to obtain... Where I is the identity matrix, and the degree matrix is ​​calculated. ; The network has 3 layers, and each layer updates node features using propagation rules, expressed by the following formula: ,in Let l be the node feature matrix of the l-th layer. Let be the weight matrix of the l-th layer. It is the ReLU activation function; Input initial node features Perform three-layer graph convolution operations to finally output the embedding representation z of each component. i ; The embedding representation of each node z i The input is fed into two separate fully connected layers: The first fully connected layer is used for classification tasks, with an output dimension of 3; The second fully connected layer is used for the regression task, outputting a single value; The trained model is used to perform forward propagation on the nodes corresponding to each component to obtain the classification results and the predicted value of the quality degradation. The process by which the intelligent quality control module for hydraulic components dynamically updates BIM model information and generates construction optimization suggestions includes: Obtain the embedded representation z of each structural component i As input data for the prediction task, the embedded representation z i Input the classification head, which is a fully connected neural network with an output dimension of 3, corresponding to three states: qualified, unqualified, and needing re-inspection. Output the probability distribution and take the state with the highest probability as the prediction result. During the training phase, the binary cross-entropy loss function is used to calculate the classification error. The embedding representation z i The input regression head is a fully connected neural network with an output dimension of 1. The output value is scaled to the [0,100] range after Sigmoid transformation and is used as the comprehensive score of component quality. During the training phase, the mean squared error loss function is used to calculate the regression error. The loss values ​​from the classification task and the regression task are added together by weights to obtain the joint loss function: ; in For classifying losses, To regress the loss, The classification loss weight is set to 1.0. The regression loss weight is set to 0.5. Create the property hasPredictedStatus to write the predicted quality status to the corresponding component's ontology instance; create the property predictedAt to write the current system time in xsd:dateTime format. Write the updated OWL ontology data into the quality attribute field of the BIM model. After the update operation is completed, the predicted state field value of the corresponding component in the BIM model will be the latest written value. Input three data items: the predicted quality status of the component, the construction stage number, and the predicted quality status of adjacent components; Matching is performed based on a preset rule table: If the predicted quality status is unacceptable, the command PauseAndInspect will be output. If the predicted quality status is that a re-inspection is required and the construction stage number is greater than 5, then the command StrengthenSupport will be output. If the status of an adjacent component is unqualified, the instruction AdjustCuringCondition will be output. If the predicted quality status is unqualified and the status of the adjacent component is that it needs to be re-inspected, then the command ReplanConstructionSequence will be output. Each component generates one instruction; if multiple rules match, the one with the highest priority is selected. Create the property hasConstructionAdvice to write the generated construction operation instructions into the corresponding component's instance; Insert a record into the scheduling system interface table, containing the following fields: component ID, operation instruction, and generation time. After the insertion operation is completed, the database transaction is committed. By integrating BIM and OWL ontology into the intelligent quality control module for hydraulic components, semantic modeling of quality control points and temporal attributes, as well as multi-source data association, are achieved, constructing a structured knowledge graph. Utilizing graph neural networks and multi-task learning, combined with component topological relationships and temporal features, the quality status and degradation degree are predicted synchronously. By dynamically updating BIM model information and generating intelligent construction suggestions, a closed loop of "perception-analysis-decision-feedback" is formed, significantly improving the intelligence, precision, and automation level of quality control in hydraulic engineering construction.

