Intelligent deepening method for partition wall

Abstract

The application belongs to the field of building engineering, and particularly relates to an intelligent partition wall deepening method, aiming to solve the problems of poor consistency of results caused by reliance on individual experience of engineers in the manual deepening process, inability to dynamically respond to changes in site process data, and lack of traceability of process constraint logic. The scheme defines a process metadata container for partition wall components in the BIM model, establishes three types of association relationships, namely structural association, mechanical and electrical association, and material attribute. Through graph matching of executable rule graph and BIM semantic graph, automatic verification of process constraint rules is realized. The application can realize millisecond-level dynamic closed loop of partition wall deepening design and site measured data, and significantly improve the accuracy and efficiency of the deepening design of the fabricated building partition wall.

Description

Technical Field

[0001] This application belongs to the field of architectural engineering, and specifically relates to a method for intelligent detailing of partition walls. Background Technology

[0002] In the field of detailed design of prefabricated building partitions, designers typically complete detailed operations such as partition positioning, joist spacing, and pipeline trenching in BIM software based on architectural drawings, structural conditions, and MEP (Mechanical, Electrical, and Plumbing) location diagrams. However, the detailed design model based on manual experience cannot respond in real time to the on-site measured data generated during the construction phase. When there are deviations in the bottom elevation of structural beams, offsets in the routing of MEP cable trays, or changes in the performance parameters of material batches, geometric inconsistencies inevitably arise between the detailed BIM model and the actual construction conditions, causing the detailed design results to lose their accurate guiding value for on-site construction.

[0003] To address the aforementioned issues, common solutions include utilizing the parametric family library function of BIM software for model updates or identifying geometric conflicts using a basic collision detection module. However, existing methods can only perform static verification based on nominal geometric dimensions during the design phase, failing to dynamically map and link on-site measured structural surface deviation meshes, equipment installation coordinates, and material batch performance parameters to the BIM geometric model in real time. Furthermore, existing rule-based verification mechanisms lack the ability to structurally model multi-source process constraint rules, making it difficult to automatically determine whether detailed nodes meet the compliance requirements of mandatory standards such as fire compartmentation, acoustic insulation, and seismic design. These dual challenges make the dynamic closed-loop design of partition walls and on-site measured data a critical problem that urgently needs to be solved in the current field of prefabricated building detailing. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for intelligent wall deepening, which can effectively solve the problems mentioned in the background.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present invention provides a method for intelligent deepening of partition walls, comprising the following steps:

[0007] A process data access middleware is constructed. The middleware is deployed on the BIM collaboration platform server and has dual-mode access capabilities of RESTful API interface and OPC UA protocol. It is used to receive in real time the point cloud data scanned by the total station, the installation space coordinates of electromechanical equipment transmitted back by IoT sensors, and the measured values ​​of the bending strength and thickness tolerance of light steel keel batches pushed by the ERP system.

[0008] Define a process metadata container for partition wall components in the BIM model. The container is embedded in the Shared Parameters of the Revit family file and includes structural association fields, MEP association fields, and material property fields.

[0009] A dynamic mapping engine for process data and geometric model is established. When the structural surface deviation mesh is received, the engine reconstructs the installation reference surface of the top and bottom guide rails of the partition wall based on Delaunay triangulation and drives the partition wall height parameters to be updated adaptively. When the coordinates and error bands of electromechanical equipment are received, the engine generates the equipment space occupant based on the conservative envelope algorithm and triggers the automatic redefinition of the partition wall slotted area. When the batch strength parameters of the keel are received, the engine calls the preset specification formula to recalculate the maximum allowable spacing of the vertical keel in real time and updates the array parameters of the keel instance in the BIM model synchronously.

[0010] All process data-driven geometric changes generate change logs with timestamps, data source identifiers, and change basis clause numbers, which are then written to the blockchain evidence storage module.

[0011] Furthermore, the total station scan point cloud data received by the process data access middleware is a structural surface deviation grid generated after registration, and the resolution of the structural surface deviation grid is a triangular mesh data of 50mm×50mm; the electromechanical equipment installation space coordinates transmitted back by the IoT sensor include XYZ three-axis error bands, and the error band range is ±2mm; the measured value of the bending strength of the light steel keel batch pushed by the ERP system is in MPa, and the thickness tolerance range is ±0.05mm.

[0012] Furthermore, the structural association field in the process metadata container is bound to the unique identifier of the nearest structural beam to establish a spatial association between the partition wall component and the main structure; the electromechanical association field is bound to the unique identifier of the cross cable tray or air duct to establish a spatial avoidance relationship between the partition wall component and the electromechanical equipment; the material attribute field references the ERP material code and corresponding mechanical parameters to establish an attribute association between the partition wall component and the material batch performance.

[0013] Furthermore, when the process data-geometric model dynamic mapping engine reconstructs the installation reference surface of the partition wall top guide rail based on Delaunay triangulation, it first performs triangulation interpolation calculation on the structural surface deviation mesh to obtain the elevation values ​​of each discrete point on the installation surface of the partition wall top guide rail. Then, it generates the spatial equation of the installation reference surface based on the elevation values ​​of each discrete point, thereby driving the partition wall height parameters to be adaptively updated according to the installation reference surface equation. When the engine generates the equipment space occupant based on the conservative envelope algorithm, it generates a spherical envelope with the equipment center point as the sphere center and the sum of the equipment size and the error band as the radius. The spherical envelope is then projected onto the plane where the partition wall is located, triggering the automatic redefinition of the partition wall slotted area to ensure that the slotted boundary and the equipment envelope maintain a preset safe distance.

[0014] Furthermore, when the engine receives the strength parameters of the keel batch, it calls the preset keel spacing calculation formula in the "Technical Specification for Cold-Formed Thin-Walled Steel". This formula takes the measured bending strength of the keel batch as the input parameter and the floor height, wall load and keel section characteristics as auxiliary calculation parameters to calculate the maximum allowable spacing of the vertical keel in real time. The calculation results are then updated to the array parameters of the corresponding keel instance in the BIM model to achieve dynamic matching between the keel spacing and the mechanical properties of the material.

