Building interior design and construction integrated platform based on BIM technology

Abstract

The invention relates to the technical field of building information processing, and discloses a building interior design and construction integrated platform based on a BIM technology, and the platform can analyze a semantic design intention label preset in a building information model, and converts the semantic design intention label into a formalized engineering constraint condition; based on the engineering constraint condition and the construction process knowledge graph, generating a final-version construction instruction packet through multi-dimensional simulation rehearsal; in the construction process, the platform collects on-site actual execution data, and compares and analyzes the actual execution data with the plan data in the instruction packet to identify systematic deviation; and finally, according to the identified systematic deviation, the platform updates the construction technology knowledge graph under a man-machine collaborative auditing mechanism. According to the method, a data closed loop from the design intention to the construction practice and then to the core knowledge base is constructed, the technical problems that design and construction are disjointed and the plan performability is low are solved, and self-evolution and continuous optimization of a construction knowledge system are achieved.

Description

Technical Field

[0001] This invention relates to the field of building information processing technology, specifically to an integrated platform for building interior design and construction based on BIM technology. Background Technology

[0002] Building Information Modeling (BIM) technology has been applied in current architectural interior design and construction practices, playing a role in 3D visualization, geometric conflict detection, and quantity surveying. However, existing technologies still have shortcomings in achieving deep integration and intelligent management of the design and construction processes.

[0003] The functional and performance requirements established during the design phase are usually in the form of text annotations or independent technical specifications, and have a weak connection with the component entities in the BIM model. This leads to on-site personnel having to rely on manual interpretation and experience to translate these high-level design intentions into specific construction operations during the construction phase. This process is prone to information distortion and omissions, resulting in deviations between the final construction outcome and the original design objectives.

[0004] Furthermore, traditional construction planning often relies on static timeframes and idealized construction environment assumptions. Such plans often fail to adequately anticipate and handle the complex dynamic changes on the construction site, such as interference between different work processes in the workspace, scheduling conflicts of critical shared equipment (e.g., construction elevators), or the impact of specific operations (e.g., spraying) on ​​nearby environmentally sensitive processes (e.g., finishing installation). This leads to frequent interruptions and adjustments to the prepared construction plans during actual execution, impacting the overall efficiency of the project.

[0005] More importantly, valuable experience data generated during construction, such as the actual performance of specific materials in non-standard environments or optimized process connections, currently lacks an effective mechanism for systematic collection, analysis, and feedback. This practical knowledge is often lost after the project ends, failing to form structured knowledge that can be reused in subsequent projects. As a result, the knowledge base on which planning and decision-making are based remains stagnant for a long time, making it impossible to achieve continuous iteration and self-improvement based on actual data.

[0006] Therefore, this invention proposes an integrated platform for architectural interior design and construction based on BIM technology to address the shortcomings of existing technologies. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides an integrated platform for architectural interior design and construction based on BIM technology. This platform solves the problems of disconnect between architectural interior design and construction processes, and the inability of construction knowledge to undergo closed-loop iteration and self-optimization based on actual on-site data.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an integrated platform for architectural interior design and construction based on BIM technology, comprising:

[0009] The knowledge graph management module is used to build and maintain a construction process knowledge graph;

[0010] The BIM model parsing module is used to parse the building information model containing building interior design information, extract the physical properties of components, and identify the preset semantic design intent tags in the building information model, and convert the semantic design intent tags into a set of formal engineering constraints.

[0011] The intelligent construction instruction set generation module is used to generate a final version of the construction instruction package based on the formalized engineering constraints and the construction technology knowledge graph.

[0012] The on-site execution and feedback module is used to send the final version of the construction instruction package to the on-site terminal equipment to present the construction instructions, and to collect the actual execution data on-site during the construction process;

[0013] The knowledge graph reverse enhancement module is used to compare and analyze the actual execution data with the planned data in the final version of the construction instruction package to identify systematic deviations, and generate knowledge update content based on the systematic deviations to update the construction process knowledge graph.

[0014] Preferably, the BIM model parsing module is specifically used for: detecting and reading semantic design intent tags attached to building information model components; querying the intent constraint mapping library configured within the platform, which stores mapping rules from semantic design intent tags to one or more sets of formal engineering constraints; and converting the semantic design intent tags into a set of formal engineering constraints according to the queried mapping rules.

[0015] Preferably, the intelligent construction instruction set generation module is specifically used for: modeling the construction planning problem as a constraint satisfaction problem, generating a candidate construction instruction package; performing multi-dimensional simulation pre-run on the candidate construction instruction package to detect dynamic execution conflicts; transforming the detected dynamic execution conflicts into new formal constraints, and using the constraint set containing the new formal constraints to re-solve the constraint satisfaction problem to generate a modified candidate construction instruction package; repeating the pre-run and re-solution process until the multi-dimensional simulation pre-run no longer detects any new dynamic execution conflicts, and using the candidate construction instruction package generated in the last solution as the final version of the construction instruction package.

[0016] Preferably, the multi-dimensional simulation pre-run includes at least one of the following simulations: spatiotemporal occupancy simulation, which is used to detect dynamic spatial conflicts of construction personnel or mobile equipment; critical resource scheduling simulation, which is used to verify the feasibility of scheduling schemes for exclusive resources; and environmental impact simulation, which is used to check whether there are cross-impact conflicts between construction tasks that generate environmental impacts and construction tasks that are sensitive to the environment.

[0017] Preferably, when identifying systematic deviations, the knowledge graph reverse enhancement module is specifically used to: aggregate and group the actual execution data and planned data according to the process entities or material entities associated with the construction task; calculate the deviation between the planned value and the actual value of each task instance for preset key performance indicators; and perform a statistical significance test on the calculated deviation sequence to determine whether the deviation is a statistically significant systematic deviation.