[0021] The water conservancy project acceptance execution module is used to automatically generate acceptance tasks based on construction progress, component quality status and dynamic acceptance strategies. It uses a rule engine to compare perceived data with standards and specifications to generate and archive digital acceptance reports. The process by which the water conservancy project acceptance execution module automatically generates acceptance tasks based on construction progress, component quality status, and dynamic acceptance strategies includes: Read the current system time and query the planned completion rate and actual completion rate corresponding to the current time from the construction plan data table; Subtract the planned completion percentage from the actual completion percentage to obtain the construction progress deviation value. Read the progress tolerance threshold, take the absolute value of the construction progress deviation value, and compare the absolute value with the progress tolerance threshold. If the absolute value is greater than the schedule tolerance threshold, the construction progress status is recorded as abnormal; if the absolute value is less than or equal to the schedule tolerance threshold, the construction progress status is recorded as normal. Read the unique identifier of the first structural component from the component list, obtain all quality index detection values ​​of the component from the quality monitoring system based on the component's unique identifier, and read the weight values ​​corresponding to each quality index from the quality weight configuration table. Multiply the measured value of each quality indicator by its corresponding weight value, add all the products together to obtain the overall quality score of the component, read the quality score pass benchmark value, and compare the overall quality score with the quality score pass benchmark value: If the overall quality score is greater than or equal to the quality score pass benchmark value, the quality status of the component is recorded as meeting the acceptance prerequisites; if the overall quality score is less than the quality score pass benchmark value, the quality status of the component is recorded as not meeting the acceptance prerequisites. Check whether the construction progress is normal and whether the quality of the components meets the acceptance requirements: If both are yes, proceed to the task generation process; if either is no, skip task generation and continue processing the next component. Create a new record, write it to the acceptance task table, fill in the unique identifier of the current component, fill in the current system time, and fill in the task status with the value "to be executed" in the new record; Read the unique identifier of the next structural component from the component list. If there are still unprocessed components in the component list, repeat all operations starting from obtaining the detection value from the quality monitoring system. If all components in the component list have been processed; The process by which the water conservancy project acceptance execution module uses a rule engine to compare perceived data with standards and specifications and generates a digital acceptance report includes: The system obtains the component's unique identifier, a list of quality indicator test values, construction progress deviation values, and comprehensive quality score as input data. Based on the component's unique identifier, it queries the minimum and maximum allowable values ​​for each quality indicator from the standard database. Call the rules engine, and input the following: quality indicator detection value, minimum allowable value, and maximum allowable value; The rule engine executes the following judgment logic: if the detected value is less than the minimum allowed value, the output judgment result is non-compliant; if the detected value is greater than the maximum allowed value, the output judgment result is non-compliant; if the detected value is greater than or equal to the minimum allowed value and less than or equal to the maximum allowed value, the output judgment result is compliant. The rules engine outputs the judgment result for each quality indicator; The rule engine calculates the results: if at least one indicator is found to be non-compliant with the specifications, the output component acceptance conclusion is non-compliant with the acceptance standards; if all indicators are found to be compliant with the specifications, the output component acceptance conclusion is compliant with the acceptance standards. Create a digital acceptance report data structure and fill in the unique identifier of the component, the acceptance conclusion of the component, the comprehensive quality score and the construction progress deviation value in the report; Perform the following operations for each quality indicator: If the detected value is greater than the maximum allowable value, calculate the detected value minus the maximum allowable value to obtain a positive deviation value and fill it in the report. If the detected value is less than the minimum allowable value, calculate the detected value minus the minimum allowable value to obtain a negative deviation value and fill it in the report. If the detected value is within the allowable range, fill in the deviation value as 0. The report data structure is serialized into a structured data format, a file name is generated, the serialized data is written into a file named with that file name, and the file is transferred to the specified storage path of the digital archive system. By integrating construction progress, component quality status, and dynamic acceptance strategies into the water conservancy project acceptance execution module, the system enables intelligent triggering and automatic generation of acceptance tasks, improving the timeliness and relevance of acceptance. Based on a rule engine, it automatically compares multi-source perception data with standards and specifications to accurately determine the compliance of indicators and the overall acceptance conclusion, ensuring objective and consistent evaluation. It supports the automated generation and archiving of digital acceptance reports, which are clear in structure and traceable, significantly improving acceptance efficiency and standardization, and promoting the digital and intelligent transformation of water conservancy project acceptance.

