Building processing quality safety monitoring system

Abstract

The invention discloses a building processing quality safety monitoring system, and belongs to the technical field of building project information management. The system comprises a data acquisition module which is responsible for acquiring original building instance data from a plurality of data sources; the data processing module is used for cleaning, standardizing and processing the original data in real time through a distributed stream processing architecture; the knowledge graph construction module is used for constructing and maintaining a knowledge graph in the building field based on the processed data, dynamically updating entity relationships and attributes, and integrating user feedback for continuous optimization; the intelligent analysis and decision support module is used for carrying out analysis by utilizing a graph neural network and a multi-objective optimization algorithm based on the knowledge graph and the dynamic data flow and providing intelligent decision suggestions; the data security module is used for providing end-to-end data security guarantee for the whole system through a block chain and a federated learning technology, and realizing non-tampering and secure sharing of data; and the man-machine interaction module is used for realizing visual display of the data.

Description

Technical Field

[0001] This application relates to the field of building project information management technology, and more specifically, to a building processing quality and safety monitoring system. Background Technology

[0002] In the modern construction industry, data collection, management, and analysis are increasingly becoming crucial factors in driving project efficiency and quality. As construction projects become more complex and larger, various types of data (such as design drawings, construction schedules, material usage, and quality inspections) are generated in diverse forms throughout the project lifecycle. Effective management and analysis of this data are essential for decision support, resource optimization, and risk control.

[0003] However, existing monitoring systems often suffer from the following problems: First, the wide range of data sources, including traditional manual entry and modern sensor equipment, makes it difficult to guarantee data quality and consistency. Second, existing systems typically lack real-time data processing capabilities, failing to respond promptly to changes at the construction site and affecting the accuracy and timeliness of decision-making. Furthermore, the construction and maintenance of knowledge graphs in the construction field are still in a relatively rudimentary stage, lacking an effective dynamic update mechanism, resulting in insufficient timeliness and applicability of knowledge.

[0004] Furthermore, data security issues are becoming increasingly prominent in the construction industry. Due to the involvement of multiple stakeholders, the complexity of data sharing and access control increases, and existing security measures often fail to meet the demands for data integrity and privacy protection. This makes achieving efficient data sharing and collaboration while ensuring data security a pressing challenge for the industry.

[0005] In conclusion, how to establish a construction processing quality and safety monitoring system that integrates data acquisition, processing, knowledge graph construction, intelligent analysis, and data security assurance to improve data quality, real-time performance, and security has become an urgent technical problem to be solved. Summary of the Invention

[0006] In order to overcome a series of defects in the existing technology, the purpose of this application is to provide a construction processing quality and safety monitoring system, which includes the following modules.

[0007] The data acquisition module is responsible for collecting raw building instance data from multiple data sources.

[0008] The data processing module cleans, standardizes, and processes raw data in real time using a distributed stream processing architecture.

[0009] The knowledge graph construction module builds and maintains a knowledge graph in the construction field based on the processed data, dynamically updates entity relationships and attributes, and integrates user feedback for continuous optimization.

[0010] The intelligent analysis and decision support module, based on knowledge graphs and dynamic data streams, utilizes graph neural networks and multi-objective optimization algorithms to perform analysis and provide intelligent decision suggestions.

[0011] The data security module provides end-to-end data security for the entire system through blockchain and federated learning technologies, enabling data immutability and secure sharing.

[0012] The human-computer interaction module provides an intuitive human-computer interaction interface through AR / VR and natural language processing technologies, enabling the visualization of data.

[0013] Furthermore, the data acquisition module includes the following components.

[0014] The data source connection unit establishes a secure connection channel with various building data sources, ensuring the reliability and integrity of the data flow.

[0015] The data capture unit automatically extracts raw building instance data from connected data sources according to a predefined scheduling plan or real-time triggering mechanism.

[0016] The metadata management unit automatically extracts and maintains metadata information from data sources, including data structure, update frequency, and data quality, providing a foundation for data traceability and quality control.

[0017] The data acquisition configuration unit centrally manages the acquisition rules, scheduling strategies, and access credentials for all data sources.

[0018] The data buffer unit uses a distributed message queue to receive and buffer the raw data stream, achieving high-throughput data entry and decoupling data acquisition from subsequent processing.

[0019] Furthermore, the data processing module includes the following components.

[0020] The data cleaning unit uses predefined rule sets and machine learning algorithms to clean the raw data and improve its quality.

[0021] Data standardization units convert cleaned data into a unified standard format and unit, ensuring that data from different sources can be compared and analyzed under the same standard.

[0022] The real-time stream processing unit builds a real-time data processing pipeline that supports low latency and high concurrency.

[0023] The data aggregation and computation unit performs real-time data aggregation, statistical analysis, and feature extraction, providing basic data support for subsequent knowledge graph construction and intelligent analysis.

[0024] The integrated monitoring and auditing unit continuously monitors the collection status, data volume, collection frequency, and integrity, accuracy, and consistency indicators of each data source during data processing, and triggers alarms when anomalies are detected; at the same time, it records logs of all data operations to support security auditing and compliance verification.

