BIM (Building Information Modeling)-based nonlinear multi-gradient terrace building construction collaborative management system

Abstract

The invention provides a BIM (Building Information Modeling)-based nonlinear multi-gradient terrace building construction collaborative management system, which comprises a data acquisition and modeling module for acquiring construction data, generating a dynamic data set through spatial registration and cleaning, constructing a BIM model, and mapping dynamic data to a corresponding position of the model; the intelligent analysis module is used for generating a construction index predicted value based on a preset prediction model and calculating the deviation between the predicted value and the measured value relative to the reference value; when the deviation exceeds a threshold value, generating a compensation parameter set by optimizing a curved surface weight coefficient; and the collaborative execution module is used for updating the BIM model and the reference value according to the compensation parameter, adaptively segmenting the construction unit based on the gradient change rate, planning an equipment path in combination with curvature constraint, and optimizing the procedure dependency relationship according to the gradient difference and the concrete initial setting time to realize parallel construction control. According to the system, the BIM model is driven to be dynamically updated through real-time data, and the construction quality and efficiency are improved.

Description

Technical Field

[0001] This invention relates to the cross-technical field of intelligent construction technology and BIM application, specifically to a BIM-based nonlinear multi-gradient floor construction collaborative management system. Background Technology

[0002] Currently, in the construction management of nonlinear multi-gradient ground structures, BIM (Building Information Modeling) technology is mainly used as a static design verification tool. However, once the model is built, it lacks the ability to interact in real-time with dynamic data from the construction site. Traditional methods struggle to effectively integrate heterogeneous data from multiple sources, such as point clouds, sensors, and equipment poses collected during construction. This results in the model failing to accurately reflect the actual working conditions during concrete pouring, leveling, and hardening. Especially in areas with complex slope variations or uneven load distribution, the deviation between the design model and the actual construction conditions accumulates continuously. Existing systems lack proactive prediction and compensation mechanisms for dynamic deviations, leading to quality problems such as inaccurate elevation control and deviations in surface shape from design requirements.

[0003] On the other hand, the unit segmentation, equipment path planning, and process coordination in floor construction rely heavily on manual experience, making it difficult to adapt to the dynamic adjustment requirements of nonlinear surfaces. Existing technologies lack adaptive granular control for dividing construction units in areas with abrupt changes in slope rate, easily leading to conflicts between equipment travel paths and surface curvature. Simultaneously, the constraints of concrete initial setting characteristics and slope differences on process logic are not quantified and incorporated into scheduling decisions, limiting the efficiency of parallel construction. This static management model struggles to achieve closed-loop control of construction quality deviations and cannot support data-driven dynamic collaborative optimization. Summary of the Invention

[0004] In order to solve the technical problems mentioned in the background art, the purpose of this invention is to provide a BIM-based nonlinear multi-gradient floor construction collaborative management system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A BIM-based collaborative management system for nonlinear multi-gradient floor construction includes: Data acquisition and modeling module: Collects real-time construction data of the floor structure, including point cloud data, sensor data, and pose data of construction equipment; performs spatial registration and cleaning processing on the real-time construction data to generate a dynamic dataset; constructs a BIM model of the multi-gradient floor structure; and maps the dynamic dataset to the corresponding position of the BIM model based on the first reference value of the gradient parameter in the BIM model. Intelligent analysis module: Input the mapped dynamic dataset into the preset prediction model to generate predicted values ​​of floor construction indicators, and calculate the prediction deviation and real-time deviation values ​​of the predicted and measured values ​​based on the first benchmark value. When the prediction deviation or real-time deviation value exceeds the preset construction indicator threshold, calculate the compensation amount of the gradient parameter and generate a compensation parameter set. Collaborative execution module: updates the BIM model based on the compensation parameter set, generates a second benchmark value based on the updated BIM model, and performs constructable unit segmentation, equipment path planning, and process optimization based on the updated BIM model and the second benchmark value.

[0006] Furthermore, the spatial registration aims to unify point cloud data, sensor data, and device pose data into a preset global coordinate system.