[0015] Furthermore, the blockchain evidence storage module adopts the Hyperledger Fabric consortium blockchain architecture. All process data-driven geometric changes generate change logs that include timestamp fields, data source identifier fields, change basis clause number fields, and parameter comparison fields before and after the change. The timestamp accuracy is in milliseconds. The data source identifier field records the name of the data source system, and the change basis clause number field records the corresponding national standard or industry standard clause number. After the change log is written to the consortium blockchain, a corresponding hash value is generated and written to the Revit model attribute column to ensure that the change record cannot be tampered with.

[0016] Furthermore, the method also includes the step of constructing an executable rule graph: constructing a three-layer rule graph structure, with the top layer being national and industry standard clause nodes, the middle layer being constraint condition nodes, and the bottom layer being geometric parameter nodes; the rule graph is matched with the BIM model semantic graph using the SPARQL query language, and a topological relationship traversal is performed for each partition wall component. When a partition wall end face is matched to a structural beam side elevation, the length of their intersection line and the curvature of their discrete intersection points are automatically extracted, and the rule engine is called to determine whether the sealant application area continuously covers the entire length of the intersection line and meets the width and depth requirements.

[0017] Furthermore, the top-level nodes of the three-layer rule diagram store national and industry standard clause numbers and clause summaries, including the provisions of Article 7.2.5 of the "Code for Acceptance of Quality of Interior Decoration and Renovation Engineering of Residential Buildings" regarding the application of elastic sealant at the junction of partition walls and structures, the provisions of the "Technical Standard for Prefabricated Steel Structure Buildings" regarding the vertical keel spacing not exceeding 600mm and the addition of reinforcing keels at the ends, and the requirements of the building fire protection design code regarding the fire resistance limit of fire-resistant partition walls; the middle-level nodes store constraint condition logical expressions, including the dimensional constraint condition that the sealant width is not less than 10mm and the depth is not less than 8mm when the junction type is rigid connection, the constraint condition comparing the keel spacing calculation result with the maximum allowable spacing, and the constraint condition comparing the fire resistance limit of the partition wall with the standard requirement value; the bottom-level nodes store geometric parameter extraction rules, including the calculation rules for the normal vector of the partition wall end face, the extraction rules for the UV coordinates of the side elevation of the structural beam, and the integral calculation rules for the intersection length.

[0018] Furthermore, the specific implementation of graph matching between the SPARQL query language and the semantic map of the BIM model is as follows: extract the entity objects of all partition wall components in the BIM model, traverse their spatial position attributes to obtain the coordinates of the vertices of the end face quadrilaterals in the world coordinate system; perform a spatial topology query on each end face vertex to return the unique identifiers of all main structural components that have a rigid connection relationship with the partition wall component; for each combination of partition wall components and main structural components, call the calculation kernel to generate the theoretical boundary of the sealant and calculate the Boolean intersection of the two to obtain the polygon of the application area; project the polygon of the application area onto the plane where the intersection line is located, and calculate the minimum bounding rectangle width and depth along the direction of the intersection line; if the width is less than 10mm or the depth is less than 8mm, then highlight the intersection line segment in red in the BIM view and pop up a compliance prompt message in the attribute panel.

[0019] The intelligent partition wall detailing method provided by this invention firstly constructs a process data access middleware and receives total station scan point cloud data, electromechanical equipment installation coordinates transmitted by IoT sensors, and light steel keel batch mechanical parameters pushed by the ERP system in real time through a dual-mode interface of RESTful API and OPC UA. Then, it defines a process metadata container for partition wall components in the BIM model and establishes three types of association relationships: structural association, electromechanical association, and material property association. Further, it establishes a process data-geometric model dynamic mapping engine. When a structural surface deviation mesh is received, the guide rail installation reference surface is reconstructed based on Delaunay triangulation, driving the partition wall height parameters to update adaptively. When electromechanical equipment coordinates are received, a space occupant is generated based on a conservative envelope algorithm, triggering automatic redefinition of the slotted area. When keel batch strength parameters are received, the specification formula is called to recalculate the maximum allowable keel spacing in real time and synchronously update the array parameters. Finally, all geometric changes generate change logs with timestamps, data source identifiers, and change basis clause numbers, and are written to a blockchain storage module. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall technical architecture of the intelligent partition wall deepening system and method according to the embodiments of this application;

[0021] Figure 2 This is a schematic diagram of the core principle framework of the process data-geometric model dynamic mapping engine according to an embodiment of this application;

[0022] Figure 3 This is a logical flowchart of the process data access and three-layer mapping mechanism according to an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of the three-layer node structure of the executable rule graph and the SPARQL graph matching verification framework according to an embodiment of this application;

[0024] Figure 5 A schematic diagram illustrating the framework for constructing a cross-disciplinary process semantic alignment layer and its process conflict tree according to embodiments of this application;

[0025] Figure 6 This is a schematic diagram of the multi-source data flow and inter-module interaction relationship in the intelligent partition wall system according to an embodiment of this application. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0027] Example 1

[0028] This embodiment details the complete technical architecture and workflow of an intelligent partition wall detailing system. Deployed on the BIM collaboration platform server, the system achieves an automated closed loop between partition wall detailing design and on-site measurement data through four core subsystems: real-time access to multi-source process data, semantic process metadata management, dynamic geometric model mapping engine, and blockchain-based evidence storage.

