Engineering project management digital model generation method and device, equipment and medium

Abstract

The invention provides a method, a device, equipment and a medium for generating an engineering project management digital model, and the method comprises the steps: collecting original data related to engineering project features, carrying out the preprocessing, forming a standardized project data set, constructing a space-time reference frame based on the data set, fusing a multi-dimensional feature tensor of multiple engineering features, and carrying out the construction of a space-time reference frame. The method comprises the steps of generating a project progress node and an integrated feature tensor, fusing the project progress node and the integrated feature tensor to generate a multi-source association graph containing entity relationships and feature semantics, generating a verified engineering project management digital model according to the project progress node, the integrated feature tensor and the association graph, and dynamically updating and optimizing the model in response to change data in project execution. According to the method, the problems of data dispersion, lack of association and static stiffness of the model in a traditional method are solved, automatic construction and dynamic evolution from multi-source heterogeneous data to an intelligent decision-making model are realized, and the accuracy, real-time performance and intelligent level of project management are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of digital management technology, and in particular to a method, apparatus, equipment and medium for generating a digital model for engineering project management. Background Technology

[0002] In traditional construction project management, management decisions heavily rely on scattered drawings and documents, isolated progress reports, and static, experience-based plans. Project data such as structural parameters, material information, labor allocation, machinery scheduling, and technological processes are typically stored in contracts, drawings, plans, and tables of varying formats, lacking unified standards and effective means of integration. This results in fragmented data, forming "information silos" that fail to reflect the overall project picture.

[0003] More importantly, traditional management methods are difficult to effectively integrate time and space dimensions, and cannot accurately locate, analyze and dynamically predict the construction process in time and space. The management models built are often static and lagging, and cannot respond to real-time changes on the construction site such as design changes, schedule delays and resource adjustments. This results in inherent defects in project management such as low transparency, poor predictability and slow regulation, which seriously restricts the improvement of efficiency and quality in the management of large and complex engineering projects. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, and medium for generating a digital model for engineering project management, in order to solve the problem of inefficiency in project management caused by data silos and static models in the prior art.

[0005] This application provides a method for generating a digital model for engineering project management. The method includes: acquiring project management requirements and collecting raw data related to engineering project characteristics; preprocessing the raw data to form a standardized project dataset; constructing a spatiotemporal reference frame based on the standardized project dataset to identify the spatiotemporal location and constraint rules of the data; performing multimodal data processing and feature fusion on the standardized project dataset to generate a multidimensional feature tensor that integrates multiple engineering features; generating a multi-source association graph containing entity relationships and feature semantics based on the spatiotemporal reference frame and the multi-dimensional feature tensor; integrating the multi-dimensional feature tensor and the multi-source association graph according to project progress nodes to generate a validated digital model for engineering project management; updating and optimizing the digital model for engineering project management in response to changes in data during project execution, and outputting the optimized digital model for engineering project management.

[0006] In one embodiment of the present invention, a spatiotemporal reference framework for identifying the spatiotemporal location and constraint rules of data is constructed, comprising: extracting data from the standardized project dataset to obtain spatial object information and time information; assigning a unique identifier to each object based on the spatial object information, establishing a mapping relationship between the data object and the spatial location, wherein the data object is determined based on the structural components, construction area, or equipment location information in the standardized project dataset; associating the time information with the spatial object to establish a spatiotemporal association relationship for recording the state of the spatial object at different time points; defining spatiotemporal constraint rules according to engineering specifications and construction plans, and integrating the mapping relationship, spatiotemporal association relationship, and constraint rules to form a spatiotemporal reference framework with the same data structure, wherein the engineering specifications and construction plans are obtained based on industry standards, project design documents, or historical construction data.

[0007] In one embodiment of the present invention, generating a multidimensional feature tensor that integrates multiple engineering features includes: identifying and transforming heterogeneous engineering data from different sources to obtain standardized multimodal data; extracting data features of different modal data from the standardized multimodal data to obtain feature representations corresponding to each modal data, and weighting and fusing the feature representations of different modalities to form cross-modal fusion features; adding spatiotemporal location encoding to the cross-modal fusion features to generate enhanced features with spatiotemporal awareness; organizing the enhanced features according to three dimensions: construction stage, spatial region, and feature category to construct an initial feature tensor; and performing feature filtering and dimensionality reduction processing on the initial feature tensor to obtain the multidimensional feature tensor.

[0008] In one embodiment of the present invention, generating a multi-source association graph containing entity relationships and feature semantics includes: creating spatiotemporal nodes representing different spatiotemporal locations based on a spatiotemporal grid defined in the spatiotemporal reference framework; extracting engineering entities from the standardized project dataset and adding the engineering entities to an initial knowledge graph; establishing multidimensional association relationships among the engineering entities in the initial knowledge graph based on the feature semantics of the multidimensional feature tensor; and semantically fusing the numerical features in the multidimensional feature tensor with the engineering entities and multidimensional association relationships in the initial knowledge graph to generate the multi-source association graph.

[0009] In one embodiment of the present invention, generating a verified digital model for engineering project management includes: fusing the multidimensional feature tensor and the multi-source correlation graph to construct a digital model for engineering project management with multiple core functional components; verifying the digital model for engineering project management and generating verification results, wherein the verification includes checking whether the data logic relationships between the components of the model are correct and verifying whether the model fully covers the required engineering elements; correcting the model based on the verification results and outputting a verified digital model for engineering project management.

[0010] In one embodiment of the present invention, updating and optimizing the digital model for engineering project management includes: receiving incremental data from a project site monitoring system and a management system, the incremental data including design change orders, progress update data, and resource adjustment records; inputting the incremental data into the digital model for engineering project management; adjusting the feature tensor values ​​and correlation graph relationships of the affected areas in the digital model for engineering project management according to the parameter changes contained in the incremental data; calling the engineering constraint rule set stored in the digital model for engineering project management to perform compliance verification and correction on the adjusted model parameters; and outputting the verified and corrected model parameters to generate the optimized digital model for engineering project management.

[0011] In one embodiment of the present invention, after generating the digital model for engineering project management, the method further includes: acquiring a standardized dataset of historical projects, and training an initial model based on the historical standardized dataset to obtain a pre-trained model; inputting project data of a new project into the pre-trained model to obtain a digital model for engineering project management of the new project and related intermediate data results; comparing the intermediate data results with preset verification rules to generate an evaluation result, and collecting project progress tracking data obtained based on a preset standard management method; comparing the evaluation result with the project progress tracking data to generate difference data for characterizing model performance; adjusting the parameters and rules of the pre-trained model based on the difference data to generate an optimized model; and repeatedly executing the steps of inputting project data of a new project to generate an optimized model until the performance of the optimized model meets preset conditions, thereby generating a target digital model for engineering project management.

