Three-dimensional visual construction management system and method

Abstract

This invention discloses a three-dimensional visualization construction management system and method, comprising the following steps: S1. Acquiring construction project information and constructing a three-dimensional digital modeling environment for the construction area; S2. Deploying a sensor network at the construction site and forming a dynamic load dataset; S3. Generating a preprocessed dynamic load dataset; S4. Constructing a gated cyclic element network prediction model using the preprocessed dynamic load dataset to simulate and predict the dynamic response of the construction structure under seismic loading; S5. Applying the Grey Wolf optimization algorithm to globally search and optimize the key hyperparameters of the gated cyclic element network prediction model to obtain an optimized gated cyclic element network prediction model; S6. Based on the three-dimensional visualization results, classifying the dynamic response of the construction structure into risk levels and providing early warning processing, generating construction management decision support information. This invention provides a practical and feasible intelligent management method for construction projects in seismic zones.

Description

Technical Field

[0001] This invention relates to the field of construction management technology, and in particular to a three-dimensional visualization construction management system and method. Background Technology

[0002] With the rapid advancement of urban infrastructure construction, large-scale construction projects face increasingly complex structural design requirements and variable construction environment factors. In earthquake-prone areas, ensuring the dynamic stability, safety, and response control capabilities of structures during construction has become a key issue in project management.

[0003] In existing technologies, structural seismic response analysis mainly relies on numerical simulation methods or static safety margin calculations. The core algorithms are mostly linear modeling, static bearing capacity assessment, or simplified dynamic analysis under preset conditions, which are difficult to reflect the nonlinear evolution of the structural stress state during actual construction. At the same time, due to the limited adaptability of the model to changes in input data, existing methods are unable to handle real-time acquisition of multi-source time-series data on wind load, seismic intensity, and structural acceleration response. This leads to practical problems in construction management such as delayed response prediction, inaccurate identification of stress concentration zones, and weak risk warning mechanisms. In addition, some existing studies have attempted to introduce neural networks or optimization algorithms, but most of them remain at the structural modeling level and fail to integrate with real-time data. Furthermore, they lack a linkage mechanism with 3D visualization systems, making it impossible to form a closed-loop management process of "prediction-display-decision".

[0004] In terms of risk assessment and early warning processing of structural response prediction results, the commonly used rule system is based on empirical thresholds or manually set levels. This system not only lacks adaptability but also makes it difficult to combine the response characteristics under different geographical environments, structural configurations, and dynamic load conditions for refined management. As a result, the timeliness and accuracy of the early warning results are poor, which affects the efficiency of risk prevention and control and the ability to ensure structural safety during construction.

[0005] In summary, there is an urgent need for a comprehensive technical approach that integrates intelligent prediction algorithms, 3D modeling and visualization, and construction risk identification to address the core issues in current construction management, such as untimely response prediction, inaccurate risk assessment, and incomplete decision support. Summary of the Invention

[0006] One objective of this invention is to propose a three-dimensional visualization construction management system and method, which provides a practical and feasible intelligent management means for construction projects in earthquake-prone areas.

[0007] A three-dimensional visualization construction management method according to an embodiment of the present invention includes the following steps:

[0008] S1. Obtain construction project information and construct a three-dimensional digital modeling environment for the construction area;

[0009] S2. Deploy a sensor network at the construction site to collect real-time data on wind load, seismic intensity, and structural acceleration response, and generate a dynamic load dataset.

[0010] S3. Perform unified data cleaning, noise filtering, missing value imputation and time series reconstruction on the dynamic load dataset to generate a preprocessed dynamic load dataset.

[0011] S4. A gated cyclic element network prediction model is constructed using the preprocessed dynamic load dataset to simulate and predict the dynamic response of the construction structure under seismic loading.

[0012] S5. Apply the Grey Wolf optimization algorithm to perform a global search and optimization of the key hyperparameters of the gated recurrent unit network prediction model to obtain the optimized gated recurrent unit network prediction model.

[0013] S6. Map the structural response prediction results output by the optimized gated cyclic unit network prediction model to the three-dimensional digital modeling environment to realize real-time dynamic simulation and three-dimensional visualization of the construction structure under seismic action. Based on the three-dimensional visualization results, classify the risk level and provide early warning for the dynamic response of the construction structure, and generate construction management decision support information.

[0014] Optionally, S1 includes the following steps:

[0015] S11. Obtain the three-dimensional structural information of the construction project. Use three-dimensional laser scanning equipment or UAV remote sensing system to collect point cloud data of the three-dimensional structural shape of the construction project and form a point cloud set P.

[0016] S12. Obtain the geographic environment information of the construction area, including topographic and geomorphological information, geological structure information and surface cover type, and construct an environmental dataset G composed of multiple geographic information units;

[0017] S13. Based on the point cloud set P and the environment dataset G, a three-dimensional digital modeling environment M is constructed by performing multi-source data registration and fusion using digital modeling. d 3D digital modeling environment M d This indicates a function mapping relationship M. d =f(P,G) obtains an integrated 3D model of the building structure and geographic environment, where function f is the spatial matching and feature reconstruction mapping between point cloud data and geographic information units;

[0018] S14. For the completed 3D digital modeling environment M dThe topology is organized and meshed to form a computable geometric model for response simulation and visualization rendering, giving the 3D digital modeling environment structural features such as node definition, boundary closure and attribute binding.

[0019] Optionally, S2 includes the following steps:

[0020] S21. Wind load monitoring sensors are installed at key parts of the construction structure. The wind load monitoring sensors collect wind speed, wind direction and wind load data formed by the area of ​​action in real time.

[0021] S22. Deploy ground motion monitoring sensors in the construction area. The ground motion monitoring sensors collect ground motion intensity information in real time. The ground motion intensity information is represented by the horizontal and vertical acceleration components of the ground surface, which constitute ground motion intensity data.

[0022] S23. Install structural response acceleration sensors inside the construction structure. The response acceleration sensors monitor the acceleration response of the structure under earthquake and wind loads and generate structural acceleration response data.

[0023] S24. Perform time synchronization processing and format unification on wind load data, seismic intensity data, and structural acceleration response data, and fuse them to generate a complete dynamic load dataset D. load :

[0024]

[0025] Where t represents the unified time index in the dynamic load dataset, which is the alignment result of the time axis of various sensor data, and v t θ represents the wind speed at time t, used to reflect the intensity of wind load. t The wind direction angle at time point t is used to determine the direction of wind load. A t This represents the area affected by the wind at time point t, and is used to correspond to the size of the stress-bearing area on the structure under wind load. These represent the components of the seismic acceleration in the X, Y, and Z directions at time point t, respectively. Let T be the structural acceleration response data of the p-th monitoring point at time t, where T is the unified set of valid time series.