[0025] The task scheduling unit dynamically adjusts the parallelism and priority of processing tasks based on data traffic and system resource status, optimizing resource utilization and ensuring processing efficiency.

[0026] Furthermore, the knowledge graph construction module includes the following components.

[0027] The entity relationship extraction unit identifies and extracts key entities from processed building instance data, and automatically extracts and defines the relationships between entities by analyzing the interactions between them, so as to construct semantic connections between entities.

[0028] The graph database management unit uses graph database technology to efficiently store and manage knowledge graph data, and supports complex graph query and analysis operations.

[0029] The ontology model management unit maintains the ontology model in the architectural domain, defines the concept hierarchy and relation types, and provides a semantic framework for the knowledge graph.

[0030] The dynamic update unit updates the entities, relationships, and attributes in the knowledge graph in real time based on new data input, ensuring the timeliness of the graph.

[0031] The user feedback integration unit collects and analyzes user feedback on the knowledge graph and applies it to continuous optimization of the graph.

[0032] Furthermore, the intelligent analysis and decision support module includes the following components.

[0033] The data fusion and knowledge reasoning unit integrates structured knowledge from the knowledge graph with real-time data streams, and performs logical reasoning based on existing knowledge and rules to discover potential new knowledge and relationships, thus enriching the content of the knowledge graph.

[0034] The graph neural network analysis unit uses graph neural network technology to perform in-depth analysis on the fused data, identifying complex relationship patterns and hidden insights.

[0035] The multi-objective optimization and decision generation unit applies a multi-objective optimization algorithm to find the optimal balance point among multiple objectives and generates preliminary decision suggestions by comprehensively analyzing the results.

[0036] The scenario simulation unit simulates different implementation scenarios based on the generated decision recommendations, and assesses potential impacts and risks.

[0037] The interpretability analysis unit provides transparency into the decision-making process, explains the analytical logic and results, and enhances understanding and trust in the decisions.

[0038] The decision recommendation optimization unit further optimizes and adjusts decision recommendations based on scenario simulation results and interpretability analysis.

[0039] Furthermore, a multi-objective optimization algorithm is applied to find the optimal balance point among multiple objectives, and the results are comprehensively analyzed to generate preliminary decision recommendations, including the following steps.

[0040] Each target indicator is quantified and formalized, an objective function expression is established, and constraints are determined.

[0041] By using the analytic hierarchy process (AHP), the relative importance of different objectives is determined, and a scientific and reasonable weight allocation scheme is established.

[0042] A multi-objective optimization algorithm is executed, and a series of non-dominated solutions, namely the Pareto optimal solution set, are obtained through iterative calculation.

[0043] The obtained Pareto optimal solution set is analyzed and evaluated, and the optimal solution that meets the actual needs is selected by combining actual constraints and decision preferences.

[0044] The selected optimal solutions are transformed into specific decision-making recommendations, detailing the expected effects, implementation steps, and potential risks of the solutions, and corresponding implementation plans are developed.

[0045] Furthermore, the data security module includes the following components.

[0046] The Access Control and Identity Management Unit implements fine-grained access control policies based on roles and attributes to ensure that users can only access the data they are authorized to access, preventing unauthorized data access.

[0047] The data encryption and key management unit is responsible for the encrypted storage and transmission of data, managing the generation, distribution, and updating of encryption keys, and ensuring the security of data during storage and transmission.

[0048] The blockchain data verification unit uses blockchain technology to record and verify the integrity of data, ensuring that data cannot be tampered with once it is written.

[0049] The Federated Learning Coordination Unit coordinates multiple data-specific parties to perform distributed machine learning, enabling them to train together without sharing the original data.

[0050] The secure multi-party computation unit allows multiple participants to jointly compute and analyze encrypted data without disclosing their original data, supporting secure data collaboration across organizations.

[0051] The privacy-protected data publishing unit applies differential privacy technology when publishing or sharing data to ensure that the published datasets provide valuable information without compromising individual privacy.

[0052] Compared with the prior art, the beneficial effects of this application are as follows.

[0053] This application utilizes distributed stream processing, knowledge graph construction and optimization, graph neural network analysis, multi-objective optimization decision-making, and blockchain security technologies to achieve efficient collection, processing, analysis, and maintenance of instance data in the construction field, providing intelligent decision support and data security assurance. Attached Figure Description

[0054] Figure 1 This is a structural diagram of a building processing quality and safety monitoring system disclosed in an embodiment of this application. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some embodiments of this invention, but not all embodiments.

[0056] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] The embodiments and directional terms described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0058] like Figure 1 As shown, a construction processing quality and safety monitoring system includes the following modules.

[0059] The data acquisition module is responsible for collecting raw building instance data from multiple data sources.

[0060] The data processing module cleans, standardizes, and processes raw data in real time using a distributed stream processing architecture.

[0061] The knowledge graph construction module builds and maintains a knowledge graph in the construction field based on the processed data, dynamically updates entity relationships and attributes, and integrates user feedback for continuous optimization.