[0012] This application provides a device for generating a digital model of engineering project management. The device includes: a data acquisition and standardization module, used to acquire project management requirements and collect raw data related to engineering project characteristics, preprocess the raw data to form a standardized project dataset; a spatiotemporal framework construction module, used to construct a spatiotemporal reference framework for identifying the spatiotemporal location and constraint rules of data based on the standardized project dataset; and to perform multimodal data processing and feature fusion on the standardized project dataset to generate a multidimensional feature tensor that integrates multiple engineering features; a feature tensor and knowledge graph construction module, used to fuse the spatiotemporal reference framework and the multidimensional feature tensor to generate a multi-source association graph containing entity relationships and feature semantics; a digital model generation module, used to integrate the multidimensional feature tensor and the multi-source association graph according to project progress nodes to generate a validated digital model of engineering project management; and a model optimization module, used to update and optimize the digital model of engineering project management in response to changes in data during project execution, and output the optimized digital model of engineering project management.

[0013] This application provides an electronic device, which includes: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, enable the electronic device to implement the method for generating a digital model for engineering project management as described above.

[0014] This application provides a computer-readable storage medium, characterized in that it stores a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the method for generating a digital model for engineering project management as described above.

[0015] The beneficial effects of this invention: The method for generating a digital model for engineering project management proposed in this invention brings significant benefits through the organic connection and synergistic effect of its core steps. First, by systematically preprocessing the raw data to form a standardized dataset, the problem of fusion of multi-source heterogeneous data is solved, laying a high-quality data foundation for model construction. Next, through the parallel and interconnected steps of constructing a spatiotemporal reference frame and generating multi-dimensional feature tensors, the processed data is endowed with precise spatiotemporal dimensions and profound feature semantics, realizing the digital representation of all elements of the project. Furthermore, by fusing these two steps to generate a multi-source correlation graph, the transformation from numerical values ​​to knowledge is achieved, establishing rich semantic relationships between entities, enabling the model to possess intelligent reasoning capabilities. Then, by integrating all information based on project progress nodes, a validated digital model is generated, ensuring the accuracy and usability of the model under complex business logic. Finally, by dynamically optimizing in response to changing data, the model breaks through the limitations of traditional static models and gains continuous evolutionary vitality. Overall, this method achieves comprehensive perception, accurate mapping, intelligent reasoning, and dynamic optimization in project management. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0017] In the attached diagram:

[0018] Figure 1 This is a schematic diagram illustrating the implementation environment of a method for generating a digital model of engineering project management, as shown in an exemplary embodiment of this application.

[0019] Figure 2 This is a flowchart illustrating a method for generating a digital model of engineering project management, as shown in an exemplary embodiment of this application;

[0020] Figure 3 This is a schematic diagram illustrating the complete steps of a method for generating a digital model of engineering project management, as shown in an exemplary embodiment of this application.

[0021] Figure 4 This is a block diagram illustrating an apparatus for generating a digital model of engineering project management, as shown in an exemplary embodiment of this application.

[0022] Figure 5 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0023] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0024] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0025] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0026] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating the implementation environment of a method for generating a digital model for engineering project management according to an embodiment of the present invention.

[0027] like Figure 1 As shown, the implementation environment for the method of generating a digital model for engineering project management includes a data acquisition module 101 and a computer device 102.

[0028] The data acquisition module 101 is used to collect raw data related to the engineering project from multiple sources in real time or periodically, including but not limited to design drawings (CAD / BIM), construction plan documents, material lists, equipment parameters, subcontracting contracts, quota databases, sensor monitoring data, and project progress reports. This module has multi-format data access, preliminary cleaning, and caching functions, providing structured and unstructured raw data input to the computer device 102.

[0029] Computer device 102 is used to receive the raw data provided therein and execute all the calculation and processing steps of the digital model generation method, including: cleaning, standardizing, vectorizing and quality checking the raw data to form a standardized project dataset; constructing a spatiotemporal reference frame; performing multimodal feature extraction, fusion and tensor construction; generating a multi-source correlation map; integrating and verifying the digital model of engineering project management; and dynamically optimizing and outputting the model in response to changing data.

[0030] This implementation environment, through integrated data acquisition and high-performance computing equipment, supports the automatic generation and updating of the entire process from raw data to intelligent decision-making digital models.

[0031] Figure 2 This is a flowchart illustrating a method for generating a digital model of engineering project management, as shown in an exemplary embodiment of this application.

[0032] like Figure 2 As shown, in an exemplary embodiment, the method for generating a digital model for engineering project management includes at least steps S210 to S250, which are described in detail below:

[0033] Step S210: Obtain project management requirements and collect raw data related to the characteristics of the engineering project. Preprocess the raw data to form a standardized project dataset.

[0034] In one embodiment of this application, the raw data is preprocessed, including at least one of the following operations: formatting heterogeneous data; normalizing data units; enhancing metadata to inject data source, confidence level, or spatiotemporal reference information; aligning data to entities based on a construction industry terminology library to resolve naming ambiguities; calibrating data to a spatiotemporal reference; and filtering and repairing data based on business rules to obtain the standardized project dataset.

[0035] In one embodiment of this application, data is collected from project documents based on five key characteristics of the engineering project: structural type, material properties, construction technology, personnel allocation, and equipment requirements. Through preprocessing processes including cleaning, standardization, vectorization, and quality verification, a standardized dataset Dstd for the project is obtained. The project documents include subcontracting contracts, project design drawings, construction plan documents, material lists, equipment parameter specifications, and quota databases.

[0036] In addition, the preprocessing of the raw data specifically includes: unifying the format of the collected heterogeneous data and converting it into a standard structured format; normalizing the units of the collected data and unifying the measurement system and time representation; enhancing the metadata by injecting key metadata such as data source, confidence level, and spatiotemporal reference system; aligning the collected data to entities and resolving naming ambiguities through a construction industry terminology library; calibrating the collected data to spatiotemporal references by converting relative time / local coordinates to absolute timestamps and global coordinate systems; and performing quality filtering and repair on the collected data by detecting and repairing missing values, outliers, and logical conflicts based on business rules.

[0037] In one specific embodiment of this application, taking the progress management of a commercial complex project as an example, the project overview is a commercial complex in the core area of ​​a city (frame-shear wall structure, 3 underground floors / 20 above-ground floors, with a total construction area of ​​120,000 square meters).

[0038] Based on the five key characteristics of the engineering project—structural type, material properties, construction technology, personnel allocation, and equipment requirements—data was collected from project documents. Through preprocessing processes including cleaning, standardization, vectorization, and quality verification, a standardized dataset for the project was obtained. Project documents included subcontracting contracts, project design drawings, construction plan documents, material lists, equipment parameter specifications, and quota databases. The resulting raw data is shown in Table 1.

[0039] Table 1

[0040]

[0041] After collecting the raw data, further preprocessing was performed. The specific preprocessing steps are as follows:

[0042] Cleaning: Remove duplicate equipment models from subcontracts (such as registering the same tower crane 3 times), and correct column dimension errors in design drawings (600mm×600mm was incorrectly labeled as 500mm×500mm);

[0043] Standardization: Material specifications are unified to national standards (steel diameter is converted from "Imperial 32mm" to "Metric 32mm") and time format is unified to ISO 8601 ("June 2025" is converted to "2025-06").

