Big data-based prefabricated building intelligent management method and system

Abstract

The application discloses an intelligent management method and system for fabricated buildings based on big data, and relates to the technical field of intelligent management of fabricated buildings.The method defines multiple-source heterogeneous data by standardization, formulates data collection formats, transmission protocols and storage specifications, collects multi-dimensional data in real time, preliminarily processes the multi-dimensional data, collects the preliminarily processed data to a fabricated building big data management platform in real time, forms a cross-stage and cross-subject unified database of the whole process, analyzes the data in the unified database in multiple dimensions, dynamically and intelligently controls the fabricated building, finally collects and feeds back the control execution result in real time, and continuously iterates and optimizes the data analysis model.The application can realize real-time synchronization of multiple-source heterogeneous data, intelligent analysis and dynamic control, reduce component rework and construction period delay, and improve overall management efficiency.

Description

Technical Field

[0001] This invention belongs to the field of intelligent management technology for prefabricated buildings, specifically relating to intelligent management methods and systems for prefabricated buildings based on big data. Background Technology

[0002] Prefabricated buildings adopt a construction mode of factory prefabrication and on-site assembly and installation, which has advantages such as high construction efficiency, energy saving and environmental protection, and controllable quality. It has become the core direction of the modernization and green and low-carbon development of the construction industry. With the large-scale promotion of prefabricated buildings, prefabricated building projects are characterized by multiple participants, multiple links, long cycles and high collaboration, covering the entire process of design deepening, component production, logistics and transportation, on-site hoisting, quality acceptance and operation and maintenance management, and the management complexity has been significantly increased. Currently, prefabricated buildings generally adopt a segmented manual management and partial information-assisted model. This model divides the project construction into stages, separating design refinement, component production, logistics and transportation, on-site assembly, quality acceptance, final acceptance, and post-construction operation and maintenance into relatively independent management units. Each unit is managed by the corresponding participating entity. The management process is driven by manual recording, manual statistics, manual transmission, and manual decision-making. Information carriers are mainly paper documents, electronic ledgers, Excel spreadsheets, and local management software. Information flow relies on manual reporting, offline communication, and point-to-point transmission. Each participating party only uses local information tools within its own business scope to complete basic operations such as data entry, query, and statistics. There is no unified data standard or real-time communication mechanism between different stages and different entities. Management decisions mainly rely on the experience of management personnel and the judgment of post-event data summary. Objectives such as progress, quality, cost, and safety are achieved through segmented control, post-event inspection, and manual correction. However, because data such as design models, production plans, logistics trajectories, hoisting progress, quality inspection records, and operation and maintenance files are scattered across different entities and systems with inconsistent formats and incompatible interfaces, it is difficult to form a unified data source across stages and entities. Design changes cannot be synchronized to the production and construction ends in real time, which can easily lead to component rework and project delays. Furthermore, management relies on manual experience and post-event review, lacking real-time aggregation, cleaning, and in-depth analysis of multi-source data. Component supply and demand matching, logistics routes, yard planning, hoisting sequences, and equipment and manpower allocation all depend on manual scheduling, which can easily lead to component backlog, waiting, errors, hoisting conflicts, and resource waste. It is also difficult to identify and proactively warn of delays, quality abnormalities, and safety hazards in advance, thus affecting overall management efficiency. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for intelligent management of prefabricated buildings based on big data, which can perform real-time synchronization of multi-source heterogeneous data, realize intelligent analysis and dynamic control, improve overall management efficiency, and solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: The intelligent management method for prefabricated buildings based on big data includes the following steps: S1. Establish a unified data standard system for the entire process of prefabricated buildings, standardize the definition of multi-source heterogeneous data, and formulate data acquisition formats, transmission protocols and storage specifications. Among them, multi-source heterogeneous data includes data on design refinement, component production, logistics and transportation, on-site hoisting, quality acceptance and operation and maintenance management. S2. A heterogeneous data acquisition architecture with multiple acquisition terminals is used to acquire multidimensional data in real time, and the multidimensional data is preliminarily processed. S3. A prefabricated building big data management platform is built through a cloud-native microservice architecture. The pre-processed data is aggregated to the prefabricated building big data management platform in real time to form a unified database that spans multiple stages and entities. S4. Conduct multi-dimensional in-depth analysis of the data in the unified database of the entire process through data analysis models. The data analysis models include one or more of the following: component supply and demand matching analysis model, logistics route optimization analysis model, yard planning analysis model, hoisting sequence simulation analysis model, equipment and manpower configuration optimization analysis model, schedule deviation analysis model, quality anomaly identification analysis model, safety hazard early warning analysis model, and building operation and maintenance performance analysis model. S5. Based on the results of multi-dimensional in-depth analysis and combined with the management objectives of prefabricated building projects, dynamic intelligent control is carried out on prefabricated buildings to coordinate and link the design, production, construction, acceptance and operation and maintenance links. S6. Real-time collection and feedback of control execution results and continuous iteration and optimization of data analysis models.

[0005] The preferred process for establishing a unified data standard system for the entire prefabricated building process is as follows: A1. Organize the business data types of the design refinement, component production, logistics and transportation, on-site hoisting, quality acceptance and operation and maintenance management links, identify the participating entities corresponding to each business link, and form a list of business scenarios for the entire assembly process and a data interaction requirement table for each participating entity; A2. Based on the sorted business scenarios and the full-process data of the participating entities, confirm the data source, generation frequency and data carrier of various types of data, and classify and grade the full-process data according to business dimensions, data structure and security level to form a data asset list; A3. Standardize the definition of data elements in the data asset list, and at the same time, build a multi-level core data indicator system in combination with the management objectives of prefabricated buildings, forming a data element standard specification for the whole process of prefabricated buildings and a core data indicator system specification for prefabricated buildings. A4. Develop separate standards for data acquisition, transmission, and storage, and create corresponding standard documents; A5. Establish a globally unique unified coding rule for the core objects of prefabricated buildings, and at the same time, formulate data interaction interface specifications and hierarchical sharing permission standards; A6. Select typical projects to conduct standard pilot verification, revise and optimize the standards based on the feedback, officially release and promote the standards through training, and establish a long-term management mechanism to achieve dynamic iteration of the standards.

[0006] Preferably, when classifying and grading, the data is first classified by business dimension, then by data structure, and finally by security level. Specifically, the business dimension is classified into six categories: design model data, component production data, logistics and transportation data, on-site hoisting data, quality acceptance data, and operation and maintenance monitoring data. Each category is further subdivided into subdivided data subsets. Data structures can be categorized into structured data, unstructured data, semi-structured data, and time-series data. The security level classification is based on the confidentiality and importance of the data, and is divided into public data, internally shared data, and core confidential data.