[0062] The intelligent analysis and decision support module, based on knowledge graphs and dynamic data streams, utilizes graph neural networks and multi-objective optimization algorithms to perform analysis and provide intelligent decision suggestions.

[0063] The data security module provides end-to-end data security for the entire system through blockchain and federated learning technologies, enabling data immutability and secure sharing.

[0064] The human-computer interaction module provides an intuitive human-computer interaction interface through AR / VR and natural language processing technologies, enabling the visualization of data.

[0065] In this embodiment, the building construction quality and safety monitoring system plays a crucial role, responsible for collecting raw building instance data from multiple data sources. Its technical advantages are reflected in two aspects: First, diversity and real-time performance. By integrating sensors, monitoring equipment, and existing databases, it can collect real-time building status data, such as temperature and humidity, structural health monitoring, and construction progress. Data diversity ensures the system can comprehensively reflect different aspects of the building, laying a solid foundation for subsequent data processing and analysis. Second, data integrity and accuracy. The use of high-precision sensors and standardized data formats ensures the high reliability of the collected data. Furthermore, intelligent algorithms can be configured to perform preliminary error detection and correction on the collected data, improving data accuracy and providing a reliable basis for subsequent analysis.

[0066] In this embodiment, the innovation of the data processing module is mainly reflected in its efficiency and scalability. First, by employing stream processing frameworks such as Apache Kafka, it can efficiently process real-time data streams, supporting low-latency data flow and analysis. Second, the distributed architecture enables the system to scale horizontally, handling the processing needs of large-scale data and ensuring system stability under high load. Simultaneously, data cleaning and standardization processes remove redundant data and fill in missing values, ensuring the high quality of the final dataset and improving the accuracy and effectiveness of subsequent knowledge graph construction and intelligent analysis.

[0067] In this embodiment, the innovation of the knowledge graph construction module lies in the richness and dynamism of the knowledge. By constructing an entity-relationship graph, various elements in a construction project (such as materials, design schemes, construction teams, etc.) can be linked to form a complete knowledge network. Simultaneously, user feedback can be integrated to achieve continuous optimization of the knowledge graph. This dynamic updating capability allows the system to promptly reflect the latest developments and technological advancements in the construction industry, enhancing the practicality and forward-looking nature of the knowledge. Furthermore, the structured data of the knowledge graph supports complex queries and reasoning analysis, providing rich foundational data for subsequent intelligent analysis and decision-making.

[0068] In this embodiment, the innovation of the intelligent analysis and decision support module lies primarily in its intelligence and accuracy. Through graph neural networks, it can deeply mine the complex entity relationships and potential correlations within the knowledge graph, identifying key factors affecting the progress, cost, and quality of construction projects. Simultaneously, the multi-objective optimization algorithm can balance multiple objectives, providing optimization suggestions and helping decision-makers make more scientific choices in resource allocation and risk management. This intelligent analysis capability greatly improves decision-making efficiency, reduces risks, and makes construction project management more scientific and refined.

[0069] In this embodiment, the innovation of the data security module lies in its security and privacy protection. Through blockchain technology, all data operations can be transparently recorded, ensuring data integrity and immutability, thereby enhancing data credibility. Simultaneously, utilizing federated learning technology, multiple data sources can collaboratively train models without disclosing individual privacy, promoting data sharing and utilization. This security mechanism allows participating parties to confidently use and share data, greatly improving the flexibility and reliability of data use and reducing the risk of data leakage.

[0070] In this embodiment, the innovation of the human-computer interaction module lies in the improved user experience and the effectiveness of information delivery. AR / VR technology can present complex building data in a three-dimensional visualization format, allowing users to browse and analyze data in a virtual reality environment, enhancing their understanding and perception of architectural examples. Simultaneously, natural language processing technology supports user interaction with the system through language commands, making data querying and operations more user-friendly. This interaction method not only increases user engagement and satisfaction but also enables non-professional users to easily understand complex data content, promoting the popularization and application of building information.

[0071] In summary, by organically combining modules for data acquisition, processing, knowledge graph construction, intelligent analysis, data security, and human-computer interaction, comprehensive management and application of building data have been achieved. Each module not only enhances the overall performance and reliability of the system but also provides solid technical support for the digital transformation of the construction industry. Through such a system, the construction field can better address complex project management needs, improve the scientific nature of decision-making, optimize resource allocation, and ultimately promote innovation and development in the construction industry.

[0072] Furthermore, the data acquisition module includes the following components.

[0073] The data source connection unit establishes a secure connection channel with various building data sources, ensuring the reliability and integrity of the data flow.

[0074] The data capture unit automatically extracts raw building instance data from connected data sources according to a predefined scheduling plan or real-time triggering mechanism.

[0075] The metadata management unit automatically extracts and maintains metadata information from data sources, including data structure, update frequency, and data quality, providing a foundation for data traceability and quality control.

[0076] The data acquisition configuration unit centrally manages the acquisition rules, scheduling strategies, and access credentials for all data sources.