[0029] At the system architecture level, the process data access middleware serves as the data entry hub for the entire system, undertaking the core functions of unified access and standardized conversion. This middleware adopts a microservice architecture deployment mode, with two built-in access channels: a RESTful API interface cluster and an OPC UA protocol gateway. The RESTful API interface cluster follows the standard interface contract defined by the OpenAPI specification, supports HTTPS two-way certificate authentication, and can stably receive structural surface deviation grid data streams from the total station scanning system under high-concurrency scenarios. The OPC UA protocol gateway adopts a Pub-Sub publish-subscribe model, specifically connecting to the coordinate sensor nodes of electromechanical equipment in the IoT sensor network to achieve industrial-grade real-time data push. Internally, the middleware maintains a high-performance message queue cluster to buffer and schedule process data packets from different data sources, ensuring peak smoothing and valley filling of data traffic.

[0030] The standardized conversion module of the process data access middleware is responsible for converting raw process data into a unified data exchange format. This module has three built-in data parsers: a total station scan point cloud data parser, an IoT sensor coordinate data parser, and an ERP system material parameter parser. The total station scan point cloud data parser receives raw point cloud files from Leica or Trimble total stations, spatially aligns them with the BIM design model using a point cloud registration algorithm, and generates a structural surface deviation mesh with a unique batch identifier. The deviation mesh output by the parser is stored using a triangular mesh topology, where each triangular facet records the three-dimensional coordinates of its three vertices in the world coordinate system and the deviation relative to the design elevation. The IoT sensor coordinate data parser subscribes to real-time location messages of electromechanical equipment pushed by the IoT platform via the OPC UA protocol. The message payload follows the sensor data exchange protocol specification and includes information such as the device's unique code, XYZ coordinate values, measurement timestamp, and sensor accuracy level. The ERP system material parameter parser calls the ERP system's material batch query interface through a RESTful API, passing in the light steel keel batch number as the query condition, and receiving the returned mechanical property parameters such as the measured value of bending strength, thickness tolerance, yield strength, and elastic modulus.

[0031] On the BIM model side, the process metadata management module constructs a semantic process metadata container for partition wall components. This container uses the Shared Parameters mechanism of Revit family files as its carrier, encapsulating structural association fields, MEP association fields, and material attribute fields into independent parameter sets through parameter group technology. When each partition wall component is instantiated, the system automatically creates a process metadata container instance for it and establishes spatial association relationships with the nearest structural components. The structural association field stores the unique identifier number of the structural beam or structural wall to which the top and bottom guide rails of the current partition wall component are attached. This number is consistent with the IfcBeam or IfcWall entity identifier in the structural BIM model. The MEP association field uses a linked list data structure to store the identifier numbers of all MEP equipment that have spatial intersections or adjacencies with the current partition wall component. Linked list nodes include equipment type fields, installation coordinate fields, and safety clearance distance fields. The material attribute fields reference the standard material coding system of the ERP system, linking to the corresponding material batch quality certificate and mechanical performance parameter table through the material code index.

[0032] The process data-geometric model dynamic mapping engine is the core engine component for achieving automated detailed design. This engine consists of four functional modules: a data receiver, a rule scheduler, a computational kernel, and a model updater. The data receiver continuously monitors the output queue of the process data access middleware. When a new structural surface deviation mesh data packet is detected, a deviation processing workflow is triggered. The rule scheduler automatically selects the corresponding mapping rule set based on the data packet type identifier and generates processing tasks to distribute to the computational kernel. The computational kernel integrates three core algorithm units: the Delaunay triangulation algorithm module, the conservative envelope algorithm module, and the procedure formula calculation module. The model updater is responsible for batch writing the geometric change parameters output by the computational kernel into the BIM model database and triggering automatic updates of the model view.

[0033] When the data receiver acquires the structural surface deviation mesh data, the deviation processing workflow is initiated. The computational kernel first calls the Delaunay triangulation algorithm module to interpolate and reconstruct the deviation mesh. The core algorithm flow of this module includes: firstly, extracting the vertex coordinates and deviations of all triangular facets in the deviation mesh to construct a Delaunay constrained triangular mesh; then, using the design projection area of ​​the partition wall top guide rail installation surface in the BIM model as the constraint boundary, performing boundary trimming on the triangular mesh; next, performing regular grid point densification interpolation within the trimmed triangular mesh, using the radial basis function interpolation method, with known deviation points as kernel function support points, to calculate the elevation correction of each discrete grid point on the installation surface; finally, performing surface fitting on all densified grid points, and using the least squares method to solve for the spatial plane equation parameters of the installation datum surface. The fitted installation datum surface equation is stored parametrically, including the three-dimensional components of the normal vector and the plane intercept value.

[0034] After reconstructing the installation reference plane, the mapping engine reads the height parameter definition of the current partition wall component, which is stored in the LevelOfDetailParameters parameter group of the partition wall family instance. The engine solves the installation reference plane equation and the bottom reference plane equation of the partition wall simultaneously to calculate the spatial coordinate trajectory of the intersection line between the two sides of the partition wall and the installation reference plane. Based on the extreme difference of the intersection trajectory coordinates, the engine calculates the actual height correction of the partition wall and writes the correction into the height parameter of the partition wall instance. Simultaneously, the engine reads the design thickness parameter of the top guide rail family of the partition wall and the installation gap parameter of the current top guide rail relative to the bottom of the structural beam, and recalculates the thickness of the infill layer between the guide rail and the structural beam based on the corrected partition wall height. If the thickness of the infill layer exceeds a preset threshold, the system automatically determines that the area needs to use high-strength grouting material to replace ordinary mortar, and writes the material replacement suggestion into the annotation parameters of the partition wall component. After the model updater completes the parameter writing, it sends a geometric change event notification to the change log storage module. The event payload includes the unique identifier of the partition wall component, a list of changed parameter names, a comparison of parameter values ​​before and after the change, and the associated structural surface deviation mesh batch number.