[0007] Preferably, the heterogeneous data acquisition architecture includes: an API interface connected to BIM design software for acquiring BIM model design parameters; The component production data acquisition module is used to collect component production process parameters and production progress data, and realizes real-time data upload through the Industrial Internet of Things. The logistics transportation data acquisition module collects real-time vehicle trajectories, transportation speeds, and the transportation environment status of components, and transmits the data back in real time via mobile networks. The on-site hoisting data acquisition module collects hoisting operation parameters as well as resource allocation and usage status data of on-site personnel or equipment; Quality testing instruments automatically collect data on the appearance, dimensions, and strength of components, and combine this with a digital data entry terminal for manual acceptance to achieve real-time collection of acceptance data; The building operation and maintenance data acquisition module collects monitoring data on structural safety, equipment operation, and energy consumption during the building operation and maintenance phase, and achieves continuous data collection through the Internet of Things.

[0008] The preferred preliminary processing flow for multidimensional data is as follows: B1. Remove abnormal data from multidimensional data and filter out duplicate data; B2. Fill in the missing data; B3. Normalize numerical data of different dimensions and ranges using the min-max normalization method; B4. Structure unstructured text, video, and model data.

[0009] Preferably, the prefabricated building big data management platform includes: a data acquisition layer that interfaces with a heterogeneous data acquisition architecture, uses a message queue to realize the real-time reception and forwarding of massive amounts of data, and writes the pre-processed data into the corresponding data storage medium according to a unified storage standard; The data storage layer builds a data warehouse, classifies and stores different types of data into matching storage media according to data storage standards, and establishes a globally unique identifier for data based on unified data coding rules to achieve physical-level data aggregation. The data processing layer extracts full-process data from various media in the storage layer through standardized data access interfaces, performs in-depth processing according to the business logic of prefabricated buildings, and forms the core data body of the unified database for the entire process. The application service layer, based on the management needs of prefabricated buildings, encapsulates the unified data source into standardized business data services, and combines data sharing permission standards to set hierarchical data source access permissions for different participating entities. The visualization layer obtains a unified data source and processed analysis results from the application service layer, presents them to all participating entities in a visual form, and receives reverse operation instructions from users. The data acquisition layer, data storage layer, data processing layer, application service layer, and visualization layer form a closed loop based on a real-time data feedback mechanism.

[0010] The preferred multi-dimensional in-depth analysis process is as follows: C1. Based on the different scenario requirements of prefabricated building management, a multi-model analysis strategy that integrates traditional algorithms and deep learning algorithms is adopted to select a suitable data analysis model and deploy the data analysis model on the data processing layer of the big data management platform. C2. Extract the feature data required by each data analysis model from a unified data source and perform feature processing. Feature processing includes feature extraction, feature selection and feature transformation. Input the feature data after feature processing into the corresponding data analysis model. C3. For single analysis scenarios, the global optimal solution is obtained through traditional algorithms, while the real-time prediction and adjustment of dynamic factors are achieved through deep learning algorithms, and multi-dimensional deep calculations are performed from the time dimension, space dimension, subject dimension, and indicator dimension. C4. Output the analysis results of the data analysis model in the form of quantitative indicators, visual charts, and early warning signals.

[0011] Preferably, when selecting a suitable data analysis model, the following selection process should be followed: C1.1 First, the management scenario of prefabricated buildings is broken down in detail to determine the core requirements of the data analysis model; C1.2 Match the characteristics of traditional algorithms and deep learning algorithms according to the needs of business scenarios: If the scenario requires finding the global optimal solution, quantitative analysis, and rule-based decision-making, traditional algorithms should be chosen first. If the scenario requires real-time dynamic prediction, complex feature extraction, image / video recognition, or non-linear data processing, deep learning algorithms should be integrated first. C1.3 For complex management scenarios, a combination of traditional and deep learning algorithms is adopted. Traditional algorithms are used to optimize the global framework, while deep learning algorithms are used to process local dynamic features, taking into account both the interpretability and intelligence of the model. C1.4. Based on the characteristics of the unified data source, verify the adaptability of the selected model to ensure that the model can extract effective feature data from the unified data source and complete the analysis and calculation.

[0012] Preferably, during dynamic intelligent control, targeted control instructions are automatically generated based on multi-dimensional in-depth analysis results and the management objectives of the building project. These control instructions are then sent to the execution terminals of each stage in real time through the application service layer. Each execution terminal completes the corresponding operation according to the sent control instructions, thereby achieving collaborative linkage between design, production, construction, acceptance and operation and maintenance stages. The control instructions include quantitative control instructions and qualitative and hierarchical control instructions. Quantitative control instructions are obtained by multi-objective weighted solution of the quantitative output results of the data analysis model to obtain the optimal control parameters and convert them into specific control instructions. Qualitative and graded control instructions are generated by first classifying the risk or deviation of the output results of the data analysis model using a fuzzy comprehensive evaluation algorithm, and then matching the classification results with a preset control rule library using a rule matching algorithm.

[0013] The big data-based intelligent management system for prefabricated buildings is used to implement the big data-based intelligent management method for prefabricated buildings described above. It includes: a data standard module, which is used to establish a unified data standard system for the entire process of prefabricated buildings, standardize the definition of multi-source heterogeneous data, and formulate data acquisition formats, transmission protocols and storage specifications. The data acquisition module includes multiple acquisition terminals deployed at various stages of prefabricated building construction, used to collect multidimensional data in real time and perform preliminary processing on the multidimensional data; The big data aggregation and storage module is electrically connected to the data acquisition module. It is used to receive and aggregate the data after preliminary processing in real time to build a unified database that spans multiple stages and entities throughout the entire process. The intelligent analysis module has a built-in preset data analysis model for in-depth analysis of data in the unified database across the entire process; The dynamic control module generates scheduling instructions based on the analysis results of the intelligent analysis module; The feedback optimization module is used to collect the control execution results in real time and feed them back to the big data aggregation and storage module for data updates. At the same time, it continuously iterates and optimizes the preset data analysis model in the intelligent analysis module based on the execution results. The terminal interaction module includes multiple management terminals configured for each participant in the prefabricated building project. These terminals are used to receive scheduling instructions and support manual data entry and operation feedback from each participant.