[0022] The water conservancy quality traceability and control module constructs a quality event map based on BIM components, associates abnormal data, early warning information and acceptance results, executes the problem handling process and performs defect clustering analysis. The water conservancy quality traceability and control module constructs a quality event map based on BIM components, and the process of linking abnormal data, early warning information, and acceptance results includes: Read the unique identifier, type, spatial location and material properties of all structural components from the BIM model, create a component node for each structural component, label the node type as "component", and fill in the unique identifier, type, spatial location and material properties in the component node; When the sensor detects that the temperature, stress or displacement data exceeds the preset threshold, an abnormal data record is generated, and the abnormal data record is associated with the component node of the corresponding component. The association type is "monitoring abnormality". When abnormal data meets the preset warning rules, perform the following operations: Read the unique identifier of the component associated with the abnormal data to determine the event type: If it is the first occurrence, the event type is one of crack, deformation or detachment. Match the classification code according to the rules, create an event node, and mark the node type as "event". Fill in the classification code, event type and occurrence time in the event node. Add a directed edge with the starting point of the component node and the ending point of the event node, and mark the edge type as "occurrence". Create a warning message containing the unique identifier of the component, the event classification code, the trigger time and the warning level. Associate the warning message with the event node, and the association type is "trigger warning". When rectifying an event, perform the following operations: read the category code and occurrence time of the event node, record the rectification start time, add a directed edge with the starting point being the component node and the ending point being the event node, and mark the edge type as "rectification"; When the rectification is completed and acceptance is performed, perform the following operations: Create an acceptance record, fill in the component unique identifier, event classification code, acceptance time and acceptance conclusion in the acceptance record, and the acceptance conclusion is "pass" or "fail"; associate the acceptance record with the component node, with the association type being "acceptance result"; associate the acceptance record with the event node, with the association type being "acceptance result". Write all nodes and edges into a graph database to form a quality event graph; The process by which the water conservancy quality traceability and control module executes the problem handling procedure and performs defect clustering analysis includes: Read the current state of all structural components from the component state table, and perform the following state update procedure for each structural component: Read whether the component has been associated with a quality event. If not, keep the status as normal. If it is associated with a quality event for the first time and the current status is normal, update the status to pending evaluation. Read whether the component has generated a warning message. If it has, and the current status is pending evaluation, update the status to warning. Read whether the component has added an edge of type "rectification". If it has, and the current status is warning, update the status to rectification. Read whether the component has associated acceptance records. If the acceptance conclusion is passed and the current status is rectification, update the status to accepted. Verify the status transition path: if the status jumps from pending evaluation to rectification, mark it as an abnormal transition. If the status rolls back from accepted to rectification, mark it as an illegal operation. All legal transitions are only allowed to proceed in the direction of normal → pending evaluation → warning → rectification → accepted. Write the updated status back to the component status table. Read all historical defect records from the quality event graph, and perform the following operations for each defect record: Extract the defect category, which is one of crack, deformation or spalling; extract the component type; extract the spatial coordinates X, Y, Z; extract the occurrence time, which is converted to the number of hours since January 1, 1970; and extract the severity level, which is an integer from 1 to 5. Each defect record is organized into a feature sample, and standardization is performed on all feature samples: Numeric fields are standardized using Z-score, while categorical fields are encoded using one-hot encoding. Set the number of clusters K to 5, and use the K-means algorithm to iteratively calculate the values ​​for the standardized feature sample set: K cluster centers are randomly initialized. Each sample is assigned to the nearest cluster center. The center point of each cluster is recalculated. The assignment and update are repeated until the cluster centers no longer change significantly or the maximum number of iterations is reached. Output the clustering results, label each defect record with its corresponding cluster number, and write the clustering results into the defect analysis results table; By constructing a quality event map integrating abnormal data, early warning information, and acceptance results through the water conservancy quality traceability and control module, the entire chain of quality issues can be visualized and traced. By defining clear state transition rules, the problem handling process is automatically executed and the compliance of operations is verified, ensuring that the control process is standardized and orderly. Cluster analysis is performed by combining the multidimensional characteristics of defects to explore the distribution patterns and potential correlations of quality issues, supporting the transformation from passive response to proactive prevention, and significantly improving the systematicness, intelligence, and scientific decision-making level of water conservancy project quality management.