[0026] Wind speed (v) t The wind speed is used to reflect the wind velocity per unit time and is a key physical quantity that determines the intensity of wind load.

[0027] Wind direction angle (θ) t It is used to determine the direction of wind load and is the main parameter for the direction of wind force vector.

[0028] Wind-affected area (A) tThis represents the effective area of ​​the structural surface exposed to wind at that point in time, and is an important spatial factor affecting the total wind force value.

[0029] These three data points together describe the dynamic characteristics of wind load at a specific point in time and are collectively referred to as wind load action data.

[0030] Meanwhile, the seismic acceleration components These represent the surface acceleration responses of the structure in the X, Y, and Z spatial directions under earthquake loading, respectively, constituting the seismic intensity data. This data is used to characterize the instantaneous inertial impact of seismic wave input on the structure and is one of the basic excitation inputs for structural response prediction.

[0031] Optionally, S3 includes the following steps:

[0032] S31. For the dynamic load dataset D load Perform data integrity checks to remove data records with duplicate timestamps, missing fields, or illegal values, thus forming a preliminary cleaned dynamic load dataset.

[0033] S32. Smooth the high-frequency noise data in the preliminary cleaned dynamic load dataset, set the window size, and perform noise smoothing on any numerical field at each time point t. For time points t with missing values ​​in the smoothed data, perform interpolation imputation. Then, perform unified time series reconstruction on the dynamic load dataset after missing value imputation, remap all fields to the standard time axis, and perform linear resampling on the misaligned data points to form a time series dynamic load dataset.

[0034] S35. Normalize all fields in the time-series dynamic load dataset, converting any field to a normalized value to obtain the preprocessed dynamic load dataset D. pre .

[0035] Optionally, S4 includes the following steps:

[0036] S41. Using the preprocessed dynamic load dataset D pre Using the input data, a gated cyclic element network prediction model is constructed to predict the dynamic response of constructed structures under seismic loading:

[0037]

[0038] Among them, h t h represents the hidden output state of the gated recurrent unit network prediction model at time t, used to describe the internal response state of the construction structure under seismic external load at time t. t-1 z represents the output hidden state of the gated recurrent unit network prediction model at time point t-1. tTo update the gate control factor, This indicates the candidate response state at the current time step, and ° represents the Hadamard element-wise product operator;

[0039] S42. Using a gated recurrent unit network prediction model, a sliding window approach is employed to train the preprocessed dynamic load dataset. Based on the current moment and historical responses, the response state of the structure is predicted at multiple future time steps, and the dynamic response prediction sequence H is output. pred ;

[0040] S43. In the process of constructing the input of the gated recurrent unit network prediction model, considering the differences in the importance of different types of load inputs to the structural response in the construction structure, a multidimensional dynamic response attention mechanism is introduced to address the input features x at each time point t. t Each characteristic component Dynamically assign attention weights to construct attention-weighted input features.

[0041]

[0042] in, This represents the feature component of the input feature in the i-th feature dimension. The attention weights represent the influence of the input features on the structural response at time point t, where d is the total dimension of the input features. The score represents the strength score of the influence of the i-th input feature on the hidden state of the structural response at time point t. W represents the strength score of the influence of the p-th input feature on the hidden state of the structural response at time t. e b e These are the weight matrix and bias vector for relevance attention, respectively;

[0043] S44. Input features after attention weighting Replace the original input feature x t The input to the gated recurrent unit network prediction model contains all information related to the update gate control factor z. t Candidate response status All calculations are based on Optimize the response capability and prediction sensitivity of the gated cyclic cell network prediction model to sudden earthquakes or severe wind loads.

[0044] Optionally, S5 includes the following steps:

[0045] S51. Construct the gray wolf optimization parameter space for the key hyperparameter Θ involved in the gated recurrent unit network prediction model, and map each set of parameters to the position X of an individual in the gray wolf population. uThe position is then input into the improved gray wolf optimization algorithm driven by three-dimensional structural response;

[0046] S52. In each iteration, calculate the structural response prediction error corresponding to the u-th gray wolf individual, and define the structural response sensitivity factor SRSF. u =(s u,1 ,s u,2 ,…,s u,d ):

[0047]

[0048] Among them, s u,j Let θ represent the structural response sensitivity factor of the u-th gray wolf individual on the j-th key hyperparameter dimension. u,j f(X) represents the value of the u-th gray wolf in the j-th key hyperparameter dimension. u ) is the structural response prediction error function, used to characterize the degree of influence of different parameter disturbances on prediction accuracy;

[0049] S53. Align the predicted structural response with the actual structural response values ​​in terms of spatial location, construct a spatiotemporal structural error distribution function, and map the spatiotemporal structural error distribution function back to the 3D digital modeling environment M. d In the spatial structural unit, a spatial mapping factor SMF(x,y,z) is constructed to reflect the contribution of response prediction error in different regions of the three-dimensional structure. A response visualization coupling fitness function f'(X) is defined. u ):

[0050]

[0051] Where λ is the spatial error weighting coefficient;

[0052] S54. Based on the error reduction trend in the gray wolf optimization iteration, define the gradient-guided dynamic step size factor γ. o And the structural response sensitivity factor SRSF u Spatial mapping factor (SMF) and gradient-guided dynamic step size factor (γ) o Integrate into the gray wolf location update to build a response-driven, improved gray wolf update mechanism:

[0053]

[0054] Among them, X u (d+1) is the position vector of the u-th gray wolf in the (d+1)-th iteration, representing the updated key hyperparameter combination, X. u (d) is the position vector of the u-th gray wolf individual in the d-th iteration, representing the current key hyperparameter combination, X. hLet h = α, β, δ represent the positions of the current best, second best, and third best individuals, respectively. h = α, β, δ are the three types of leader individuals in the Gray Wolf algorithm, and w represents the indices of the globally best α, second best β, and third best δ individuals, respectively. h The corresponding guiding weight;

[0055] S55. Iterate through S51-S54 until convergence, obtaining the optimal set of key hyperparameters Θ. opt The optimal set of key hyperparameters is applied to the gated recurrent unit network prediction model to construct an optimized GRU prediction model. opt .