[0044] Vectorization: Construction technology adopts unique thermal coding ("rotary drilling" → [1,0,0], "percussion pile" → [0,1,0]). Personnel skill levels are generated into 128-dimensional vectors using Word2Vec (e.g., "senior welder" → [0.72, -0.35, ...]).

[0045] Quality verification and cross-validation: Comparing the material list with the incoming inspection form, it was found that the deviation in the amount of concrete admixtures exceeded 5% (triggering the data correction process);

[0046] Logical verification: Check the time interval between "pile foundation construction → foundation slab pouring" in the construction plan (required to be ≥28 days, actually calculated to be 30 days, which complies with the specifications).

[0047] Step S220: Based on the standardized project dataset, construct a spatiotemporal reference frame for identifying the spatiotemporal location and constraint rules of the data; and perform multimodal data processing and feature fusion on the standardized project dataset to generate a multidimensional feature tensor that integrates multiple engineering features.

[0048] In one embodiment of this application, a spatiotemporal reference framework for identifying the spatiotemporal location and constraint rules of data is constructed, including: extracting data from a standardized project dataset to obtain spatial object information and time information; assigning a unique identifier to each object based on the spatial object information, establishing a mapping relationship between data objects and spatial locations, wherein the data objects are determined based on structural components, construction areas, or equipment location information in the standardized project dataset; associating time information with spatial objects to establish a spatiotemporal association relationship for recording the state of spatial objects at different points in time; defining spatiotemporal constraint rules according to engineering specifications and construction plans, and integrating the mapping relationship, spatiotemporal association relationship, and constraint rules to form a unified spatiotemporal reference framework data structure, wherein the engineering specifications and construction plans are obtained based on industry standards, project design documents, or historical construction data.

[0049] In one embodiment of this application, based on the aforementioned standardized project dataset, a spatiotemporal reference framework for the project is obtained through geospatial indexing technology. This includes: selecting a corresponding indexing algorithm based on the spatial information in the standardized data; assigning a unique identifier to each spatial object; establishing a "data-location" mapping relationship; associating the temporal information in the standardized data with the spatial index; establishing a "spatial object-time node" association table to record the state of each object at different times; defining spatiotemporal constraint rules according to specifications and construction plans; and defining the spatiotemporal reference framework, which includes a database or data structure comprising a spatial index structure, time-series data, and a spatiotemporal association table. The spatiotemporal reference framework is defined as follows:

[0050] F st =(T,G,M) ST Equation (1)

[0051] Where T is the time set, G is the spatial grid, and M is the time grid. ST The correlation matrix is ​​specifically an n×m sparse matrix used to mark the spatiotemporal grid where the data points are located.

[0052] Specifically, the parameters in equation (1) above are expressed as follows:

[0053]

[0054] Where n is the number of discrete time periods, determined by milestone events or node events; and Let Δt be the start and end times of the i-th time period, respectively; i For managing time intervals.

[0055] G={g xyz |x∈[1,X],y∈[1,Y],z∈[1,Z]} Formula (3)

[0056] Where x, y, z are the spatial grid dimensions, and g is the spatial grid dimension. xyz It is a spatial grid unit.

[0057]

[0058] Where, δ ij As an indicator function, when data j simultaneously belongs to Δt i and g xyz When the value is 1, it is 0 otherwise; m is the total amount of data.

[0059] In one specific embodiment of this application, based on the aforementioned standardized project dataset, a spatiotemporal reference framework for the project is obtained through geospatial indexing technology. Taking spatiotemporal management during the pile foundation construction phase as an example, the specific steps are as follows:

[0060] First, complete the spatial index construction and data mapping, including:

[0061] Spatial Information Extraction: Extract the following spatial features from the standardized dataset:

[0062] Structural components: piles (1200), columns (800), beams (2400), etc.

[0063] Construction area: Divided into 12 major areas (such as pile foundation construction area and main structure area) and 48 sub-areas (such as the first basement level east area) according to construction stage;

[0064] Equipment locations: Tower cranes (4 units), construction elevators (3 units), material storage areas (10 locations);

[0065] Indexing algorithm selection: Select the R-tree indexing algorithm according to the characteristics of the project, assign a unique identifier to each spatial object (such as Column_1F_001 representing column 1 of the first floor), and establish a "data-location" mapping relationship.

[0066] Secondly, implement time information association and status recording, including:

[0067] Time dimension definition: Time intervals are divided by key milestone events: T = {T: pile foundation construction (January-March), T: underground structure (April-June), T: main structure (July-15), T: decoration and finishing (June-16-20), T: equipment installation (June-18-22), T: final acceptance (June-23-24)};

[0068] Spatiotemporal relationship table establishment: Records the state of each spatial object at different times, as shown in Table 2:

[0069] Table 2

[0070]

[0071] Finally, the spatiotemporal constraint rule definition is implemented, including:

[0072] Based on the constraints defined by construction specifications and plans, such as: spatial constraints: no fixed material storage areas shall be set up within the tower crane's operating radius; the time interval between concrete pouring of adjacent columns shall be ≥24 hours; time constraints: waterproofing construction must begin within 48 hours after the base layer has passed inspection; and the dismantling of external scaffolding shall begin within 30 days after the main structure is capped.

[0073] The spatiotemporal reference frame data structure is implemented using a relational database and includes the following core table structures: a spatial index table, used to store the geometric information and attributes of spatial objects; a time interval table, used to record the start and end times and descriptions of each time interval; a spatiotemporal association table (n×m sparse matrix), used to mark the spatiotemporal grid where data points are located; and a constraint rule table, used to store various spatiotemporal constraints.

[0074] In one embodiment of this application, generating a multidimensional feature tensor that integrates multiple engineering features includes: identifying and transforming heterogeneous engineering data from different sources to obtain standardized multimodal data; extracting data features of different modal data from the standardized multimodal data to obtain feature representations corresponding to each modal data, and weighting and fusing the feature representations of different modalities to form cross-modal fusion features; adding spatiotemporal location encoding to the cross-modal fusion features to generate enhanced features with spatiotemporal awareness; organizing the enhanced features according to three dimensions: construction stage, spatial region, and feature category to construct an initial feature tensor; and performing feature filtering and dimensionality reduction processing on the initial feature tensor to obtain a multidimensional feature tensor.

[0075] In one embodiment of this application, when performing dimensionality reduction on the initial feature tensor, a sparsification computation mechanism is adopted to reduce computational complexity based on spatiotemporal neighborhood relationships, and multi-scale fusion is performed on spatiotemporal features of different resolutions.

[0076] In one embodiment of this application, based on the aforementioned standardized project dataset, a five-dimensional feature tensor for construction engineering projects is obtained by using multimodal deep learning technology through multimodal data processing, cross-modal feature fusion, spatiotemporal feature enhancement, five-dimensional feature tensor construction, feature optimization and acceleration. This includes multimodal data processing, cross-modal feature fusion, spatiotemporal feature enhancement, five-dimensional feature tensor construction, and feature optimization and acceleration.