[0007] Furthermore, the steps for constructing the BIM model are as follows: 1) Use the preset geometric surface as the geometric expression tool of the BIM model. The geometric surface refers to the non-uniform rational B-spline surface, and use the global coordinate system as the coordinate system of the BIM model. 2) Map the parameter domain in the geometric surface to a three-dimensional surface in physical space. The physical space refers to the physical space defined based on the global coordinate system, and the points in the parameter domain are parameter coordinates. The shape of the geometric surface is determined by the surface construction point grid and weight coefficients. The functional expression formula for the geometric surface is as follows: in, Let be the three-dimensional coordinates of the geometric surface at the parametric coordinates (u,v). and These represent the number of rows and columns of the point mesh used to construct the surface, respectively. The weight coefficient is the weight coefficient of the m-th row and n-th column. Let be the coordinates of the construction point in the m-th row and n-th column; 3) Transform the construction design requirements into gradient parameters of the geometric surface, including elevation level, slope change rate, and load bearing capacity; The elevation level is the Z-axis coordinate of the construction point; The slope change rate is at the construction point and Partial derivatives in the direction; Load bearing capacity is an attribute field associated with the construction point, including concrete thickness, reinforcement ratio, and material strength parameters.

[0008] Furthermore, the first benchmark value includes an elevation level benchmark value, a slope change rate benchmark value, and a load bearing capacity benchmark value; The calculation steps for the elevation benchmark value are as follows: 1) Within the parameter domain of the geometric surface, generate the coordinates of sampling points at fixed intervals; 2) Substitute the coordinates of each sampling point into the geometric surface function to calculate its corresponding Z-axis coordinate value in physical space; 3) Obtain the elevation benchmark value by taking the arithmetic mean of the Z coordinate values ​​of all sampling points. ; The slope change rate benchmark value By calculating geometric surfaces and The maximum gradient of the direction is obtained; The load bearing capacity benchmark value Based on concrete compressive strength and effective thickness The formula for calculation using the limit state design method is as follows: in, This is the load distribution factor.

[0009] Furthermore, the mapping of the dynamic dataset to the BIM model includes inverse calculation of parameter coordinates and attribute association; The parameter coordinate inverse calculation is based on Newton's iteration method to inversely calculate the parameter coordinates of each data point in the dynamic dataset; The attribute association binds the geometric data in the dynamic dataset to the geometric position of the parametric coordinates; and associates the physical data in the dynamic dataset with the attribute fields in the corresponding parametric coordinates in the BIM model.

[0010] Furthermore, the prediction deviation value The calculation formula is as follows: in, , and These are the deviation values ​​for elevation level, slope change rate, and predicted load bearing capacity, respectively. , and These are the predicted values ​​of elevation level, slope change rate, and load bearing capacity at time t, respectively. Real-time deviation value The calculation formula is as follows: in, , and These are the real-time deviations of elevation level, slope change rate, and load bearing capacity, respectively. , and These represent the actual values ​​of elevation level, slope change rate, and load bearing capacity at time t, respectively.

[0011] Furthermore, when the prediction deviation value or real-time deviation value Any value exceeding the preset construction indicator threshold The generation of the compensation parameter set is triggered at certain times. The specific steps are as follows: 1) Combine the prediction deviation with the real-time deviation to form 2) Weighting coefficients for points constructed using BIM model surfaces To optimize the variables, an objective function is established: in, Adjustment amount for weighting coefficients The changes in the predicted and real-time deviations caused by the error are obtained iteratively using the damped Gauss-Newton method. 3) The final Generate a compensation parameter set by associating it with the coordinates of the out-of-limit location: in, This is a subset of construction points affected by the overlimit event.

[0012] Furthermore, the steps for updating the BIM model are as follows: Based on the compensation parameter set The weight coefficients of the construction points in the BIM model are adjusted using the following formula: in, These are the adjusted weighting coefficients. These are the original weighting coefficients. This is the adjustment amount for the weighting coefficient; The geometric surface function is recalculated based on the adjusted weighting coefficients, and the gradient parameters are dynamically corrected. The second baseline value is generated by recalculating the baseline value based on the updated BIM model.