[0077] The data buffer unit uses a distributed message queue to receive and buffer the raw data stream, achieving high-throughput data entry and decoupling data acquisition from subsequent processing.

[0078] The specific implementation methods for establishing secure connection channels between the data source connection unit and various building data sources include: firstly, identifying the available service endpoints of the target data source through a service discovery mechanism and standardized API interfaces; then, establishing an encrypted transmission channel based on the TLS 1.3 protocol to ensure the confidentiality and integrity of data transmission; and during the connection establishment process, using the OAuth 2.0 authorization framework combined with the JWT token authentication mechanism to authenticate the data source through pre-configured client credentials.

[0079] The specific process by which the data crawling unit automatically extracts raw building instance data from connected data sources according to a predefined scheduling plan or real-time triggering mechanism includes: parsing a preset time plan based on Cron expressions and intelligently adjusting the crawling frequency and time window by combining the historical update patterns and current load of the data source; adopting an incremental crawling strategy during the crawling process, and only acquiring data records that have changed since the last crawl by maintaining the last update timestamp and version number of each data source; responding to data source change events in a timely manner by listening to Webhook notifications from the data source or polling specific status interfaces during the real-time triggering mechanism; and implementing intelligent rate limiting control during the crawling process, dynamically adjusting the request frequency according to the service level agreement and response capacity of the data source to avoid being denied service by the data source due to excessive access.

[0080] The data buffer unit adopts a distributed message queue architecture based on Apache Kafka. Configuration parameters include: 16 partitions, a replication factor of 3 to ensure high availability and fault tolerance; a message retention time of 168 hours, a maximum single message size limit of 1MB, and a batch size of 16384 bytes. Producer-side configurations include an acknowledgment mechanism in all mode, 3 retries, and a batch processing latency of 5 milliseconds. Consumer-side configurations include a session timeout of 10 seconds, a heartbeat interval of 3 seconds, a maximum of 500 records to pull, a manual offset commit strategy, and a commit interval of 1 second. A dynamic scaling mechanism is also implemented; when the queue lag exceeds a set threshold of 10,000 messages, the number of consumer instances is automatically increased to ensure stable system operation under high concurrency scenarios.

[0081] In summary, the data acquisition module achieves efficient and reliable data acquisition of building examples through the collaborative work of multiple key components. The data source connection unit ensures connection security and compatibility, the data capture unit improves the automation and real-time performance of data acquisition, the metadata management unit provides assurance for data quality and traceability, the data acquisition configuration unit realizes centralized and flexible management, and the data buffer unit improves scalability and stability through decoupling design.

[0082] Furthermore, the data processing module includes the following components.

[0083] The data cleaning unit uses predefined rule sets and machine learning algorithms to clean the raw data and improve its quality.

[0084] Data standardization units convert cleaned data into a unified standard format and unit, ensuring that data from different sources can be compared and analyzed under the same standard.

[0085] The real-time stream processing unit builds a real-time data processing pipeline that supports low latency and high concurrency.

[0086] The data aggregation and computation unit performs real-time data aggregation, statistical analysis, and feature extraction, providing basic data support for subsequent knowledge graph construction and intelligent analysis.

[0087] The integrated monitoring and auditing unit continuously monitors the collection status, data volume, collection frequency, and integrity, accuracy, and consistency indicators of each data source during data processing, and triggers alarms when anomalies are detected; at the same time, it records logs of all data operations to support security auditing and compliance verification.

[0088] The task scheduling unit dynamically adjusts the parallelism and priority of processing tasks based on data traffic and system resource status, optimizing resource utilization and ensuring processing efficiency.

[0089] The real-time stream processing unit adopts the Apache Flink-based stream computing framework, with the following technical parameters configured: parallelism is set to 1 times the number of CPU cores, checkpoint interval is configured to 5 minutes, the state backend uses RocksDB, and incremental checkpointing is configured to improve performance; memory configuration includes TaskManager heap memory set to 4GB, network buffer size to 32MB, and managed memory ratio set to 0.4; time semantics adopts event-time mode, the watermark generator is configured as BoundedOutOfOrdernessWatermarks, and the maximum allowed out-of-order time is 5 minutes; window operation configuration includes sliding window size of 15 minutes, sliding step size of 5 minutes, and session window timeout set to 30 minutes; the fault tolerance mechanism adopts at-least-once semantic guarantee, the restart strategy is configured as exponential delayed restart, the maximum number of restarts is 3, the initial restart interval is 10 seconds, and a backpressure detection mechanism is implemented, triggering a dynamic adjustment algorithm when the backpressure exceeds the threshold.

[0090] In summary, the data processing module, through the collaborative work of multiple key components, ensures the efficient cleaning, standardization, and processing of building instance data. The data cleaning unit improves data quality, the data standardization unit ensures data consistency and comparability, the real-time stream processing unit enhances the timeliness of data processing, the data aggregation and computation unit uncovers the potential value of the data, and the comprehensive monitoring and auditing unit provides guarantees for the transparency and security of data processing. Finally, the dynamic resource management mechanism of the processing task scheduling unit ensures the efficient operation of the system under various conditions.