[0018] Preferably, when updating the construction process knowledge graph, the knowledge graph reverse enhancement module is specifically used to: automatically generate a knowledge update proposal containing suggested attribute values ​​based on the identified systematic deviations; push the knowledge update proposal to a preset review queue for review by users with corresponding permissions; and execute the update operation on the construction process knowledge graph after receiving the user's approval or modification instruction.

[0019] Preferably, the actual execution data collected by the on-site execution and feedback module is multimodal data, which includes: task status data submitted by construction personnel through on-site terminal equipment; on-site process data in the form of images or videos collected by the built-in camera of the on-site terminal equipment; and environmental data automatically collected by IoT sensing devices deployed on the construction site.

[0020] Preferably, the construction process knowledge graph includes: a set of nodes representing process entities, material entities, tool entities, quality standard entities, or environmental constraint entities; and a set of edges representing the semantic relationships between entity nodes, including prerequisite dependencies, material compatibility relationships, or tool requirement relationships.

[0021] Preferably, before generating the final version of the construction instruction set intelligent generation module, it is further configured to: identify the relevant knowledge domains required to complete the project, and calculate the knowledge entropy of each knowledge domain to quantify the degree of uncertainty of the knowledge domain; when the knowledge entropy of any knowledge domain is greater than a preset entropy threshold, it is determined that the knowledge domain has high uncertainty, and one or more active exploration protocols are generated for the knowledge domain, wherein the active exploration protocol is a field test task for accurately collecting decision data.

[0022] Preferably, the platform is deployed on a cloud server and interacts with one or more field terminal devices via a network; the field terminal devices are smartphones, tablets, or augmented reality glasses.

[0023] This invention provides an integrated platform for architectural interior design and construction based on BIM technology. It offers the following advantages:

[0024] 1. This invention, through a BIM model parsing module, can automatically query and transform semantic design intent tags such as "high-level sound insulation" added by designers in the model into a set of clear, machine-readable formal engineering constraints. For example, it can mandate that the filling material be "high-density rock wool" and the joint treatment process be "staggered installation and use of sound insulation sealant". This process directly links high-level design requirements with low-level construction specifications, effectively avoiding design requirement downgrades or construction errors caused by deviations in manual interpretation and information transmission, and ensuring a high degree of consistency between the final delivery and the initial design goals.

[0025] 2. This invention utilizes an intelligent construction instruction set generation module. After generating the construction plan, it further performs multi-dimensional simulations and pre-runs on the spatial and temporal occupancy, key resource scheduling, and environmental impact. This module can identify and resolve dynamic execution conflicts that may occur during actual construction in advance, such as conflicts in the work paths of two different work teams in a confined space, or the impact of volatile gases generated by spraying operations on the precision finishing installation tasks in adjacent areas. This mechanism significantly improves the actual executability of the construction plan by transforming conflicts into new constraints and iteratively solving them, reducing on-site delays and resource waste caused by inadequate planning.

[0026] 3. This invention establishes a feedback loop from on-site execution to the core knowledge base. Through the knowledge graph reverse enhancement module, it compares and analyzes planned data with actual execution data collected on-site to identify statistically significant systematic deviations. For example, if the data shows that the actual drying time of a specific process under a specific humidity environment is generally longer than the standard value in the knowledge graph, the system will generate a knowledge update proposal. This mechanism corrects the construction process knowledge graph after the identified deviations are reviewed by human-machine collaboration, enabling the platform's knowledge system to evolve and continuously optimize itself based on actual construction data, allowing historical project experience to be quantified, accumulated, and used to guide future engineering practices. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0028] Figure 2 This is a schematic diagram of the method flow of the present invention;

[0029] Figure 3This is a schematic diagram illustrating the transformation of semantic design intent tags into formal engineering constraints in this invention.

[0030] Figure 4 This is a schematic diagram of the closed-loop optimization process of the construction instruction set based on simulation pre-drilling according to the present invention.

[0031] Figure 5 This is a schematic diagram of the on-site multimodal data acquisition and structured alignment of the present invention;

[0032] Figure 6 This is a schematic diagram of the knowledge graph evolution process of human-machine collaboration according to the present invention.

[0033] The module includes: 100 Knowledge Graph Management Module; 200 BIM Model Parsing Module; 300 Intelligent Generation Module for Construction Instruction Sets; 400 On-site Execution and Feedback Module; 500 Knowledge Graph Reverse Enhancement Module; 600 Cloud Server; and 700 On-site Terminal Equipment. Detailed Implementation

[0034] The technical solutions in 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. 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.

[0035] See attached document Figure 1 , Figure 1 This is a schematic diagram of the system architecture of a BIM-based integrated platform for architectural interior design and construction, according to an embodiment of the present invention. The present invention provides a BIM-based integrated platform for architectural interior design and construction, which is deployed on one or more cloud servers 600 and interacts with one or more field terminal devices 700. The platform may include: a knowledge graph management module 100, a BIM model parsing module 200, a construction instruction set intelligent generation module 300, a field execution and feedback module 400, and a knowledge graph reverse enhancement module 500.

[0036] See attached document Figure 1 The hardware architecture of the platform in this embodiment of the invention includes a cloud server 600, an Internet of Things (IoT) sensing device deployed at the construction site, and a field terminal device 700 for use by construction and management personnel. The cloud server 600 is used to carry the core computing and data storage functions of the platform. The field terminal device 700 includes, but is not limited to, smartphones, tablets, or augmented reality (AR) glasses, for receiving and presenting instructions and collecting and reporting field data.

[0037] The initial input to the process is a Building Information Model (BIM) containing architectural and interior design information; users upload the BIM model to the platform deployed on the cloud server 600 via on-site terminal device 700 or other clients.