[0013] Furthermore, the specific steps for dividing the construction unit are as follows: In the parameter domain, the partial derivative magnitude of the geometric surface is used as the slope change rate, and a non-uniform grid is divided based on the slope change rate. The parameter domain is divided into a first region and a second region based on a preset rate of change threshold. The region where the slope change rate is greater than the change rate threshold is the first region, and the grid cells are generated using fine-grained segmentation parameter domains. The region where the slope change rate is greater than the change rate threshold is the second region, and the grid cells are generated using coarse-grained segmentation parameter domains. For the subset of construction points affected by the over-limit event, compensation is performed by expanding the grid outward by twice the grid step size with its parameter coordinates as the center, and forced fine-grained segmentation is performed to ensure compensation accuracy. The mesh element is defined by a rectangular mesh divided in the parameter domain, and its boundary is determined by the parameter coordinate range. and Determine and map the three-dimensional curved surface region of the physical space; The steps for device path planning are as follows: Based on the boundaries of the grid cells, the travel path of the construction equipment is planned; Along the parameter domain or Direction generates the main path, and the device's turning radius. Curvature constraints must be met: in, The preset equipment safety factor; If the path conflicts with the pose data of unconstructed areas or construction equipment, the A* algorithm is used to replan in the parameter domain, with Manhattan distance as the cost function; finally, a sequence of equipment paths is generated, each path containing the start and end coordinates and travel direction angle of the grid cell; The specific steps for optimizing the process are as follows: If the slope change rate of adjacent grid cells differs If the construction process depends on the slope, then the construction should proceed with the work on the slope with the gentlest slope. If the grid unit Concrete pouring completion time and Planned start time satisfy ,in If the initial setting time of the concrete is taken into account, then the process dependency is eliminated, and parallel construction is allowed.

[0014] Furthermore, the prediction model is constructed based on a Long Short-Term Memory (LSTM) network and uses mean squared error as the loss function.

[0015] Compared with the prior art, the advantages of the present invention are as follows: 1. This invention uses an updated BIM model to intelligently divide construction units based on the slope change rate, plans equipment safety paths in conjunction with surface curvature constraints, and optimizes process logic based on slope differences and concrete initial setting time, significantly improving the adaptability of nonlinear curved surface construction. 2. This invention utilizes an LSTM prediction model to predict construction indicators in advance, and proactively triggers a compensation mechanism before measured deviations occur, achieving dual prevention and control of "predicted deviations and real-time deviations". It shifts from passive correction to proactive prevention, effectively suppressing the accumulation of quality defects. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the system workflow of the present invention; Figure 2 This is a schematic diagram of the spatial registration workflow of the present invention; Figure 3 This is a schematic diagram illustrating the generation of the reference value in this invention. Detailed Implementation

[0018] To achieve the above objectives, this invention provides a BIM-based nonlinear multi-gradient floor construction collaborative management system. Please refer to [link to relevant documentation]. Figure 1 The system includes: Data acquisition and modeling module: Collects real-time construction data of the floor structure, including point cloud data, sensor data, and pose data of construction equipment; performs spatial registration and cleaning processing on the real-time construction data to generate a dynamic dataset; constructs a BIM model of the multi-gradient floor structure; and maps the dynamic dataset to the corresponding position of the BIM model based on the first reference value of the gradient parameter in the BIM model. The real-time construction data collection process is as follows: Point cloud data is collected by fixed or mobile 3D laser scanners (such as ground stations, station racks, or scanners mounted on mobile robots) deployed at the construction site to scan the floor construction area. Sensor data is collected by embedding a sensor network in the foundation layer or the concrete layer before initial setting. The sensors include: Tilt sensor: Monitors slope changes and levelness in localized areas of the ground; Stress sensors: monitor the internal stress state of concrete during the hardening process and the load distribution in the future load-bearing area (construction period and early curing period). Temperature and humidity sensors: monitor the temperature and humidity inside the concrete casting and the surrounding environment to provide a basis for analyzing changes in material properties (such as setting time and strength development); Strain gauges: used to monitor deformation in critical areas or complex gradient transition zones; Coordinates are measured using a total station and sensors. Pose data is obtained in real time by adding positioning (such as RTK-GNSS) and attitude sensing (IMU) devices to construction equipment (such as laser leveling machines, pavers, and compaction equipment); Data synchronization: The time source of all acquisition devices is synchronized through an NTP server deployed on site to ensure that point cloud data, sensor data and pose data have a unified timestamp reference.