[0091] Furthermore, data aggregation can be expressed by the following formula: Where A(t) represents the aggregated building instance data value at time t; D i (t) represents the original building instance data collected by the i-th data source at time t; N is the total number of data sources.

[0092] Statistical analysis can be expressed by the following formula: ,in, This represents the mean of the building instance data within the time window [t0, t]. Indicates time The data is the building instance data; [t0,t] is the specified time window used to calculate the average value of the data over a period of time.

[0093] Feature extraction can be expressed by the following formula: R(t) represents the rate of change of the building instance data at time t; D(t) represents the building instance data at time t, and D(t-1) represents the data at the previous time point.

[0094] In summary, the detailed explanation of the above formulas demonstrates the crucial roles of data aggregation, statistical analysis, and feature extraction in the processing of architectural case data. Data aggregation provides a holistic view by integrating results from different data sources; statistical analysis helps understand the overall trend of data within a specific time window; and feature extraction provides important signals of change for real-time monitoring. These operations complement each other, collectively enhancing the intelligence level of the data processing module and making the management of architectural projects more efficient and scientific.

[0095] Furthermore, the degree of parallelism for dynamically adjusting the processing tasks is expressed by the following formula: Where P(t) is the task parallelism at time t, representing the number of tasks processed simultaneously; F(t) is the data flow at time t, representing the amount of data to be processed per unit time; S(t) is the available system resources at time t; C represents the system resources required for each task to process a unit of data; and M represents the amount of system resources required for each task.

[0096] The dynamic adjustment of task priority is expressed by the following formula: Where Q(t) is the task priority at time t; F max The maximum data flow that the system can handle; S max E represents the system's maximum resource usage; E represents the task's urgency; α, β, and γ are weighting coefficients used to balance the impact of various factors on task priority.

[0097] In summary, dynamically adjusting the parallelism and priority of processing tasks provides flexibility and efficiency to the data processing module. Dynamic adjustment of parallelism ensures adaptability based on real-time data traffic and resource availability, maintaining good processing performance under both high and low load conditions. Dynamic adjustment of priority ensures that critical tasks are processed first under resource constraints or high traffic, thus meeting the system's high availability and real-time requirements in building instance data processing. Combining these two adjustment mechanisms, the building data processing system achieves more intelligent and efficient data management, providing strong support for decision-making in the construction industry.

[0098] Furthermore, the knowledge graph construction module includes the following components.

[0099] The entity relationship extraction unit identifies and extracts key entities from processed building instance data, and automatically extracts and defines the relationships between entities by analyzing the interactions between them, so as to construct semantic connections between entities.

[0100] The graph database management unit uses graph database technology to efficiently store and manage knowledge graph data, and supports complex graph query and analysis operations.

[0101] The ontology model management unit maintains the ontology model in the architectural domain, defines the concept hierarchy and relation types, and provides a semantic framework for the knowledge graph.

[0102] The dynamic update unit updates the entities, relationships, and attributes in the knowledge graph in real time based on new data input, ensuring the timeliness of the graph.

[0103] The user feedback integration unit collects and analyzes user feedback on the knowledge graph and applies it to continuous optimization of the graph.

[0104] The specific process of the entity relationship extraction unit in identifying and extracting key entities from processed building instance data includes: First, feature extraction of text data is performed using a pre-trained language model based on the Transformer architecture, converting the text into a 768-dimensional vector representation; a deep learning model based on BiLSTM-CRF is trained using labeled data from the building domain, enabling the identification of entity types such as building projects, contractors, materials, equipment, and addresses; a neural network model based on an attention mechanism is used to analyze the semantic relationships between entities, supporting the identification of 15 common relationship types such as contracting relationships, supply relationships, and dependency relationships; by calculating the cosine similarity of the entity context, newly identified entities are matched with existing entities in the knowledge graph, with a similarity threshold set to 0.8; a confidence score between 0 and 1 is calculated for each extraction result, and only entities and relationships with a confidence score exceeding 0.75 are automatically added to the knowledge graph, while results below this threshold require manual review and confirmation.

[0105] The graph database management unit uses Neo4j graph database, with configuration parameters including: initial heap memory size set to 8GB, maximum heap memory limit to 16GB, page cache configured to 25% of available memory, and transaction log size limit to 1GB; index configuration includes btree index, fulltext index, and range index, with a synchronous update strategy and an index refresh interval of 1 minute; query optimizer configuration includes a cost estimation model based on statistical information, a query plan cache size of 1000, and a query timeout limit of 60 seconds; cluster configuration adopts a causal cluster mode, with 3 core servers, 2 read replicas, an election timeout of 7 seconds, and a heartbeat interval of 2 seconds; the data sharding strategy uses consistent hashing based on entity type and attributes, with a replication factor of 3 to ensure high data availability and balanced distribution of query performance.