[0038] The BIM model parsing module 200 receives the BIM model. The BIM model parsing module 200 first extracts the geometric information, dimensions, materials and other physical properties of the components in the model. At the same time, the BIM model parsing module 200 identifies and parses the semantic design intent tags that are pre-annotated in the model to describe functional or performance requirements, and converts these tags into a set of formalized engineering constraints that can be processed by a computer.

[0039] The intelligent construction instruction set generation module 300 receives model data and engineering constraints processed by the BIM model parsing module 200. The intelligent construction instruction set generation module 300 first calls the knowledge graph management module 100 to retrieve knowledge such as process dependencies, material properties, and quality standards related to the current engineering task from the construction process knowledge graph stored therein.

[0040] Subsequently, the intelligent construction instruction set generation module 300 performs calculations based on the retrieved knowledge and input constraints to generate a candidate construction instruction package (CIP) containing detailed procedures, resource requirements, and temporal relationships. The intelligent construction instruction set generation module 300 further performs multi-dimensional simulation pre-runs on the candidate instruction package to verify and resolve potential spatiotemporal, resource, or environmental conflicts. After conflict resolution, a final version of the construction instruction package is generated.

[0041] The cloud server 600 distributes the final version of the construction instruction package to one or more field terminal devices 700 via the network; the field execution and feedback module 400 deployed on the field terminal device 700 presents the contents of the instruction package to the field construction personnel in the form of a visual step list or a 3D model overlay to guide their work.

[0042] During construction, the on-site execution and feedback module 400 continuously collects actual execution data on-site. This data is multimodal, including progress information and quality acceptance photos submitted by construction personnel through terminal devices, as well as environmental parameters automatically uploaded by on-site IoT sensing devices. This actual data, which includes time and space stamps, is transmitted back to the cloud server 600.

[0043] The knowledge graph reverse enhancement module 500 receives the actual execution data returned; the knowledge graph reverse enhancement module 500 compares and analyzes the actual data with the planned data in the final version of the construction instruction package to identify the systematic deviations between the two.

[0044] When the knowledge graph reverse enhancement module 500 identifies a significant systematic deviation, it generates a knowledge update proposal. After the proposal is reviewed and confirmed by an administrator with the appropriate permissions, it is executed by the knowledge graph management module 100 to correct the relevant parameters or rules in the construction process knowledge graph. This step completes the data loop from construction practice to the core knowledge base, enabling the platform's knowledge system to iterate and optimize.

[0045] See attached document Figure 2 , Figure 2 This is a schematic diagram of a BIM-based integrated method for architectural interior design and construction according to an embodiment of the present invention. The present invention provides a BIM-based integrated method for architectural interior design and construction, which may include the following steps:

[0046] S100, Obtain the Building Information Model (BIM) containing architectural interior design information.

[0047] S200 parses the building information model, extracts the physical properties of components, and identifies the pre-defined semantic design intent tags in the model, transforming them into a set of formalized engineering constraints.

[0048] S300, based on engineering constraints and a pre-set construction technology knowledge graph, generates candidate construction instruction packages through multi-constraint solving, and performs multi-dimensional simulation and pre-playing of the candidate construction instruction packages to resolve conflicts, thereby generating the final version of the construction instruction package.

[0049] S400 sends the final version of the construction instruction package to at least one field terminal device, and presents the construction instructions to the construction personnel through the field terminal device.

[0050] During construction, the S500 collects actual execution data on-site through field terminal equipment and / or IoT sensing equipment, and then transmits the actual execution data back.

[0051] S600 compares and analyzes the actual execution data with the planned data in the final version of the construction instruction package to identify systematic deviations. When a systematic deviation is identified, knowledge update content is generated based on the deviation, and the construction process knowledge graph is updated using the knowledge update content under a preset review mechanism.

[0052] The following will provide a more detailed description of each step in the method of the embodiments of the present invention, as well as the platform functional modules that implement these steps.

[0053] The knowledge graph management module 100 in this embodiment of the invention is used to construct and maintain a construction process knowledge graph that serves as the basis for platform reasoning and decision-making. The implementation of the knowledge graph management module 100 includes the extraction and representation of construction knowledge, and the formal modeling and storage of knowledge.

[0054] The knowledge graph management module 100 processes multi-source heterogeneous data to obtain structured construction knowledge. These data sources include, but are not limited to, national or industry building construction specifications, technical specifications for specific materials, and internal project technical briefing documents and construction logs. The acquisition process may include the following steps:

[0055] S110 preprocesses the raw text data, including text cleaning, sentence segmentation, and word segmentation, to form standardized text units.

[0056] S120, using a named entity recognition model, the preprocessed text units are processed to identify and extract predefined construction entities; construction entity types include: process entities, such as wall putty application; material entities, such as water-resistant putty; tool entities, such as scraper; quality standard entities, such as surface flatness; environmental constraint entities, such as construction ambient temperature; the named entity recognition model used to perform this step can be a deep learning-based model, such as a BERT model fine-tuned on a professional corpus in the construction field, which is a well-known technology in the field and will not be described in detail here.

[0057] S130 uses a relation extraction model to process the text units of the identified construction entities in order to identify and extract the semantic relationships between the entities. The types of semantic relationships include: prerequisite dependencies, which describe the sequence of processes; material compatibility, which describes whether different materials can be used together; and tool requirement, which describes the tools required for a specific process.

[0058] The extracted construction entities and semantic relationships will be used to construct a formalized knowledge graph of construction processes. This construction process knowledge graph The structure can be defined as: ;

[0059] in, It is a set of nodes in the diagram, representing the various construction entities extracted in the preceding steps; set Each node in Each is a structured data object, and its data structure can be represented as follows: ;in, It is the unique identifier of that node; It is the entity type of this node, and its value is one of the predefined types such as process entity, material entity, tool entity, etc. It is a set of attributes used to describe the specific attributes of the node in the form of key-value pairs. For example, the attribute set of a material entity can be {name: "light steel keel", specification: "DC50", fire rating: "Class A"}.