[0014] The intelligent management method and system for prefabricated buildings based on big data proposed in this invention have the following advantages compared with existing technologies: 1. This invention establishes a unified data standard system for the entire process, standardizes the definition of multi-source heterogeneous data in design, production, logistics, construction, acceptance, and operation and maintenance, and forms a unified data source across stages and entities. Information such as design changes can be synchronized to each link in real time, reducing component rework and schedule delays, and improving overall management efficiency. 2. This invention relies on the full-process data gathered by the prefabricated building big data management platform, and conducts multi-dimensional in-depth analysis through multiple types of data analysis models, thereby realizing scientific and precise decision-making for each stage of prefabricated building construction. 3. Based on data analysis results, this invention enables dynamic and intelligent control of the design, production, construction, acceptance, and operation and maintenance processes, achieving coordinated linkage among all processes, reducing component backlog, waiting, errors, omissions, hoisting conflicts, and other issues, thereby further improving overall management efficiency. 4. This invention, through a data analysis model, can identify and proactively warn of quality, safety, and schedule issues during the construction and operation of prefabricated buildings, effectively reducing quality and safety risks and ensuring that the construction period proceeds as planned. 5. This invention continuously iterates and optimizes the data analysis model by providing real-time feedback and data updates on the control execution results, enabling the management method to adapt to the dynamic changes of prefabricated building projects and meet the management needs of prefabricated building projects of different scales and types. Attached Figure Description

[0015] Figure 1 A flowchart of a method according to an embodiment of the present invention is shown; Figure 2 A flowchart of multi-dimensional depth analysis according to an embodiment of the present invention is shown; Figure 3 A structural block diagram of a prefabricated building big data management platform according to an embodiment of the present invention is shown; Figure 4 A system block diagram of a system according to an embodiment of the present invention is shown. Detailed Implementation

[0016] 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. The specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] This invention provides, for example Figure 1-3 The big data-based intelligent management method for prefabricated buildings shown includes the following steps: S1. Establish a unified data standard system for the entire process of prefabricated buildings, standardize the definition of multi-source heterogeneous data, and formulate data acquisition formats, transmission protocols and storage specifications. Among them, multi-source heterogeneous data includes data from design refinement, component production, logistics and transportation, on-site hoisting, quality acceptance and operation and maintenance management.

[0018] The process for establishing a unified data standard system for the entire prefabricated building process is as follows: A1. Analyze the business data types in the design refinement, component production, logistics and transportation, on-site hoisting, quality acceptance, and operation and maintenance management stages; identify the participating entities in each business stage; confirm the relevant needs of each entity for data generation, use, and flow; and form a list of business scenarios for the entire assembly process and a data interaction requirement table for each participating entity. For example, the design refinement stage includes BIM model design, component attribute definition, and assembly relationship rules; the on-site hoisting stage includes hoisting sequence rules, equipment resource allocation, and operation process monitoring, etc.

[0019] A2. Based on the analyzed business scenarios and the full-process data of the participating entities, identify the data sources, generation frequency, and data carriers of various types of data, and classify and grade the full-process data according to business dimensions, data structure, and security level to form a data asset list; among which, data sources include BIM software, MES system, GPS terminal, sensor, testing instrument, etc., and data carriers include structured tables, unstructured files, semi-structured data, and time-series data streams; When classifying and grading, we first classify by business dimension, then by data structure, and finally by security level. Specifically, the business dimension classification is divided into six categories: design model data, component production data, logistics and transportation data, on-site hoisting data, quality acceptance data, and operation and maintenance monitoring data. Each category is further broken down into sub-subsets, such as production data being divided into process parameters, production progress, and finished product inspection data. Data structures are classified into structured data, unstructured data, semi-structured data, and time-series data. Structured data includes production schedules and inspection data; unstructured data includes BIM models, on-site videos, and acceptance reports; semi-structured data includes component attributes in XML format and sensor data in JSON format; and time-series data includes temperature, humidity, and settlement data monitored in real time by sensors. The security level classification is based on the confidentiality and importance of the data, and is divided into public data, internally shared data, and core confidential data. Core confidential data includes project cost data and core process parameters.

[0020] A3. Standardize the definition of data elements in the data asset list, clarify the unified attributes of data elements, and construct a multi-level core data indicator system in conjunction with the management objectives of prefabricated buildings. Clarify the relevant rules of the indicators and form a standard specification for data elements and a standard specification for the core data indicator system of prefabricated buildings throughout the entire process. When standardizing the definition of data elements, clarify the attributes of data elements such as Chinese name, English name, data code, data type, data length, precision requirements, and value range. For example, the data element component casting temperature is defined as a numerical type with a precision of 1 decimal place, a value range of 0-100℃, and a unit of measurement of ℃. The same data element is uniquely defined and has unified attributes throughout the entire process and among all participating entities to avoid the phenomenon of the same number with different names and the same name with different numbers. The management objectives of prefabricated buildings include schedule, quality, cost, safety, and operation and maintenance. When constructing a multi-level core data indicator system, a quantifiable and interoperable core data indicator system is built from different business dimensions. This core data indicator system is divided into primary indicators, secondary indicators, and tertiary indicators, and the statistical scope, calculation method, data source, and statistical period of each indicator are clearly defined. For example, the primary indicator of quality is the component production qualification rate, and the secondary indicator is the composite slab strength qualification rate. The statistical scope is the number of qualified composite slabs or the total production quantity, and the data source is the production-end testing instruments.

[0021] A4. Develop separate standards for data collection, transmission and storage, clarify the specific rules and mechanisms for each stage, and form corresponding standard documents; Specifically, the data acquisition standards clarify the requirements for acquisition methods, acquisition frequencies, acquisition interfaces, and data format uniformity for different types and sources. For structured data, the unified acquisition format is CSV / JSON, and the field mapping rules for the acquisition interface are specified. For BIM model data, the unified IFC standard format is adopted, and the attribute mapping and encoding rules for model components are specified. For IoT sensor data, the unified acquisition frequency is specified, and the field specifications for data reporting are specified, with acquisition frequencies such as a real-time data acquisition interval of ≤5 seconds and a near-real-time data acquisition interval of ≤1 minute. For unstructured data, the naming rules and format requirements for videos and documents are specified, such as videos being MP4 and reports being PDF. Data transmission standards are used to define cross-entity and cross-system data transmission protocols, transmission methods, transmission security, and anomaly handling standards. Specifically, the MQTT lightweight data transmission protocol is uniformly adopted for real-time IoT data, the HTTPS secure protocol is uniformly adopted for business system data, and the FTP / SFTP protocol is adopted for batch transmission of massive amounts of data. The applicable scenarios for transmission methods such as real-time push, periodic retrieval, and batch synchronization must be clearly defined. Encrypted transmission is adopted for core confidential data, and the identity authentication rules for data transmission are clearly defined to ensure transmission security. In terms of anomaly handling, emergency handling rules and data retransmission mechanisms for data transmission interruption, data loss, and data errors are formulated. Data storage standards establish unified standards for storage media, storage structures, storage periods, and data backup based on data types, characteristics, and usage requirements. Specifically, structured data is stored in relational databases, time-series data in time-series databases, unstructured data in distributed file systems, and massive aggregated data in data warehouses. The storage structure must clearly define the table structure, field definitions, and indexing rules for each database, and be consistent with the data element standards. The storage period must clearly define the retention period for different types of data. Regarding data backup, the frequency and rules for off-site backup and scheduled backup are established.

