Building information multidimensional semantic space and geometric space-time domain bidirectional projection mapping method

Abstract

The application belongs to the technical field of computer aided design, and discloses a building information multi-dimensional semantic space and geometric space-time domain bidirectional projection mapping method, which comprises the following steps: constructing a building information multi-dimensional semantic space, extracting component topology chain and progress node logical dependence weight to generate a topology feature description operator, determining a projection residual vector by comparing real-time and historical operators through bit operation, positioning a topology change node and determining a semantic node influence range according to the projection residual vector, and executing local data projection reorganization by using a node correlation weight model. The application changes the mapping relationship from a static index into a structure perception mechanism, ensures that model evolution maintains semantic logic continuity, avoids the mapping breakage problem caused by the traditional identifier binding mode, accurately identifies and incrementally updates the change area, eliminates the dependence of the system on full-amount data synchronization, and improves the logic convergence capability in large-scale building information processing.

Description

Technical Field

[0001] This invention relates to a method for bidirectional projection mapping of building information in a multidimensional semantic space and a geometric spatiotemporal domain, belonging to the field of computer-aided design technology. Background Technology

[0002] As the core carrier in the field of computer-aided design, Building Information Modeling (BIM) currently carries multi-dimensional heterogeneous data of geometric components and business logic. It establishes the association between three-dimensional geometric entities and schedule plans and cost lists through globally unique identifiers, forming the basis for collaborative management of building projects. The evolution of BIM is accompanied by asynchronous changes in geometric topology and business logic. When model components undergo partial splitting, logical merging, or spatial displacement, the matching method based on static pointers cannot perceive the continuity of spatial relationships of components. This leads to logical breaks in the association relationship when the topological structure of the physical space changes, resulting in a large number of isolated nodes.

[0003] Besides the limitations of static association architecture, there are also shortcomings at the software control level. For example, Chinese invention patent CN115906228A discloses a method for generating building information models. It identifies building objects by comparing two-dimensional digital building plans with three-dimensional scanning data and integrates them into the BIM model. This solution's mapping logic is limited to geometric feature alignment and lacks structural awareness of component topology chains and logical dependencies of progress nodes. Faced with high-frequency deformation of non-standard components in large-scale models or nonlinear disturbances in business logic, it is difficult to maintain semantic logic topology fidelity. It also lacks a state increment protection mechanism, triggering full data reconstruction calculations, resulting in computational redundancy and failing to achieve logical convergence in high-frequency fine-tuning scenarios. To maintain data consistency, existing technologies trigger full reconstruction calculations after data changes. Existing technologies mainly have the following defects: 1. The mapping mechanism lacks topology fidelity capability, resulting in mapping state offset when components deform; 2. The two-way linkage response is lagging, and the synchronous calculation process of large-scale data causes computational redundancy; 3. Due to the lack of a state increment protection mechanism, the system is difficult to achieve logical convergence when handling high-frequency fine-tuning.

[0004] Therefore, the technical problem to be solved by this invention is how to implement a mechanism with topological fidelity characteristics, which eliminates logical deviations in the dynamic collaboration process of heterogeneous data streams through incremental perception and bidirectional distribution of spatial and semantic state changes without relying on full data re-transformation. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A method for bidirectional projection mapping of building information in a multidimensional semantic space and a geometric spatiotemporal domain, comprising the following steps: Step 101: Retrieve the heterogeneous building dataset containing geometric component topology data, progress node timing data and engineering business attributes from the computer-aided design system memory, and construct a multidimensional semantic space for building information in memory. The multidimensional semantic space for building information consists of multiple interconnected semantic nodes. Step 102: parse the geometric constraint topology chain between geometric components and extract the logical dependency weights between progress nodes to generate a topological feature descriptor operator that characterizes the structural features of the heterogeneous building dataset. Step 103: Perform bitwise operations to compare the currently extracted real-time topology feature description operator with the historical topology feature description operator in the version index library. By calculating the difference between the hash components of the current state and the historical state under the preset mapping operator, the projection residual vector representing the deviation of the data structure is obtained. Step 104: Based on the feature dimension index where the magnitude is not zero in the projection residual vector, locate the topology change node in the geometric model that has undergone a state jump, and search the influence range of the semantic node driven by the logic of the topology change node based on topology invariance. Step 105: Using a preset node association weight model, update the association strength of semantic nodes within the influence range of semantic nodes based on the number of path topology steps, and perform local data projection and reorganization according to the updated association strength to realize local incremental data update between the geometric spatiotemporal domain and the multidimensional semantic space of building information. The single calculation cycle of incremental update is limited by the local topology depth within the influence range of semantic nodes.