[0019] Spatial registration aims to unify point cloud data, sensor data, and device pose data into a preset global coordinate system (defined as the origin). , and The axis is parallel to the boundary of the construction area. (The axis is vertically upward), the specific steps are as follows, please refer to [link / reference]. Figure 2 : 1) Use a total station to set up at least 3 reference control points (coordinate accuracy ≤ ±2mm) on the construction site based on the global coordinate system to form a global coordinate system reference network; The coordinates of the reference control point are marked as , serving as the absolute reference for all data registration; The coordinates of the reference control points include GNSS coordinates. and global coordinates ; in , and These are longitude, latitude, and elevation, respectively. 2) Register the point cloud data to the global coordinate system using the Iterative Closest Point (ICP) algorithm: in, For rotation matrix, It is a translation vector. For the point cloud data j The coordinates of the points The coordinates of the control point closest to the j-th point are... This represents the total number of coordinate points in the point cloud data. The accuracy of registration is verified based on the root mean square (RMSE) metric, requiring the RMSE to be less than or equal to... ; 3) Transfer sensor data (tilt angle) ,stress Temperature and humidity and ,strain This is associated with spatial location to form a data packet. 4) Pose data is transformed to the global coordinate system using the Bursa-Bursa seven-parameter method: in, , and These are the coordinates in the global coordinate system. This is the scale factor (obtained by inverse calculation of the baseline control points). For rotation matrix (based on) , and (The three basic rotation matrices of the axis are combined.) The difference in longitude between the longitude of the construction equipment and the longitude of the origin. The difference in latitude between the construction equipment latitude and the origin latitude. This refers to the difference in elevation between the construction equipment elevation and the origin elevation. , and The translation amount is obtained by combining least squares adjustment based on the origin and global coordinates. The origin is calculated using a weighted average method based on the GNSS coordinates of multiple reference control points. The attitude angles in the pose data are directly correlated to the coordinates of the center point of the construction equipment. Rotation matrix based on , and The three basic rotation matrices of the axes are combined, for example, the rotation matrix of the X-axis. : in, for X The rotation angle of the axis (the pitch deviation angle between the X-axis of the global coordinate system and the corresponding axis of the GNSS coordinate system).

[0020] The specific steps for cleaning are as follows: 1) Remove outliers from point cloud data based on a statistical outlier removal algorithm; 2) Replace tilt angle values ​​that are outside the effective range with the effective tilt angle value of the previous moment (the time difference must be ≤ 1 second) based on the preset effective tilt angle range; Stress values ​​that exceed the preset effective stress range are marked as invalid values. Based on the preset effective temperature and humidity range, temperature and humidity values ​​that exceed the effective temperature and humidity range are eliminated, and linear interpolation compensation is performed based on the nearest effective temperature and humidity values. Based on the elastic modulus of concrete Calculate theoretical strain data: in, The measured stress value If theory Data that is flagged as suspicious is removed.

[0021] The steps for building a BIM model are as follows: 1) Use the preset geometric surface as the geometric expression tool of the BIM model. The geometric surface refers to the non-uniform rational B-spline surface, and use the global coordinate system as the coordinate system of the BIM model. 2) Map the parameter domain ((u,v)∈[0,1]×[0,1]) in the geometric surface to the three-dimensional surface in the physical space. The points in the parameter domain are parameter coordinates, which are used to define the parameterized position identifier of any position on the surface. For example, (u,v)=(0.5,0.5) corresponds to the center point of the surface parameter domain. The shape of a geometric surface is determined by the following two key parameters: Curved surface construction point grid: Three-dimensional spatial points arranged in a two-dimensional grid, derived from elevation points and key points of slope change in the construction design drawings; Weighting coefficient: A value associated with each control point, used to adjust the local shape of the surface. The weighting coefficient is determined based on the design values ​​of concrete thickness and reinforcement ratio (e.g., the weight is increased in high-load areas). The functional expression formula for a geometric surface is as follows: in, For geometric surfaces in parametric coordinates (u,v) The three-dimensional coordinates of the location (i.e., the actual coordinates of the point on the ground surface). and These represent the number of rows and columns of the surface construction point mesh, respectively, determined by the number of construction points in the design drawings (e.g., ...). × =10×10 grid). The weight coefficient is the weight coefficient of the m-th row and n-th column. For the first m Line number nThe coordinates of the column's construction point; 3) Transform construction design requirements into gradient parameters of geometric surfaces, including elevation level, slope change rate, and load bearing capacity; The elevation level is the Z-axis coordinate of the construction point; The slope change rate is at the construction point and Partial derivatives in the direction and , used to control the curvature of geometric surfaces; Load bearing capacity is an attribute field associated with the construction point, including concrete thickness, reinforcement ratio, and material strength parameters (such as C30 / C40 concrete grade).