[0022] A5. Establish globally unique unified coding rules for core objects of prefabricated buildings, and at the same time, formulate data interaction interface specifications and hierarchical sharing permission standards, clarifying interface parameters, calling rules and approval processes for data access and sharing; among them, the unified coding rules must include business attribute information to ensure cross-entity identification, such as the coding rule for prefabricated components being component type code-project code-building number-floor number-sequence number, and the coding rule for equipment being equipment type code-construction unit code-sequence number; The data interaction interface specification includes interface type, interface parameters, request / response format, interface call frequency, and interface return code rules to achieve seamless integration between systems. The hierarchical sharing permission standard combines data security level classification to formulate data sharing scope, sharing method, and access permission standards, and establishes a hierarchical authorization access control mechanism. For example, public data can be shared by all entities, internally shared data is limited to project participants, and core confidential data is limited to access by designated personnel. At the same time, the application and approval process for data sharing is clearly defined.

[0023] A6. Select typical projects to conduct standard pilot verification, revise and optimize the standards based on the feedback, officially release and promote training, and establish a long-term management mechanism to achieve dynamic iteration of standards. Specifically, prefabricated building projects, such as residential and office buildings, will be selected as pilot projects. All participating entities will conduct practical business verification according to the established complete set of data standards. Problems and feedback will be collected during the pilot process, with a focus on verifying the feasibility, adaptability, and practicality of the standards. For example, verifying the consistency of data coding in production, logistics, and construction stages, and verifying the smoothness of data interface connections between various systems. In response to the problems discovered during the pilot program, such as some data element definitions not matching actual business needs, insufficient extensibility of coding rules, and missing interface parameters, experts and technicians from the fields of design, production, and construction were organized to revise the standards, optimize and improve various standard specifications, and ensure that the standards are highly matched with actual business needs. The revised and improved complete set of data standards will be officially released, forming enterprise / industry-level standard documents; at the same time, all participating entities will be organized to carry out standard promotion and training to ensure that relevant personnel are proficient in the standard content and implementation requirements; Establish a long-term management mechanism for the data standard system, set up a standard maintenance team, and dynamically update and iterate the data standard system based on the technological development of the prefabricated building industry, such as BIM technology upgrades, the application of new IoT devices, business model innovations, and new needs of actual projects, to ensure the timeliness and advancement of the system.

[0024] S2. A heterogeneous data acquisition architecture with multiple acquisition terminals is used to acquire multidimensional data in real time, and the multidimensional data is preliminarily processed. The heterogeneous data acquisition architecture includes an API interface for connecting to BIM design software, a component production data acquisition module, a logistics and transportation data acquisition module, an on-site hoisting data acquisition module, quality testing instruments, and a building operation and maintenance data acquisition module. Among them, the component production data acquisition module includes industrial sensors, PLC controllers, and a production management system (MES) deployed in the production workshop to collect component production process parameters and production progress data, and realize real-time data upload through the Industrial Internet of Things. The logistics transportation data acquisition module includes a GPS positioning terminal installed on the transport vehicle and temperature and humidity sensors and vibration sensors installed on the components. These sensors collect the real-time trajectory of the vehicle, the transport speed, and the transport environment status of the components, and transmit the data back in real time via a mobile network. The on-site hoisting data acquisition module includes working condition sensors installed on the hoisting equipment and cameras deployed on the construction site. The cameras are used for video monitoring and personnel positioning, and collect hoisting operation parameters as well as resource configuration and usage status data of on-site personnel or equipment. Quality testing instruments include ultrasonic flaw detectors and rebound hammers. The digital interface of the quality testing instruments automatically collects test data such as the appearance, size, and strength of the components. Combined with the digital reporting terminal for manual acceptance, the acceptance data is collected in real time. The building operation and maintenance data acquisition module includes temperature sensors, settlement sensors, and energy consumption sensors installed on the building structure. It collects monitoring data on structural safety, equipment operation, and energy consumption during the building operation and maintenance phase, and achieves continuous data collection through the Internet of Things.

[0025] The preliminary processing flow for multidimensional data is as follows: B1. Remove abnormal data from multidimensional data, such as extreme values ​​caused by sensor failure, invalid encoded data, and filter out duplicate data. B2. Fill in the missing data. Specifically, use the mean to fill in the numerical data, the mode to fill in the categorical data, and the interpolation method to fill in the time series data. B3. Normalize numerical data of different dimensions and ranges using the min-max normalization method to lay the foundation for subsequent data analysis model calculations. B4. Perform structuring processing on unstructured text, video, and model data, such as extracting key information about hoisting operations from videos and parsing BIM model data into structured component attribute tables.