[0035] When the data receiver receives the electromechanical equipment installation coordinate data packet, the conservative envelope algorithm module initiates the equipment space occupant generation workflow. The core processing logic of this module is based on the principle of three-dimensional geometric envelope: extracting the three-dimensional coordinates of the equipment center point in the BIM world coordinate system as the sphere's center position, reading the equipment's external dimensions as the base radius, and superimposing the equipment coordinate measurement error band as the safety margin radius to generate a complete spherical envelope. The mathematical expression of the spherical envelope is an implicit surface equation with the equipment center as the origin and the safety radius as the threshold. The module projects this spherical envelope along a direction perpendicular to the partition wall plane onto the geometric plane where the partition wall component is located, generating a projected outline polygon. After extending the boundary of the projected outline polygon outward by a preset safety distance threshold, a Boolean difference operation is performed with the current slotted area definition of the partition wall component. If the Boolean operation results in an intersection area, this intersection area is marked as the new slotted boundary range. The mapping engine reads the slotting parameter definition of the partition wall component, writes the newly calculated slotting boundary coordinates into the slotting area parameter group, and synchronously updates the depth parameters of the slotting area to ensure that the minimum design-specified spacing is maintained between the electromechanical equipment envelope and the slotting boundary. After the model updater completes parameter synchronization, it sends a geometric change event notification to the change log storage module. The event payload includes the unique identifier of the partition wall component, the list of geometric coordinates of the slotting area, the unique equipment code, and the clause number on which the change is based.

[0036] When the data receiver receives the batch strength parameter data package of the light steel keel, the specification formula calculation module initiates the keel spacing recalculation workflow. This module has a pre-set standardized formula set for calculating the vertical keel spacing from the "Technical Specification for Cold-Formed Thin-Walled Steel". The module reads the measured bending strength value of the current batch of light steel keel pushed by the ERP system as the core input parameter of the formula, and extracts the floor design height parameter, wall surface density calculation parameter, and keel cross-sectional geometric characteristic parameter from the BIM model as auxiliary input. The module calls the specification formula to calculate the maximum allowable spacing of the vertical keel in the construction area in real time, and compares the calculation result with the current array spacing parameter of the keel family instance in the BIM model. If the calculation result is less than the current array spacing, the system automatically triggers the keel array parameter update workflow. The model updater reads the array definition parameters of all vertical keel family instances in the partition wall component, rewrites the array spacing parameter to the maximum allowable spacing value, and recalculates the number of keel arrays based on the new spacing value. If no reinforcing keel is set at the end of the keel and the new spacing value triggers the reinforcement setting conditions in the specification, the system automatically adds a reinforcing keel family instance at the end of the keel array. After the parameter update is completed, the model updater sends a geometric change event notification to the change log storage module. The event load includes the unique identifier of the partition wall component, the list of keel family instances, the measured value of batch strength, the calculated allowable spacing, the comparison of array spacing before and after the change, and the referenced clause number.

[0037] The Change Log Evidence Storage Module is responsible for recording and permanently storing all process data-driven geometric change operation logs. This module uses the Hyperledger Fabric consortium blockchain as its underlying evidence storage infrastructure, building a dedicated channel for in-depth change management. The module maintains a change log data structure, including a timestamp field, a data source identifier field, a change-based clause number field, a parameter comparison field before and after the change field, and an associated entity identifier field. The timestamp field uses the ISO 8601 standard format, with millisecond precision, ensuring the time-series traceability of change operations. The data source identifier field records the name of the original data source system that triggered the change, such as a total station scanning system, an IoT sensor network, or an ERP material management system. The change-based clause number field stores the clause number of the national or industry standard upon which the change is based, following the standard code plus clause number format. The parameter comparison field before and after the change records the original and new values ​​of each changed parameter in JSON format. The associated entity identifier field stores a unique list of BIM model components involved in the change.

[0038] Upon receiving a geometric change event notification, the change log notarization module first constructs a complete change log data structure locally, then calls the Fabric SDK to write the log data to the consortium blockchain channel. The write operation employs a three-node consensus mechanism to ensure data consistency; upon successful consensus, the Fabric network returns a transaction number hash value. The module converts this hash value into a readable format and writes it to a dedicated parameter field in the Revit model's attribute panel for the corresponding partition wall component, achieving a two-way binding between the BIM model and the blockchain notarization record. Any attempt to tamper with the notarization record will result in hash verification failure, ensuring the auditability and immutability of the detailing process.

[0039] The executable rule graph verification mechanism, as an important extended function of the system, is embedded in the collaborative verification process of the BIM model. This mechanism constructs a three-layer rule graph structure, achieving structured modeling and automatic verification of national and industry standard clauses through semantic rule expressions. The top-level rule nodes store the standardized expressions of national and industry standard clauses. Each clause is parsed into a node object containing a clause number, clause name, clause summary, and mandatory attribute identifier. The clause number follows the national standard code specification, and the mandatory attribute identifier uses a Boolean type to distinguish between mandatory and non-mandatory clauses. The top-level nodes and the middle-level constraint nodes are connected by "deductive" relationship edges, which record the logical derivation path from clause provisions to constraint conditions.

[0040] The mid-level constraint nodes store the specific constraint logic expressions derived from the top-level clauses. Each constraint expression has a condition variable on the left and a condition threshold on the right; the entire expression is encapsulated as an executable decision function. Mid-level nodes are divided into three subclasses based on the nature of the constraints: dimensional constraints, condition comparisons, and compliance judgments. The dimensional constraint subclass defines the comparison relationship between two-dimensional or three-dimensional dimensional parameters and thresholds, such as when the joint type is equal to a rigid connection, the sealant application width is greater than or equal to a set number of millimeters and the depth is greater than or equal to a set number of millimeters. The condition comparison subclass defines the inclusion relationship between the calculation result and the allowable extreme value, such as when the calculated keel spacing is less than or equal to the maximum allowable spacing in the code. The compliance judgment subclass defines composite judgment conditions, such as when the fire resistance rating of the partition wall is greater than or equal to the fire resistance rating required by the fire compartment design and the acoustic sound insulation is greater than or equal to the design sound insulation threshold. Mid-level nodes are connected to the bottom-level geometric parameter nodes through "reference" relationships, which record the source path of the geometric parameters required for the constraint conditions.

