Construction control method based on three-dimensional dynamic model

Through the construction control method based on three-dimensional dynamic model, combined with IMU sensors and drone image data, real-time monitoring and dynamic prediction of power facilities construction site are achieved, and the problem of insufficient real-time construction site data in the existing technology is solved, and the problem of improvement of the real-time performance of construction site data is improved. Efficiency and environmental change monitoring capabilities of construction management.

CN119048719AInactive Publication Date: 2024-11-29YUNCHENG POWER SUPPLY COMPANY OF STATE GRID SHANXI ELECTRIC POWER +5

Patent Information

Application Number
CN202411529752.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-30
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the prior art, the construction site data of the power facility relies on manual measurement and static data analysis, and cannot provide real-time feedback, resulting in difficult real-time adjustment of construction progress and quality control, which may lead to construction quality problems and project delays.

Method used

Using a construction control method based on three-dimensional dynamic model, through data reading and the establishment of digital twin models, combined with real-time data of IMU sensors and positioning sensors, the prediction fitting network is activated to predict embedded parts position and attitude, and collected through drone The image data updates the model in real time and builds a virtual reality display.

Benefits of technology

Real-time digitalization and dynamic prediction of construction site status are realized, real-time and decision-making efficiency of construction management are improved, problems caused by errors are reduced, construction process is optimized, and the consistency between the model and the actual environment is ensured.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048719A_ABST
    Figure CN119048719A_ABST
Patent Text Reader

Abstract

The invention provides a construction control method based on a three-dimensional dynamic model, and relates to the technical field of construction control, and the method comprises the steps: carrying out the data reading of a construction site, building a digital twin model, reading a construction task, and carrying out the model region identification; a construction task is analyzed, an IMU sensor and a positioning sensor are configured in an embedded part, and a three-dimensional dynamic model is established; the monitoring data set is read in real time, construction control information is obtained, a prediction fitting network is activated, the position and posture of the embedded part are predicted, and a position prediction result and a posture prediction result are established; activating the unmanned aerial vehicle to perform image data acquisition, establishing embedded part environment characteristics, and updating the digital twin model in real time; and virtual reality display is constructed. The technical problem that in the prior art, manual measurement and static data analysis are usually adopted in construction management, changes in the construction process are difficult to monitor in real time, and consequently the construction quality is unstable is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of construction control, and in particular to a construction control method based on a three-dimensional dynamic model. Background Art

[0002] There are many components that need to be buried underground during the construction of power facilities to ensure the safety, reliability and aesthetics of the power system, such as cables for transmitting high-voltage electricity, optical fibers for power system control and communication, as well as grounding devices, cable trenches and pipelines, underground transformer foundations, junction boxes and distribution boxes. The correct installation and maintenance of these underground components are crucial to the stable operation of the power system. However, in the existing technology, the construction site data of power facilities usually relies on manual measurement and static data analysis. This method has delays and cannot provide real-time feedback, making it difficult to adjust the construction progress and quality control in real time, which may lead to construction quality problems and project delays; and during excavation and filling construction, the model may not be updated in time, resulting in the construction plan not being consistent with the actual situation, affecting the construction progress and quality control. Summary of the invention

[0003] This application provides a construction control method based on a three-dimensional dynamic model, aiming to solve the technical problem that the existing technology usually uses manual measurement and static data analysis in construction management, which makes it difficult to monitor changes in the construction process in real time, resulting in unstable construction quality.