[0026] S3. A prefabricated building big data management platform is built through a cloud-native microservice architecture. The pre-processed data is aggregated to the prefabricated building big data management platform in real time to form a unified database that spans multiple stages and entities. The prefabricated building big data management platform includes a data acquisition layer, a data storage layer, a data processing layer, an application service layer, and a visualization layer. The data acquisition layer connects to a heterogeneous data acquisition architecture and uses a message queue to achieve real-time reception and forwarding of massive amounts of data. The pre-processed data is written to the corresponding data storage medium according to a unified storage standard, realizing the integrated aggregation of design, production, logistics, construction, acceptance, and operation and maintenance data. The data storage layer builds a data warehouse, classifying and storing different types of data into matching storage media according to data storage standards. For example, structured production progress data is stored in MySQL, BIM models are stored in HDFS, and time-series data of settlement monitoring is stored in InfluxDB. At the same time, a globally unique identifier for the data is established based on unified data coding rules, so that the scattered data has a basis for being correlated and realizes the physical convergence of data. The data processing layer extracts end-to-end data from various media in the storage layer through standardized data access interfaces and performs in-depth processing according to the business logic of prefabricated buildings: First, it performs fine cleaning and standardization of data to eliminate data noise; second, based on unified data coding and data element standards, it associates and matches scattered data across links and entities, such as associating production data, logistics data, and hoisting data of prefabricated components through unique component codes; third, it constructs data models according to business subject domains such as design, production, and logistics, integrating physically scattered data into logically unified and related datasets to form the core data body of the unified end-to-end database. The data processing layer transmits the integrated unified data source to the application service layer. Based on the management needs of prefabricated buildings, the application service layer encapsulates the unified data source into standardized business data services. At the same time, combined with the data sharing permission standard, it sets hierarchical data source access permissions for different participating entities, allowing each entity to call the unified data source within its own permission scope through standardized interfaces, thereby realizing the business transformation and secure sharing of the unified data source. The visualization layer obtains unified data sources and processed analysis results from the application service layer and presents them to all participating entities in a visual form. At the same time, it receives reverse operation instructions from users, such as querying the full-process data of a certain type of component or filtering the progress data of a certain stage. The instructions are then passed to the application service layer, which triggers the data processing layer to extract the corresponding data from the unified data source. This achieves two-way linkage between data presentation, user interaction, and data retrieval, enabling the unified data source to truly serve the actual management decisions of prefabricated buildings and realize the value of the unified data source. The data acquisition layer, data storage layer, data processing layer, application service layer, and visualization layer form a closed loop based on a real-time data feedback mechanism: the application service layer and visualization layer feed back the data needs and data problems encountered by users during the process to the data processing layer, which then optimizes the data fusion rules accordingly; at the same time, the acquisition layer continuously receives real-time incremental data from the six stages and constantly supplements the storage layer and processing layer, realizing the real-time updating and dynamic improvement of the unified data source, and ensuring the timeliness and integrity of the unified data source.

[0027] S4. Conduct multi-dimensional in-depth analysis of data in the unified database across the entire process using data analysis models; Specifically, the data analysis model includes one or more of the following: component supply and demand matching analysis model, logistics route optimization analysis model, yard planning analysis model, hoisting sequence simulation analysis model, equipment and manpower configuration optimization analysis model, schedule deviation analysis model, quality anomaly identification analysis model, safety hazard early warning analysis model, and building operation and maintenance performance analysis model; The component supply and demand matching analysis model combines a time series forecasting model with an inventory optimization model. The time series forecasting model is an ARIMA model, used to predict the demand for components in the prefabricated building construction process, obtaining the quantity and timing of component demand at a future stage. This provides a basis for production capacity planning. The formula for the ARIMA model is as follows: , In the formula, For a d-order difference operator, Let be the component demand data at time t, c be a constant term, and p be the autoregressive order. Here, q represents the autoregressive coefficients, and q represents the order of the moving average. The moving average coefficient is... It is a white noise sequence. The component demand data is the historical time point i-th unit prior to time t. Let be the error value predicted by the model for the historical time point j units backward from time t.

[0028] The inventory optimization model is either the Economic Order Quantity (EOQ) model or the (s,S) dynamic inventory model. The EOQ model calculates the optimal production or supply quantity of components based on the component demand predicted by the ARIMA model, achieving quantity matching between component supply and demand and avoiding component backlog or shortage. The formula for the EOQ model is as follows: , In the formula, For the optimal economic order quantity, D is the total annual or phased demand for components, C is the fixed cost of each order, and H is the annual or phased holding cost per component. The (s,S) dynamic inventory model adapts to the dynamic changes in demand for prefabricated building components, realizes dynamic control of inventory, ensures the supply of components at the construction end, and reduces inventory holding costs. The core logic of the (s,S) dynamic inventory model is to set a lower limit s and an upper limit S of inventory. When the inventory level of a component is lower than the lower limit s, it needs to be replenished to the upper limit S immediately. When the inventory level is between the lower limit s and the upper limit S, no replenishment is required.

[0029] The logistics route optimization analysis model is a VRP vehicle route model optimized by a genetic algorithm. It is used to optimize the transportation vehicle route of prefabricated components from the manufacturer to the construction site, reduce transportation time, lower transportation costs, improve the efficiency of component logistics transportation, and ensure that components arrive on time. The yard planning and analysis model is a prefabricated component yard layout and scheduling optimization model that integrates genetic algorithm and tabu search. It is used to integrate discrete event simulation algorithm for simulation verification. It is used to scientifically plan the spatial layout, component stacking sequence and yard resource allocation of the prefabricated component yard at the construction site, improve the yard space utilization rate, reduce secondary handling of components, and avoid hoisting delays caused by yard chaos. The hoisting sequence simulation analysis model is a model that integrates the critical path method and the genetic algorithm. In addition, the hoisting sequence simulation analysis model also incorporates a discrete event simulation algorithm to simulate and predict the hoisting operation process. This model is used to simulate, predict, and optimize the hoisting sequence of prefabricated components at the construction site, determine the optimal hoisting sequence, avoid hoisting operation conflicts, improve hoisting operation efficiency, and ensure the progress of hoisting construction. The equipment and manpower configuration optimization analysis model is a dynamic configuration model for construction resources that integrates mixed integer programming algorithm and particle swarm optimization algorithm. It is combined with the time quota method for basic quantification and is used to scientifically configure and dynamically adjust resources such as hoisting equipment, transportation equipment, and construction personnel on the construction site. This achieves the optimal matching of equipment and manpower, reduces resource idleness and waste, and improves the efficiency of construction resource utilization.

[0030] The schedule deviation analysis model uses the earned value management model quantitative algorithm to quantitatively analyze the entire process of prefabricated building projects, accurately calculates the schedule deviation value, identifies the lagging / advancing links, provides data support for schedule correction, and ensures that the project progresses as planned. The quality anomaly identification and analysis model employs a fusion of XGBoost classification, PCA principal component analysis, and SVM support vector machine for the production end, and a CNN convolutional neural network model for the construction end. Furthermore, the model combines an Apriori association rule mining model to uncover the correlation patterns of quality problems, enabling intelligent analysis of quality data from both the production and construction ends of prefabricated buildings. This allows for accurate identification of quality anomalies, uncovering correlation patterns, and achieving early identification and source tracing of quality anomalies, thereby reducing the quality defect rate. The safety hazard early warning analysis model includes a threshold-based equipment anomaly detection model and a YOLO target detection model. Furthermore, the safety hazard early warning analysis model constructs a safety risk evaluation index system through the analytic hierarchy process and the fuzzy comprehensive evaluation model to achieve graded early warning of safety risks and reduce the incidence of safety accidents. The building operation and maintenance performance analysis model is a comprehensive analysis model for building operation and maintenance that integrates LSTM time series forecasting, analytic hierarchy process (AHP), and fuzzy comprehensive evaluation. It combines entropy weight method to optimize index weights and is used to continuously monitor and analyze structural safety, equipment operating status, energy consumption, and environmental indicators during the operation and maintenance phase of prefabricated buildings. It evaluates building operation and maintenance performance, provides data support for operation and maintenance scheme optimization, equipment maintenance, and energy-saving renovation, and improves the level of building operation and maintenance management.