[0006] Preferably, in step 104, by mapping the projection residual vector to the topological correlation matrix, the target component index corresponding to the non-zero element in the projection residual vector is identified, and the first-order correlation logical node of the target component index in the multidimensional semantic space of building information is retrieved. The target component index and the first-order correlation logical node are jointly defined as the semantic node influence range. When the geometric model undergoes component splitting or merging, the semantic attributes of the new component after splitting or merging are redistributed through the structure perception mechanism of the topological feature description operator.

[0007] Preferably, in step 102, by identifying the start time, end time, and logical predecessor relationship of each task node in the schedule, the schedule is transformed into a directed acyclic graph data structure, and the node degree distribution and path centrality of the directed acyclic graph under each time slice are extracted as the input basis for calculating the topological feature descriptor.

[0008] Preferably, each semantic node in the multidimensional semantic space of building information stores a semantic feature vector, which includes the material physical properties of geometric components, cost quota data, and construction zoning codes.

[0009] Preferably, during the local data incremental update process, if the magnitude of the projection residual vector exceeds the preset reorganization threshold, a global topology verification procedure is initiated to perform logical address reset for conflicting nodes in the multidimensional semantic space of building information.

[0010] Preferably, in step 105, the node association weight model follows the following quantization rules: ,in, The updated weight values, The preset initial weight score, The preset correlation strength attenuation coefficient, The number of topological path steps between nodes within the influence range of a semantic node, and the logical diffusion of the influence range of a semantic node is stopped when the updated weight value is lower than the preset truncation threshold.

[0011] Preferably, performing local incremental data updates includes the following steps: Step 701, writing the change information within the influence range of the semantic node into the 3D spatial index cache table; Step 702, extracting the corresponding attribute descriptors from the 3D spatial index cache table according to the system's view transformation parameters; Step 703, distributing the attribute descriptors to the display buffer and performing a refresh of the rendering state of the corresponding region in the geometric spatiotemporal domain.

[0012] Preferably, in step 102, the geometric constraint topology chain between geometric components is determined by calculating the contact area of ​​the component surfaces, the center distance offset, and the degree of freedom of the connection pair constraints.

[0013] Preferably, after step 105, the method further includes the following steps: Step 901, real-time monitoring of the access frequency of each semantic node in the multidimensional semantic space of building information; Step 902, when the access frequency exceeds... When the time comes, the data of the corresponding node will be stored in the prefetch cache.

[0014] Preferably, the bidirectional projection mapping update maintains the logical temporal sequence of the topological feature description operator records in the version index library, supporting the execution of historical state backtracking based on topological paths in the multidimensional semantic space of building information.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the multidimensional semantic space of building information, topological feature fingerprints extracted from spatial association constraints and key path dependencies are transformed from static indexes to structure-aware association mechanisms. This ensures that the topological continuity at the logical level is maintained within the semantic space when 3D components are split, merged, or nonlinear displacements occur during progress tasks. This avoids the mapping breakage or isolated node problems that are prone to occur in the model evolution process of traditional unique identifier binding methods.

[0016] 2. By using the projection residual vector calculated from the historical state fingerprint recorded by the hash chain list and the current state fingerprint, the system can accurately identify the changing areas of heterogeneous data streams. By incrementally updating only the affected local topology, the system breaks the dependence on full data synchronization in computer-aided design systems, enabling the system to have millisecond-level logical convergence capability when processing large-scale building information data, and effectively reducing computational redundancy in high-frequency fine-tuning scenarios.