[0041] The bottom-level geometric parameter nodes store the specific rules and calculation methods for extracting geometric parameters from the semantic graph of the BIM model. Each bottom-level node defines a parameter extraction function, which, by inputting the component identifier and parameter name, calls the BIM model API interface to obtain the corresponding geometric or semantic data. The bottom-level nodes are divided into three subcategories based on parameter type: spatial location, geometric topology, and dimensional measurement. Spatial location nodes define the rules for extracting the coordinates of components in the world coordinate system, such as the world coordinate extraction function for the vertices of the quadrilateral on the end face of a partition wall. Geometric topology nodes define the rules for determining and querying the spatial relationships between components, such as the spatial query syntax for the adjacency relationship between the end face of a partition wall and the side elevation of a structural beam. Dimensional measurement nodes define dimensional measurement rules based on the geometric calculation engine, such as the integral algorithm for calculating the intersection length based on BRep topology.

[0042] The executable rule graph performs semantic-level graph matching operations with the BIM model semantic graph using the SPARQL query language. The semantic graph uses IFC4.3 Schema as the ontology model, exporting the Revit BIM model into a semantic graph format. Each building component is mapped to an entity node in the graph, and the spatial and semantic relationships between components are mapped to edges between entity nodes. The SPARQL query interface of the rule graph receives the identifier of the partition wall component to be verified as an input parameter. First, it generates a top-level clause traversal query statement, locating all national and industry standard clause nodes related to the partition wall component in the rule graph. Then, based on the deductive relationship edges between clause nodes and constraint nodes, it traverses and obtains all associated constraint expressions. Finally, based on the geometric parameter nodes referenced by the constraint expressions, it dynamically generates a set of parameter extraction query statements, and obtains the actual parameter values ​​from the BIM semantic graph through graph matching operations.

[0043] In the scenario of verifying the continuity and dimensional compliance of the elastic sealant application area, the rule graph matching engine executes the following workflow: The engine first generates an IfcWall entity traversal query statement, extracting all entity objects and their spatial location attributes of partition wall components from the BIM semantic graph, and obtaining a list of vertex coordinates of the quadrilateral end face of each partition wall component in the world coordinate system. Then, a spatial topology query is performed on the end face vertices of each partition wall component. The query statement traverses all main structural components that have a geometrical adjacency relationship with the end face of that partition wall through the IfcConnectedFaces interface. The topology query results are filtered by type, retaining only structural beam entities of type IfcBeam, and obtaining the unique identifier number of all structural beams that have a rigid connection relationship with the partition wall.

[0044] For each combination of partition wall components and structural beam components, the engine invokes the OpenCASCADE geometry calculation kernel to execute a polygon generation algorithm for the sealant application area. The algorithm first performs a BRep offset operation on the partition wall end face, with an offset distance equal to half the designed sealant thickness, generating the theoretical boundary surface of the sealant on the partition wall side. Then, it performs a BRep offset operation of the same distance on the structural beam side elevation, generating the theoretical boundary surface of the sealant on the structural beam side. Finally, it performs a Boolean intersection operation on the two offset surfaces, and projects the boundary line of the intersection result onto the plane containing the partition wall end face, generating the polygon for the sealant application area.

[0045] The engine projects the polygon of the application area onto the plane containing the intersection line and calculates the minimum enclosing rectangle of the polygon along the intersection line direction. The width of the minimum enclosing rectangle is defined as the maximum projected length along the intersection line direction, and the depth is defined as the projected width perpendicular to the intersection line direction. The engine compares the calculated width and depth values ​​with the size thresholds defined in the middle-level constraint nodes. If either size is less than the threshold, a non-compliance judgment result is generated. The non-compliance judgment result includes the unique identifier of the partition wall component, the unique identifier of the structural beam, the intersection line segment identifier, the measured width value, the measured depth value, the non-compliant clause number, and the compliance prompt text. The engine pushes the non-compliance judgment result to the BIM model view rendering engine, highlights the intersection line segment in red in the 3D view, and pops up a compliance prompt dialog box in the component properties panel.

[0046] The cross-disciplinary process semantic alignment layer serves as an auxiliary analysis function of the system, constructing a process conflict tree based on three types of semantic tags: installation stage, tolerance type, and process dependency. This semantic alignment layer builds a complete causal traceability graph for each partition wall detailing issue. The graph's nodes include three categories: problem phenomenon nodes, cause classification nodes, and data source nodes. Edges include three categories: direct causal edges, indirect causal edges, and related evidence edges. Problem phenomenon nodes record various conflict events detected during the detailing process, such as grooved areas intruding into the keel web, discontinuous sealant application, or keel spacing exceeding allowable values. Cause classification nodes are organized into three levels: direct cause, indirect cause, and root cause. Direct cause nodes record process operations or data deviations that directly lead to the problem; indirect cause nodes record upstream factors affecting the direct cause; and root cause nodes record systemic design or management defects. Data source nodes record all traceable data identifiers related to the problem, including original measurement data identifiers, material batch data identifiers, and design change record identifiers.

[0047] In a specific application example, a prefabricated affordable housing project is used to illustrate the complete workflow of the system. The project, located in a district of a city, comprises three residential buildings with a steel frame-braced structure. The interior partition walls utilize an ALC panel and light steel keel composite system. During the third-floor construction phase, the total station scanning system performed a structural surface deviation measurement scan on the east side of the corridor in Area B. The scan results showed that the measured elevation of the bottom of the structural beams in this area was generally lower than the elevation in the BIM design model, with a maximum deviation of 18 millimeters. The process data access middleware receives the deviation mesh data package from the total station scanning system via a RESTful API interface. After standardization and conversion, it generates a structural surface deviation mesh file with batch identification and pushes it to the mapping engine.