[0106] The specific implementation mechanism of the dynamic update unit in real-time updating entities, relationships, and attributes in the knowledge graph based on new data input includes: establishing an event-driven update architecture, where corresponding update events are generated and added to the event queue when new data arrives; sorting events according to their priority and timestamps to ensure the orderliness and consistency of update operations; maintaining version history for each entity and relationship in the knowledge graph, supporting incremental updates and rollback operations; automatically identifying potential data conflicts during the update process; employing multiple methods, including weighted voting based on data source credibility, the latest-first principle based on timestamps, and intelligent merging based on expert rules; and ensuring that when an entity or relationship changes, the related derived information and inference results are also updated in a timely manner.

[0107] In summary, the knowledge graph construction module, through the collaborative efforts of multiple components, forms an efficient and dynamic knowledge management mechanism. The entity relation extraction unit provides the basic knowledge structure for the knowledge graph; the graph database management unit ensures efficient data storage and retrieval; the ontology model management unit provides a semantic framework, enhancing the expressive power of knowledge; the dynamic update unit ensures the timeliness of knowledge; and the user feedback integration unit continuously optimizes the knowledge graph to better meet actual needs.

[0108] Furthermore, the intelligent analysis and decision support module includes the following components.

[0109] The data fusion and knowledge reasoning unit integrates structured knowledge from the knowledge graph with real-time data streams, and performs logical reasoning based on existing knowledge and rules to discover potential new knowledge and relationships, thus enriching the content of the knowledge graph.

[0110] The graph neural network analysis unit uses graph neural network technology to perform in-depth analysis on the fused data, identifying complex relationship patterns and hidden insights.

[0111] The multi-objective optimization and decision generation unit applies a multi-objective optimization algorithm to find the optimal balance point among multiple objectives and generates preliminary decision suggestions by comprehensively analyzing the results.

[0112] The scenario simulation unit simulates different implementation scenarios based on the generated decision recommendations, and assesses potential impacts and risks.

[0113] The interpretability analysis unit provides transparency into the decision-making process, explains the analytical logic and results, and enhances understanding and trust in the decisions.

[0114] The decision recommendation optimization unit further optimizes and adjusts decision recommendations based on scenario simulation results and interpretability analysis.

[0115] The graph neural network analysis unit adopts a PyTorch Geometric-based graph neural network architecture with the following specific configuration parameters: the network structure uses a hybrid architecture of Graph Convolutional Network (GCN) and Graph Attention Network (GAT), containing 3 convolutional layers with feature dimensions of 256, 128, and 64 for each node; the activation function is ReLU, the dropout rate is set to 0.3, the initial learning rate is 0.001, and the Adam optimizer is used; the batch size is configured to 32 subgraphs, the gradient accumulation step is set to 2, the maximum number of training epochs is 500, and the early stopping mechanism is configured to stop when the validation set loss has not improved for 10 consecutive epochs; the graph sampling strategy uses the FastGCN sampling method with a sampling rate of 0.2 and a subgraph size limit of 5000 nodes; the model regularization uses L2 regularization with a weight decay coefficient of 0.0001, and a graph structure attention mechanism is implemented with 4 attention heads to enhance the model's ability to understand complex graph structures.

[0116] In summary, the intelligent analysis and decision support module provides powerful intelligent decision support for the construction industry through the synergistic effect of its various components. The data fusion and knowledge reasoning unit ensures the integration and enrichment of basic data; the graph neural network analysis unit identifies complex relationships through deep analysis; the multi-objective optimization and decision generation unit achieves comprehensive consideration of multiple factors; the scenario simulation unit helps evaluate implementation effects and risks; the interpretability analysis unit ensures the transparency of the decision-making process; and the decision suggestion optimization unit achieves continuous improvement through dynamic feedback. Through these technological effects, the module provides the construction industry with scientific, rational, and transparent decision support, greatly enhancing the level of intelligence in project management.

[0117] Furthermore, complex relationship patterns and hidden insights can be identified using the following formula: Where P represents the relationship patterns and insights obtained through deep analysis; f is a learning function responsible for extracting important information from node representations; L is the total number of layers in the graph neural network; H （L） It is the node representation matrix of the Lth layer, which contains the feature representation of each node after processing through the Lth layer, and can capture the complex relationships between nodes in the graph; For activation functions; H is the normalized adjacency matrix, representing the connection relationships between nodes. （0） W is the initial node feature matrix, representing the initial features of each node in the graph. （0） W （1） , ...,W （L-1） This is the weight matrix, used to perform linear transformations on the node features at each layer.

[0118] In summary, the stepwise learning and feature aggregation of node representations at each layer, the nonlinear introduction of activation functions, and the dynamic adjustment of weight matrices constitute the technical foundation of this process in the above formula. Such analysis not only enriches the understanding and insight into architectural projects but also provides reliable support for project management and decision-making, ensuring the successful implementation of projects in dynamic environments.

[0119] Furthermore, a multi-objective optimization algorithm is applied to find the optimal balance point among multiple objectives, and the results are comprehensively analyzed to generate preliminary decision recommendations, including the following steps.