[0022] Based on the gradient parameters of the BIM model, a first reference value is defined in the global coordinate system, including the elevation level reference value, the slope change rate reference value, and the load bearing capacity reference value. Please refer to [link / reference]. Figure 3 ; The steps for calculating the elevation benchmark value are as follows: 1) Within the parameter domain of the geometric surface, at fixed intervals (e.g. , Generate sampling point coordinates The sampling density is determined based on the construction accuracy requirements (such as areas with drastic slope changes requiring more frequent sampling). 2) Set the coordinates of each sampling point Substitute into geometric surface function Calculate its corresponding physical space Z Axis coordinate values: in, For surface functions Z Axis coordinate values; 3) Take the arithmetic mean of the Z coordinate values ​​of all sampling points to obtain the elevation benchmark value. The formula is as follows: in, This represents the total number of sampling points.

[0023] Slope change rate benchmark value By calculating geometric surfaces and The maximum gradient in the direction is obtained using the following formula: in, and Geometric surfaces in and Partial derivatives in the direction (in vector form) , ); Load bearing capacity benchmark value Based on concrete compressive strength and effective thickness The formula for calculation using the limit state design method is as follows: in, This is the load distribution factor (determined by design specifications).

[0024] Mapping dynamic datasets to BIM models includes parametric coordinate inversion and attribute association; Parametric coordinates are derived by using each data point in the dynamic dataset. G (The smallest unit in a dynamic dataset, i.e., independent data with spatial coordinates collected by sensors, scanning devices, or positioning devices) The specific steps for inversely calculating the parameter coordinates are as follows: The parameter coordinates are solved using Newton's iteration method, as shown in the following formula: in, The coordinates of the parameters to be solved; Attribute association binds geometric data (e.g., point clouds and device poses) in a dynamic dataset to the geometric location of parametric coordinates; it associates physical data (e.g., tilt angle, stress, and temperature and humidity) with attribute fields in the corresponding parametric coordinates of the BIM model (e.g., tilt angle values ​​are written to the slope attribute set).

[0025] Intelligent analysis module: Input the mapped dynamic dataset into the preset prediction model to generate predicted values ​​of floor construction indicators, and calculate the prediction deviation and real-time deviation values ​​of the predicted and measured values ​​based on the first benchmark value. When the prediction deviation or real-time deviation value exceeds the preset construction indicator threshold, calculate the compensation amount of the gradient parameter and generate a compensation parameter set. A prediction model is built based on a Long Short-Term Memory (LSTM) network; Set the window length to (For example, (each sampling period) outputs the future (For example, The predicted values ​​of construction indicators (for each cycle) include the predicted values ​​of elevation level, slope change rate, and load bearing capacity. The core of the Long Short-Term Memory (LSTM) network is the cell state, a vertical state vector spanning the time series, which undertakes the function of long-term memory storage. Its design uses three gating mechanisms to coordinate and regulate the information flow: Forget gate: Based on the current input vector and the hidden state of the previous time step, a forget gate vector is generated through a fully connected layer and a sigmoid activation function. This vector acts on the previous cell state and determines the degree of retention of each dimension (0 means completely discarded, 1 means completely retained). The input gate consists of two parts: Gating vector generation: Similar to the forget gate, it generates an input gating vector in the interval [0,1] to filter the new information to be written; Candidate memory generation: A candidate memory cell content vector (within the range of [-1,1]) is generated through a fully connected layer and a tanh activation function. This vector contains potential new information after fusing the current input and the previous hidden state. Output gate: Generates an output gating vector in the interval [0,1] using the current input and the previous hidden state, which controls the exposure range of the updated cell state; The prediction model is trained using historical construction data, with the mean squared error (MSE) used as the loss function to optimize the weights, as shown in the following formula: in, The total number of training samples. The length of the time window, For the first t Time of the first a The true value of similar construction indicators, For the first t Time of the first a Predicted values ​​for construction indicators.