[0048] The mapping engine's data receiver identifies the data packet type as a structural surface deviation mesh, triggering the deviation processing workflow. The Delaunay triangulation algorithm module of the computation kernel loads the deviation mesh file, extracts the vertex coordinates and deviation data of all triangular faces in the mesh, and performs boundary clipping using the design projection area of ​​the partition wall GW-03-12 as the constraint boundary. Within the clipped area, the module performs regular grid point densification interpolation at 50 mm intervals, and uses the radial basis function interpolation method to calculate the correction amount of each densified grid point relative to the design elevation. Finally, the spatial plane equation parameters of the top guide rail installation reference surface in this area are generated through least squares surface fitting.

[0049] The mapping engine reads the current height parameter design value of partition wall GW-03-12 and calculates the spatial coordinates of the intersection lines between the two facades and the reference plane based on the installation reference plane equation. The extreme value difference of the Z-axis component of the intersection line coordinates is calculated to be 32 mm, which is an 18 mm correction compared to the original design height parameter of 50 mm. The engine simultaneously calculates the thickness of the infill layer between the top guide rail and the bottom surface of structural beam BL-03-07. The original design thickness was 15 mm, and the corrected thickness is 33 mm. The engine determines that the infill layer thickness exceeds the applicable range of ordinary mortar and indicates in the partition wall component annotation parameters that high-strength grout with a strength grade of not less than C40 should be used to replace ordinary mortar. The model updater batch writes the corrected height parameters and infill layer thickness parameters into the BIM model database, completing parameter synchronization.

[0050] The mapping engine sends a geometric change event notification to the change log storage module. The event payload includes the partition wall component identifier GW-03-12, the changed parameter names (height and infill layer thickness), a comparison of parameter values ​​before and after the change, the associated structural surface deviation mesh batch identifier BSCAN-03-B-EAST, and the change-based clause number GB50210-2018, Article 6.3.4. The change log storage module constructs a complete change log record, writes it to the Hyperledger Fabric consortium blockchain, obtains the transaction hash value, and writes the hash value to the Revit model attribute column of partition wall GW-03-12. The entire processing flow is completed within seconds, achieving a millisecond-level dynamic closed loop between detailed design and on-site measurement data.

[0051] Example 2

[0052] This embodiment describes an intelligent method for detailed processing of partition walls in areas with dense electromechanical pipelines. This method expands the electromechanical equipment processing capabilities of the process data-geometric model dynamic mapping engine in Embodiment 1, adding multi-device collaborative envelope calculation and automatic conflict resolution functions.

[0053] In areas with dense electromechanical equipment, calculating the spatial footprint of a single device may not meet the comprehensive avoidance requirements of complex pipeline layouts. This solution proposes a multi-device collaborative envelope algorithm based on spatial segmentation and weighted merging. The core processing flow of this algorithm includes: first, extracting the installation coordinates and external dimensions of all electromechanical equipment within the processing area to construct an equipment list. Then, generating an independent spherical envelope for each device in the list, with the device center as the sphere and the sum of the device size and error band as the radius. Next, using an axial segmentation strategy, spatially partitioning all spherical envelopes, with the segmentation plane perpendicular to the partition wall plane, dividing the equipment group into two groups: a forward equipment group and a backward equipment group. Finally, performing a Boolean union operation on all spherical envelopes in the forward equipment group to generate a merged forward comprehensive envelope; performing the same Boolean union operation on the backward equipment group to generate a merged backward comprehensive envelope.

[0054] After the merged composite envelope is projected onto the partition wall plane, an irregularly shaped composite projection profile is generated. This profile, after convex hull simplification, is then subjected to a Boolean difference operation with the current slotted area definition of the partition wall. If the Boolean operation results in an intersection, the system automatically calculates the geometric center point and major axis direction of this intersection region, generating an optimized list of slotted area boundary coordinates. The optimization algorithm minimizes the slotted area to reduce the impact on the overall structural performance of the partition wall while ensuring safe clearance for mechanical and electrical equipment.

[0055] In a specific application scenario, consider the electromechanical pipe gallery area on the basement floor of a commercial complex project. The partition walls in this area need to traverse multiple parallel fire sprinkler main pipes, ventilation ducts, and electrical cable trays, with various equipment positioned close together. The IoT sensor network transmits the installation coordinates and error band data of each device in real time. Process data is received by the middleware, standardized, and then pushed to the mapping engine.

[0056] The multi-device collaborative envelope algorithm module of the mapping engine loads the entire equipment list, totaling twelve fire sprinkler pipes, four ventilation ducts, and eight electrical cable trays. The module first generates an independent cylindrical or cuboid basic envelope for each pipe or cable tray, then superimposes error bands to generate an extended envelope. Next, a horizontal segmentation strategy is used to divide the equipment group along the pipe gallery into multiple processing sections, each three meters long, to ensure a balance between computational accuracy and performance. Boolean union operations are performed on the extended envelopes of the equipment within each processing section to generate a section-level comprehensive envelope. These section-level comprehensive envelopes are then sequentially spliced ​​along the pipe gallery direction to generate a complete comprehensive envelope corridor.

[0057] After the corridor envelope is projected onto the partition wall plane, the final slotted area boundary is generated through convex hull simplification and safety distance expansion. The mapping engine automatically adjusts the boundary coordinate list, depth parameters, and reinforcing rib parameters in the partition wall slotting parameter group. The slotting depth is determined based on the maximum protrusion of the equipment, and the reinforcing ribs are evenly distributed along the long side of the slot to ensure the integrity of the partition wall structure. After the model updater completes parameter synchronization, it sends a geometric change event notification to the change log storage module. The event payload includes the unique identifier of the partition wall component, the geometric parameters of the slotted area, the equipment list summary data, and the clause number on which the change is based.

[0058] An executable rule-based graph verification mechanism automatically verifies the compliance of the detailed design results. The rule-based graph matching engine extracts the geometric parameters of the partition wall slotted area and their spatial relationship with adjacent joists, and calls the compliance judgment module to verify the compliance of fireproof sealing nodes. The verification includes whether the selection of fireproof sealing materials in the slotted area meets the fire resistance limit requirements of the fire compartment, whether the width-to-depth ratio of the slot meets the specifications of the construction method atlas, and whether the spacing of the reinforcing ribs meets the structural calculation requirements. The verification results are presented visually in the BIM view, with non-compliant nodes highlighted in yellow with compliance rectification suggestions.