[0056] Optionally, S6 includes the following steps:

[0057] S61. The optimized GRU prediction model is used. opt Output structural response prediction sequence H pred Matched to the 3D digital modeling environment M by time index d For the corresponding structural unit in the model, construct the structural response mapping function R(x,y,z,t), which represents the predicted response intensity at time t at the three-dimensional coordinate point (x,y,z).

[0058] S62. Color-encode and deformation-driven conversion are performed on the structural response mapping function R(x,y,z,t) to convert the response intensity value into the dynamic deformation form, vibration frequency visualization and stress intensity heat map in the three-dimensional structural model, and dynamic rendering is performed in combination with the time axis to generate a three-dimensional visualization animation of the structural response.

[0059] S63. Construct a response threshold classification system, set a structural response intensity classification range, and classify the risk level of the response prediction value of each unit in the three-dimensional structural model.

[0060] S64. Based on the three-dimensional visualization animation of structural response and the risk level classification results, construction management decision support information is generated, including the current overall safety status report of the construction structure, the spatial distribution map of the risk level of each structural unit, the visualization early warning prompts and key node response trend predictions, as well as corresponding construction suggestions and scheduling optimization strategies.

[0061] Optionally, the risk level classification rules for the response threshold grading system are as follows:

[0062] Safe zone: The structural unit is classified as a safe zone when the value of the structural response mapping function is less than or equal to the risk classification threshold θ1 of the structural response intensity, and both the unit structural deformation and unit stress are less than 40% of the structural bearing capacity standard.

[0063] Region of interest: The structural unit is classified as the region of interest if the value of the structural response mapping function is greater than θ1 and less than or equal to the risk classification critical value θ2 of the structural response intensity, and the unit structural deformation and unit stress are in the range of 40% to 70% of the structural bearing capacity standard.

[0064] Warning zone: The value of the structural response mapping function is greater than θ2 and less than or equal to the risk classification critical value θ3 of the structural response intensity, and the unit structural deformation and unit stress are in the range of 70% to 90% of the structural bearing capacity standard. The corresponding structural unit is divided into the warning zone, and the system will automatically prompt to suspend construction and generate response optimization suggestions.

[0065] Hazardous Area: When the value of the structural response mapping function is greater than the risk classification threshold θ3 of the structural response intensity, and the unit structural deformation and unit stress exceed 90% of the structural bearing capacity standard, the corresponding structural unit is classified as a hazardous area. The system automatically triggers the early warning mechanism and recommends the implementation of emergency control and construction suspension measures.

[0066] A 3D visualization construction management system, applied to a 3D visualization construction management method, includes:

[0067] The data acquisition module is used to acquire construction project information, build a three-dimensional digital modeling environment for the construction area, and collect data in real time, including wind load, seismic intensity and structural acceleration response, to form a dynamic load dataset.

[0068] The data preprocessing module is used to preprocess the dynamic load dataset and generate a preprocessed dynamic load dataset.

[0069] The structural response prediction module is used to construct a prediction model for a gated cyclic element network using a preprocessed dynamic load dataset.

[0070] The parameter optimization module is used to apply the Grey Wolf optimization algorithm to perform a global search and optimization of the key hyperparameters in the gated recurrent unit network prediction model, so as to obtain the optimized gated recurrent unit network prediction model.

[0071] The 3D visualization simulation module is used to map the structural response prediction results output by the optimized gated cyclic unit network prediction model to the 3D digital modeling environment, thereby realizing the 3D visualization of the structural response.

[0072] The risk identification and decision support module is used to analyze the structural response mapping function of each structural unit based on the 3D visualization results, classify the risk level according to the relative relationship between the response intensity and the structural bearing capacity standard, and generate the risk level assessment results of the construction structure.

[0073] The beneficial effects of this invention are:

[0074] (1) In the gray wolf optimization process, the present invention introduces a structural response sensitivity factor and dynamically weights the gradient sensitivity of each parameter dimension according to the structural response prediction error function. This effectively guides the search direction to converge toward the key hyperparameter dimension that has the greatest impact on prediction accuracy, strengthens the responsive control of the update gate, reset gate, and candidate state core structure, and improves the adaptability and optimization efficiency of the model under nonlinear seismic response conditions.

[0075] (2) This invention defines a structural response mapping function to spatially map and align the multidimensional time series response results output by the gated cyclic unit network with the structural units in the three-dimensional digital modeling environment. Based on stress intensity and deformation amplitude response indicators, it constructs real-time deformation rendering, heat map display and dynamic animation, which significantly enhances the interpretability of the construction structure behavior, breaks through the limitations of the traditional two-dimensional chart display method, realizes the full-time and full-domain response visualization of the structure under seismic load, and improves the engineer's perception and judgment accuracy of complex construction conditions.

[0076] (3) Based on the structural response prediction results, this invention integrates three indicators: response amplitude, unit stress, and structural deformation, and establishes a refined classification standard with four risk levels. It also links with a three-dimensional visualization system to realize dynamic risk classification, distribution map generation, and management suggestion output at the structural unit level. This breaks through the extensive safety control method in traditional construction management that is mainly based on experience thresholds. The system can automatically output construction scheduling suggestions and emergency response measures according to the risk level, which greatly improves the response speed and adaptability of the construction site. It is superior to the existing system in terms of the timeliness of risk identification before and after the occurrence of earthquake disturbance. Attached Figure Description

[0077] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0078] Figure 1 This is a flowchart of a three-dimensional visualization construction management system and method proposed in this invention. Detailed Implementation

[0079] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0080] refer to Figure 1 A three-dimensional visualization construction management method includes the following steps:

[0081] S1. Obtain construction project information and construct a three-dimensional digital modeling environment for the construction area;

[0082] S2. Deploy a sensor network at the construction site to collect real-time data on wind load, seismic intensity, and structural acceleration response, and generate a dynamic load dataset.

[0083] S3. Perform unified data cleaning, noise filtering, missing value imputation and time series reconstruction on the dynamic load dataset to generate a preprocessed dynamic load dataset.

[0084] S4. A gated cyclic element network prediction model is constructed using the preprocessed dynamic load dataset to simulate and predict the dynamic response of the construction structure under seismic loading.

[0085] S5. Apply the Grey Wolf optimization algorithm to perform a global search and optimization of the key hyperparameters of the gated recurrent unit network prediction model to obtain the optimized gated recurrent unit network prediction model.

[0086] S6. Map the structural response prediction results output by the optimized gated cyclic unit network prediction model to the three-dimensional digital modeling environment to realize real-time dynamic simulation and three-dimensional visualization of the construction structure under seismic action. Based on the three-dimensional visualization results, classify the risk level and provide early warning for the dynamic response of the construction structure, and generate construction management decision support information.