[0017] 3. By combining the influence range boundary determined by the correlation strength attenuation model with the asynchronous distribution mechanism of the virtual display state cache pool, the semantic analysis results are projected into three-dimensional space, time axis and business billboard with differentiated projection strategies. The system only sends attribute change instructions to local components with non-zero residuals, avoiding graphics engine display lag caused by full refresh, and ensuring the consistency and real-time response of multi-dimensional visualization carriers under complex linkage interactions. Attached Figure Description

[0018] Figure 1 This is the main flowchart of the bidirectional projection mapping and local incremental update of building information in the multidimensional semantic space and geometric spatiotemporal domain of the present invention. Figure 2 This is a time sequence diagram of the dynamic identification of the influence range of semantic nodes and the update of association strength based on the node association weight model of this invention. Detailed Implementation

[0019] The following detailed description is intended to further illustrate the present invention and is not intended to limit the scope of protection of the present invention. Unless otherwise specified, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0020] A method for bidirectional projection mapping between a multidimensional semantic space and a geometric-spatiotemporal domain of building information includes retrieving a heterogeneous building dataset from the memory of a computer-aided design system. This dataset encompasses topological data of geometric components, time-series data of progress nodes, and engineering business attributes. Three-dimensional components are described using boundary representation data, and construction plans are associated using directed graph logic. Addressing the mapping discontinuity problem caused by the lack of a unified link between heterogeneous data in building lifecycle management, this invention constructs a multidimensional semantic space of building information in memory, consisting of multiple associated semantic nodes. Each semantic node stores a semantic feature vector. ,in, The semantic feature vector contains the material physical properties, cost quota data, and construction zoning codes of the corresponding geometric components. By integrating scattered schedule, cost, and geometric data into a node network with a unified logical starting point, a foundation for the connection between physical space and business space is established. When processing large-scale building information, the spatial constraints between geometric components and the logical dependencies between schedule tasks present a complex coupling situation. To address the technical bottleneck of insufficient structural perception capabilities for high-dimensional datasets, the system parses the topological chains of geometric constraints between geometric components, determining them based on the contact area of ​​the component surfaces, the center distance offset, and the degrees of freedom of the connection pair constraints. Simultaneously, the system identifies the start time, end time, and logical predecessor relationships of each task node in the schedule plan, converting the schedule plan into a directed acyclic graph (DAG) data structure. It extracts the node degree distribution and path centrality of this DAG under each time slice, thereby calculating and generating a topological feature descriptor operator that characterizes the structural features of the heterogeneous building dataset. This operator serves as a digital fingerprint characterizing spatial association constraints and critical path dependencies, used for subsequent incremental state comparisons.

[0021] During large-scale data synchronization, full reconstruction computation leads to computational redundancy. To identify changed areas in the data stream, the system uses a hash chain to record the consistency status of components in each dimension. It performs bitwise operations to compare the currently extracted real-time topology feature description operator with the historical topology feature description operator in the version index. Specifically, it calculates the difference between the hash components of the current and historical states under a preset mapping operator to obtain the projected residual vector representing the deviation of the data structure. ,in, The system uses the projection residual vector as its basis. The feature dimension index with a non-zero mid-model length is used to locate topological change nodes in the geometric model that undergo state transitions. Based on topological invariance, the influence range of semantic nodes driven by the logic of these topological change nodes is searched, and the projection residual vector is used to... Mapping to the topological incidence matrix to identify the projected residual vector The target component index corresponding to the non-zero element is determined, and the first-order association logical node of the target component index in the multidimensional semantic space of building information is retrieved. The target component index and the first-order association logical node are jointly defined as the semantic node influence range; the projection residual vector is calculated. Beforehand, the contact area of ​​the component surface is pre-calculated. and center distance offset Rounding to three decimal places eliminates numerical noise from floating-point arithmetic; the recombination threshold is determined based on the magnitude distribution of the projection residual vector of the historical sample set. Recombination threshold Satisfy the formula ,in, Indicates the recombination threshold. This represents the arithmetic mean of the residual moduli of the test set. It represents the standard deviation.