[0120] Each target indicator is quantified and formalized, an objective function expression is established, and constraints are determined.

[0121] By using the analytic hierarchy process (AHP), the relative importance of different objectives is determined, and a scientific and reasonable weight allocation scheme is established.

[0122] A multi-objective optimization algorithm is executed, and a series of non-dominated solutions, namely the Pareto optimal solution set, are obtained through iterative calculation.

[0123] The obtained Pareto optimal solution set is analyzed and evaluated, and the optimal solution that meets the actual needs is selected by combining actual constraints and decision preferences.

[0124] The selected optimal solutions are transformed into specific decision-making recommendations, detailing the expected effects, implementation steps, and potential risks of the solutions, and corresponding implementation plans are developed.

[0125] In conclusion, the application of multi-objective optimization algorithms has made the decision-making process in the construction field more scientific, rational, and efficient. It not only improves the accuracy of objective setting and optimization but also ensures the feasibility and practicality of the final decision. From the quantification and formalization of objectives to the scientific allocation of weights, the generation and analysis of non-dominated solutions, and finally the transformation into specific decision recommendations, the entire process constitutes a complete optimization system, providing strong support for the successful implementation of construction projects.

[0126] The detailed implementation steps for determining the relative importance of different objectives and establishing a scientific and reasonable weight allocation scheme through the analytic hierarchy process (AHP) include: constructing a three-layer hierarchical model to decompose the complex multi-objective decision-making problem into an objective layer, a criterion layer, and a scheme layer. The objective layer includes the overall objective of project success; the criterion layer includes 4-8 evaluation indicators such as cost control, schedule management, quality assurance, and safety management; and the scheme layer includes 3-10 specific implementation schemes. A pairwise comparison matrix is ​​established, and 3-7 experts in the construction field are invited to conduct pairwise comparisons of each element at the same level. The 1-9 scale method proposed by Satie is used to quantify expert judgments, where 1 represents equal importance and 9 represents extreme importance. The logical consistency of the judgment matrix is ​​verified by calculating the consistency ratio (CR). When CR ≤ 0.10, the judgment matrix is ​​considered to have acceptable consistency; when CR > 0.10, the experts need to readjust the judgment results. The weight calculation uses the geometric mean method to solve for the normalized eigenvector corresponding to the largest eigenvalue λmax of the judgment matrix, thus obtaining the relative weights of each element.

[0127] Furthermore, the data security module includes the following components.

[0128] The Access Control and Identity Management Unit implements fine-grained access control policies based on roles and attributes to ensure that users can only access the data they are authorized to access, preventing unauthorized data access.

[0129] The data encryption and key management unit is responsible for the encrypted storage and transmission of data, managing the generation, distribution, and updating of encryption keys, and ensuring the security of data during storage and transmission.

[0130] The blockchain data verification unit uses blockchain technology to record and verify the integrity of data, ensuring that data cannot be tampered with once it is written.

[0131] The Federated Learning Coordination Unit coordinates multiple data-specific parties to perform distributed machine learning, enabling them to train together without sharing the original data.

[0132] The secure multi-party computation unit allows multiple participants to jointly compute and analyze encrypted data without disclosing their original data, supporting secure data collaboration across organizations.

[0133] The privacy-protected data publishing unit applies differential privacy technology when publishing or sharing data to ensure that the published datasets provide valuable information without compromising individual privacy.

[0134] In summary, access control and identity management ensure the security of data access; data encryption and key management maintain data confidentiality; blockchain technology enhances data integrity; and federated learning and secure multi-party computation promote data collaboration and privacy protection. Finally, privacy-preserving data release ensures the security of data during sharing. This comprehensive security framework enhances system security and user trust, ensuring compliance and efficiency in data management within the construction industry.

[0135] 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A construction processing quality and safety monitoring system, characterized in that, Includes the following modules: The data acquisition module is responsible for collecting raw building instance data from multiple data sources; The data processing module cleans, standardizes, and processes raw data in real time using a distributed stream processing architecture. The knowledge graph construction module builds and maintains a knowledge graph in the construction field based on the processed data, dynamically updates entity relationships and attributes, and integrates user feedback for continuous optimization. The intelligent analysis and decision support module, based on knowledge graphs and dynamic data streams, utilizes graph neural networks and multi-objective optimization algorithms to perform analysis and provide intelligent decision suggestions. The data security module provides end-to-end data security for the entire system through blockchain and federated learning technologies, enabling data immutability and secure sharing. The human-computer interaction module provides an intuitive human-computer interaction interface through AR / VR and natural language processing technologies, enabling the visualization of data.

2. The building processing quality and safety monitoring system according to claim 1, characterized in that, The data acquisition module includes the following components: The data source connection unit establishes a secure connection channel with various building data sources to ensure the reliability and integrity of the data flow; The data capture unit automatically extracts raw building instance data from the connected data source according to a predefined scheduling plan or real-time triggering mechanism; The metadata management unit automatically extracts and maintains metadata information from data sources, including data structure, update frequency, and data quality, providing a foundation for data traceability and quality control. The data acquisition configuration unit centrally manages the acquisition rules, scheduling strategies, and access credentials for all data sources; The data buffer unit uses a distributed message queue to receive and buffer the raw data stream, achieving high-throughput data entry and decoupling data acquisition from subsequent processing.