[0059] Example 3

[0060] This embodiment describes a dynamic adaptation scheme for material batch performance in a light steel keel partition wall system. This scheme expands the intelligence level of the keel spacing recalculation workflow in Embodiment 1, and adds risk assessment and preventive adjustment functions based on the material batch variation coefficient.

[0061] In light steel keel partition wall systems, the measured bending strength of the same batch of light steel keels exhibits a certain degree of dispersion. Traditional detailed design methods use standard values ​​or batch averages to calculate keel spacing, which may lead to insufficient load-bearing capacity of the partition wall in extreme cases where the bending strength is too low. This solution proposes a preventive adjustment algorithm based on the material batch variation coefficient. This algorithm introduces a safety margin coefficient based on the calculation results of the specification formula to achieve preventive conservative adjustment of the keel spacing.

[0062] The core processing flow of this algorithm includes: First, receiving the complete set of mechanical performance parameters for the current batch of light steel keel from the ERP system, including the measured average value of bending strength, the measured standard deviation of bending strength, the thickness tolerance range, and the mean value of elastic modulus. Then, calculating the coefficient of variation for this batch of materials, defined as the ratio of the standard deviation to the mean, used to characterize the dispersion of material properties. Next, obtaining the corresponding safety margin coefficient from a table based on the coefficient of variation. The safety margin coefficient increases with the increase of the coefficient of variation, ensuring that more conservative design parameters are used when the material properties have large dispersion.

[0063] After obtaining the safety margin coefficient, the algorithm performs a weighted calculation with the result of the standard formula to generate the maximum allowable spacing after preventative adjustments. The adjustment strategy adopts a reduction method, that is, the maximum allowable spacing is equal to the value calculated by the standard formula multiplied by the safety margin coefficient. If the adjusted maximum allowable spacing conflicts with the keel array spacing in the current BIM model, the system automatically triggers the keel array parameter update workflow and rearranges the keel array according to the adjusted maximum allowable spacing. If, during the update process, it is found that the number of added keels causes the total width of the partition wall to exceed the design limit, the system automatically evaluates and proposes alternative solutions such as adding horizontal reinforcing keels or adjusting the thickness of the partition wall for the designer to decide.

[0064] In a specific application scenario, let's take a residential project as an example. This project uses a light steel keel and gypsum board interior partition wall system. The keel material is the same batch of C-type keel supplied by the same supplier. When the third batch of keel material arrived, the batch parameters pushed by the ERP system showed that the measured average bending strength of this batch of keel was 235 MPa, the standard deviation was 12 MPa, and the coefficient of variation was approximately 5.1%. The system looked up the safety margin coefficient of 0.95 based on the coefficient of variation.

[0065] The mapping engine's formula calculation module uses 235 MPa as input parameters and calculates based on a floor height of 3 meters, a wall surface density of 25 kg / m², and a C75 keel cross-section, yielding a maximum permissible keel spacing of 450 mm. A preventative adjustment algorithm multiplies this value by a safety margin factor of 0.95, resulting in an adjusted maximum permissible spacing of 427.5 mm. The system automatically adjusts the keel array spacing parameter from 450 mm to 420 mm and notes "Preventative adjustment considering material batch dispersion" in the partition wall component annotation.

[0066] After the model updater completes parameter synchronization, it sends a geometric change event notification to the change log storage module. The change log records detailed batch material parameters, coefficient of variation, safety margin coefficient, calculated values ​​from regulations, adjusted calculated values, and referenced clause numbers, ensuring the traceability and compliance of design changes.

[0067] An executable rule-based graph verification mechanism automatically verifies the compliance of the detailed design results. The rule-based graph matching engine reads the adjusted keel array spacing parameters and calls the compliance judgment module to verify the spacing compliance. The verification includes whether the adjusted spacing meets the maximum spacing requirements allowed by the regulations, whether the end reinforcement keel settings meet the specifications, and whether changes in the number of keels affect the overall stability calculation results of the partition wall. After the verification passes, the system generates a verification pass report and writes it to the change log, completing the entire process of material-driven detailed design adjustments for this batch.

[0068] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the technical solutions of the present invention are not limited to the specific embodiments described above; the present invention covers all technical solutions formed by novel concepts proposed within its spirit and substance. The scope of protection of the present invention is determined by the contents of the claims.

Claims

1. A method for intelligent detailing of partition walls, characterized in that... Includes the following steps: A process data access middleware is constructed. The middleware is deployed on the BIM collaboration platform server and has dual-mode access capabilities of RESTful API interface and OPC UA protocol. It is used to receive in real time the point cloud data scanned by the total station, the installation space coordinates of electromechanical equipment transmitted back by IoT sensors, and the measured values ​​of the bending strength and thickness tolerance of light steel keel batches pushed by the ERP system. Define a process metadata container for partition wall components in the BIM model. The container is embedded in the SharedParameters of the Revit family file and includes structural association fields, MEP association fields, and material property fields. A dynamic mapping engine for process data and geometric model is established. When the structural surface deviation mesh is received, the engine reconstructs the installation reference surface of the top and bottom guide rails of the partition wall based on Delaunay triangulation and drives the partition wall height parameters to be updated adaptively. When the coordinates and error bands of electromechanical equipment are received, the engine generates the equipment space occupant based on the conservative envelope algorithm and triggers the automatic redefinition of the partition wall slotted area. When the batch strength parameters of the keel are received, the engine calls the preset specification formula to recalculate the maximum allowable spacing of the vertical keel in real time and updates the array parameters of the keel instance in the BIM model synchronously. All process data-driven geometric changes generate change logs with timestamps, data source identifiers, and change basis clause numbers, which are then written to the blockchain evidence storage module.