[0077] Specifically, multimodal data processing includes: identifying and preprocessing heterogeneous data from multiple sources such as CAD drawings, BIM models, engineering texts, and monitoring forms, and converting them into a unified data representation format; CAD drawings are converted into pixel matrices, BIM models are exported as point cloud data, engineering texts are generated into word vectors using Word2Vec / BERT, and monitoring forms are kept in numerical matrix form, forming a standardized dataset.

[0078] Cross-modal feature fusion specifically includes: extracting features using adapted neural networks tailored to the characteristics of data from different modalities; extracting geometric features from images using CNNs, parsing text semantics using Transformer / LSTM, and processing structured data using fully connected layers or GNNs; and achieving dynamic fusion of cross-modal features through a multi-head attention mechanism, constructing multi-layer neural networks to extract low-level visual or semantic features and high-level cross-modal correlation features, generating multi-dimensional feature tensors to represent various aspects of engineering project information. The specific attention fusion formula is as follows:

[0079]

[0080] Where m∈{1,2,3,4} represents the modalities in CAD, BIM, text, and table 4; W a ∈R da×dm Here is the attention weight matrix; q∈R da This is a learnable query vector.

[0081] Spatiotemporal feature enhancement specifically includes: further enhancing feature representation from a spatiotemporal dimension based on feature extraction and fusion. This involves adding spatiotemporal location information to feature vectors through position encoding, enabling the model to perceive the specific location and time of features; and utilizing dynamic coordinate transformation of the homogeneous transformation matrix to adapt to spatial reference differences in different construction areas. Spatiotemporal grid mapping is performed to construct a feature tensor with spatiotemporal awareness capabilities, and the corresponding position encoding formula is as follows:

[0082]

[0083] Where t is the time position, the construction stage index; i is the dimension index (i = 1, ..., d / 2); d represents the feature dimension.

[0084] The corresponding dynamic coordinate change formula is then obtained as follows:

[0085]

[0086] The construction of the five-dimensional feature tensor specifically includes: based on the spatiotemporally enhanced feature tensor, clarifying the engineering semantics of the five-dimensional features (structure, materials, process, labor, and equipment) based on the number of construction stages, the number of grid cells, and the five dimensions of structure, materials, process, labor, and equipment, and constructing the five-dimensional feature tensor. Through a multi-head attention mechanism, the importance weights of features at different spatiotemporal locations are dynamically calculated, focusing on key areas and time points. Multi-source feature vectors from design parameters, construction data, and monitoring data are mapped to a unified multi-source feature representation space, eliminating modal differences and achieving cross-modal feature fusion, as shown below:

[0087]

[0088] Where Τ[p,g,k] represents the tensor value of the p-th construction stage, the g-th spatial grid, and the k-th feature channel; Ω pg This represents the construction space area corresponding to stage pq; |Ω pg |For region Ω pg The number of grid cells within; Let q be the attribute value of the mesh q on the k-th feature channel; feature channel, k=0 is the structural feature, k=1 is the material feature, k=2 is the process feature, k=3 is the artificial feature, and k=4 is the equipment feature.

[0089] Feature optimization and acceleration specifically include: optimizing and accelerating the constructed five-dimensional feature tensor. The SHAP / LIME algorithm is used to analyze feature importance, select key features, and utilize PCA to reduce dimensionality and remove redundancy, verifying the semantic rationality and discriminative power of the feature tensor. To meet the real-time requirements of engineering projects, a sparse attention mechanism is adopted, calculating attention based on spatiotemporal neighborhood to reduce computational complexity. Multi-scale fusion is used to integrate spatiotemporal features of different resolutions, achieving upsampling and feature fusion. Block-based computation is employed to process the spatiotemporal grid in blocks, improving computational efficiency, capturing long-distance spatiotemporal dependencies, and ultimately obtaining a five-dimensional feature tensor with spatiotemporal awareness capabilities, providing core data support for digital modeling of engineering projects.

[0090] Finally, features with SHAP values ​​exceeding a threshold τ are selected for dimensionality reduction. The specific expression for the key feature set is as follows:

[0091] F key ={f i |SHAP(f i )>τ}

[0092]

[0093] Where |S| represents the size of subset S; |F| represents the total number of features; The weights are used to ensure that the probabilities of each subset are equal; v(S) represents the model's predicted value on the feature subset S, and the difference v(S∪{i})-v(S) represents the sum of the values ​​added to f. i Contribution to prediction.

[0094] In one specific embodiment of this application, based on the aforementioned standardized project dataset, and utilizing multimodal deep learning technology, a five-dimensional feature tensor for construction engineering projects is obtained through multimodal data processing, cross-modal feature fusion, spatiotemporal feature enhancement, five-dimensional feature tensor construction, feature optimization, and acceleration. The specific steps are as follows:

[0095] First, the data undergoes multimodal data processing. The specific processing methods for different types of data are as follows:

[0096] CAD drawings: Convert the underground structure plan into a 1000×800 pixel matrix and mark the geometric positions of components such as pile foundations and pile caps;

[0097] BIM model: Export the point cloud data of the main structure (including coordinates of 8000+ structural components), and add attributes such as concrete strength (C35) and steel grade (HRB400);

[0098] Engineering text: The safety logs are vectorized using the BERT model, and the embedding vectors (768 dimensions) of risk words such as "empty job" and "edge protection" are extracted.

[0099] Monitoring Forms: Convert progress reports into a time series matrix (24×150 dimensions, 24 months × 150 process indicators).

[0100] Secondly, cross-modal feature fusion is performed on the data from each modality. A CNN network (ResNet50) is used to extract the spatial distribution features of components from CAD drawings (2048 dimensions), a Transformer model is used to parse the semantics of the construction plan text and generate a process logic feature vector (512 dimensions), a GNN graph neural network is used to process BIM point cloud data and construct a component connection relationship graph (adjacency matrix 1000×1000), and then a multi-head attention mechanism is used to fuse the features of the four modalities.

[0101] Then, spatiotemporal feature enhancement processing is performed on the obtained cross-modal fusion features. For example, spatiotemporal coding is added to the feature vector of the main construction stage (months 4-12). When the time position t=8 (month 8), the coding value of dimension i=1 is: the spatial position is converted into a grid cell (G10-03, East Zone, 10th Floor) through the BIM coordinate system (X=56, Y=48, Z=45); to address the coordinate system differences of different construction areas (underground / above ground), a homogeneous transformation matrix is ​​used to unify the coordinates.

[0102] Next, an initial feature tensor is constructed, which in this embodiment is a five-dimensional feature tensor. Where T∈R 6 ×200×5 (6 construction stages × 200 spatiotemporal grids × 5 types of features), and then dynamically allocate weights through a multi-head attention mechanism to focus on key features such as formwork turnover rate (weight 0.35) and tower crane utilization rate (weight 0.28) in the core area construction.

[0103] Finally, the obtained feature tensors are optimized and accelerated, specifically including key feature selection and sparse attention computation. Key feature selection uses the SHAP algorithm to calculate feature importance, selecting key features with SHAP values ​​> 0.1 (such as 35 core indicators like "concrete strength" and "number of workers"). Dimensionality reduction is achieved by PCA, reducing the feature dimension from 1200 to 200 while retaining 95% of the information entropy. Sparse attention computation restricts the spatiotemporal neighborhood to within ±2 layers of the current grid, reducing the computational complexity from k = 10 neighborhoods to 12 minutes, thus shortening the feature computation time for the main construction phase from 45 minutes.