[0022] To address the potential state drift risk that may arise when semantic analysis results are projected into the concrete space, this invention controls information diffusion through a node association weight model. The system updates the association strength of semantic nodes within the influence range of semantic nodes based on the number of path topology steps. The node association weight model follows the following quantization rules: ,in, The updated weight values, The preset initial weight score, The preset correlation strength attenuation coefficient, The number of topological path steps between nodes within the influence range of a semantic node, when the updated weight value... When the threshold is lower than the preset truncation threshold, the logical diffusion of the semantic node's influence range is stopped. Local data projection and reconstruction are performed based on the updated association strength. Local incremental data updates are then performed between the geometric spatiotemporal domain and the multidimensional semantic space of building information. The single calculation cycle of the incremental update is limited by the local topological depth within the semantic node's influence range. If the projection residual vector... If the modulus exceeds the preset reorganization threshold, the system initiates a global topology verification procedure to perform logical address reset for conflicting nodes in the building information multidimensional semantic space. During the reset process, the system maintains an address mapping buffer with a capacity of 2048 rows, where each row stores the correspondence between the unique identifier of the semantic node and its 32-bit offset starting address in physical memory. When a global conflict is detected, the system calls a linear addressing algorithm to rescan the heterogeneous dataset, and within 10 milliseconds, overwrites the corresponding index row in the buffer with the detected new physical address, completing the address remapping through atomic write instructions to ensure that subsequent rendering instructions can directly jump to the correct data storage area. The global topology verification procedure traverses all semantic nodes in the building information multidimensional semantic space, retrieves the physical memory address mapping table corresponding to the current timestamp in the version index library, and identifies the projection residual vector. Pointing to the failed logical address, the topological feature description operator is regenerated based on the current time. The linear addressing algorithm is used to locate the new starting offset of the corresponding geometric component in memory. The failed logical address is updated to the new starting offset through pointer rewriting operation, thus completing the remapping of the physical association path between semantic nodes and geometric entities.

[0023] To ensure consistency in multi-carrier visualization and avoid stuttering caused by graphics engine refresh, the system asynchronously distributes rendering commands through a 3D spatial index cache table. This includes writing change information within the semantic node's influence range into the 3D spatial index cache table, extracting corresponding attribute descriptors based on the system's view transformation parameters, distributing the attribute descriptors to the display buffer, refreshing the rendering state of the corresponding area in the geometric spatiotemporal domain, and performing bidirectional projection mapping to distribute commands in reverse from the multi-dimensional semantic space of building information to the geometric spatiotemporal domain, extracting the changed semantic feature vector. The construction zoning code or cost quota data is used to retrieve the 3D spatial index cache table through the target component index in the topological feature description operator. The view transformation matrix is ​​used to calculate the bounding box coordinates of the target component in the current view area, and the semantic feature vector is then used. The data is converted into rendering attribute descriptors and distributed to the asynchronous refresh queue of the display buffer. This drives the rendering thread to update the rendering state of the corresponding area in the geometric model. Simultaneously, the access frequency of each semantic node in the multi-dimensional semantic space of the building information is monitored. When the access frequency exceeds [a certain threshold], [the system will take action]. At Hz, the data of the corresponding node is stored in the prefetch cache. Bidirectional projection mapping updates maintain the logical timing of topological feature description operators recorded in the version index, supporting historical state backtracking based on topological paths within the multidimensional semantic space of building information. When processing large-scale graphics rendering tasks, the 3D spatial index cache table uses a hierarchical spatial bounding box data structure to record geometric attribute changes within the influence range of semantic nodes. The system extracts the current view transformation matrix to determine the viewport clipping range, extracts attribute descriptors intersecting with the current viewport from the 3D spatial index cache table, and pushes them into the asynchronous refresh queue of the display buffer. The rendering execution unit performs state updates in a rendering thread independent of logical operations. If the fluctuation rate of the access frequency of a specific semantic node in the multidimensional semantic space of building information exceeds... If this occurs, the system triggers a dynamic capacity expansion procedure for the prefetch cache, using a replacement algorithm based on data access frequency to preload component data in high-frequency interaction areas.

[0024] Example 1: During the detailed construction phase of the steel structure roof of a large hub airport terminal, due to limitations in the on-site lifting capacity, the main truss model was split into four secondary truss components. This change in component topology caused the globally unique identifiers in the original mapping table to become invalid, resulting in a logical break in the relationship between schedule nodes and cost quota items. The system, by parsing the geometric constraint topology chain in the heterogeneous building dataset, extracts a real-time topology feature description operator composed of the degrees of freedom of the component connection pair constraints and the surface contact area of ​​the component. The system performs bitwise operations to compare the extracted real-time topology feature description operator with the historical topology feature description operators in the version index library, and obtains the projected residual vector by calculating the hash component difference. ,in, The projection residual vector; based on the projection residual vector The feature dimension index is used to locate the topology change node affected by the split, and the scope of the semantic node driven by the change is retrieved in the multidimensional semantic space of building information. This scope covers the construction flow section code, steel purchase order and node welding time quota.