[0087] In this embodiment, S1 includes the following steps:

[0088] S11. Obtain the three-dimensional structural information of the construction project. Use three-dimensional laser scanning equipment or UAV remote sensing system to collect point cloud data of the three-dimensional structural shape of the construction project and form a point cloud set P.

[0089] S12. Obtain the geographic environment information of the construction area, including topographic and geomorphological information, geological structure information and surface cover type, and construct an environmental dataset G composed of multiple geographic information units;

[0090] S13. Based on the point cloud set P and the environment dataset G, a three-dimensional digital modeling environment M is constructed by performing multi-source data registration and fusion using digital modeling. d 3D digital modeling environment M d This indicates a function mapping relationship M. d =f(P,G) obtains an integrated 3D model of the building structure and geographic environment, where function f is the spatial matching and feature reconstruction mapping between point cloud data and geographic information units;

[0091] S14. For the completed 3D digital modeling environment M d The topology is organized and meshed to form a computable geometric model for response simulation and visualization rendering, giving the 3D digital modeling environment structural features such as node definition, boundary closure and attribute binding.

[0092] The formation of a computable geometric model for response simulation and visualization rendering specifically includes:

[0093] Topological structure organization involves closing boundaries, removing redundant points, and confirming connectivity of the structural outline generated by point cloud reconstruction, ensuring that each 3D structural unit has clear geometric boundaries, connection surface information, and topological adjacency relationships.

[0094] The node and edge definitions define key points in the structural model as computational nodes, and construct an edge set between every two connected nodes for subsequent physical attribute binding and response path analysis.

[0095] The structural domain partitioning and mesh generation, based on the differences in structural complexity and response characteristics, uses the finite volume method or tetrahedral partitioning algorithm to divide the structure into multiple mesh elements. Each element is a finite substructure, which has the characteristic of being discretizable and solvable.

[0096] Attribute binding: Each structural unit is bound to physical material properties, boundary constraints, and mechanical load transfer paths, providing basic parameters for subsequent response analysis models.

[0097] This implementation method spatially registers and integrates point cloud data with geographic information data to construct a three-dimensional digital modeling environment with node definition and attribute binding. This significantly improves the integrated expression capability between the construction structure and the geographic environment, providing a highly realistic and computable spatial foundation for subsequent seismic response simulation and visualization.

[0098] In this embodiment, S2 includes the following steps:

[0099] S21. Wind load monitoring sensors are installed at key parts of the construction structure. The wind load monitoring sensors collect wind speed, wind direction and wind load data formed by the area of ​​action in real time.

[0100] S22. Deploy ground motion monitoring sensors in the construction area. The ground motion monitoring sensors collect ground motion intensity information in real time. The ground motion intensity information is represented by the horizontal and vertical acceleration components of the ground surface, which constitute ground motion intensity data.

[0101] S23. Install structural response acceleration sensors inside the construction structure. The response acceleration sensors monitor the acceleration response of the structure under earthquake and wind loads and generate structural acceleration response data.

[0102] S24. Perform time synchronization processing and format unification on wind load data, seismic intensity data, and structural acceleration response data, and fuse them to generate a complete dynamic load dataset D. load :

[0103]

[0104] Where t represents the unified time index in the dynamic load dataset, which is the alignment result of the time axis of various sensor data, and v t θ represents the wind speed at time t, used to reflect the intensity of wind load. t The wind direction angle at time point t is used to determine the direction of wind load. A t This represents the area affected by the wind at time point t, and is used to correspond to the size of the stress-bearing area on the structure under wind load. These represent the components of the seismic acceleration in the X, Y, and Z directions at time point t, respectively. Let T be the structural acceleration response data of the p-th monitoring point at time t, where T is the unified set of valid time series.

[0105] This implementation method constructs a structural dynamic load dataset by unifying the time and format of three types of sensor data: wind load, ground motion, and structural response. This provides high-quality input for time series modeling and response prediction, and improves the stability of the data-driven model and the perception accuracy of load state changes.

[0106] In this embodiment, S3 includes the following steps:

[0107] S31. For the dynamic load dataset D load Perform data integrity checks to remove data records with duplicate timestamps, missing fields, or illegal values, thus forming a preliminary cleaned dynamic load dataset.

[0108] S32. Smooth the high-frequency noise data in the preliminary cleaned dynamic load dataset, set the window size, and perform noise smoothing on any numerical field at each time point t. For time points t with missing values ​​in the smoothed data, perform interpolation imputation. Then, perform unified time series reconstruction on the dynamic load dataset after missing value imputation, remap all fields to the standard time axis, and perform linear resampling on the misaligned data points to form a time series dynamic load dataset.

[0109] S35. Normalize all fields in the time-series dynamic load dataset, converting any field to a normalized value to obtain the preprocessed dynamic load dataset D. pre .

[0110] This implementation performs integrity detection, missing data completion, and time series reconstruction on multi-source dynamic load data. It also forms a structured input set through normalization preprocessing, which significantly reduces sensor data noise and redundant interference, and improves the training quality and generalization ability of the prediction model.

[0111] In this embodiment, S4 includes the following steps:

[0112] S41. Using the preprocessed dynamic load dataset D pre Using the input data, a gated cyclic element network prediction model is constructed to predict the dynamic response of constructed structures under seismic loading:

[0113]

[0114] Among them, h t h represents the hidden output state of the gated recurrent unit network prediction model at time t, used to describe the internal response state of the construction structure under seismic external load at time t. t-1 z represents the output hidden state of the gated recurrent unit network prediction model at time point t-1. t To update the gate control factor, This indicates the candidate response status at the current time step. This represents the Hadamard element-wise product operator;

[0115] S42. Using a gated recurrent unit network prediction model, a sliding window approach is employed to train the preprocessed dynamic load dataset. Based on the current moment and historical responses, the response state of the structure is predicted at multiple future time steps, and the dynamic response prediction sequence H is output. pred ;

[0116] S43. In the process of constructing the input of the gated recurrent unit network prediction model, considering the differences in the importance of different types of load inputs to the structural response in the construction structure, a multidimensional dynamic response attention mechanism is introduced to address the input features x at each time point t. t Each characteristic component Dynamically assign attention weights to construct attention-weighted input features.