[0060] in, It is a set of edges in a graph, representing the semantic relationships between entity nodes; set Each edge in Each is a triplet, and its data structure can be represented as follows: ;in, It is the head entity node of the relationship. It is the tail entity node of the relation. It refers to the type of relationship between the two; for example, a prerequisite dependency relationship can be represented as (wall puttying, prerequisite dependency, wall base treatment), which indicates that the wall puttying process can only begin after the wall base treatment process is completed.

[0061] To achieve efficient storage and retrieval of the construction process knowledge graph, the knowledge graph management module 100 can use a graph database for data persistence; those skilled in the art can use graph database products such as Neo4j; graph databases can natively support graph structures of nodes and edges, and provide high-performance query capabilities for complex relationships, making them suitable for storing and managing the construction process knowledge graph in this embodiment of the invention.

[0062] See attached document Figure 3 , Figure 3 This is a schematic diagram illustrating the conversion of semantic design intent tags in a BIM model into formal engineering constraints according to an embodiment of the present invention. The BIM model parsing module 200 in this embodiment of the invention functions to parse the input Building Information Model (BIM) to extract the physical attribute information and formal engineering constraints required for subsequent steps.

[0063] The BIM model parsing module 200 receives BIM model files stored in the Industrial Foundation Class (IFC) standard format or a specific software native format. For the extraction of the geometric information, dimensions, materials, location and other physical attributes of components in the BIM model, those skilled in the art can use existing BIM parsing libraries or software development kits. The specific implementation is a well-known technology in the field and will not be described in detail here.

[0064] In addition, the BIM model parsing module 200 is also used to perform semantic parsing and constraint transformation of design intent. To achieve this function, designers can select one or more design intent tags from a preset tag library through a dedicated plug-in tool integrated with the BIM design software during the BIM design phase, and attach them as custom attributes to specific components of the BIM model. Design intent tags are high-level semantic identifiers used to describe the functional, performance, or effect requirements of components, such as "high-level sound insulation", "cinema-grade sound absorption effect", or "seamless finish".

[0065] The specific steps that the BIM model parsing module 200 takes when performing parsing may include:

[0066] S210 iterates through all components in the BIM model, extracting their physical properties and detecting and reading the design intent tags attached to the components.

[0067] S220: For each read design intent tag, a query is performed in an intent constraint mapping library configured within the platform. This intent constraint mapping library stores a series of mapping rules, each of which defines the correspondence between a design intent tag and one or more sets of engineering constraints.

[0068] S230, based on the retrieved mapping rules, transforms high-level, unstructured design intent tags into one or more sets of formalized engineering constraints that can be directly processed by a computer; each engineering constraint... Each is constructed as a structured data object, whose data structure can be represented as follows: ;in: For constraint types, their values ​​are selected from a predefined set of types, including: material constraints, process constraints, lower limits of quality standards, performance index requirements, etc. For the purpose of constraint, it is used to specify the specific engineering attributes to which the constraint applies, such as the type of wall framing, the number of coats of putty, and the surface flatness tolerance of the wall. For constraint operators, their values ​​are selected from a predefined set of mathematical or logical operators, which includes: equal to, greater than or equal to, less than or equal to, contained in a set, etc. The constraint value is the specific value or range of values ​​for the constraint target.

[0069] For example, for a wall component with a "high-level sound insulation" design intent label, the intent constraint mapping library may contain the following mapping rule: convert the "high-level sound insulation" label into a set containing two engineering constraints; the first engineering constraint is (material forced, filling material, equal to "high-density rock wool"), and the second engineering constraint is (process forced, joint treatment, equal to "staggered installation and use of sound insulation sealant").

[0070] S240, associate the resulting set of engineering constraints with the unique identifier of the component (e.g., its globally unique identifier GUID).

[0071] Finally, the BIM model parsing module 200 outputs the parsed component physical attribute information and the formalized engineering constraint set associated with each component, for use by the subsequent intelligent construction instruction set generation module 300.

[0072] Before formally solving the construction instruction set, the intelligent construction instruction set generation module 300 in this embodiment of the invention first executes an active exploration mechanism based on knowledge entropy evaluation to pre-evaluate and complete the certainty of the knowledge on which the generated instructions depend.

[0073] The mechanism aims to identify knowledge domains in the construction process knowledge graph that are relevant to the current project and have high uncertainty. The uncertainty stems from the sparsity of historical execution data, conflicts in historical data results, or the introduction of new materials or processes not covered by the knowledge graph in the project.

[0074] The specific implementation of this mechanism may include the following steps:

[0075] S310, after the construction instruction set intelligent generation module 300 receives the data output by the BIM model parsing module 200, it first identifies all relevant knowledge domains required to complete the project; one knowledge domain It can be defined as a subgraph in the construction process knowledge graph that is related to a specific construction task, such as all nodes and relationships related to "construction process of a specific brand of waterproof coating".

[0076] S311, for each identified knowledge domain Calculate its knowledge entropy Knowledge Entropy The computational model used to quantify the degree of uncertainty in this knowledge domain is as follows:

[0077] ;

[0078] in: , , The preset weight coefficients represent the importance weights of data sparsity, data conflict, and knowledge novelty, respectively, and satisfy the following conditions: ; Data sparsity entropy is used for measurement and knowledge domains. The number of relevant historically successful execution cases; this value is inversely proportional to the number of cases, and its calculation formula is as follows: ,in, For the knowledge domain The total number of cases successfully executed in the historical project database. This is a positive constant used to control the decay rate; Data conflict entropy measures the consistency of key performance indicator results across historical cases. Its value is proportional to the dispersion of the indicator results, and its calculation formula is as follows: ,in, It is related to the knowledge domain Random variables related to one or more key performance indicators, such as labor hours or material loss rate; This indicates that the variance of the variable is calculated. This indicates the calculation of its expected value; Knowledge novelty entropy is used to measure the knowledge domain. Whether it is entirely new content in the knowledge graph; its calculation formula can be an indicator function: Among them, when the knowledge domain When the technology or materials involved have no corresponding record in the construction technology knowledge graph, ,otherwise .