[0025] The system uses a node association weight model to evaluate the path topology of semantic nodes within their influence range. The node association weight model follows the following quantification rules: ,in, The updated weight values, The preset initial weight score, The preset correlation strength attenuation coefficient, Define the number of topological path steps between nodes within the influence range of a semantic node; set initial weights. The correlation strength attenuation coefficient is 1.0. The value is 0.5 when the number of topological path steps is... When increasing, the weight value It exhibits exponential decay when the weight value When the weight falls below the truncation threshold of 0.1, the system stops updating the weights. Since the truss splitting action only changes local connectivity, the weight value changes after the topology path steps exceed level 3. Below the truncation threshold, incremental updates are restricted to directly related procurement and hoisting logic nodes. Based on the updated association strength, the system performs local data projection and reorganization, mapping and refreshing the geometric parameters of the four secondary trusses with the start and end times of the progress nodes. The construction partition coding and cost data in the building information multidimensional semantic space are kept consistent with the geometric model after the change. This process solves the mapping breakage problem caused by component splitting through the structural perception mechanism, realizing incremental synchronization between the geometric spatiotemporal domain and the building information multidimensional semantic space.

[0026] Example 2: In a building digital model simulation environment containing 200,000 geometric components and 5,000 progress node tasks, the test platform is equipped with a system that supports floating-point operations and has a video memory bandwidth of not less than [missing information]. The system retrieves boundary representation data from the building information model and the directed acyclic graph structure from the schedule from the computational nodes, and superimposes a signal-to-noise ratio of [value missing] on the information transmission link. Gaussian white noise is used to simulate unstable communication conditions in engineering sites, with a sampling frequency of... The setting is affected by the component change rate. Drive, in which, Sampling frequency, in units of , The component change rate is expressed in units per second. When it is greater than 50, the sampling frequency Determined as Under normal operating conditions in this experiment, the sampling frequency is... Selected as Furthermore, the experimental data originates from publicly available heterogeneous datasets of building engineering standards. The sample group of this invention adopts a bidirectional projection mapping method of building information multidimensional semantic space and geometric spatiotemporal domain, while the control group adopts a full data reconstruction method based on global unique identifier comparison. When the local position offset and topological attribute reorganization of 200 beam and column components occur at the simulated construction site, the system extracts real-time topological feature description operators and performs bit operation comparison.

[0027] Table 1: Number of steps in different topologies The system response parameters below Referring to Table 1, the sample group of this invention calculates the projection residual vector. Lock the changed area, among which, Let the projection residual vector be the correlation strength attenuation coefficient. The updated weight value is given by a weight of 0.5. With the number of steps in the topology path The increase in exhibits an exponential decline, where , The updated weight values, The preset initial weight score, The preset correlation strength attenuation coefficient, The number of topological path steps between nodes within the influence range of a semantic node, when When the weight reaches 4, the weight value By reducing the value to 0.14, the diffusion of local computational load is limited, and the average response time remains at [value missing]. Level, while the out-of-range control group showed that when When the value is below 0.1, the system fails to truncate the association calculations of irrelevant nodes, resulting in a processing time exceeding [a certain threshold]. ,when When the value is set above 0.9, the influence range of semantic nodes is compressed to a single node, causing the mapping accuracy to drop. In the following operating conditions with noise interference, the system uses bitwise operations to extract the hash component difference and projects the residual vector. Unstructured attribute fluctuations were filtered out, and the rate of increase in computation time tended towards a preset reorganization threshold as the density of component changes increased. This is consistent with the physical law that weight components in the node association weight model saturate with the number of path steps. This was achieved by analyzing the semantic feature vectors. Incremental refresh, among which, Representing semantic feature vectors, physical changes in the geometric spatiotemporal domain are incrementally synchronized in real time within the multidimensional semantic space of building information.