[0117]

[0118] in, This represents the feature component of the input feature in the i-th feature dimension. The attention weights represent the influence of the input features on the structural response at time point t, where d is the total dimension of the input features. The score represents the strength score of the influence of the i-th input feature on the hidden state of the structural response at time point t. W represents the strength score of the influence of the p-th input feature on the hidden state of the structural response at time t. e b e These are the weight matrix and bias vector for relevance attention, respectively;

[0119] The formula's mechanism enables the model to automatically identify and amplify the most critical input features for seismic response prediction, such as sudden changes in wind speed or Z-axis acceleration. Compared to the original input, attention-weighted input significantly enhances the model's ability to perceive sudden, multi-source dynamic load characteristics, thereby improving the accuracy of response prediction.

[0120] S44. Input features after attention weighting Replace the original input feature x t The input to the gated recurrent unit network prediction model contains all information related to the update gate control factor z. t Candidate response status All calculations are based on Optimize the response capability and prediction sensitivity of the gated cyclic cell network prediction model to sudden earthquakes or severe wind loads.

[0121] This implementation constructs a gated recurrent unit network prediction model with an attention mechanism. The system can assign response weights to different load types, improve the model's ability to identify key structural responses, and enhance the model's prediction accuracy and controllability of structural responses under multiple time-series dynamic loads.

[0122] In this embodiment, S5 includes the following steps:

[0123] S51. Construct the gray wolf optimization parameter space for the key hyperparameter Θ involved in the gated recurrent unit network prediction model, and map each set of parameters to the position X of an individual in the gray wolf population. u The position is then input into the improved gray wolf optimization algorithm driven by three-dimensional structural response;

[0124] S52. In each iteration, calculate the structural response prediction error corresponding to the u-th gray wolf individual, and define the structural response sensitivity factor SRSF. u =(s u,1 ,s u,2 ,…,s u,d ):

[0125]

[0126] Among them, s u,j Let θ represent the structural response sensitivity factor of the u-th gray wolf individual on the j-th key hyperparameter dimension. u,j f(X) represents the value of the u-th gray wolf in the j-th key hyperparameter dimension. u ) is the structural response prediction error function, used to characterize the degree of influence of different parameter disturbances on prediction accuracy;

[0127] The formula quantifies the local variation trend of the model prediction error across various parameter dimensions, giving the Grey Wolf algorithm a differentiated search orientation. Compared to the traditional uniform weight update strategy, this method focuses more on key structural control parameters, improving convergence speed and local optimum escape capability.

[0128] S53. Align the predicted structural response with the actual structural response values ​​in terms of spatial location, construct a spatiotemporal structural error distribution function, and map the spatiotemporal structural error distribution function back to the 3D digital modeling environment M. d In the spatial structural unit, a spatial mapping factor SMF(x,y,z) is constructed to reflect the contribution of response prediction error in different regions of the three-dimensional structure. A response visualization coupling fitness function f'(X) is defined. u ):

[0129]

[0130] Where λ is the spatial error weighting coefficient;

[0131] The formula is an error reconstruction mechanism that introduces the contribution of three-dimensional spatial response. By binding the prediction error to the structural position mapping, the algorithm is guided to prioritize reducing the response uncertainty in areas with high error incidence during the optimization process. The mechanism strengthens the error-sensitive modeling of high-risk structural units, enabling the model to have a higher accuracy in response prediction in earthquake-vulnerable areas.

[0132] S54. Based on the error reduction trend in the gray wolf optimization iteration, define the gradient-guided dynamic step size factor γ. o And the structural response sensitivity factor SRSF u Spatial mapping factor (SMF) and gradient-guided dynamic step size factor (γ) o Integrate into the gray wolf location update to build a response-driven, improved gray wolf update mechanism:

[0133] X u (d+1)=X u (d)+γ o ·∑ h＝α,β,δ w h ·(X h -X u (d))·SRSF u ;

[0134] Among them, X u (d+1) is the position vector of the u-th gray wolf in the (d+1)-th iteration, representing the updated key hyperparameter combination, X. u (d) is the position vector of the u-th gray wolf individual in the d-th iteration, representing the current key hyperparameter combination, X. hLet h = α, β, δ represent the positions of the current best, second best, and third best individuals, respectively. h = α, β, δ are the three types of leader individuals in the Gray Wolf algorithm, and w represents the indices of the globally best α, second best β, and third best δ individuals, respectively. h This corresponds to the guiding weight.

[0135] The update method significantly improves the model's stable convergence performance under nonlinear complex loads and enhances its ability to control the uncertainty of the structural response error space. In simulation experiments, under the same dataset conditions, the prediction performance is improved by 8.5%, and the convergence time is shortened by nearly 25%. It also demonstrates better global exploration and local convergence balance capabilities in multi-objective optimization scenarios.

[0136] S55. Iterate through S51-S54 until convergence, obtaining the optimal set of key hyperparameters Θ. opt The optimal set of key hyperparameters is applied to the gated recurrent unit network prediction model to construct an optimized GRU prediction model. opt .

[0137] This implementation introduces a structural response sensitivity factor during the gray wolf optimization process. Based on the structural response prediction error function, the gradient sensitivity of each parameter dimension is dynamically weighted, effectively guiding the search direction to converge toward the key hyperparameter dimension that has the greatest impact on prediction accuracy. This strengthens the responsive control of the update gate, reset gate, and candidate state core structure, and improves the model's adaptability and optimization efficiency under nonlinear seismic response conditions.

[0138] In this embodiment, S6 includes the following steps:

[0139] S61. The optimized GRU prediction model is used. opt Output structural response prediction sequence H pred Matched to the 3D digital modeling environment M by time index d The corresponding structural unit in the model is used to construct a structural response mapping function R(x,y,z,t), which represents the predicted response intensity at time t at the three-dimensional coordinate point (x,y,z).

[0140] S62. Color-encode and deformation-driven conversion are performed on the structural response mapping function R(x,y,z,t) to convert the response intensity value into the dynamic deformation form, vibration frequency visualization and stress intensity heat map in the three-dimensional structural model, and dynamic rendering is performed in combination with the time axis to generate a three-dimensional visualization animation of the structural response.

[0141] The structural response 3D visualization animation is generated by expanding the structural response mapping function R(x,y,z,t) in continuous time frames, combining vertex displacement-driven mesh deformation and color-coded heatmap expression, and using a time-axis-controlled frame rendering method to generate a structural response visualization animation that can be dynamically played over time.

[0142] S63. Construct a response threshold classification system, set a structural response intensity classification range, and classify the risk level of the response prediction value of each unit in the three-dimensional structural model.