2. The intelligent partition wall deepening method according to claim 1, characterized in that, The process data access middleware receives total station scan point cloud data which is a structural surface deviation grid generated after registration. The structural surface deviation grid has a resolution of 50mm×50mm triangular mesh data. The electromechanical equipment installation space coordinates transmitted by the IoT sensor include XYZ three-axis error bands, with an error band range of ±2mm. The measured value of the bending strength of the light steel keel batch pushed by the ERP system is in MPa, and the thickness tolerance range is ±0.05mm.

3. The intelligent partition wall deepening method according to claim 1, characterized in that, The structural association field in the process metadata container is bound to the unique identifier of the nearest structural beam to establish the spatial association between the partition wall component and the main structure; the electromechanical association field is bound to the unique identifier of the cross cable tray or air duct to establish the spatial avoidance relationship between the partition wall component and the electromechanical equipment; the material attribute field references the ERP material code and the corresponding mechanical parameters to establish the attribute association between the partition wall component and the material batch performance.

4. The intelligent partition wall deepening method according to claim 3, characterized in that, When the process data-geometric model dynamic mapping engine reconstructs the installation reference surface of the partition wall top guide rail based on Delaunay triangulation, it first performs triangulation interpolation calculation on the structural surface deviation mesh to obtain the elevation values ​​of each discrete point on the installation surface of the partition wall top guide rail. Then, it fits the elevation values ​​of each discrete point to generate the spatial equation of the installation reference surface, thereby driving the partition wall height parameters to be adaptively updated according to the installation reference surface equation. When the engine generates the equipment space occupant based on the conservative envelope algorithm, it generates a spherical envelope with the equipment center point as the sphere center and the sum of the equipment size and the error band as the radius. The spherical envelope is then projected onto the plane where the partition wall is located, triggering the automatic redefinition of the partition wall slotted area to ensure that the slotted boundary and the equipment envelope maintain a preset safe distance.

5. The intelligent partition wall deepening method according to claim 1, characterized in that, When the engine receives the strength parameters of the keel batch, it calls the keel spacing calculation formula in the "Technical Specification for Cold-Formed Thin-Walled Steel". This formula takes the measured bending strength of the keel batch as the input parameter and the floor height, wall load and keel section characteristics as auxiliary calculation parameters to calculate the maximum allowable spacing of the vertical keel in real time. The calculation results are then updated to the array parameters of the corresponding keel instance in the BIM model to achieve dynamic matching between the keel spacing and the mechanical properties of the material.

6. The intelligent partition wall deepening method according to claim 1, characterized in that, The blockchain evidence storage module adopts the Hyperledger Fabric consortium blockchain architecture. All process data-driven geometric changes generate change logs that include timestamp fields, data source identifier fields, change basis clause number fields, and parameter comparison fields before and after the change. The timestamp accuracy is in milliseconds. The data source identifier field records the name of the data source system, and the change basis clause number field records the corresponding national or industry standard clause number. After the change log is written to the consortium blockchain, a corresponding hash value is generated and written to the Revit model attribute column to ensure that the change record cannot be tampered with.

7. The intelligent partition wall deepening method according to claim 1, characterized in that, The method also includes the step of constructing an executable rule graph: constructing a three-layer rule graph structure, with the top layer being national and industry standard clause nodes, the middle layer being constraint condition nodes, and the bottom layer being geometric parameter nodes; The rule graph uses the SPARQL query language to perform graph matching with the semantic graph of the BIM model. It performs topological relationship traversal for each partition wall component. When the partition wall end face is matched to the side elevation of the structural beam, it automatically extracts the length of the intersection line and the curvature of the discrete points of the intersection line, and calls the rule engine to determine whether the sealant application area continuously covers the entire length of the intersection line and meets the width and depth size requirements.

8. The intelligent partition wall deepening method according to claim 1, characterized in that, The top-level nodes of the three-layer rule diagram store national and industry standard clause numbers and clause summaries, including Article 7.2.5 of the "Code for Acceptance of Quality of Interior Decoration and Renovation Engineering of Residential Buildings" which stipulates that elastic sealant should be installed at the junction of partition walls and structures; the "Technical Standard for Prefabricated Steel Structure Buildings" which stipulates that the vertical keel spacing should not exceed 600mm and that reinforcing keels should be added at the ends; and the requirements for the fire resistance limit of fire-resistant partition walls in the Code for Fire Protection Design of Buildings. The middle-level nodes store constraint condition logical expressions, including the dimensional constraint that the sealant width should not be less than 10mm and the depth should not be less than 8mm when the junction type is rigid connection; the constraint that compares the calculated keel spacing with the maximum allowable spacing; and the constraint that compares the fire resistance limit of the partition wall with the standard requirement value. The rules for extracting geometric parameters stored in the underlying nodes include the rules for calculating the normal vector of the partition wall end face, the rules for extracting the UV coordinates of the side elevation of the structural beam, and the rules for calculating the integral length of the intersection line.

9. The intelligent partition wall deepening method according to claim 1, characterized in that, The specific implementation method for graph matching between the SPARQL query language and the semantic map of the BIM model is as follows: Extract the entity objects of all partition wall components in the BIM model, traverse their spatial position attributes to obtain the coordinates of the vertices of the end face quadrilaterals in the world coordinate system; perform a spatial topology query on each end face vertex to return the unique identifiers of all main structural components that have a rigid connection relationship with the partition wall component; for each combination of partition wall components and main structural components, call the calculation kernel to generate the theoretical boundary of the sealant and calculate the Boolean intersection of the two to obtain the polygon of the application area; project the polygon of the application area onto the plane where the intersection line is located, and calculate the width and depth of the minimum enclosing rectangle along the direction of the intersection line; if the width is less than 10mm or the depth is less than 8mm, highlight the intersection line segment in red in the BIM view and pop up a compliance prompt message in the attribute panel.

Images