[0079] S312, calculate the knowledge entropy Compared with the preset entropy threshold Compare; if If the uncertainty is high, then the knowledge domain is deemed insufficient to support the generation of highly reliable construction instructions.

[0080] S313 For each knowledge domain identified as having high uncertainty, the system generates one or more Active Exploration Protocols (AEPs); an Active Exploration Protocol is a structured, small-scale, low-risk field testing task whose sole purpose is to accurately collect decision data for high-entropy knowledge points.

[0081] The data structure of an active exploration protocol may include: a unique task identifier, a test target, a list of required materials, detailed test steps, data recording requirements, and task termination conditions. For example, for a novel quick-drying adhesive material (corresponding to a knowledge domain with high novelty entropy), the system can generate an active exploration protocol, the contents of which are: applying three different thicknesses of adhesive to a standard sample substrate, and using a timer and sensor to continuously record its curing time and bonding strength under specific temperature and humidity conditions until the strength reaches a preset standard.

[0082] The generated proactive exploration protocol will be integrated as a separate, high-priority task into the preparation phase of the subsequently generated construction plan to ensure that its knowledge uncertainty has been effectively reduced before the commencement of relevant large-scale construction.

[0083] In this embodiment of the invention, the intelligent construction instruction set generation module 300, after completing the knowledge entropy assessment, executes a construction instruction set generation method coupled with multidimensional constraints to transform the construction planning problem into a formalized mathematical problem that can be solved by a computer; the method models the process of generating the construction plan as a constraint satisfaction problem (CSP).

[0084] A constraint satisfaction problem can be formally defined as a triple. .in, It is a set of variables; in this embodiment, this set Let represent the n independent construction tasks required to complete the project; each construction task Each contains a set of variables to be determined, primarily including the planned start time of the task. and the set of resources allocated to this task. .in, It is the set of domains, which contains sets. The possible value range for each variable; for example, for the variable "planned start time". Its domain is all valid time points within the project duration; for the variable resource set Its domain is the combination of all available workers, equipment, and tools in the enterprise resource pool. It is a set of constraints, containing a series of conditions that must be satisfied; these conditions are derived from the construction process knowledge graph, BIM model analysis results, and general project management rules; this set of constraints... Specifically, constraints may include the following types:

[0085] S321, Temporal Constraint; This constraint defines the temporal order of different construction tasks, primarily based on the "predependency" relationships defined in the construction technology knowledge graph; for any two construction tasks... and If there exists from arrive The "predependency" relationship must satisfy the following conditions: ;in, and These are the construction tasks and The planned start time, It is a construction task The planned duration.

[0086] S322, Resource Constraint; This constraint ensures that at any given time, the total demand for any type of resource does not exceed the total availability of that resource; for any given time... and any type of resource (For example, workers in a specific job or a specific type of equipment) must meet the following conditions:

[0087] ;

[0088] in, It is a resource type Total available quantity; It is assigned to the construction task A collection of resources; It is the calculation that assigns tasks to construction projects. The resource type is A function of the quantity of resources; It is an indicator function; its value is 1 when its internal condition is true, and 0 otherwise.

[0089] S323, Spatial Constraint; This constraint ensures that at any given time, the workspaces of any two concurrently executed construction tasks will not physically interfere with each other; for any two construction tasks that overlap in time... and The following conditions must be met: ;

[0090] in, Represents construction tasks At any moment The bounding box of the three-dimensional space occupied; the bounding box includes not only the geometric space of the component itself, but also the worker operating space, equipment space and temporary material storage space required to perform the task.

[0091] S324, Project-Specific Constraint; this constraint originates from the formalized project constraints output by the BIM model parsing module 200; for example, a project constraint of type "Material Mandatory" will directly limit the construction task. variable resource set The domain of discourse, that is, the set must contain the specified material type.

[0092] The intelligent construction instruction set generation module 300 uses a constraint solver to solve the constraint satisfaction problem constructed above. For solving constraint satisfaction problems, those skilled in the art can employ algorithms such as backtracking search and constraint propagation; their specific implementations are well-known technologies in the field and will not be elaborated upon here. The goal of the solution is to find a set of pairs of variables. The assignment of values ​​to all variables in the constraint set makes the constraint set All constraints are satisfied simultaneously.

[0093] Once the solution is successful, the result is a candidate construction instruction package. This candidate construction instruction package contains the clear start and end times of all construction tasks, resource allocation schemes, and logically satisfies all known engineering constraints.

[0094] See attached document Figure 4 , Figure 4 This is a schematic diagram of a closed-loop optimization process for a construction instruction set based on simulation pre-drafting, according to an embodiment of the present invention. After generating candidate construction instruction packages, the intelligent construction instruction set generation module 300 further performs instruction set verification and closed-loop optimization based on simulation pre-drafting to verify and resolve potential dynamic execution conflicts that are difficult to detect at the logical level.

[0095] This process takes place in a four-dimensional (4D) virtual construction environment built on a BIM model; this environment integrates the project's three-dimensional spatial information with the time planning information of candidate construction instruction packages; the process may include the following steps:

[0096] S331, perform multi-dimensional simulation pre-run on candidate construction instruction packages; this pre-run is not a single-dimensional check, but includes multiple levels of parallel simulation analysis:

[0097] Spatiotemporal occupancy simulation: This simulation models construction workers, mobile devices, etc., as dynamic agents with preset movement paths and speeds; the simulation engine extrapolates the spatiotemporal volume of these dynamic agents on the time axis to detect whether there are dynamic spatial conflicts such as path intersections, traffic congestion, or insufficient safety distance.