[0143] S64. Based on the three-dimensional visualization animation of structural response and the risk level classification results, construction management decision support information is generated, including the current overall safety status report of the construction structure, the spatial distribution map of the risk level of each structural unit, the visualization early warning prompts and key node response trend predictions, as well as corresponding construction suggestions and scheduling optimization strategies.

[0144] This implementation method defines a structural response mapping function to spatially map and align the multidimensional time-series response results output by the gated cyclic unit network with the structural units in the 3D digital modeling environment. Based on response indicators such as stress intensity and deformation amplitude, it constructs real-time deformation rendering, heat map display, and dynamic animation, which significantly enhances the interpretability of the construction structure behavior, breaks through the limitations of traditional two-dimensional chart display methods, realizes full-time and full-domain visualization of the structure's response under seismic loads, and improves engineers' perception and judgment accuracy of complex construction conditions.

[0145] In this embodiment, the risk level classification rules of the response threshold grading system are as follows:

[0146] Safe zone: The structural unit is classified as a safe zone when the value of the structural response mapping function is less than or equal to the risk classification threshold θ1 of the structural response intensity, and both the unit structural deformation and unit stress are less than 40% of the structural bearing capacity standard.

[0147] Region of interest: The structural unit is classified as the region of interest if the value of the structural response mapping function is greater than θ1 and less than or equal to the risk classification critical value θ2 of the structural response intensity, and the unit structural deformation and unit stress are in the range of 40% to 70% of the structural bearing capacity standard.

[0148] Warning zone: The value of the structural response mapping function is greater than θ2 and less than or equal to the risk classification critical value θ3 of the structural response intensity, and the unit structural deformation and unit stress are in the range of 70% to 90% of the structural bearing capacity standard. The corresponding structural unit is divided into the warning zone, and the system will automatically prompt to suspend construction and generate response optimization suggestions.

[0149] Hazardous Area: When the value of the structural response mapping function is greater than the risk classification threshold θ3 of the structural response intensity, and the unit structural deformation and unit stress exceed 90% of the structural bearing capacity standard, the corresponding structural unit is classified as a hazardous area. The system automatically triggers the early warning mechanism and recommends the implementation of emergency control and construction suspension measures.

[0150] Based on the structural response prediction results, this implementation method integrates three indicators—response amplitude, unit stress, and structural deformation—to establish a refined classification standard with four risk levels. It also links with a three-dimensional visualization system to achieve dynamic risk classification, distribution map generation, and management suggestion output at the structural unit level. This breaks through the extensive safety control method in traditional construction management that relies mainly on experience thresholds. The system can automatically output construction scheduling suggestions and emergency response measures according to the risk level, which greatly improves the response speed and adaptability of the construction site. It is superior to existing systems in terms of the timeliness of risk identification before and after earthquake disturbances.

[0151] A 3D visualization construction management system, applied to a 3D visualization construction management method, includes:

[0152] The data acquisition module is used to acquire construction project information, build a three-dimensional digital modeling environment for the construction area, and collect data in real time, including wind load, seismic intensity and structural acceleration response, to form a dynamic load dataset.

[0153] The data preprocessing module is used to preprocess the dynamic load dataset and generate a preprocessed dynamic load dataset.

[0154] The structural response prediction module is used to construct a prediction model for a gated cyclic element network using a preprocessed dynamic load dataset.

[0155] The parameter optimization module is used to apply the Grey Wolf optimization algorithm to perform a global search and optimization of the key hyperparameters in the gated recurrent unit network prediction model, so as to obtain the optimized gated recurrent unit network prediction model.

[0156] The 3D visualization simulation module is used to map the structural response prediction results output by the optimized gated cyclic unit network prediction model to the 3D digital modeling environment, thereby realizing the 3D visualization of the structural response.

[0157] The risk identification and decision support module is used to analyze the structural response mapping function of each structural unit based on the 3D visualization results, classify the risk level according to the relative relationship between the response intensity and the structural bearing capacity standard, and generate the risk level assessment results of the construction structure.

[0158] Example 1:

[0159] In October 2024, a municipal traffic interchange project located in District A entered the main structural construction phase. Situated in the intersection of a basin's seismic belt, the project area has complex geology and is surrounded by densely populated residential areas and major roads, placing extremely high demands on structural stability and safety during construction. The project's structural design includes four overpass bridges, one semi-underground box culvert, and four high-pier joint structures, with a planned construction period of 14 months. During the foundation construction and main pier formwork erection phases, the construction management team adopted this invention for continuous monitoring, prediction, and management control of the structure at different stages.

[0160] Forty-two multi-dimensional sensor nodes were deployed at the construction site, located in the main piers, bridge piers, steel cages, steel structure installation sections, and foundation slab areas. Sensor types included seismic monitoring instruments (24Hz sampling frequency), multi-directional accelerometers (supporting X / Y / Z triaxial acquisition), wind load locators, and concrete stress sensors. Data acquisition was conducted continuously for 24 hours, with a uniform frequency of once every 10 seconds.

[0161] During the construction of the 3D digital modeling environment, the project team used drones equipped with LiDAR to collect point clouds of the construction structure. This data was then combined with construction drawings to generate a structural point cloud set, and local topographic information was used to construct a 3D geographic information set. A 3D visualization modeling environment was generated through data fusion and topology reconstruction, and structural units were linked to actual construction numbers.

[0162] On the fifth day of construction, a minor earthquake of magnitude 3.4 occurred in the Chengdu area, with the epicenter 22 km from the construction site. The system captured ground motion data in real time, extracting a peak horizontal acceleration of 0.072g and a peak vertical acceleration of 0.034g. The peak structural acceleration response occurred at the main pier number G-07, with a maximum response of 0.109g. Simultaneously, this acceleration data was automatically correlated with the structural mapping function of the construction model, generating a thermo-deformation diagram which was then visualized on the construction management platform.

[0163] The system then initiated training of the gated cyclic unit network structural response prediction model, using earthquake monitoring data, environmental load data, and structural phase response data from the first three months of the project as training samples, totaling 28,800 samples. The input fields included wind speed, wind direction, horizontal and vertical earthquake acceleration, concrete temperature gradient, and structural deformation, while the output fields were structural displacement response and stress response.

[0164] In this invention, the key hyperparameters of the gated recurrent unit network are globally optimized using the Grey Wolf optimization algorithm. During the optimization process, the optimization path is dynamically adjusted based on the structural response gradient sensitivity factor, ultimately converging to an error value of 0.0028, achieving a model accuracy of 94.2%. On the same dataset, using a traditional long short-term memory neural network for training, the optimal model achieves a prediction accuracy of 86.3% with 70 training rounds, while this invention requires only 43 rounds.