[0104] Step S230: Based on the spatiotemporal reference frame and multidimensional feature tensor, generate a multi-source association graph containing entity relationships and feature semantics.

[0105] In one embodiment of this application, generating a multi-source association graph containing entity relationships and feature semantics includes: creating spatiotemporal nodes representing different spatiotemporal locations based on a spatiotemporal grid defined in a spatiotemporal reference framework; extracting engineering entities from a standardized project dataset and adding the engineering entities to an initial knowledge graph; establishing multidimensional association relationships between engineering entities in the initial knowledge graph based on the feature semantics of a multidimensional feature tensor; and semantically fusing the numerical features in the multidimensional feature tensor with the engineering entities and multidimensional association relationships in the initial knowledge graph to generate a multi-source association graph.

[0106] In one embodiment of this application, based on the project spatiotemporal reference framework and the project five-dimensional feature tensor (similar to the aforementioned multi-dimensional feature tensor), a multi-source association graph and a spatiotemporal feature tensor are obtained through knowledge graph embedding technology. This includes: first, enhancing the features through knowledge graph embedding technology and constructing spatiotemporal nodes based on the spatiotemporal reference framework; second, adding engineering entities and forming a multi-source association spatiotemporal knowledge graph structure by establishing spatial adjacency, establishing temporal continuity (based on construction time sequence and time series dependency), adding entity relationships, and constructing entity knowledge graph edges; and third, optimizing the knowledge graph embedding technology by integrating the feature tensor with graph semantics and graph relationships to feed back the feature enhancement, thereby obtaining the multi-source association graph and the spatiotemporal feature tensor.

[0107] In one specific embodiment of this application, based on the project spatiotemporal reference framework and the project feature tensor, the spatiotemporal feature tensor and multi-source association graph of the project are obtained through knowledge graph embedding technology. The specific steps are as follows:

[0108] First, construct spatiotemporal nodes, which are 200 spatiotemporal grids based on a spatiotemporal reference frame. Create a "spatiotemporal node" for each grid (e.g., T08-G10-03 represents the 10th floor of the East Zone in August) and associate it with all construction activities (concrete pouring, rebar tying, etc.) within that grid.

[0109] Secondly, relationship modeling is determined. On the one hand, spatial adjacency is established, for example, establishing a "spatial adjacency" relationship between Column_1F_001 and Beam_1F_002 (adjacent beams); on the other hand, time dependency is established, for example, establishing a "sequential" relationship between process W05 (tying) → W06 (formwork) (time interval ≥ 4h); it also includes determining resource consumption, for example, establishing a "consumption" relationship between work group G01 and material B02 (15 people per work group for every 10 tons of steel bars).

[0110] Finally, graph embedding is achieved by using the TransR algorithm to embed the five-dimensional feature tensor into the knowledge graph, thereby realizing feature feedback and graph enhancement.

[0111] Step S240: Based on the project progress nodes, integrate multi-dimensional feature tensors and multi-source correlation graphs to generate a validated digital model for engineering project management.

[0112] In one embodiment of this application, generating a verified digital model for engineering project management includes: fusing multidimensional feature tensors and multi-source correlation graphs to construct a digital model for engineering project management with multiple core functional components; verifying the digital model for engineering project management and generating verification results, wherein verification includes checking whether the data logic relationships between the components of the model are correct and verifying whether the model fully covers the required engineering elements; correcting the model based on the verification results and outputting a verified digital model for engineering project management.

[0113] In one embodiment of this application, the verification of the digital model for engineering project management includes data alignment checks, semantic association verification, and state rule verification to complete the consistency and integrity verification of the model. Specifically, integrity verification includes checking the field integrity of the spatiotemporal reference frame, detecting isolated nodes in the multi-source association graph, displaying missing data in the multi-dimensional feature tensor, and statistically analyzing the attachment status of project deliverables to ensure that the model comprehensively reflects all elements of the project.

[0114] In one embodiment of this application, milestone events in progress management are used as time nodes to divide management intervals. A digital model for engineering project management is constructed by combining the project's spatiotemporal feature tensor and multi-source correlation graph. Third-party tools are then used to verify the model's consistency and integrity. Specific details include the following:

[0115] The digital model for project management adopts a five-dimensional spatiotemporal fusion architecture consisting of structure, materials, processes, manpower, and equipment, and has five core functional components: spatiotemporal benchmark framework, five-dimensional feature tensor, knowledge graph integration, state evolution engine, and deliverable attachment system.

[0116] Spatiotemporal reference framework: In the time dimension, project phases are divided by milestone events, supporting schedule management and time series analysis; in the spatial dimension, spatial management is achieved through three-dimensional grid spatial positioning and topology analysis.

[0117] Five-dimensional feature tensor: Construct a five-dimensional data cube that integrates information on time, space, structure, materials, processes, human factors, and equipment. Integrate multi-source heterogeneous data to provide standardized input for deep learning models to support spatiotemporal prediction.

[0118] Knowledge graph integration: Establish a knowledge graph that includes engineering entities (such as components, processes, materials, etc.) and their relationships (such as belonging, consumption, use, etc.) to achieve semantic-level data association and support relational reasoning and semantic query.

[0119] State Evolution Engine: Simulates the dynamic change of model state over time, drives model evolution through state transition equations and business rules, and supports simulation and dynamic deduction of "if-then" type scenarios.

[0120] Deliverable Linking System: Manages delivery documents throughout the project lifecycle, establishes the link between documents and milestones, tracks document requirements and progress through a "milestone-document-status" matrix; integrates spatiotemporal indexes and model data to generate structured deliverables, ensuring document compliance and traceability of digital assets during the operation and maintenance phase, and automatically generates digital delivery packages and spatiotemporal index as-built models that conform to the ISO 19650 standard.

[0121] During the model validation phase, third-party tools are first used to perform consistency verification. Specifically, this includes: using Python and Pandas for data alignment checks; verifying semantic associations with graph databases using SPARQL; matching feature vectors with entity semantics using knowledge graph embedding tools such as PyKEEN; and verifying state evolution rules using a Python unit testing framework. Consistency verification aims to eliminate "spatiotemporal silos" and semantic contradictions, ensuring that the relationships in the knowledge graph are consistent with the material feature values ​​in the feature tensor, guaranteeing model logical consistency, and avoiding prediction bias caused by data disconnect between components. This verification ensures logical matching of each component, including the consistency of data existence between the spatiotemporal baseline index and the five-dimensional feature tensor, the semantic correspondence between knowledge graph entity attributes and feature tensor numerical values, and the correct driving force of business rules on state evolution.