[0098] Critical Resource Scheduling Simulation: This simulation focuses on validating scheduling schemes for exclusive resources used serially on a construction site (such as tower cranes and construction elevators). The simulation engine examines the time and physical feasibility of state transitions (e.g., a crane moving from point A to point B and adjusting its lifting gear) between consecutive tasks assigned to this type of resource in order to identify unrealistic scheduling arrangements.

[0099] Environmental Impact Simulation: This simulation is used to assess the environmental impact of a specific construction task on the surrounding area; it is suitable for construction tasks that generate dust, volatile gases, high humidity, or noise. (For example, spraying, welding), the system defines a time-varying area of ​​influence for it during execution. Meanwhile, for construction tasks that are sensitive to the environment... (For example, precision finishing installation, final paint application), the system checks its workspace. Is it related to the affected area? Overlap occurs; that is, the following conditions must be met: ;in, It is a construction task At any moment The workspace enclosure box, It is a construction task At any moment The environmental impact area; if this condition is not met, it is determined that there is a conflict of environmental cross-impact.

[0100] S332 transforms any conflict events detected during simulation into new, formalized constraints; this transformation process is crucial for achieving closed-loop optimization. For example:

[0101] If two construction tasks are detected in the spatiotemporal occupancy simulation and The path of the executor at any time If a conflict occurs, the system can generate a new timing constraint that stipulates that the two tasks cannot be executed in parallel. or .

[0102] If a construction task is detected in the environmental impact simulation Received To mitigate interference, the system also generates a timing constraint that prevents the two from executing in parallel.

[0103] S333, Add the newly generated constraints to the existing constraint set. In this process, an updated set of constraints is formed. .

[0104] S334, the intelligent construction instruction set generation module 300 uses the updated constraint set. The aforementioned constraint satisfaction problem is re-solved to generate a revised, new candidate construction instruction package.

[0105] S335, repeat steps S331 to S334 until no new conflicts are detected in a complete simulation pre-run; this iterative process ensures the systematic investigation and correction of various dynamic conflicts.

[0106] When the iteration terminates, the candidate construction instruction package generated by the last solution is determined as the final version of the construction instruction package. This final version of the construction instruction package not only logically satisfies all initial constraints, but also passes multi-dimensional simulation verification at the dynamic execution level, and has higher actual executability.

[0107] See attached document Figure 5 , Figure 5This is a schematic diagram of on-site multimodal data acquisition and structured alignment according to an embodiment of the present invention. In this embodiment, the on-site execution and feedback module 400 is deployed on one or more on-site terminal devices 700 to receive and present construction instructions, while simultaneously collecting actual on-site execution data to form a feedback loop.

[0108] The on-site execution and feedback module 400 receives the final version of the construction instruction package issued by the cloud server 600. This module 400 parses the structured instruction data and visualizes it through the user interface of the on-site terminal device 700. When construction personnel log into their accounts, the on-site execution and feedback module 400 can display their personal task list for the day or week. For each construction task in the list, the module 400 provides a linked view with the BIM 3D model. When a construction personnel selects a task, the system will highlight the corresponding BIM component in the 3D view and clearly display the work requirements in the form of step-by-step instructions, a material list, and quality standards. For on-site terminal devices 700 equipped with augmented reality capabilities, the on-site execution and feedback module 400 can also overlay construction steps, virtual models, and positioning information onto the real construction scene, providing construction personnel with more intuitive guidance.

[0109] To achieve reverse enhancement of the construction process knowledge graph, the on-site execution and feedback module 400 also performs knowledge-enhanced on-site multimodal data acquisition. This acquisition process aims to obtain high-quality, structured actual execution data that is precisely aligned with the planned data with minimal personnel burden. This process may include the following steps:

[0110] S410 collects task status data submitted by construction personnel through the user interface of the field terminal device 700; construction personnel can operate through the buttons on the interface when the task starts, pauses, resumes and is completed; when the field execution and feedback module 400 receives the operation instruction, it automatically records the current operation type and the precise timestamp, and associates it with the corresponding task unique identifier.

[0111] S410 uses the built-in camera of the field terminal device 700 to collect on-site process data in the form of images or videos. For example, after a key process is completed or before a concealed project is closed, the system will prompt the construction personnel to take photos as process records or quality acceptance certificates. The captured image data is also associated with a unique task identifier and timestamp. Some image data can be uploaded to the cloud server 600 for secondary processing by computer vision analysis services, such as automatically assessing the flatness of the finish or the accuracy of component installation through image recognition technology.

[0112] The S420 receives and integrates environmental data automatically collected by IoT sensing devices deployed at the construction site; these devices continuously monitor parameters such as temperature, humidity, and volatile organic compound (VOC) concentration in the air in specific areas; each sensor reading is recorded along with the sensor identifier, timestamp, and spatial location information.

[0113] S430 performs unified structuring and alignment processing on all multimodal data collected through the above methods; each independent actual data point All are formatted as a standard data tuple, whose data structure can be represented as:

[0114] ;

[0115] in: It is the unique identifier of the construction task associated with the data point. This identifier comes directly from the final version of the construction instruction package and is the key to achieving accurate alignment between planned data and actual data. This is the timestamp of the data point being collected; It is the spatial location information of the data point, which can be derived from the Global Positioning System (GPS), indoor positioning system or the spatial coordinates of its associated BIM components; It refers to the data type of this data point, such as task status, ambient temperature, quality acceptance photos, etc. This refers to the specific value of the data point. For example, the task status could be "completed", the temperature could be "21.5℃", and the photo could be its storage address. It is the source identifier of the data point, such as a specific construction worker ID, IoT sensor ID, or the name of a computer vision analytics service.

[0116] The actual execution data stream, after being structured, is transmitted back to the cloud server 600 in real time or in batches by the on-site execution and feedback module 400.