[0165] The structural response 3D visualization module maps the predicted response results, dynamically displaying the structural response trend changes over the next 3 minutes, including deformation zone expansion, displacement path trends, and stress variation range. The system automatically marks nodes whose response intensity exceeds the critical value as "warning areas," with G-07 and G-08 added to the warning list. Subsequently, the project construction management platform automatically generates construction suggestions based on this system: postpone concrete pouring time, optimize steel structure support angles, and strengthen temporary support node connections to avoid secondary disaster risks.

[0166] Furthermore, based on the structural risk level assessment results generated by the system, the system was able to accurately identify 12 medium- and high-risk component units. In subsequent post-earthquake investigations and tests, 9 structural units were found to have stress concentration and loose connections, with an early warning identification accuracy rate of 75%, which is significantly better than traditional early warning methods based on experience.

[0167] In summary, this embodiment effectively verifies the accuracy of structural response prediction, the timeliness of risk assessment, and the intelligence of construction management decision support in large and complex construction projects in seismically active areas. It solves the problems of delayed response prediction and inability to link visualization and early warning strategies in traditional methods, and provides a practical and feasible intelligent management method for construction projects in seismic zones.

[0168] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A three-dimensional visualization construction management method, characterized in that, Includes the following steps: S1. Obtain construction project information and construct a three-dimensional digital modeling environment for the construction area; S2. Deploy a sensor network at the construction site to collect real-time data on wind load, seismic intensity, and structural acceleration response, and generate a dynamic load dataset. S3. Perform unified data cleaning, noise filtering, missing value imputation and time series reconstruction on the dynamic load dataset to generate a preprocessed dynamic load dataset. S4. A gated cyclic element network prediction model is constructed using the preprocessed dynamic load dataset to simulate and predict the dynamic response of the construction structure under seismic loading. S5. Apply the Grey Wolf optimization algorithm to perform a global search and optimization of the key hyperparameters of the gated recurrent unit network prediction model to obtain the optimized gated recurrent unit network prediction model. S6. Map the structural response prediction results output by the optimized gated cyclic unit network prediction model to the three-dimensional digital modeling environment to realize real-time dynamic simulation and three-dimensional visualization of the construction structure under seismic action. Based on the three-dimensional visualization results, classify the risk level and provide early warning for the dynamic response of the construction structure, and generate construction management decision support information. S4 includes the following steps: S41. Using the preprocessed dynamic load dataset Using the input data, a gated cyclic element network prediction model is constructed to predict the dynamic response of constructed structures under seismic loading: in, This indicates that the gated recurrent unit network prediction model is at time point The output hidden state is used to describe the construction structure at a given time point. The internal response state under the action of external seismic load. This indicates that the gated recurrent unit network prediction model is at time point The output of the hidden state To update the gate control factor, This indicates the candidate response status at the current time step. This represents the Hadamard element-wise product operator; S42. Using a gated recurrent unit network prediction model, a sliding window approach is employed to train the preprocessed dynamic load dataset. Based on the current moment and historical responses, the response state of the structure is predicted at multiple future time steps, and the dynamic response prediction sequence of the structure is output. ; S43. In the process of constructing the gated cyclic unit network prediction model, considering the differences in the importance of different types of load inputs to the structural response in the construction structure, a multidimensional dynamic response attention mechanism is introduced to address the input at each time point. Input features Each characteristic component Dynamically assign attention weights to construct attention-weighted input features. : in, Indicates the input feature at the th Feature components in each feature dimension Indicates the input features at time points Attention weights for the impact on structural response. The total dimension of the input features. Indicates a point in time Next The strength score of the influence of each input feature on the hidden state of the structural response. Indicates a point in time Next The strength score of the influence of each input feature on the hidden state of the structural response. , These are the weight matrix and bias vector for relevance attention, respectively; S44. Input features after attention weighting Replace the original input features Input into the gated recurrent unit network prediction model, all information related to updating the gate control factor Candidate response status All calculations are based on To optimize the response capability and prediction sensitivity of the gated cyclic unit network prediction model to sudden earthquakes or severe wind loads; S5 includes the following steps: S51. Key Hyperparameters Involved in Constructing the Prediction Model of Gated Recurrent Unit Network The gray wolf optimization parameter space maps each parameter combination to the location of an individual in the gray wolf population. The position is then input into the improved gray wolf optimization algorithm driven by three-dimensional structural response; S52. In each iteration, calculate the... The structural response prediction error is defined for each individual gray wolf, and a structural response sensitivity factor is defined. : in, This represents the structural response sensitivity factor of the u-th gray wolf individual on the j-th key hyperparameter dimension. Indicates the first The gray wolf in the first The values ​​of each key hyperparameter dimension This is the structural response prediction error function, used to characterize the degree of influence of different parameter disturbances on prediction accuracy; S53. Align the predicted structural response with the actual structural response values ​​in terms of spatial location, construct a spatiotemporal structural error distribution function, and map the spatiotemporal structural error distribution function back to the 3D digital modeling environment. Spatial structural units in the space are used to construct spatial mapping factors. This reflects the contribution of response prediction errors to different regions of the three-dimensional structure, and defines a response visualization coupled fitness function. : in, These are the spatial error weighting coefficients; S54. Based on the error reduction trend in the gray wolf optimization iteration, define a gradient-guided dynamic step size factor. and the structural response sensitivity factor Spatial mapping factor With gradient-guided dynamic step size factor Integrate into the gray wolf location update to build a response-driven, improved gray wolf update mechanism: ； in, Let be the position vector of the u-th gray wolf individual in the (d+1)-th iteration, representing the updated combination of key hyperparameters. Let be the position vector of the u-th gray wolf individual in the d-th iteration, representing the current key hyperparameter combination. These represent the positions of the current best, second best, and third best individuals, respectively. In the Grey Wolf Algorithm, there are three types of leader individuals, each representing the globally optimal... Second best and the third best Individual index, The corresponding guiding weight; S55. Iterate through S51-S54 until convergence, obtaining the optimal set of key hyperparameters. The optimal set of key hyperparameters is applied to the gated recurrent unit network prediction model to construct an optimized gated recurrent unit network prediction model. .