[0023] Example 2: The technical solution of this embodiment of the invention differs from that of Example 1 in that: like Figure 2 As shown, the water conservancy acceptance strategy optimization module integrates construction progress, environmental conditions, material status and historical acceptance data based on the BIM model to build an acceptance strategy parameter adjustment model, dynamically generate detection frequency, detection point spatial layout scheme and key indicator judgment threshold, and output the optimized strategy to the water conservancy project acceptance execution module to guide the execution of dynamic acceptance tasks. The process of optimizing water conservancy acceptance strategies, based on a BIM model, integrating construction progress, environmental conditions, material status, and historical acceptance data, and constructing an acceptance strategy parameter adjustment model includes: Read the integrated component dataset from the BIM model. The dataset contains the construction progress status, environmental conditions, material status and historical acceptance data of each component. Standardize the construction progress status and convert the original values ​​into dimensionless values ​​using the mean and standard deviation of the historical data. Temperature, humidity, and wind speed in the environmental conditions are standardized, and then the arithmetic mean of the three is calculated to generate a comprehensive environmental index. The material condition is structured as follows: the material arrival time is converted into the material age, the average temperature during storage is compared with the standard storage temperature to calculate the temperature deviation, the average humidity during storage is compared with the standard storage humidity to calculate the humidity deviation, the material batch number is mapped to the historical quality score of that batch, the age, temperature deviation, humidity deviation and quality score are weighted and summed to generate a comprehensive material condition score, and the comprehensive score is standardized. Processing historical acceptance data: Read the number of times the same type of component has passed acceptance and the total number of acceptances in historical projects, calculate the acceptance pass rate, standardize the pass rate, and generate historical acceptance indicators; The standardized value of construction progress status, comprehensive environmental index, comprehensive material status score, and historical acceptance index are used as four input features. The historical acceptance frequency adjustment value is read from the acceptance execution record and used as the model output label. The output label is normalized to make it within a uniform numerical range. Define the structure of the acceptance strategy parameter adjustment model: the model type is a linear regression model, the input is four features, and the output is the acceptance frequency adjustment coefficient; The least squares method was used to perform regression analysis on the input feature matrix and the output label vector, and four regression coefficients were calculated, which correspond to the influence weights of construction progress status, environmental conditions, material status and historical acceptance data, respectively. The four regression coefficients were written into the model parameter table. Write the model structure information into the model metadata table, including the model type, input feature name, and output variable name. Associate the model parameters with the model structure information through the model number. Set the model status to "built". The process by which the water conservancy acceptance strategy optimization module dynamically generates the detection frequency, spatial layout plan of detection points, and key indicator judgment thresholds includes: The system acquires the detection task trigger signal, reads the current construction completion percentage from the progress management system, reads the baseline detection frequency from the system configuration table, reads the frequency adjustment coefficient from the system configuration table, transforms the current construction completion percentage using the inverse function of the standard normal distribution, multiplies the transformation result by the frequency adjustment coefficient, and adds the product to the baseline detection frequency to obtain the current detection frequency. The detection frequency generation model is expressed by the following formula: Where f(t) is the detection frequency at the current moment, μ is the reference detection frequency, and σ is the frequency adjustment coefficient. R(t) is the inverse function of the standard normal distribution, and R(t) is the current percentage of construction completed. Generate inspection frequency parameters, with parameter values ​​being the calculation results. Read the geometric center coordinates of the component to be inspected from the BIM model, read the risk level of the component to be inspected from the risk level table, and determine the number of inspection points based on the risk level: risk level 1, number 1; risk level 2, number 2; risk level 3, number 3; risk level 4, number 4; risk level 5, number 5. Based on the geometric center, the positions of detection points are generated on the surface of the component in a symmetrical distribution. A unique identifier and three-dimensional spatial coordinates are generated for each detection point. A spatial layout scheme for detection points is generated, which includes the identifiers and coordinates of all detection points. The spatial layout scheme for detection points is then bound to the unique identifier of the component. The initial judgment threshold of key indicators is read from the design specification library, the material state influence coefficient is read from the system configuration table, the current material state is read from the material management system, the deviation value between the material state and the standard state is calculated, and the material state influence coefficient and the material state deviation value are multiplied to obtain the material correction item. The environmental condition influence coefficient is read from the system configuration table, and the current environmental conditions are read from the environmental monitoring system. The deviation value from the ideal state is calculated. The environmental condition influence coefficient is multiplied by the environmental condition deviation value to obtain the environmental correction term. The initial judgment threshold, material correction term, and environmental correction term are added together to obtain the current key indicator judgment threshold. The dynamic adjustment model of the judgment threshold is expressed by the following formula: ,in The threshold value at the current moment. This is the initial threshold for judgment. The influence coefficient of material condition. This represents the deviation between the material's condition and its standard condition. This is the environmental condition influence coefficient. This represents the deviation between environmental conditions and the ideal state. Generate key indicator judgment threshold parameters, with parameter values ​​being the calculation results. Generate a detection task plan data package, which includes: detection frequency parameters, detection point spatial layout plan, and key indicator judgment threshold parameters. Write the detection task plan data package into the detection task configuration table and mark the plan status as generated. By integrating construction progress, environmental, material, and historical data through the water conservancy acceptance strategy optimization module, a quantifiable and interpretable acceptance strategy parameter adjustment model is constructed to achieve dynamic optimization of detection frequency, detection point layout, and judgment threshold. Through a data-driven approach, the scientific nature and adaptability of acceptance resource allocation are improved, and the responsiveness to complex working conditions is enhanced. The output optimization strategy effectively supports the intelligent decision-making of the water conservancy project acceptance execution module, promoting the transformation of acceptance from static and fixed to dynamic and precise, and significantly improving acceptance efficiency, pertinence, and the level of project quality assurance.