[0122] Secondly, integrity verification is performed. JSON Schema is used to verify the field integrity of the spatiotemporal baseline framework; the Neo4j APOC library is used to detect isolated nodes in the knowledge graph; NumPy and data visualization tools are used to display missing data in the feature tensors; and Excel / Pandas and regular expressions are used to statistically analyze the attachment status of deliverables. Integrity verification identifies and locates missing data, such as using heatmaps to locate missing manual features within a specific time period, or using graph analysis to find isolated process nodes without associated equipment, thereby ensuring that the model comprehensively covers all elements of the project and provides complete data support for decision-making. This verification ensures that there is no missing data in the model components, specifically reflected in: time segmentation covering the entire project lifecycle; complete spatial grid division of regions; no NaN values ​​for the structure and material features of the five-dimensional feature tensors at each spatiotemporal point (time × space); complete key entities and relationships in the knowledge graph; and all necessary milestone documents are attached and their status is updated in real time.

[0123] In one specific implementation of this application, milestone events in schedule management are used as time nodes to divide the management interval. A digital model for engineering project management is constructed by combining the project's spatiotemporal characteristic tensor and multi-source correlation graph. Third-party tools are then used to verify the consistency and integrity of the model.

[0124] The core components include the following examples: A spatiotemporal baseline framework uses six key milestones, including completion of the zero-level structure (T3) and main structure completion (T6), to divide the project into multiple phases. Each phase covers thirty to fifty spatiotemporal grids, supporting weekly or monthly progress projections. A five-dimensional feature tensor constructs a 24×100×5 data cube, corresponding to twenty-four time months, one hundred spatial partitions, and five types of feature indicators. It integrates various information sources, such as design parameters like column grid dimensions, construction data like shift attendance, and monitoring data like tower crane load. The knowledge graph integration component establishes over two thousand eight hundred "component-process-equipment" triples, enabling efficient semantic queries. For example, it can find all processes that simultaneously use a T03 tower crane and consume B02 steel reinforcement, with a query response time of less than two seconds.

[0125] Furthermore, during the model validation phase, consistency verification was first performed. Pandas was used to check the consistency between the spatiotemporal grid and the five-dimensional feature tensor in terms of coordinates. Then, SPARQL was used to query the association relationship between process W05 and equipment T03 in the knowledge graph to ensure that the association weight of 0.32 for the equipment features extracted from the tensor was consistent with the strength of the graph relationship. Subsequently, integrity verification was performed. The completeness of the fields of the spatiotemporal baseline framework was checked using JSON Schema. It was found that the fireproof material grade feature was missing in the decoration stage. After completing it, the integrity reached 100%. At the same time, the APOC library of Neo4j was used to detect three isolated process nodes of unrelated work groups. After supplementing the work group relationships, the overall connectivity of the graph improved to 98%.

[0126] In one embodiment of this application, after generating the digital model for engineering project management, the method further includes: pre-training the model based on historical pre-processed data; processing the project data provided by the construction party using the pre-trained model to obtain the digital model for construction project management and related intermediate results such as the five-dimensional feature tensor and multi-source correlation graph of the standardized project dataset. Specifically, this includes: acquiring a standardized dataset of historical projects and training the initial model based on the historical standardized dataset to obtain a pre-trained model; inputting the project data of the new project into the pre-trained model to obtain the digital model for engineering project management of the new project and related intermediate data results; comparing the intermediate data results with preset verification rules to generate results, and collecting project progress tracking data obtained based on preset standard management methods; comparing the evaluation results with the project progress tracking data to generate difference data to characterize model performance; adjusting the parameters and rules of the pre-trained model based on the difference data to generate an optimized model; and repeatedly executing the steps of inputting project data of the new project to generate the optimized model until the performance of the optimized model meets preset conditions, thereby generating the target digital model for engineering project management.

[0127] In one specific embodiment of this application, a large-scale bridge construction project is used as an example for illustration. This method continuously improves the model's accuracy and practicality through continuous learning and comparative verification.

[0128] First, historical data was extracted from three completed bridge projects of the same type, including structural drawings, construction logs, material testing reports, and personnel and equipment records. These raw materials underwent systematic cleaning, standardization, and vectorization transformation to form a standardized historical dataset. This dataset was then used to train an initial digital model, resulting in a pre-trained model. This model already possesses the ability to identify and integrate five core characteristics of bridge engineering: structure, materials, processes, personnel, and equipment.

[0129] Next, the original data of the new bridge project, such as the BIM model, construction plan, and contract documents, are input into the pre-trained model. The model automatically processes the data and outputs a series of intermediate results, including the standardized dataset of the new project, a five-dimensional feature tensor, and a multi-source correlation graph, thus initially constructing a digital model for the engineering project management of the new project.

[0130] Then, the intermediate results are compared with preset verification rules. For example, the integrity of the data structure is checked using JSON Schema, the connectivity in the knowledge graph is detected using the Neo4j graph database, and the consistency of feature tensors is verified using Pandas tools. Based on this, an integrity verification report is generated. At the same time, actual project progress data obtained using traditional management methods such as Gantt charts combined with manual inspections are collected.

[0131] The project compares the progress data predicted by the model with the actual progress data recorded by traditional methods, calculates key indicators such as schedule deviation rate and resource allocation error, and forms a difference dataset. Based on this difference data, the feature weights, spatiotemporal constraint rules, and attention mechanism parameters in the pre-trained model are adjusted to generate an optimized new model.

[0132] Repeat the above process of processing new project data, verification and comparison, and model optimization. After processing data from two more new bridge projects in sequence, the model undergoes multiple rounds of iterative optimization, ultimately achieving stable performance and meeting preset conditions. The resulting digital model for target engineering project management can then be used for refined, full-cycle management of similar projects.

[0133] Step S250: In response to changes in data during project execution, the digital model for engineering project management is updated and optimized, and the optimized digital model for engineering project management is output. Specifically, when design changes or schedule updates occur, the feature tensors within the affected time interval and spatial region are located and updated; based on dynamic evolution equations, real-time field data, engineering constraint rules, and noise simulation are integrated to perform gradient updates on the model parameters, and the optimized digital model is output.

[0134] In one embodiment of this application, updating and optimizing a digital model for engineering project management includes: receiving incremental data from a project site monitoring system and a management system, the incremental data including design change orders, progress update data, and resource adjustment records; inputting the incremental data into the digital model for engineering project management; adjusting the feature tensor values ​​and correlation graph relationships of the affected areas in the digital model for engineering project management based on the parameter changes contained in the incremental data; calling the set of engineering constraint rules stored in the digital model for engineering project management to perform compliance verification and correction on the adjusted model parameters; and outputting the verified and corrected model parameters to generate an optimized digital model for engineering project management.

[0135] In one embodiment of this application, a dynamic update mechanism is established to address potential changes during project execution. After optimizing the original model, a dynamically optimized version of the digital model for construction project management is obtained. This includes updating the feature tensors of the affected time interval and spatial region when encountering local changes such as design changes or schedule updates. The specific dynamic evolution equation is as follows:

[0136]

[0137] Where M represents the digital model for project engineering management; α represents the model's update rate over time; α, β, and γ are weight coefficients that control the intensity of the influence of data, constraints, and noise, respectively. Real-time incremental data from the field is used to update model parameters through gradient updates, ensuring that the model is synchronized with the field. To impose engineering constraints, the model parameters are forced to evolve within a legal range; To address unpredictable factors and random noise, the model's adaptability to complex scenarios is improved.