[0028] Example 3: This example combines Figures 1 to 2 This section describes a method for bidirectional projection mapping between the multidimensional semantic space and the geometric-spatiotemporal domain of building information, such as... Figure 1 As shown, a heterogeneous building dataset containing geometric component topology, progress node timing, and engineering business attributes is retrieved. After constructing a multidimensional semantic space of building information composed of multiple interconnected semantic nodes, the logical dependency weights of component topology chains and progress nodes are extracted to generate a topological feature description operator. By performing bitwise operations to compare the historical state input of this real-time operator with the historical topological feature description operator recorded in the version index library, the hash component difference is calculated and the difference is output to determine the projection residual vector representing the deviation of the data structure. The topology change node is located based on the non-zero modulus index, and the influence range of the semantic node is determined based on topological invariance. The association strength is updated by combining the constraint input of the node association weight model, thereby performing local data projection and reorganization on the region limited by the local topological depth. Finally, the update is completed to achieve synchronization between the geometric spatiotemporal domain and the semantic space, so that the system achieves a state of structural awareness, logical continuity, and incremental convergence.

[0029] like Figure 2 As shown, the topology change node submits the change node index to the topology association matrix. After identifying the components corresponding to non-zero elements and retrieving the target component index, the matrix returns a first-order association logical node to the influence range controller to initialize the semantic node influence range. The controller requests an association strength assessment from the association weight model. During the traversal of nodes within the influence range, the topology path steps are calculated and a decay coefficient is applied to calculate the weight. If the weight value is higher than the truncation threshold, the system marks it as a valid influence node and expands the search to the next-level adjacent nodes. If the weight value is lower than the truncation threshold, the logical diffusion of the path is stopped to limit the calculation boundary. Finally, the influence range controller outputs the final influence range to the semantic node set and outputs the association strength of each node to the update output module for subsequent local projection reorganization.

[0030] Example 4: During the detailed construction of a large stadium's spatial irregular steel grid shell structure, the number of node connections exceeded 50,000. When adjusting the eccentricity of the ball joints of the support nodes, the static mapping method based on identifiers failed due to the lack of recognition of spatial topological continuity, causing the detailed drawing index to fail. The system retrieved a heterogeneous building dataset containing the grid shell node connection relationships from memory, constructed a multi-dimensional semantic space of building information composed of multiple associated semantic nodes in memory, and performed structure-aware topological feature fingerprint extraction to obtain the target components. Extract the target component from the geometric constraint data. The constraint degree of freedom vector between the connection pairs and adjacent components Calculate the target component Center distance offset between it and its first-order adjacent member and the angle difference of the normal vector of the contact surface By using a nonlinear mapping function, physical quantities are converted into fixed-length hash component strings, generating a real-time topological feature descriptor operator that characterizes the local structural features of the node.

[0031] The system compares the generated real-time topology feature description operator with historical records in the version index, and obtains the projection residual vector characterizing the strength of topology deviation by taking the L2 norm of the XOR result vector of the hash component. ,in, Let the projection residual vector be the projection residual vector; when the projection residual vector is... The modulus is greater than 0 and less than the recombination threshold. When this occurs, the system determines that the change is a local physical parameter drift, based on the projected residual vector. The system uses the location index of the non-zero components to retrieve affected semantic nodes in the topological association matrix, determining the influence range of semantic nodes covering node welding stress parameters and installation deviation limits; the system executes parameter calibration procedures for the density of the reticulated shell structure, and associates the strength attenuation coefficient. The value is determined based on the average node degree within the semantic node's influence range. Confirmed, the calculation formula is as follows: ,in, The correlation strength attenuation coefficient, The preset cutoff threshold, The expected maximum number of logical diffusion steps; under this reticulated shell structure condition, the correlation strength attenuation coefficient is set. The initial weight is 0.45. The system assigns a weight of 1.0 to the updated weight value. Perform local data projection and reconstruction, and synchronize the change in the ball hinge eccentricity to the hoisting safety evaluation vector in the building information multidimensional semantic space. When component replacement occurs at the construction site, resulting in a projection residual vector... The modulus length exceeds the recombination threshold When this happens, the system stops incremental updates, triggers a global topology verification procedure, traverses all semantic nodes in the multidimensional semantic space of building information, performs address reset for nodes that cause logical conflicts, and repairs the data link under extreme change conditions.