[0117] The knowledge graph reverse enhancement module 500 in this embodiment of the invention is used to analyze the actual execution data transmitted back from the site in order to continuously optimize the construction process knowledge graph; the knowledge graph reverse enhancement module 500 first performs the identification of the systematic deviation between the plan and the actual situation.

[0118] This identification process aims to distinguish between random, unpredictable deviations that occur accidentally during construction and statistically significant systematic deviations caused by imprecise knowledge parameters or rules in the knowledge graph. The process may include the following steps:

[0119] S510, Data Aggregation and Grouping; The knowledge graph reverse enhancement module 500 receives the structured actual execution data stream returned by the on-site execution and feedback module 400, and retrieves the corresponding final version of the construction instruction package as the plan data from the database of the cloud server 600; The knowledge graph reverse enhancement module 500 aggregates and groups discrete actual data points according to the process entities, material entities, or environmental constraint entities in the construction process knowledge graph associated with the construction task, forming a dataset for statistical analysis; For example, the actual execution data of all construction tasks related to the "specific brand putty application" process entity will be grouped into the same dataset.

[0120] S511, calculate the deviation index; for each aggregated dataset, the knowledge graph reverse enhancement module 500 calculates the deviation between the planned value and the actual value of each task instance for each preset key performance index; taking the duration of a construction task as an example, for the first task in the dataset... The duration deviation of each task instance It can be calculated as: ;in, The actual duration of the task instance is obtained by subtracting the start time from the "start" time. The planned duration for this task instance in the final version of the construction instruction package.

[0121] S512, Perform a statistical significance test; to determine whether the calculated bias is systematic, the knowledge graph reverse enhancement module 500 performs a bias index sequence (e.g., ...) on each dataset. Perform a hypothesis test; the null hypothesis of this test. The population mean of this group of deviations is zero, meaning there is no systematic difference between planned and actual values; alternative hypothesis The population mean is not zero.

[0122] The Knowledge Graph Reverse Enhancement Module 500 can use a one-sample t-test to perform this step; first, calculate the t-statistic, the formula of which is: ;in: For all biases in this dataset The sample mean; For all biases in this dataset The sample standard deviation; This represents the number of samples in the dataset, i.e., the total number of times this type of construction task has been executed.

[0123] Subsequently, the knowledge graph reverse enhancement module 500 calculates the corresponding p-value based on the calculated t-statistic and degrees of freedom n−1; compares the p-value with the preset significance level β (e.g., 0.05); if p<β, the null hypothesis is rejected and the bias is determined to be a statistically significant systematic bias.

[0124] S513, perform correlation analysis to explore the root causes of deviations; for each identified systematic deviation, the knowledge graph reverse enhancement module 500 further analyzes its correlation with other environmental variables or contextual factors in the construction process; for example, the knowledge graph reverse enhancement module 500 can calculate the task duration deviation. Average ambient humidity during the mission The Pearson correlation coefficient between them; if the correlation is significant, it indicates that environmental humidity may be an important factor contributing to this systematic bias.

[0125] S514, Generate a systematic deviation record; for each systematic deviation verified through the above steps, the knowledge graph reverse enhancement module 500 generates a structured systematic deviation record; this record includes: the identifier of the affected knowledge entity, the key performance indicator of the deviation, the magnitude of the deviation (e.g., average deviation value), the statistical confidence level (e.g., p-value), and any identified strongly correlated influencing factors. This record will serve as direct input for subsequent generation of knowledge update proposals.

[0126] See attached document Figure 6 , Figure 6 This is a schematic diagram of a human-machine collaborative knowledge graph evolution workflow according to an embodiment of the present invention. After the knowledge graph reverse enhancement module 500 identifies a systematic deviation, in order to ensure the accuracy and reliability of the knowledge graph update, this embodiment of the present invention further provides a human-machine collaborative knowledge graph evolution mechanism.

[0127] This mechanism aims to combine data-driven analysis results with the experiential knowledge of domain experts, enabling secure and traceable updates to the construction process knowledge graph through a closed-loop workflow that includes review and verification steps. The specific implementation of this mechanism may include the following steps:

[0128] S520 automatically generates knowledge update proposals based on identified systematic deviation records; the knowledge graph reverse enhancement module 500 parses each systematic deviation record and transforms it into a structured knowledge update proposal; this proposal is a data object, and its data structure may include:

[0129] Unique identifier for the proposal;

[0130] Target knowledge entity: Identifier of the specific node or edge in the construction technology knowledge graph that is targeted by this update;

[0131] Target attribute: Specifies the specific attribute of the target knowledge entity that needs to be modified, such as the "standard working hours" attribute of a certain process node;

[0132] Current attribute value: The existing value of the target attribute read from the knowledge graph;

[0133] Suggested attribute values: New attribute values ​​calculated based on the quantitative analysis results of systematic deviations; for example, the new standard working hours can be the sum of the original standard working hours and the identified average time deviation.

[0134] Basis for change: Related systematic deviation records, which contain statistical evidence to support this change, such as sample size, mean deviation, p-value, and analysis results of relevant influencing factors.

[0135] S521 pushes the generated knowledge update proposal to the preset review queue and notifies the user with the corresponding review authority; the user, such as the enterprise's process expert or senior project manager, can access the review queue through the management interface provided by the platform.

[0136] S522, the reviewer reviews the knowledge update proposal through the management interface; the management interface presents all the information of the proposal in a visual way, including the numerical comparison before and after the change and the summary of statistical data used as the basis for the change; the reviewer judges the proposal based on his professional knowledge and engineering experience and performs one of the following actions: approve, reject or modify.

[0137] If the reviewer selects "Approve", it means that they fully agree with the suggested attribute values ​​generated by the system.

[0138] If the reviewer selects "Reject", the proposal will be archived and the knowledge graph will remain unchanged. The system will then require the reviewer to fill in the reason for rejection, which will be recorded for future improvements to the deviation identification model.