[0026] Prediction deviation value The calculation formula is as follows: in, , and These are the predicted deviations for elevation level, slope change rate, and load bearing capacity, respectively. , and They are time t The predicted values ​​of elevation level, slope change rate, and load bearing capacity at that time; Real-time deviation value The calculation formula is as follows: in, , and These are the real-time deviations of elevation level, slope change rate, and load bearing capacity, respectively. , and They are time t The actual values ​​of elevation level, slope change rate and load bearing capacity at that time (obtained based on sensor data). When the prediction deviation value or real-time deviation value Any value exceeding the preset construction indicator threshold ( , and The generation of compensation parameter sets is triggered when the corresponding elevation threshold, slope threshold, and load bearing threshold are set respectively. The specific steps are as follows: Elevation threshold Based on the requirements for floor flatness in the "Code for Design of Building Ground" (e.g., 3mm / 2m), and combined with the accuracy setting of the laser leveling machine (±1.5mm), a 20% safety redundancy is reserved; Slope threshold The design is set based on the dual constraints of the minimum slope of the drainage design (e.g., 0.5%) and the maximum curvature of the surface construction point (the critical curvature to prevent concrete cracking). Load bearing threshold Based on the structural design safety factor, the minimum value of the design load and the converted value of the 28-day standard compressive strength of concrete is taken as the load bearing threshold. 1) Combine the prediction bias and the real-time bias as follows: 2) Weighting coefficients for points constructed using BIM model surfaces To optimize the variables, an objective function is established: in, Adjustment amount for weighting coefficients The changes in the predicted and real-time deviations caused by these deviations are solved iteratively through the following steps: initialization ( (zero-dimensional vector); Calculate the deviation Jacobian matrix Based on the gradient parameters of the BIM model Sensitivity analysis was conducted to obtain the results. Iterative updates: Where l is the number of iterations. It is the pseudo-inverse of the Jacobian matrix; when Or it may terminate when the preset maximum number of iterations is reached; 3) The final Generate a compensation parameter set by associating it with the coordinates of the out-of-limit location: in, A subset of construction points affected by the over-limit event (e.g., the area surrounding the slope over-limit zone). (Construction points within the mesh).

[0027] Collaborative execution module: updates the BIM model based on the compensation parameter set, generates a second benchmark value based on the updated BIM model, and performs constructable unit segmentation, equipment path planning, and process optimization based on the updated BIM model and the second benchmark value; Gradient parameter correction is based on the compensation parameter set Adjust the weight coefficients of construction points in the BIM model using the following formula: in, These are the adjusted weighting coefficients. These are the original weighting coefficients. This is the adjustment amount for the weighting coefficient; Recalculate the geometric surface function based on the adjusted weighting coefficients. This allows for dynamic correction of gradient parameters; A second baseline value is generated by recalculating the baseline value based on the updated BIM model; Based on the updated BIM model and the second benchmark, constructable unit segmentation, equipment path planning, and process optimization are performed. The specific steps for dividing the construction unit are as follows: In the parameter domain Above, the partial derivatives of the geometric surface and The modulus length is used as the slope change rate, and a non-uniform grid is divided based on the slope change rate. Based on a preset rate of change threshold (Based on construction safety, material properties, and equipment performance settings) the parameter domain is divided into a first region and a second region; satisfy or The first region is divided using fine-grained segmentation (e.g., mesh step size). )Parameter domain generates mesh elements; satisfy and The region is the second region, which is segmented using coarse-grained methods (e.g., )Parameter domain generates mesh elements; For the subset of construction points affected by the overlimit event (Right now Compensation is performed on the region using its parameter coordinates. Expand the grid outward by twice the grid step size from the center, and force fine-grained segmentation to ensure compensation accuracy.

[0028] Each grid cell Defined by rectangular meshes divided within the parameter domain, the boundaries of which are determined by the parameter coordinate range. and Define and map the three-dimensional curved surface region to the physical space.