[0031] Furthermore, the process for multi-dimensional in-depth analysis is as follows: C1. Based on the different scenario requirements of prefabricated building management, a multi-model analysis strategy that integrates traditional algorithms and deep learning algorithms is adopted to select a suitable data analysis model, and the data analysis model is deployed on the data processing layer of the big data management platform.

[0032] When selecting a suitable data analysis model, follow the selection process below: C1.1 First, the prefabricated building management scenario is broken down into detailed components to clarify the core management objectives of the scenario, such as progress control, quality inspection, resource optimization, and safety early warning. The core requirements of the data analysis model are then determined, such as quantitative calculation, real-time prediction, dynamic optimization, and image recognition. C1.2 Match the characteristics of traditional algorithms and deep learning algorithms according to the needs of business scenarios; If the scenario requires finding the global optimal solution, quantitative analysis, and rule-based decision-making, such as batch calculation of supply and demand, quantification of schedule deviation, and optimization of logistics routes, traditional algorithms should be selected first to ensure the accuracy and interpretability of the results. If the scenario requires real-time dynamic prediction, complex feature extraction, image / video recognition, and nonlinear data processing, such as quality anomaly image detection, safety hazard target identification, and dynamic prediction of component requirements, deep learning algorithms should be integrated first to improve the model's adaptive and accurate recognition capabilities. C1.3 For complex management scenarios, such as quality anomaly identification and hoisting sequence simulation, a combination of traditional and deep learning algorithms is adopted. Traditional algorithms are used to optimize the global framework, while deep learning algorithms are used to process local dynamic features, taking into account both the interpretability and intelligence of the model. C1.4. Based on the characteristics of the unified data source, such as data type, data volume, and real-time performance, verify the adaptability of the selected model to ensure that the model can extract effective feature data from the unified data source and complete the analysis and calculation.

[0033] C2. Extract the feature data required by each data analysis model from a unified data source and perform feature processing. Feature processing includes feature extraction, feature selection, and feature transformation. Input the feature data after feature processing into the corresponding data analysis model.

[0034] C3. For single analysis scenarios, the global optimal solution is obtained through traditional algorithms, while the real-time prediction and adjustment of dynamic factors are achieved through deep learning algorithms, performing multi-dimensional deep calculations from the time dimension, spatial dimension, subject dimension, and indicator dimension.

[0035] C4. Output the analysis results of the data analysis model in the form of quantitative indicators, visual charts, and early warning signals, such as quantitative values ​​of schedule deviations, optimization schemes for logistics routes, and graded early warning signals for safety hazards.

[0036] S5. Based on the results of multi-dimensional in-depth analysis and combined with the management objectives of prefabricated building projects, dynamic intelligent control is carried out on prefabricated buildings to coordinate and link the design, production, construction, acceptance and operation and maintenance links. During dynamic intelligent control, targeted control instructions are automatically generated based on multi-dimensional in-depth analysis results and the management objectives of the construction project. These instructions clearly define the control targets, measures, and target values. The control instructions are then distributed in real-time through the application service layer to execution terminals at each stage, such as the BIM design terminal of the design unit, the MES system of the manufacturer, the smart construction site management terminal, and the transportation dispatch terminal of the logistics unit. Each execution terminal completes the corresponding operation based on the issued control instructions, achieving collaborative linkage between design, production, construction, acceptance, and operation and maintenance. For example, the route of transport vehicles is adjusted based on logistics route optimization results; the equipment and personnel configuration for on-site hoisting are adjusted based on hoisting sequence simulation analysis results; and the process parameters in the production workshop are adjusted based on quality anomaly identification results. Finally, the execution process of the control instructions at each stage is monitored in real-time through the prefabricated building big data management platform, collecting data during the execution process to ensure the effective implementation of control measures.

[0037] Specifically, the control instructions include quantitative control instructions and qualitative and hierarchical control instructions. Quantitative control instructions are obtained by multi-objective weighted solving of the quantitative output of the data analysis model to obtain the optimal control parameters, which are then transformed into specific control instructions. The formula for multi-objective weighted solving is as follows: , Constraints: , In the formula, The final quantitative control parameters are defined as follows: argmax is the control parameter that maximizes the objective function. The value of , The control parameters to be optimized The range of values ​​for the control parameter is given, and M represents the number of positive control targets. Let m be the weight of the m-th positive control target. Let N be the benefit function of the m-th positive control objective, and N be the quantity of negative control costs. The weight of the nth negative control cost. Let be the loss function of the nth negative control cost. and These represent the minimum and maximum values ​​of the control parameter, respectively, and K represents the number of actual constraints in the project. The constraint equation is the actual constraint of the k-th project.

[0038] Qualitative and graded control instructions are generated by first classifying the risk or deviation of the data analysis model output results using a fuzzy comprehensive evaluation algorithm, and then matching the classification results with a preset control rule base using a rule matching algorithm. The formula expression for the fuzzy comprehensive evaluation algorithm is as follows: R, B*=argmax(B), Where B is the membership vector of risk or deviation levels, and A is the weight of the evaluation index calculated by AHP or entropy weight method. To achieve the fusion calculation of weights and evaluation matrices, R is the membership matrix of each evaluation index to different risk or deviation levels, B* is the final determined risk or deviation level, and argmax(B) is the risk or deviation level corresponding to the maximum value in the membership vector B.

[0039] The formula for the rule matching algorithm is: I=f(B*,L), where I is the final generated qualitative and hierarchical control instruction, L is the specific business scenario characteristics of prefabricated buildings, and f(.) is the rule matching function, i.e. the preset control rule matching mapping relationship.

[0040] S6. Real-time collection and feedback of control execution results and continuous iteration and optimization of data analysis models; Specifically, the results of the regulation and control implementation include data on progress, quality, safety, and resource allocation indicators after regulation and control. These data are fed back to the prefabricated building big data management platform. By establishing an evaluation index system for the regulation and control effect, the actual indicators after regulation and control are compared with the target values ​​to analyze the effectiveness of the regulation and control measures and identify the reasons for not achieving the regulation and control targets, such as unreasonable model parameters and data collection errors. Finally, the new data of the regulation and control implementation results are updated to the unified database of the entire process. The data analysis model is iteratively optimized using online learning, adjusting the parameters, feature weights, and algorithm thresholds of the data analysis model. For data analysis models with unsatisfactory results, the model is retrained or a more suitable algorithm is replaced to achieve continuous optimization of the data analysis model and improve the accuracy of subsequent analysis and regulation.