[0138] In a specific embodiment of this application, a dynamic update mechanism is established to address potential changes during project execution. By optimizing the original model, a dynamically optimized version of the digital model for building construction project management is obtained. For example, a change request is "the owner proposed adjusting the shop floor height in the 10th month, which resulted in an increase of 10 centimeters in the beam height of the third to fifth floors." The corresponding dynamic evolution process is described by equation (9), and its specific instantiation is as follows:

[0139] M(t+Δt)=M(t)+αD+βC+γN Formula (10)

[0140] In this formula, α is 0.6, representing the weight of data influence; the new design change order data D contains 15 feature adjustments, with a constraint weight of 0.3; at the same time, a specification constraint C is applied, that is, the thickness of the fireproof coating needs to be increased accordingly after the floor height is adjusted, and a noise weight of 0.1 is introduced to simulate unforeseen factors N such as material transportation delays.

[0141] After dynamic adjustment, the model automatically recalculated the time schedule for the 20 affected processes, optimizing the project time forecast from a 10-day delay to a 6-day delay.

[0142] Figure 3 This is a schematic diagram illustrating the complete steps of a method for generating a digital model of engineering project management, as shown in an exemplary embodiment of this application.

[0143] In one complete embodiment of this application, the method for generating a digital model for engineering project management proposed in this application is specifically as follows: Figure 3 As shown, this method integrates multi-source data with artificial intelligence technology to construct an intelligent management model with spatiotemporal awareness capabilities. The following section uses a large commercial complex project as an example to explain the implementation steps of this method in detail.

[0144] First, raw data was collected from various project documents. These documents included BIM models, CAD design drawings, construction schedules, material procurement lists, equipment ledgers, and labor subcontracting contracts, covering five major characteristic categories: structural type, material properties, construction techniques, personnel allocation, and equipment requirements. The collected data underwent a series of preprocessing steps, including removing duplicates, correcting annotation errors, standardizing units of measurement and time formats, and converting textual information into numerical vectors, ultimately forming a standardized project dataset.

[0145] Next, a spatiotemporal reference frame and a multidimensional feature tensor are constructed based on a standardized dataset. Spatial and temporal information is extracted from the dataset, and a unique identifier is assigned to each component using a spatial indexing algorithm, establishing a mapping relationship between it and its three-dimensional spatial location. Simultaneously, temporal information is associated with spatial objects, recording the object state at different time points, and spatiotemporal constraint rules are defined according to engineering specifications. On the other hand, multimodal deep learning technology is used to process data in different formats such as drawings, text, and tables, extracting and fusing their features, then enhancing their semantics through spatiotemporal encoding, ultimately organizing them into a multidimensional feature tensor with engineering significance.

[0146] Subsequently, a multi-source association graph is generated. This step creates spatiotemporal nodes based on a spatiotemporal reference frame and extracts engineering entities as graph nodes. By analyzing the semantic information in the feature tensor, various associations such as spatial adjacency, temporal order, and resource consumption are established between entities, forming a semantically rich structured knowledge network.

[0147] Then, based on the project schedule nodes, feature tensors and relationship graphs are integrated to generate a preliminary digital model. This model includes five core components: spatiotemporal baseline, feature tensors, knowledge graph, state evolution engine, and deliverable management. Subsequently, third-party tools are used to perform consistency and integrity verification on the model to ensure its logical correctness and data integrity, ultimately outputting a workable digital model for engineering project management.

[0148] Finally, to address dynamic changes during project implementation, a model update mechanism is established. When design changes or schedule adjustments occur, incremental on-site data is received, the affected areas are automatically located, model parameters and map relationships are updated, compliance checks are performed in conjunction with predefined constraint rules, and an optimized digital model is output, providing reliable support for real-time decision-making in project management.

[0149] It should be noted that the method for generating a digital model for engineering project management proposed in this application constructs a digital model with spatiotemporal awareness capabilities through deep fusion and intelligent processing of multi-source heterogeneous data, significantly improving the refinement and intelligence level of engineering project management. This method effectively eliminates data silos, achieving unified representation and deep correlation mining of multimodal data from CAD drawings and BIM models to text and tables, providing a comprehensive and reliable data foundation for management decisions. Simultaneously, the model possesses dynamic response and self-optimization capabilities, enabling real-time adjustments based on project progress and changes, ensuring its practicality and accuracy in real-world environments. Through built-in consistency and integrity verification mechanisms, the model's logical self-consistency and element completeness are ensured, greatly improving the reliability of the output results. Ultimately, the model supports full-process semantic querying, dynamic deduction, and standard delivery, not only improving management efficiency at each stage but also strengthening cross-stage collaboration capabilities, providing solid technical support for promoting the digital transformation and upgrading of the construction industry.

[0150] Figure 4 This is a block diagram illustrating an apparatus for generating a digital model of engineering project management, as shown in an exemplary embodiment of this application. This apparatus can be applied to… Figure 1 The implementation environment shown is illustrated. This device can also be applied to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the implementation environment to which the device is applicable.

[0151] like Figure 4 As shown, the exemplary apparatus for generating a digital model of engineering project management includes:

[0152] The system includes the following modules: a data acquisition and standardization module 410, which acquires project management requirements and collects raw data related to project characteristics, preprocesses the raw data to form a standardized project dataset; a spatiotemporal framework construction module 420, which constructs a spatiotemporal reference framework based on the standardized project dataset to identify the spatiotemporal location and constraint rules of the data, and performs multimodal data processing and feature fusion on the standardized project dataset to generate a multidimensional feature tensor that integrates multiple engineering features; a feature tensor and knowledge graph construction module 430, which integrates the spatiotemporal reference framework and the multidimensional feature tensor to generate a multi-source association graph containing entity relationships and feature semantics; a digital model generation module 440, which integrates the multidimensional feature tensor and the multi-source association graph according to project progress nodes to generate a validated digital model for project management; and a model optimization module 450, which updates and optimizes the digital model for project management in response to changes in data during project execution and outputs the optimized digital model for project management.

[0153] It should be noted that the apparatus for generating a digital model of engineering project management provided in the above embodiments and the method for generating a digital model of engineering project management provided in the above embodiments belong to the same concept. The specific ways in which each module and unit performs operations have been described in detail in the method embodiments, and will not be repeated here. In practical applications, the apparatus for generating a digital model of engineering project management provided in the above embodiments can allocate the above functions to different functional modules as needed, that is, divide the internal structure of the apparatus into different functional modules to complete all or part of the functions described above, and this is not a limitation.

[0154] Embodiments of this application also provide an electronic device, including: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, enable the electronic device to implement the method for generating a digital model for engineering project management provided in the above embodiments.