[0029] The steps for device path planning are as follows: Based on the boundaries of the construction unit, plan the travel path of the construction equipment; Along the parameter domain or Direction generates the main path, and the device's turning radius. Curvature constraints must be met: in, The preset equipment safety factor (based on engineering safety specifications, equipment performance boundaries, and dynamic operating condition verification). If the path conflicts with unconstructed areas or equipment pose data, the A* algorithm is used to replan in the parameter domain, with Manhattan distance as the cost function. Generate a sequence of device paths, each path containing the start and end coordinates of the grid cell and the direction angle of travel; The specific steps for process optimization are as follows: If the slope change rate of adjacent grid cells differs If the construction process depends on the slope, the construction should be carried out first to ensure that the concrete flows smoothly at the joints and avoid segregation or cracking caused by sudden changes in elevation. If grid cells Concrete pouring completion time and Planned start time satisfy If the initial setting time of concrete is determined, then the process dependency is eliminated, and parallel construction is allowed.

[0030] In summary, this invention integrates heterogeneous data from multiple sources, including point clouds, sensors, and equipment poses, and generates a dynamic dataset through spatial registration and cleaning. This dataset is then mapped to the NURBS surface parameter coordinates of the BIM model. This achieves real-time synchronization between the construction status and the design model, solving the problem that traditional static BIM models cannot reflect actual working conditions.

[0031] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A BIM-based nonlinear multi-gradient floor construction collaborative management system, characterized in that, include: Data acquisition and modeling module: Collects real-time construction data of the ground structure, including point cloud data, sensor data and pose data of construction equipment; Spatial registration and cleaning are performed on the real-time construction data to generate a dynamic dataset; Construct a BIM model of a multi-gradient ground floor building; based on the first baseline value of the gradient parameters in the BIM model, map the dynamic dataset to the corresponding location in the BIM model; Intelligent analysis module: Input the mapped dynamic dataset into the preset prediction model to generate predicted values ​​of floor construction indicators, and calculate the prediction deviation and real-time deviation values ​​of the predicted and measured values ​​based on the first benchmark value. When the prediction deviation or real-time deviation value exceeds the preset construction indicator threshold, calculate the compensation amount of the gradient parameter and generate a compensation parameter set. Collaborative execution module: updates the BIM model based on the compensation parameter set, generates a second benchmark value based on the updated BIM model, and performs constructable unit segmentation, equipment path planning, and process optimization based on the updated BIM model and the second benchmark value.

2. The system according to claim 1, characterized in that, The spatial registration aims to unify point cloud data, sensor data, and device pose data into a preset global coordinate system.

3. The system according to claim 2, characterized in that, The steps for constructing the BIM model are as follows: 1) Use the preset geometric surface as the geometric expression tool of the BIM model. The geometric surface refers to the non-uniform rational B-spline surface, and use the global coordinate system as the coordinate system of the BIM model. 2) Map the parameter domain in the geometric surface to a three-dimensional surface in physical space. The physical space refers to the physical space defined based on the global coordinate system, and the points in the parameter domain are parameter coordinates. The shape of the geometric surface is determined by the surface construction point grid and weight coefficients. The functional expression formula for the geometric surface is as follows: in, For the geometric surface in parametric coordinates ( u,v The three-dimensional coordinates at point ) and These represent the number of rows and columns of the point mesh used to construct the surface, respectively. For the first m Line number n Column weighting coefficients For the first m Line number n The coordinates of the column's construction point; 3) Transform the construction design requirements into gradient parameters of the geometric surface, including elevation level, slope change rate, and load bearing capacity; The elevation level is the Z-axis coordinate of the construction point; The slope change rate is at the construction point and Partial derivatives in the direction; Load bearing capacity is an attribute field associated with the construction point, including concrete thickness, reinforcement ratio, and material strength parameters.

4. The system according to claim 3, characterized in that, The first benchmark value includes the elevation level benchmark value, the slope change rate benchmark value, and the load bearing capacity benchmark value; The calculation steps for the elevation benchmark value are as follows: 1) Within the parameter domain of the geometric surface, generate the coordinates of sampling points at fixed intervals; 2) Substitute the coordinates of each sampling point into the geometric surface function to calculate its corresponding Z-axis coordinate value in physical space; 3) For all sampling points Z The arithmetic mean of the coordinate values ​​is used to obtain the elevation benchmark value. ; The slope change rate benchmark value By calculating geometric surfaces and The maximum gradient of the direction is obtained; The load bearing capacity benchmark value Based on concrete compressive strength and effective thickness The formula for calculation using the limit state design method is as follows: in, This is the load distribution factor.