[0032] Example 5: During the construction preparation stage of the super high-rise frame core tube structure, in order to build an initial consistency benchmark for the version index library, the system retrieves model data from the computer-aided design system and executes the topology fingerprint pre-set procedure. By traversing all beam-column members and shear wall entities in the model, the system extracts the geometric constraint topology chain of their initial state, calculates the distance matrix between the centroid coordinates of each geometric member and the centroids of adjacent members, and uses the normalized matrix as the input parameter of the discretization mapping operator to generate a historical topology feature description operator. At the same time, in order to establish the initial strength of the associated nodes in the multidimensional semantic space of building information, the system adjusts the attenuation coefficient of the association strength according to the member hierarchical attribute and construction zoning code. Gradient presetting is performed, and the correlation strength attenuation coefficient is determined by calculating the component connectivity distribution within a unit space. The baseline value is used to complete the filing of all topological features before the system officially starts bidirectional mapping linkage.

[0033] Once the architectural model enters the detailed design phase, for local deformation conditions in non-standard floor irregular steel structure areas, the system executes an on-site deployment calibration process to fine-tune the local mapping accuracy and identify the projection residual vector of the current detailed model relative to the filing benchmark version. Based on the projection residual vector The modulus distribution is used to calculate the correlation strength bias correction for the current local region. Correlation strength bias correction amount Follow the calculation procedure below: ,in, This is the correlation strength bias correction amount. The preset calibration gain coefficient, Let L be the norm of the projected residual vector. The average topological degree of semantic nodes within the affected region is determined by adjusting the association strength bias. Compensation to correlation strength attenuation coefficient In the middle, the system completes the setting of the recombination threshold. Physical boundary locking ensures that the delineation of the influence range of semantic nodes within a specific area conforms to engineering logic, and that the geometric spatiotemporal domain and the multidimensional semantic space of building information are logically aligned.

[0034] Example 6: In the initial deployment phase of a large-scale prefabricated hospital building complex system, to determine the initial consistency benchmark of the version index library, the system executes a standardized offline calibration and baseline construction procedure, and reads the initial geometric parameter set of geometric components through the computer-aided design system interface. ,in, Including the surface contact area of ​​the components Center distance offset With connection pair constraint degree vector The system retrieves a preset nonlinear discretization mapping operator to perform a hash transformation on the above parameters, and calculates the hash components corresponding to the topological features between related components using the following formula. : ,in, For the generated hash components, The contact area of ​​the component surface. This is the center distance offset. This is the area weighting coefficient. This is the distance weighting coefficient. The preset hash space modulus is used to convert the physical space quantity into discrete topological fingerprints and store them in the initial slots of the version index.

[0035] After completing the initial baseline filing, in order to determine the critical point that triggers global topology verification, the system performs a reassembly threshold based on the perturbation test set under simulation conditions. The calibration process involves injecting predefined random displacements into local parts of the model and calculating the corresponding projection residual vectors. Modulus distribution, and the recombination threshold is determined based on the principle of statistical significance. Determined as the projected residual vector in the test sample The formula for calculating the sum of the mean and three standard deviations of the modulus is as follows: ,in, The recombination threshold, To test the arithmetic mean of the residual moduli of the test set, The standard deviation is used to determine the recombination threshold. lie in to Within the specified interval, the system determines whether the identified physical deformation falls within the scope of incremental updates. When a drastic topological change occurs due to component type replacement, a logical address reset procedure is triggered to maintain the convergence performance of the multidimensional semantic space of building information in a heterogeneous environment. After completing the initial baseline filing of the hospital building complex, the version index library records a series of topological feature description operator sequences using timestamp indexes. ,in, For the first The topological feature description operator for each version slice, when construction changes involve tracing back to a specific historical node. At that time, the historical operator corresponding to the system index Then, perform a bitwise XOR operation with the current real-time operator to obtain the projected residual vector representing the changed inversion path. Based on the projection residual vector The entry in the topological association matrix points to the semantic node in the multidimensional semantic space of building information, which drives the execution of logical address reset, restoring the association between business attributes and geometric entities to a consistent state at a historical moment.