3. The building processing quality and safety monitoring system according to claim 1, characterized in that, The data processing module includes the following components: The data cleaning unit cleans the raw data using predefined rule sets and machine learning algorithms to improve data quality. Data standardization units convert cleaned data into a unified standard format and unit, ensuring that data from different sources can be compared and analyzed under the same standard. The real-time stream processing unit builds a real-time data processing pipeline that supports low latency and high concurrency. The data aggregation and computing unit performs real-time data aggregation, statistical analysis, and feature extraction, providing basic data support for subsequent knowledge graph construction and intelligent analysis. The integrated monitoring and auditing unit continuously monitors the collection status, data volume, collection frequency, and integrity, accuracy, and consistency indicators of each data source during data processing, and triggers alarms when anomalies are detected; it also records logs of all data operations to support security auditing and compliance verification. The task scheduling unit dynamically adjusts the parallelism and priority of processing tasks based on data traffic and system resource status, optimizing resource utilization and ensuring processing efficiency.

4. The building processing quality and safety monitoring system according to claim 1, characterized in that, The knowledge graph construction module includes the following components: The entity relationship extraction unit identifies and extracts key entities from processed building instance data, and automatically extracts and defines the relationships between entities by analyzing the interactions between them, so as to construct semantic connections between entities; The graph database management unit uses graph database technology to efficiently store and manage knowledge graph data, and supports complex graph query and analysis operations; The ontology model management unit maintains the ontology model in the architectural domain, defines the concept hierarchy and relation types, and provides a semantic framework for the knowledge graph. The dynamic update unit updates the entities, relationships, and attributes in the knowledge graph in real time based on new data input, ensuring the timeliness of the graph; The user feedback integration unit collects and analyzes user feedback on the knowledge graph and applies it to continuous optimization of the graph.

5. The building processing quality and safety monitoring system according to claim 1, characterized in that, The intelligent analysis and decision support module includes the following components: The data fusion and knowledge reasoning unit integrates structured knowledge in the knowledge graph with real-time data streams, and performs logical reasoning based on existing knowledge and rules to discover potential new knowledge and relationships, thereby enriching the content of the knowledge graph. The graph neural network analysis unit uses graph neural network technology to perform in-depth analysis on the fused data, identifying complex relationship patterns and hidden insights; The multi-objective optimization and decision generation unit applies a multi-objective optimization algorithm to find the optimal balance point among multiple objectives and comprehensively analyzes the results to generate preliminary decision suggestions. The scenario simulation unit simulates different implementation scenarios based on the generated decision recommendations, and assesses potential impacts and risks; The interpretability analysis unit provides transparency to the decision-making process, explains the analysis logic and results, and enhances understanding and trust in the decisions. The decision recommendation optimization unit further optimizes and adjusts decision recommendations based on scenario simulation results and interpretability analysis.

6. The building processing quality and safety monitoring system according to claim 5, characterized in that, A multi-objective optimization algorithm is applied to find the optimal balance point among multiple objectives, and the results are comprehensively analyzed to generate preliminary decision recommendations, including the following steps: Each target indicator is quantified and formalized, an objective function expression is established, and constraints are determined. By using the analytic hierarchy process (AHP), the relative importance of different objectives can be determined, and a scientific and reasonable weight allocation scheme can be established. By executing a multi-objective optimization algorithm, a series of non-dominated solutions, i.e., the Pareto optimal solution set, are obtained through iterative calculation. The obtained Pareto optimal solution set is analyzed and evaluated, and the optimal solution that meets the actual needs is selected by combining the actual constraints and decision preferences. The selected optimal solutions are transformed into specific decision-making recommendations, detailing the expected effects, implementation steps, and potential risks of the solutions, and corresponding implementation plans are developed.

7. The building processing quality and safety monitoring system according to claim 1, characterized in that, The data security module includes the following components: The Access Control and Identity Management Unit implements fine-grained access control policies based on roles and attributes to ensure that users can only access the data they are authorized to access and to prevent unauthorized access to data. The data encryption and key management unit is responsible for the encrypted storage and transmission of data, managing the generation, distribution and updating of encryption keys, and ensuring the security of data during storage and transmission. The blockchain data verification unit uses blockchain technology to record and verify the integrity of data, ensuring that data cannot be tampered with once it is written. The Federated Learning Coordination Unit coordinates multiple data-specific parties to perform distributed machine learning, enabling them to train together without sharing the original data. The secure multi-party computation unit allows multiple participants to jointly compute and analyze encrypted data without disclosing their original data, supporting secure data collaboration across organizations. The privacy-preserving data publishing unit applies differential privacy technology when publishing or sharing data to ensure that the published datasets provide valuable information without compromising individual privacy.

Images