[0041] This invention also provides a big data-based intelligent management system for prefabricated buildings, used to implement the big data-based intelligent management method for prefabricated buildings described above, such as... Figure 4 As shown, the intelligent management system includes a data standard module, a data acquisition module, a big data aggregation and storage module, an intelligent analysis module, a dynamic control module, a feedback optimization module, and a terminal interaction module. The data standard module is used to establish a unified data standard system for the entire process of prefabricated buildings, to standardize the multi-source heterogeneous data of each link of design refinement, component production, logistics and transportation, on-site hoisting, quality acceptance, and operation and maintenance management, and to formulate data acquisition formats, transmission protocols and storage specifications. The data acquisition module includes multiple acquisition terminals deployed at various stages of prefabricated building construction. These terminals are used to collect multidimensional data in real time, including design model data, component production process and progress data, logistics trajectory and transportation status data, on-site hoisting operation and resource allocation data, quality acceptance and testing data, and building operation and maintenance monitoring data. The module also has data format conversion and cleaning functions to perform preliminary processing of multidimensional data and convert the raw data into standardized data that conforms to a unified data standard system. The big data aggregation and storage module is electrically connected to the data acquisition module to receive and aggregate the standardized data in real time, build a unified database for the entire process across stages and subjects, and realize centralized storage and real-time communication of data. The intelligent analysis module has a built-in preset data analysis model and is electrically connected to the big data aggregation and storage module. It is used to perform in-depth analysis of data in the unified database of the whole process, including component supply and demand matching analysis, logistics route optimization analysis, yard planning analysis, hoisting sequence simulation analysis, equipment and manpower configuration optimization analysis, schedule deviation analysis, quality anomaly identification analysis, and safety hazard early warning analysis. The dynamic control module is electrically connected to the intelligent analysis module. Based on the analysis results of the intelligent analysis module, it generates scheduling instructions, early warning information and optimization suggestions, and pushes them to the management terminals of each participant in real time, so as to realize dynamic intelligent control of each stage of prefabricated building and collaborative linkage of each participant. The feedback optimization module is electrically connected to the dynamic control module and the intelligent analysis module respectively. It is used to collect the control execution results of each link in real time and feed them back to the big data aggregation and storage module for data update. At the same time, it continuously iterates and optimizes the preset data analysis model in the intelligent analysis module based on the execution results. The terminal interaction module is electrically connected to the dynamic control module. The terminal interaction module includes multiple management terminals configured for each participant in the prefabricated building project. These terminals are used to receive scheduling instructions, early warning information, and optimization suggestions, while also supporting manual data entry and operation feedback from each participant.

[0042] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A big data-based intelligent management method for prefabricated buildings, characterized by: Includes the following steps: S1. Establish a unified data standard system for the entire process of prefabricated buildings, standardize the definition of multi-source heterogeneous data, and formulate data acquisition formats, transmission protocols and storage specifications. Among them, multi-source heterogeneous data includes data on design refinement, component production, logistics and transportation, on-site hoisting, quality acceptance and operation and maintenance management. S2. A heterogeneous data acquisition architecture with multiple acquisition terminals is used to acquire multidimensional data in real time, and the multidimensional data is preliminarily processed. S3. A prefabricated building big data management platform is built through a cloud-native microservice architecture. The pre-processed data is aggregated to the prefabricated building big data management platform in real time to form a unified database that spans multiple stages and entities. S4. Conduct multi-dimensional in-depth analysis of the data in the unified database of the entire process through data analysis models. The data analysis models include one or more of the following: component supply and demand matching analysis model, logistics route optimization analysis model, yard planning analysis model, hoisting sequence simulation analysis model, equipment and manpower configuration optimization analysis model, schedule deviation analysis model, quality anomaly identification analysis model, safety hazard early warning analysis model, and building operation and maintenance performance analysis model. S5. Based on the results of multi-dimensional in-depth analysis and combined with the management objectives of prefabricated building projects, dynamic intelligent control is carried out on prefabricated buildings to coordinate and link the design, production, construction, acceptance and operation and maintenance links. S6. Real-time collection and feedback of control execution results and continuous iteration and optimization of data analysis models.

2. The intelligent management method for prefabricated buildings based on big data according to claim 1, characterized in that: The process for establishing a unified data standard system for the entire prefabricated building process is as follows: A1. Organize the business data types of the design refinement, component production, logistics and transportation, on-site hoisting, quality acceptance and operation and maintenance management links, identify the participating entities corresponding to each business link, and form a list of business scenarios for the entire assembly process and a data interaction requirement table for each participating entity; A2. Based on the sorted business scenarios and the full-process data of the participating entities, confirm the data source, generation frequency and data carrier of various types of data, and classify and grade the full-process data according to business dimensions, data structure and security level to form a data asset list; A3. Standardize the definition of data elements in the data asset list, and at the same time, build a multi-level core data indicator system in combination with the management objectives of prefabricated buildings, forming a data element standard specification for the whole process of prefabricated buildings and a core data indicator system specification for prefabricated buildings. A4. Develop separate standards for data acquisition, transmission, and storage, and create corresponding standard documents; A5. Establish a globally unique unified coding rule for the core objects of prefabricated buildings, and at the same time, formulate data interaction interface specifications and hierarchical sharing permission standards; A6. Select typical projects to conduct standard pilot verification, revise and optimize the standards based on the feedback, officially release and promote the standards through training, and establish a long-term management mechanism to achieve dynamic iteration of the standards.

3. The intelligent management method for prefabricated buildings based on big data according to claim 2, characterized in that: When classifying and grading, we first classify by business dimension, then by data structure, and finally by security level. Specifically, the business dimension classification is divided into six categories: design model data, component production data, logistics and transportation data, on-site hoisting data, quality acceptance data, and operation and maintenance monitoring data. Each category is further broken down into subdivided data subsets. Data structures can be categorized into structured data, unstructured data, semi-structured data, and time-series data. The security level classification is based on the confidentiality and importance of the data, and is divided into public data, internally shared data, and core confidential data.