[0155] Figure 5 A schematic diagram of a computer system suitable for implementing the embodiments of this application is shown. It should be noted that... Figure 5 The computer system 500 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0156] like Figure 5As shown, the computer system 500 includes a Central Processing Unit (CPU) 501, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 502 or programs loaded from storage portion 508 into Random Access Memory (RAM) 503, such as performing the methods described in the above embodiments. The RAM 503 also stores various programs and data required for system operation. The CPU 501, ROM 502, and RAM 503 are interconnected via a bus 504. An Input / Output (I / O) interface 505 is also connected to the bus 504.

[0157] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. Removable media 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 510 as needed so that computer programs read from them can be installed into storage section 508 as needed.

[0158] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by central processing unit (CPU) 501, it performs various functions defined in the system of this application.

[0159] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0160] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0161] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0162] Another aspect of this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the method for generating a digital model for engineering project management as described above. This computer-readable storage medium may be included in the electronic device described in the above embodiments, or it may exist independently and not assembled into the electronic device.

[0163] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method for generating a digital model for engineering project management provided in the various embodiments described above.

[0164] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for generating a digital model for engineering project management, characterized in that, The method includes: Obtain project management requirements and collect raw data related to the characteristics of engineering projects. Preprocess the raw data to form a standardized project dataset. Based on the standardized project dataset, a spatiotemporal reference framework is constructed to identify the spatiotemporal location and constraint rules of the data; and multimodal data processing and feature fusion are performed on the standardized project dataset to generate a multidimensional feature tensor that integrates multiple engineering features. Based on the spatiotemporal reference framework and the multidimensional feature tensor, a multi-source association graph containing entity relationships and feature semantics is generated. Based on the project schedule milestones, the multidimensional feature tensor and multi-source correlation graph are integrated to generate a validated digital model for engineering project management. In response to changes in data during project execution, the digital model for project management is updated and optimized, and an optimized digital model for project management is output.

2. The method for generating a digital model for engineering project management according to claim 1, characterized in that, Construct a spatiotemporal reference framework for identifying the spatiotemporal location and constraint rules of data, including: Data extraction is performed on the standardized project dataset to obtain spatial object information and temporal information; Based on the spatial object information, a unique identifier is assigned to each object, and a mapping relationship between data objects and spatial locations is established. The data objects are determined based on the structural components, construction areas, or equipment location information in the standardized project dataset. The time information is associated with the spatial object to establish a spatiotemporal relationship for recording the state of the spatial object at different points in time. Based on the engineering specifications and construction plan, spatiotemporal constraint rules are defined, and the mapping relationship, spatiotemporal correlation relationship and constraint rules are integrated to form a spatiotemporal reference framework with the same data structure. The engineering specifications and construction plan are based on industry standards, project design documents or historical construction data.

3. The method for generating a digital model for engineering project management according to claim 1, characterized in that, Generate a multidimensional feature tensor that integrates multiple engineering features, including: Identify and transform heterogeneous engineering data from different sources to obtain standardized multimodal data; Data features of different modal data are extracted from the standardized multimodal data to obtain the feature representations corresponding to each modal data, and the feature representations of different modalities are weighted and fused to form cross-modal fused features; Spatiotemporal location encoding is added to the cross-modal fusion features to generate enhanced features with spatiotemporal awareness; The enhanced features are organized according to three dimensions: construction stage, spatial region, and feature category, to construct an initial feature tensor; The initial feature tensor is subjected to feature filtering and dimensionality reduction to obtain the multidimensional feature tensor.

4. The method for generating a digital model for engineering project management according to claim 1, characterized in that, Generate a multi-source association graph containing entity relationships and feature semantics, including: Based on the spatiotemporal grid defined in the spatiotemporal reference framework, spatiotemporal nodes representing different spatiotemporal locations are created; Engineering entities are extracted from the standardized project dataset, and these engineering entities are added to the initial knowledge graph. Based on the feature semantics of the multidimensional feature tensor, a multidimensional association relationship is established among the engineering entities in the initial knowledge graph; The numerical features in the multidimensional feature tensor are semantically fused with the engineering entities and multidimensional relationships in the initial knowledge graph to generate the multi-source association graph.

5. The method for generating a digital model for engineering project management according to claim 1, characterized in that, Generate validated digital models for engineering project management, including: The multidimensional feature tensor and multi-source correlation graph are fused to construct a digital model for engineering project management with multiple core functional components; The digital model for project management is validated, and validation results are generated. The validation includes checking whether the data logic relationships between the components of the model are correct, and verifying whether the model fully covers the required engineering elements. The model is revised based on the validation results, and a validated digital model for engineering project management is output.

6. The method for generating a digital model for engineering project management according to any one of claims 1-5, characterized in that, The digital model for project management is updated and optimized, including: Receive incremental data from the project site monitoring system and management system, including design change orders, progress update data, and resource adjustment records; The incremental data is input into the digital model for project management. Based on the parameter changes contained in the incremental data, adjust the feature tensor values ​​and correlation graph relationships of the affected areas in the digital model of engineering project management; The set of engineering constraint rules stored in the digital model of project management is invoked to perform compliance verification and correction on the adjusted model parameters; The validated and corrected model parameters are output to generate the optimized digital model for project management.

7. The method for generating a digital model for engineering project management according to claim 6, characterized in that, After generating the digital model for project management, the method further includes: Obtain a standardized dataset of historical projects, and train the initial model based on the standardized historical dataset to obtain a pre-trained model; The project data of the new project is input into the pre-trained model to obtain the digital model of the engineering project management of the new project and related intermediate data results; The intermediate data results are compared with the preset verification rules to generate evaluation results, and project progress tracking data obtained based on the preset standard management method is collected. The evaluation results are compared with the project progress tracking data to generate difference data to characterize the model performance; Based on the discrepancy data, the parameters and rules of the pre-trained model are adjusted to generate an optimized model; Repeat the steps of inputting project data for a new project to generate an optimized model until the performance of the optimized model meets the preset conditions, thus generating a digital model for the management of the target project.

8. A device for generating a digital model for engineering project management, characterized in that, The device includes: The data acquisition and standardization module is used to obtain project management requirements and collect raw data related to the characteristics of engineering projects. The raw data is preprocessed to form a standardized project dataset. The spatiotemporal framework construction module is used to construct a spatiotemporal reference framework for identifying the spatiotemporal location and constraint rules of the data based on the standardized project dataset; and to perform multimodal data processing and feature fusion on the standardized project dataset to generate a multidimensional feature tensor that integrates multiple engineering features. The feature tensor and knowledge graph construction module is used to generate a multi-source association graph containing entity relationships and feature semantics based on the spatiotemporal reference framework and the multidimensional feature tensor. The digital model generation module is used to integrate the multidimensional feature tensor and multi-source correlation graph according to the project progress nodes to generate a verified digital model for engineering project management. The model optimization module is used to update and optimize the digital model of engineering project management in response to changes in data during project execution, and output the optimized digital model of engineering project management.

9. An electronic device, characterized in that, The electronic device includes: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the electronic device to implement the method for generating a digital model of engineering project management as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by the computer's processor, causes the computer to perform the method for generating the digital model of engineering project management as described in any one of claims 1 to 7.

Images