[0024] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and its improved concept, should be covered within the scope of protection of the present invention.

Claims

1. A BIM-based digital acceptance management system for water conservancy project construction quality, characterized in that: include: The data acquisition and processing module is used to collect multi-source sensing data from various automated monitoring points deployed during the construction of water conservancy projects, and to preprocess the collected multi-source sensing data. The intelligent quality control module for hydraulic components uses an OWL ontology built on a BIM model to express quality control points and temporal attributes. It integrates pre-processed multi-source perception data, uses graph neural networks and multi-task learning methods to predict the quality status of components, dynamically updates BIM model information, and generates construction optimization suggestions. The water conservancy project acceptance execution module is used to automatically generate acceptance tasks based on construction progress, component quality status and dynamic acceptance strategies. It uses a rule engine to compare perceived data with standards and specifications to generate and archive digital acceptance reports. The water conservancy quality traceability and control module constructs a quality event map based on BIM components, associates abnormal data, early warning information and acceptance results, executes the problem handling process and performs defect clustering analysis. The water conservancy acceptance strategy optimization module integrates construction progress, environmental conditions, material status, and historical acceptance data based on the BIM model to construct an acceptance strategy parameter adjustment model, dynamically generating inspection frequency, inspection point spatial layout scheme, and key indicator judgment thresholds.

2. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 1, characterized in that, The process by which the intelligent quality control module for hydraulic components constructs an OWL ontology based on the BIM model to express quality control points and temporal attributes, and integrates preprocessed multi-source sensing data, includes: Extract the unique identifier, type, spatial location, and material properties of structural units from the BIM model, and create a construction structural unit class instance for each structural unit; Create one or more quality control point class instances for each construction structure unit class instance. For each quality control point class instance, use components to contain quality control point attributes. Create quality status instances for each quality control point class instance at different time points. Obtain the preprocessed multi-source sensing data, extract the preprocessed multi-source sensing data related to each quality control point instance, and aggregate the data by time point; Create a graph structure with node types of construction structure unit class and quality control point class. Use the generated feature vector as the input feature of the quality control point class node, and use the output graph structure as the input of the graph neural network.

3. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 2, characterized in that, The process by which the intelligent quality control module for hydraulic components predicts the quality status of components using graph neural networks and multi-task learning methods includes: Extract unique identifiers for all structural components from the BIM model, and create a graph node for each component, forming a node set V={v i }; Analyze the physical connections between components in the BIM model. If there is a physical connection between component A and component B, then at the corresponding node v i and v j Add an edge between them to form the edge set E={(v i ,v j )}; For each component c i Obtain historical data for all quality control points, and calculate the time-weighted average of these data as the feature vector f for this node. i ; The network has three layers. Each layer updates node features using a propagation rule. The initial node features are input and subjected to a three-layer graph convolution operation. The final output is the embedding representation z of each component. i ; The embedding representation of each node z i The input is fed into two independent fully connected layers, and the trained model is used to perform forward propagation on the nodes corresponding to each component to obtain the classification results and the predicted value of the quality degradation degree, respectively.

4. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 3, characterized in that, The process by which the intelligent quality control module for hydraulic components dynamically updates BIM model information and generates construction optimization suggestions includes: Obtain the embedded representation z of each structural component i , embedding representation z i Input the classification header, output the state corresponding to the highest probability as the prediction result, and embed the representation z. i Input the regression head, and output the value as the overall quality score of the component; The loss values ​​of the classification task and the regression task are added together by weight to obtain the joint loss function. An attribute hasPredictedStatus is created to write the predicted quality status into the ontology instance of the corresponding component. An attribute predictedAt is created to write the current system time in xsd:dateTime format. Write the updated OWL ontology data into the quality attribute field of the BIM model, and input three data items: the predicted quality status of the component, the construction stage number, and the predicted quality status of adjacent components. Matching is performed according to a preset rule table. Each component generates one instruction. If multiple rules match, the one with the highest priority is selected. Create the attribute hasConstructionAdvice, write the generated construction operation instructions into the corresponding component's ontology instance, insert a record into the scheduling system interface table, and commit the database transaction after the insertion operation is completed.

5. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 1, characterized in that, The process by which the water conservancy project acceptance execution module automatically generates acceptance tasks based on construction progress, component quality status, and dynamic acceptance strategies includes: Read the current system time and retrieve the planned completion rate and actual completion rate corresponding to the current time from the construction plan data table; Subtract the planned completion percentage from the actual completion percentage to obtain the construction progress deviation value. Read the progress tolerance threshold, take the absolute value of the construction progress deviation value, and determine whether the construction progress status is normal by comparing the absolute value with the progress tolerance threshold. Read the unique identifier of the first structural component from the component list, and obtain all quality index test values ​​of the component from the quality monitoring system based on the component's unique identifier; The measured value of each quality indicator is multiplied by its corresponding weight value, and all the product results are added together to obtain the overall quality score of the component. The quality score pass benchmark value is read, and the overall quality score is compared with the quality score pass benchmark value to determine whether the quality status of the component meets the acceptance conditions. Check whether the construction progress status is normal and whether the component quality status meets the acceptance requirements. Create a new record and write it into the acceptance task table. Read the unique identifier of the next structural component from the component list. If there are still unprocessed components in the component list, repeat all operations starting from obtaining the detection value from the quality monitoring system.

6. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 5, characterized in that, The process by which the water conservancy project acceptance execution module uses a rule engine to compare perceived data with standards and specifications and generates a digital acceptance report includes: The system obtains the component's unique identifier, a list of quality indicator test values, construction progress deviation values, and comprehensive quality score as input data. Based on the component's unique identifier, it queries the minimum and maximum allowable values ​​for each quality indicator from the standard database. The rules engine is invoked, and it outputs the judgment results for each quality indicator, creating a digital acceptance report data structure and serializing the report data structure into a structured data format.

7. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 1, characterized in that, The water conservancy quality traceability and control module constructs a quality event map based on BIM components, and the process of linking abnormal data, early warning information and acceptance results includes: Read the unique identifiers, types, spatial locations, and material properties of all structural components from the BIM model, and create a component node for each structural component; When the sensor detects that the temperature, stress or displacement data exceeds the preset threshold, an abnormal data record is generated. When the abnormal data meets the preset warning rules, corresponding operations are performed for processing. When the event is rectified, corresponding operations are performed for processing. When the rectification is completed and acceptance is carried out, the corresponding operations are performed to process the data, and all nodes and edges are written into the graph database to form a quality event graph.

8. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 7, characterized in that, The process by which the water conservancy quality traceability and control module executes the problem handling procedure and performs defect clustering analysis includes: Read the current status of all structural components from the component status table, execute the corresponding status update process for each structural component, read all historical defect records from the quality event graph, and perform corresponding operations to process each defect record. Set the number of clusters K, use the K-means algorithm to iteratively calculate the standardized feature sample set, output the clustering results, and write the clustering results into the defect analysis result table.

9. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 1, characterized in that, The process by which the water conservancy acceptance strategy optimization module integrates construction progress, environmental conditions, material status, and historical acceptance data based on the BIM model and constructs an acceptance strategy parameter adjustment model includes: Read the integrated component dataset from the BIM model, standardize the construction progress status, and generate standardized construction progress status values. Temperature, humidity, and wind speed in environmental conditions are standardized to generate comprehensive environmental indicators. The material condition is structured to generate a comprehensive material condition score, and historical acceptance data is processed to generate historical acceptance indicators. The standardized value of construction progress status, comprehensive environmental index, comprehensive score of material status, and historical acceptance index are used as four input features. The historical acceptance frequency adjustment value is read from the acceptance execution record and used as the model output label. Define the acceptance strategy parameters to adjust the model structure, and use the least squares method to perform regression analysis on the input feature matrix and the output label vector; Write the model structure information into the model metadata table, and establish a relationship between the model parameters and the model structure information through the model number.

10. The BIM-based digital acceptance management system for water conservancy project construction quality according to claim 9, characterized in that, The process by which the water conservancy acceptance strategy optimization module dynamically generates the detection frequency, the spatial layout plan of detection points, and the judgment thresholds for key indicators includes: The system acquires the detection task trigger signal, reads the current construction completion percentage from the progress management system, reads the baseline detection frequency from the system configuration table, and obtains the current detection frequency through a series of processes. Generate inspection frequency parameters, read the geometric center coordinates of the component to be inspected from the BIM model, read the risk level of the component to be inspected from the risk level table, and determine the number of inspection points based on the risk level; Based on the geometric center, the locations of detection points are generated on the surface of the component in a symmetrical manner. A unique identifier and three-dimensional spatial coordinates are generated for each detection point, and a spatial arrangement scheme for the detection points is generated. The initial judgment thresholds of key indicators are read from the design specification library, the material state influence coefficient is read from the system configuration table, the current material state is read from the material management system, and the material correction items are calculated. The environmental condition influence coefficient is read from the system configuration table, the current environmental conditions are read from the environmental monitoring system, the environmental correction item is calculated, and the initial judgment threshold, material correction item, and environmental correction item are added together to obtain the current key indicator judgment threshold. Generate key indicator judgment threshold parameters, generate detection task plan data package, and write the detection task plan data package into the detection task configuration table.

Images