[0139] If the reviewer selects "Modify", they can enter a more accurate correction value based on the suggested attribute value, and attach the reason for the modification.

[0140] S523, upon receiving the "approval" or "modification" instruction from the reviewer, the system executes an update operation on the construction process knowledge graph. This update operation is initiated by the knowledge graph reverse enhancement module 500 to the knowledge graph management module 100, and follows a preset secure update protocol to ensure the consistency and maintainability of the knowledge base. This protocol may include:

[0141] Version control: Before any knowledge entity is modified, the system automatically creates and archives a version snapshot of the entity's current state;

[0142] Transaction execution: Modification operations on the knowledge graph are executed within a database transaction, ensuring the atomicity of the operation; if any error occurs during the operation, the entire transaction will be rolled back, and the knowledge graph will be restored to its state before the operation.

[0143] Log recording: Every successful update operation is recorded in detail in the system log, including the unique identifier of the proposal, the time of execution of the update, the reviewer of the operation, the attribute value before the change, and the attribute value after the change, forming a complete and traceable change history.

[0144] Through the aforementioned human-machine collaborative workflow, it is ensured that every update of the construction process knowledge graph is based on actual on-site data and has been confirmed by domain experts, thereby achieving continuous and reliable self-evolution of the knowledge base.

Claims

1. A BIM-based integrated platform for architectural interior design and construction, characterized in that, include: The knowledge graph management module is used to build and maintain a construction process knowledge graph; The BIM model parsing module is used to parse the building information model containing building interior design information, extract the physical properties of components, and identify the preset semantic design intent tags in the building information model, and transform the semantic design intent tags into a set of formal engineering constraints. The intelligent construction instruction set generation module is used to generate a final version of the construction instruction package based on the formalized engineering constraints and the construction technology knowledge graph. The on-site execution and feedback module is used to send the final version of the construction instruction package to the on-site terminal equipment to present the construction instructions, and to collect the actual execution data on-site during the construction process; The knowledge graph reverse enhancement module is used to compare and analyze the actual execution data with the planned data in the final version of the construction instruction package to identify systematic deviations, and generate knowledge update content based on the systematic deviations to update the construction process knowledge graph.

2. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, The BIM model parsing module is specifically used for: Detect and read semantic design intent tags attached to building information model components; The intent constraint mapping library configured within the platform is used for querying. The intent constraint mapping library stores the mapping rules from semantic design intent tags to one or more sets of formal engineering constraints. Based on the queried mapping rules, the semantic design intent tags are transformed into a set of formalized engineering constraints.

3. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, The intelligent construction instruction set generation module is specifically used for: The construction planning problem is modeled as a constraint satisfaction problem, and a candidate construction instruction package is generated. Multi-dimensional simulation and pre-run of candidate construction instruction packages are conducted to detect dynamic execution conflicts; The detected dynamic execution conflicts are transformed into new formal constraints, and the constraint satisfaction problem is resolved using the constraint set containing the new formal constraints to generate a modified candidate construction instruction package. Repeat the process of pre-simulation and re-solving until the multi-dimensional simulation pre-simulation no longer detects any new dynamic execution conflicts, and use the candidate construction instruction package generated by the last solution as the final version of the construction instruction package.

4. The integrated platform for architectural interior design and construction based on BIM technology according to claim 3, characterized in that, The multidimensional simulation pre-run includes at least one of the following simulations: Spatiotemporal occupancy simulation is used to detect dynamic spatial conflicts between construction workers or mobile equipment. Critical resource scheduling simulation is used to verify the feasibility of scheduling schemes for exclusive resources. Environmental impact simulation is used to examine whether there are overlapping impacts or conflicts between construction tasks that generate environmental impacts and environmentally sensitive construction tasks.

5. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, The knowledge graph reverse enhancement module is specifically used to identify systematic biases when: The actual execution data and the planned data are aggregated and grouped according to the process entities or material entities associated with the construction tasks; For each preset key performance indicator, calculate the deviation between the planned value and the actual value of each task instance; Perform a statistical significance test on the calculated deviation sequence to determine whether the deviation is a statistically significant systematic deviation.

6. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, The knowledge graph reverse enhancement module, when updating the construction technology knowledge graph, is specifically used for: Based on the identified systematic biases, a knowledge update proposal containing suggested attribute values ​​is automatically generated; The knowledge update proposal is pushed to a pre-set review queue for review by users with the appropriate permissions. Upon receiving user approval or modification instructions, the system will update the construction process knowledge graph.

7. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, The actual execution data collected by the on-site execution and feedback module is multimodal data, which includes: Task status data submitted by construction personnel through on-site terminal equipment; On-site process data in the form of images or videos collected by the built-in cameras of on-site terminal equipment; Environmental data automatically collected by IoT sensing devices deployed at the construction site.

8. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, The construction process knowledge graph includes: A set of nodes representing process entities, material entities, tool entities, quality standard entities, or environmental constraint entities; A set of edges representing semantic relationships between entity nodes, including prerequisite dependencies, material compatibility relationships, or tool requirement relationships.

9. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, Before generating the final version of the construction instruction set, the intelligent generation module is also used for: Identify the relevant knowledge domains required to complete the project and calculate the knowledge entropy of each knowledge domain to quantify the degree of uncertainty in the knowledge domain; When the knowledge entropy of any knowledge domain exceeds a preset entropy threshold, it is determined that the knowledge domain has high uncertainty, and one or more proactive exploration protocols are generated for the knowledge domain. The proactive exploration protocols are field test tasks used to accurately collect decision data.

10. The integrated platform for architectural interior design and construction based on BIM technology according to claim 1, characterized in that, The platform is deployed on a cloud server and interacts with one or more field terminal devices via a network; the field terminal devices are smartphones, tablets, or augmented reality glasses.

Images