2. The three-dimensional visualization construction management method according to claim 1, characterized in that, S1 includes the following steps: S11. Obtain the three-dimensional structural information of the construction project. Use three-dimensional laser scanning equipment or UAV remote sensing system to collect point cloud data of the three-dimensional structural shape of the construction project and form a point cloud set. ; S12. Obtain the geographic environment information of the construction area, including topographic and geomorphological information, geological structure information and surface cover type, and construct an environmental dataset G composed of multiple geographic information units; S13. Based on point cloud collections With environmental datasets A 3D digital modeling environment is constructed by registering and fusing multi-source data using digital modeling. 3D digital modeling environment This indicates a mapping relationship through functions. The resulting integrated 3D model of the building structure and geographical environment, function For spatial matching and feature reconstruction mapping between point cloud data and geographic information units; S14. The completed 3D digital modeling environment The topology is organized and meshed to form a computable geometric model for response simulation and visualization rendering, giving the 3D digital modeling environment structural features such as node definition, boundary closure and attribute binding.

3. The three-dimensional visualization construction management method according to claim 2, characterized in that, S2 includes the following steps: S21. Wind load monitoring sensors are installed at key parts of the construction structure. The wind load monitoring sensors collect wind speed, wind direction and wind load data formed by the area of ​​action in real time. S22. Deploy ground motion monitoring sensors in the construction area. The ground motion monitoring sensors collect ground motion intensity information in real time. The ground motion intensity information is represented by the horizontal and vertical acceleration components of the ground surface, which constitute ground motion intensity data. S23. Install structural response acceleration sensors inside the construction structure. The response acceleration sensors monitor the acceleration response of the structure under earthquake and wind loads and generate structural acceleration response data. S24. Perform time synchronization processing and format unification on wind load data, seismic intensity data, and structural acceleration response data, and fuse them to generate a complete dynamic load dataset. : in, This represents a unified time index in the dynamic load dataset, indicating the alignment of the time axes for various sensor data. Indicates a point in time The wind speed is used to represent the intensity of the wind load. Indicates a point in time The wind direction angle is used to determine the direction of wind load. Indicates a point in time The area affected by the wind is used to determine the size of the stress-bearing area on the structure corresponding to the wind load. Representing time points The components of seismic acceleration in the X, Y, and Z directions. For the first Each monitoring point at the time point The structural acceleration response data is given above, where T is the unified set of effective time series.

4. The three-dimensional visualization construction management method according to claim 3, characterized in that, S3 includes the following steps: S31. Dynamic load dataset Perform data integrity checks to remove data records with duplicate timestamps, missing fields, or illegal values, thus forming a preliminary cleaned dynamic load dataset. S32. Smooth the high-frequency noise data in the preliminary cleaned dynamic load dataset, set the window size, and smooth the data at each time point. Noise smoothing is performed on any numerical field, interpolation is performed to fill in the missing values ​​at time points t in the smoothed data, and a unified time series reconstruction is performed on the dynamic load dataset after missing value filling. All fields are remapped to the standard time axis, and linear resampling is performed on the unaligned data points to form a time series dynamic load dataset. S35. Normalize all fields in the time-series dynamic load dataset, converting any field to a normalized value to obtain the preprocessed dynamic load dataset. .

5. The three-dimensional visualization construction management method according to claim 4, characterized in that, S6 includes the following steps: S61. The optimized gated recurrent unit network prediction model Output structural response prediction sequence Matched to the 3D digital modeling environment by time index Construct the structural response mapping function based on the corresponding structural units in the code. , indicating the three-dimensional coordinate point Above the time point The predicted response intensity; S62. Structural response mapping function Color coding and deformation-driven conversion are performed to transform the response intensity values ​​into dynamic deformation morphology, vibration frequency visualization, and stress intensity heatmap in the three-dimensional structural model. Combined with the time axis, dynamic rendering is performed to generate a three-dimensional visualization animation of the structural response. S63. Construct a response threshold classification system, set a structural response intensity classification range, and classify the risk level of the response prediction value of each unit in the three-dimensional structural model. S64. Based on the three-dimensional visualization animation of structural response and the risk level classification results, construction management decision support information is generated, including the current overall safety status report of the construction structure, the spatial distribution map of the risk level of each structural unit, the visualization early warning prompts and key node response trend predictions, as well as corresponding construction suggestions and scheduling optimization strategies.

6. The three-dimensional visualization construction management method according to claim 5, characterized in that, The risk level classification rules of the response threshold grading system are as follows: Safe zone: The value of the structural response mapping function is less than or equal to the risk classification threshold of the structural response intensity. Furthermore, the unit structural deformation and unit stress are both lower than 40% of the structural bearing capacity standard, and the corresponding structural units are divided into safe zones. Region of interest: The value of the structural response mapping function is greater than And less than or equal to the risk classification threshold of structural response intensity Furthermore, the unit structural deformation and unit stress are within the range of 40% to 70% of the structural bearing capacity standard, and the corresponding structural units are divided into areas of interest. Warning zone: The value of the structural response mapping function is greater than... And less than or equal to the risk classification threshold of structural response intensity Furthermore, if the unit structural deformation and unit stress are within 70% to 90% of the structural bearing capacity standard, the corresponding structural unit is divided into an early warning area, and the system automatically prompts to suspend construction and generates response optimization suggestions. Hazardous area: The value of the structural response mapping function is greater than the risk classification threshold of the structural response intensity. Furthermore, if the unit structural deformation and unit stress exceed 90% of the structural bearing capacity standard, the corresponding structural unit is classified as a dangerous area, and the system automatically triggers an early warning mechanism and recommends the implementation of emergency control and construction suspension measures.

7. A three-dimensional visualization construction management system, applied to the three-dimensional visualization construction management method according to any one of claims 1-6, characterized in that, include: The data acquisition module is used to acquire construction project information, build a three-dimensional digital modeling environment for the construction area, and collect wind load data, seismic intensity data and structural acceleration response data in real time, and form a dynamic load dataset. The data preprocessing module is used to preprocess the dynamic load dataset and generate a preprocessed dynamic load dataset. The structural response prediction module is used to construct a prediction model for a gated cyclic element network using a preprocessed dynamic load dataset. The parameter optimization module is used to apply the Grey Wolf optimization algorithm to perform a global search and optimization of the key hyperparameters in the gated recurrent unit network prediction model, so as to obtain the optimized gated recurrent unit network prediction model. The 3D visualization simulation module is used to map the structural response prediction results output by the optimized gated cyclic unit network prediction model to the 3D digital modeling environment, thereby realizing the 3D visualization of the structural response. The risk identification and decision support module is used to analyze the structural response mapping function of each structural unit based on the 3D visualization results, classify the risk level according to the relative relationship between the response intensity and the structural bearing capacity standard, and generate the risk level assessment results of the construction structure.

Images