5. The system according to claim 4, characterized in that, The mapping of the dynamic dataset to the BIM model includes parametric coordinate inverse calculation and attribute association; The parameter coordinate inverse calculation is based on Newton's iteration method to inversely calculate the parameter coordinates of each data point in the dynamic dataset; The attribute association binds the geometric data in the dynamic dataset to the geometric position of the parametric coordinates; and associates the physical data in the dynamic dataset with the attribute fields in the corresponding parametric coordinates in the BIM model.

6. The system according to claim 5, characterized in that, The prediction deviation value The calculation formula is as follows: in, , and These are the deviation values ​​for elevation level, slope change rate, and predicted load bearing capacity, respectively. , and They are time t The predicted values ​​of elevation level, slope change rate, and load bearing capacity at that time; Real-time deviation value The calculation formula is as follows: in, , and These are the real-time deviations of elevation level, slope change rate, and load bearing capacity, respectively. , and They are time t The actual values ​​of elevation level, slope change rate, and load bearing capacity at that time.

7. The system according to claim 6, characterized in that, When the prediction deviation value or real-time deviation value Any value exceeding the preset construction indicator threshold The generation of the compensation parameter set is triggered at certain times. The specific steps are as follows: 1) Combine the prediction deviation with the real-time deviation to form 2) Weighting coefficients for points constructed using BIM model surfaces To optimize the variables, an objective function is established: in, Adjustment amount for weighting coefficients The changes in the predicted and real-time deviations caused by the error are obtained iteratively using the damped Gauss-Newton method. 3) The final Generate a compensation parameter set by associating it with the coordinates of the out-of-limit location: in, This is a subset of construction points affected by the overlimit event.

8. The system according to claim 7, characterized in that, The steps for updating the BIM model are as follows: Based on the compensation parameter set The weight coefficients of the construction points in the BIM model are adjusted using the following formula: in, These are the adjusted weighting coefficients. These are the original weighting coefficients. This is the adjustment amount for the weighting coefficient; The geometric surface function is recalculated based on the adjusted weighting coefficients, and the gradient parameters are dynamically corrected. The second baseline value is generated by recalculating the baseline value based on the updated BIM model.

9. The system according to claim 8, characterized in that, The specific steps for dividing the construction unit are as follows: In the parameter domain, the partial derivative magnitude of the geometric surface is used as the slope change rate, and a non-uniform grid is divided based on the slope change rate. The parameter domain is divided into a first region and a second region based on a preset rate of change threshold. The region where the slope change rate is greater than the change rate threshold is the first region, and the grid cells are generated using fine-grained segmentation parameter domains. The region where the slope change rate is greater than the change rate threshold is the second region, and the grid cells are generated using coarse-grained segmentation parameter domains. For the subset of construction points affected by the over-limit event, compensation is performed by expanding the grid outward by twice the grid step size with its parameter coordinates as the center, and forced fine-grained segmentation is performed to ensure compensation accuracy. The mesh element is defined by a rectangular mesh divided in the parameter domain, and its boundary is determined by the parameter coordinate range. and Determine and map the three-dimensional curved surface region of the physical space; The steps for device path planning are as follows: Based on the boundaries of the grid cells, the travel path of the construction equipment is planned; Along the parameter domain or Direction generates the main path, and the device's turning radius. Curvature constraints must be met: in, The preset equipment safety factor; If the path conflicts with the pose data of unconstructed areas or construction equipment, the A* algorithm is used to replan in the parameter domain, with Manhattan distance as the cost function; finally, a sequence of equipment paths is generated, each path containing the start and end coordinates and travel direction angle of the grid cell; The specific steps for optimizing the process are as follows: If the slope change rate of adjacent grid cells differs If the construction process depends on the slope, then the construction should proceed with the work on the slope with the gentlest slope. If the grid unit Concrete pouring completion time and Planned start time satisfy ,in If the initial setting time of the concrete is taken into account, then the process dependency is eliminated, and parallel construction is allowed.

10. The system according to claim 1, characterized in that, The prediction model is built on a Long Short-Term Memory (LSTM) network and uses mean squared error as the loss function.

Images