[0036] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0037] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for bidirectional projection mapping of building information in a multidimensional semantic space and a geometric spatiotemporal domain, characterized in that, Includes the following steps: Step 101: Retrieve the heterogeneous building dataset containing geometric component topology data, progress node timing data and engineering business attributes from the computer-aided design system memory, and construct a multidimensional semantic space for building information in memory. The multidimensional semantic space for building information consists of multiple interconnected semantic nodes. Step 102: parse the geometric constraint topology chain between geometric components and extract the logical dependency weights between progress nodes to generate a topological feature descriptor operator that characterizes the structural features of the heterogeneous building dataset. Step 103: Perform bitwise operations to compare the currently extracted real-time topology feature description operator with the historical topology feature description operator in the version index library. By calculating the difference between the hash components of the current state and the historical state under the preset mapping operator, the projection residual vector representing the deviation of the data structure is obtained. Step 104: Based on the feature dimension index where the magnitude is not zero in the projection residual vector, locate the topology change node in the geometric model that has undergone a state jump, and search the influence range of the semantic node driven by the logic of the topology change node based on topology invariance. Step 105: Using a preset node association weight model, update the association strength of semantic nodes within the influence range of semantic nodes based on the number of path topology steps, and perform local data projection and reorganization according to the updated association strength to realize local incremental data update between the geometric spatiotemporal domain and the multidimensional semantic space of building information. The single calculation cycle of incremental update is limited by the local topology depth within the influence range of semantic nodes.

2. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, In step 104, by mapping the projection residual vector to the topological correlation matrix, the target component index corresponding to the non-zero element in the projection residual vector is identified, and the first-order correlation logical node of the target component index in the multidimensional semantic space of building information is retrieved. The target component index and the first-order correlation logical node are jointly defined as the semantic node influence range. When the geometric model undergoes component splitting or merging, the semantic attributes of the split or merged new components are reassigned through the structure perception mechanism of the topological feature description operator.

3. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, In step 102, by identifying the start time, end time, and logical predecessor relationships of each task node in the schedule, the schedule is transformed into a directed acyclic graph data structure, and the node degree distribution and path centrality of the directed acyclic graph under each time slice are extracted as the input basis for calculating the topological feature descriptor.

4. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, Each semantic node in the multidimensional semantic space of building information stores a semantic feature vector, which contains the material physical properties of geometric components, cost quota data, and construction zoning codes.

5. The bidirectional projection mapping method for multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, During the local data incremental update process, if the magnitude of the projection residual vector exceeds the preset reorganization threshold, the global topology verification procedure is initiated to perform logical address reset for conflicting nodes in the building information multidimensional semantic space.

6. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, In step 105, the node association weight model follows the following quantization rules: ,in, The updated weight values, The preset initial weight score, The preset correlation strength attenuation coefficient, The number of topological path steps between nodes within the influence range of a semantic node, and the logical diffusion of the influence range of a semantic node is stopped when the updated weight value is lower than the preset truncation threshold.

7. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, Performing local incremental data updates includes the following steps: Step 701, writing the change information within the scope of the semantic node into the 3D spatial index cache table; Step 702, extracting the corresponding attribute descriptors from the 3D spatial index cache table according to the system's view transformation parameters; Step 703, distributing the attribute descriptors to the display buffer and performing a refresh of the rendering state of the corresponding region in the geometric spatiotemporal domain.

8. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, In step 102, the geometric constraint topology chain between geometric components is determined by calculating the contact area of ​​the component surfaces, the center distance offset, and the degree of freedom of the connection pair constraints.

9. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, Following step 105, the following steps are also included: Step 901, real-time monitoring of the access frequency of each semantic node in the multidimensional semantic space of building information; Step 902, when the access frequency exceeds... When the time comes, the data of the corresponding node will be stored in the prefetch cache.

10. The method for bidirectional projection mapping of multidimensional semantic space and geometric spatiotemporal domain of building information according to claim 1, characterized in that, Bidirectional projection mapping updates maintain the logical temporal order of topological feature description operator records in the version index, enabling historical state backtracking based on topological paths to be performed in the multidimensional semantic space of building information.

Images