4. The intelligent management method for prefabricated buildings based on big data according to claim 1, characterized in that: The heterogeneous data acquisition architecture includes: an API interface connected to BIM design software for acquiring BIM model design parameters; The component production data acquisition module is used to collect component production process parameters and production progress data, and realizes real-time data upload through the Industrial Internet of Things. The logistics transportation data acquisition module collects real-time vehicle trajectories, transportation speeds, and the transportation environment status of components, and transmits the data back in real time via mobile networks. The on-site hoisting data acquisition module collects hoisting operation parameters as well as resource allocation and usage status data of on-site personnel or equipment; Quality testing instruments automatically collect data on the appearance, dimensions, and strength of components, and combine this with a digital data entry terminal for manual acceptance to achieve real-time collection of acceptance data; The building operation and maintenance data acquisition module collects monitoring data on structural safety, equipment operation, and energy consumption during the building operation and maintenance phase, and achieves continuous data collection through the Internet of Things.

5. The intelligent management method for prefabricated buildings based on big data according to claim 4, characterized in that: The preliminary processing flow for multidimensional data is as follows: B1. Remove abnormal data from multidimensional data and filter out duplicate data; B2. Fill in the missing data; B3. Normalize numerical data of different dimensions and ranges using the min-max normalization method; B4. Structure unstructured text, video, and model data.

6. The intelligent management method for prefabricated buildings based on big data according to claim 5, characterized in that: The prefabricated building big data management platform includes: a data acquisition layer that interfaces with a heterogeneous data acquisition architecture, uses a message queue to realize the real-time reception and forwarding of massive data, and writes the pre-processed data into the corresponding data storage medium in accordance with a unified storage standard. The data storage layer builds a data warehouse, classifies and stores different types of data into matching storage media according to data storage standards, and establishes a globally unique identifier for data based on unified data coding rules to achieve physical-level data aggregation. The data processing layer extracts full-process data from various media in the storage layer through standardized data access interfaces, performs in-depth processing according to the business logic of prefabricated buildings, and forms the core data body of the unified database for the entire process. The application service layer, based on the management needs of prefabricated buildings, encapsulates the unified data source into standardized business data services, and combines data sharing permission standards to set hierarchical data source access permissions for different participating entities. The visualization layer obtains a unified data source and processed analysis results from the application service layer, presents them to all participating entities in a visual form, and receives reverse operation instructions from users. The data acquisition layer, data storage layer, data processing layer, application service layer, and visualization layer form a closed loop based on a real-time data feedback mechanism.

7. The intelligent management method for prefabricated buildings based on big data according to claim 1, characterized in that: The workflow for multi-dimensional in-depth analysis is as follows: C1. Based on the different scenario requirements of prefabricated building management, a multi-model analysis strategy that integrates traditional algorithms and deep learning algorithms is adopted to select a suitable data analysis model and deploy the data analysis model on the data processing layer of the big data management platform. C2. Extract the feature data required by each data analysis model from a unified data source and perform feature processing. Feature processing includes feature extraction, feature selection and feature transformation. Input the feature data after feature processing into the corresponding data analysis model. C3. For single analysis scenarios, the global optimal solution is obtained through traditional algorithms, while the real-time prediction and adjustment of dynamic factors are achieved through deep learning algorithms, and multi-dimensional deep calculations are performed from the time dimension, space dimension, subject dimension, and indicator dimension. C4. Output the analysis results of the data analysis model in the form of quantitative indicators, visual charts, and early warning signals.

8. The intelligent management method for prefabricated buildings based on big data according to claim 7, characterized in that: When selecting a suitable data analysis model, follow the selection process below: C1.1 First, the management scenario of prefabricated buildings is broken down in detail to determine the core requirements of the data analysis model; C1.2 Match the characteristics of traditional algorithms and deep learning algorithms according to the needs of business scenarios: If the scenario requires finding the global optimal solution, quantitative analysis, and rule-based decision-making, traditional algorithms should be chosen first. If the scenario requires real-time dynamic prediction, complex feature extraction, image / video recognition, or non-linear data processing, deep learning algorithms should be integrated first. C1.3 For complex management scenarios, a combination of traditional and deep learning algorithms is adopted. Traditional algorithms are used to optimize the global framework, while deep learning algorithms are used to process local dynamic features, taking into account both the interpretability and intelligence of the model. C1.

4. Based on the characteristics of the unified data source, verify the adaptability of the selected model to ensure that the model can extract effective feature data from the unified data source and complete the analysis and calculation.

9. The intelligent management method for prefabricated buildings based on big data according to claim 1, characterized in that: During dynamic intelligent control, targeted control instructions are automatically generated based on multi-dimensional in-depth analysis results and the management objectives of the building project. These instructions are then distributed in real time to the execution terminals at each stage through the application service layer. Each execution terminal completes the corresponding operation according to the distributed control instructions, thereby achieving collaborative linkage between design, production, construction, acceptance, and operation and maintenance. The control instructions include quantitative control instructions and qualitative and hierarchical control instructions. Quantitative control instructions are obtained by multi-objective weighted solution of the quantitative output results of the data analysis model to obtain the optimal control parameters and convert them into specific control instructions. Qualitative and graded control instructions are generated by first classifying the risk or deviation of the output results of the data analysis model using a fuzzy comprehensive evaluation algorithm, and then matching the classification results with a preset control rule library using a rule matching algorithm.

10. A big data-based intelligent management system for prefabricated buildings, used to implement the big data-based intelligent management method for prefabricated buildings as described in any one of claims 1-9, characterized in that: include: The data standards module is used to establish a unified data standards system for the entire process of prefabricated buildings, to standardize the definition of multi-source heterogeneous data, and to formulate data acquisition formats, transmission protocols and storage specifications. The data acquisition module includes multiple acquisition terminals deployed at various stages of prefabricated building construction, used to collect multidimensional data in real time and perform preliminary processing on the multidimensional data; The big data aggregation and storage module is electrically connected to the data acquisition module. It is used to receive and aggregate the data after preliminary processing in real time to build a unified database that spans multiple stages and entities throughout the entire process. The intelligent analysis module has a built-in preset data analysis model for in-depth analysis of data in the unified database across the entire process; The dynamic control module generates scheduling instructions based on the analysis results of the intelligent analysis module; The feedback optimization module is used to collect the control execution results in real time and feed them back to the big data aggregation and storage module for data updates. At the same time, it continuously iterates and optimizes the preset data analysis model in the intelligent analysis module based on the execution results. The terminal interaction module includes multiple management terminals configured for each participant in the prefabricated building project. These terminals are used to receive scheduling instructions and support manual data entry and operation feedback from each participant.

Images