[0004] The present application discloses a construction control method based on a three-dimensional dynamic model, the method comprising: reading data from a construction site, establishing a digital twin model of the construction site, and reading construction tasks, and identifying a model area of ​​the digital twin model based on the construction tasks; parsing the construction tasks, configuring IMU sensors and positioning sensors on embedded parts based on the parsing results, and establishing a three-dimensional dynamic model in the digital twin model based on basic information of the embedded parts, the three-dimensional dynamic model communicating with the IMU sensors and positioning sensors in real time digitally; reading monitoring data sets of the IMU sensors and positioning sensors in real time, and obtaining construction control information with time series identifiers, activating a prediction fitting network in the three-dimensional dynamic model, predicting the position and posture of the embedded parts based on the prediction fitting network according to the construction control information and the monitoring data set, and establishing position prediction results and posture prediction results; activating a drone based on the model area identification, collecting image data of the embedded parts area of ​​the construction site through the drone, establishing embedded parts environmental characteristics, and updating the digital twin model in real time based on the embedded parts environmental characteristics; feeding back the position prediction results and posture prediction results to the updated digital twin model to construct a virtual reality display.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages: By reading data from the construction site and establishing a digital twin model, the real-time digitization of the construction site status is achieved. The data from the IMU sensor and positioning sensor is transmitted to the three-dimensional dynamic model in real time, ensuring the instant update of the construction data, improving the real-time and accuracy of the data, and enabling the construction management to grasp the on-site situation in real time, thereby improving the efficiency of construction control and the timeliness of decision-making; by activating the prediction fitting network, the position and posture of the embedded parts are predicted based on the monitoring data of the IMU sensor and positioning sensor. By combining real-time data and prediction models, dynamic prediction of the construction status is achieved, so that the position and posture during the construction process can be predicted in advance, thereby reducing the error caused by The problems caused by this problem were solved and the construction process was optimized. Image data of the construction site was collected by drones, the environmental characteristics of embedded parts were established, and the digital twin model was updated in real time based on these characteristics, ensuring that the model can promptly reflect changes in the on-site environment, improving the monitoring capability of environmental changes, ensuring the consistency of the digital twin model with the actual construction environment, and enabling construction management to quickly respond to environmental changes. The position prediction results and posture prediction results were fed back to the updated digital twin model, and a virtual reality display was constructed. Through virtual reality technology, an intuitive construction progress and status display platform was provided, which strengthened the visual management of the construction status and improved the convenience and decision-making efficiency of construction management.

[0006] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A schematic diagram of a construction control method based on a three-dimensional dynamic model is provided for the embodiment of the present application; Figure 2 A schematic flow chart of establishing embedded parts environment characteristics in a construction control method based on a three-dimensional dynamic model is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0008] The embodiment of the present application solves the technical problem that the prior art usually uses manual measurement and static data analysis in construction management, which makes it difficult to monitor changes in the construction process in real time and leads to unstable construction quality by providing a construction control method based on a three-dimensional dynamic model.

[0009] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0010] like Figure 1As shown, the embodiment of the present application provides a construction control method based on a three-dimensional dynamic model, the method comprising: The data of the construction site is read, a digital twin model of the construction site is established, the construction tasks are read, and the model area of ​​the digital twin model is identified based on the construction tasks.

[0011] Data is read from the construction site of power facilities, including but not limited to power facility information data, geographic information data, on-site sensor data, image data, environmental data, etc. 3D modeling tools are used to integrate the construction site data and establish a digital twin model of the construction site. The digital twin model refers to a digital mirror image of the construction site. Through the combination of real-time data and historical data, a comprehensive and real-time digital expression of the construction site is achieved.

[0012] The construction management system reads construction tasks, including construction objectives, construction areas, schedules, resource allocation, and operating procedures. Specific areas in the digital twin model are marked according to the construction tasks. Specifically, the construction tasks are analyzed to determine the specific areas that need to be constructed. In the digital twin model, color marking, labels, etc. are used to mark the areas involved in the task, and the construction tasks are associated with the marked areas so that they can be quickly located in subsequent operations.

[0013] Analyze the construction tasks, configure IMU sensors and positioning sensors on the embedded parts based on the analysis results, and establish a three-dimensional dynamic model in the digital twin model according to the basic information of the embedded parts. The three-dimensional dynamic model communicates digitally with the IMU sensors and positioning sensors in real time.

[0014] Analyze the construction tasks to determine the specific location, quantity, and type of sensors that need to be installed. IMU sensors (inertial measurement unit sensors) are used to measure dynamic information such as acceleration and angular velocity of objects, and provide posture and motion information of objects. Positioning sensors are used to determine the specific location of objects and provide spatial position coordinates.

[0015] Embedded parts are components that need to be buried underground during the construction of power facilities to ensure the safety, reliability and aesthetics of the power system. Embedded parts can be cables for transmitting high-voltage power, optical fibers for power system control and communication, as well as grounding devices, cable trenches and pipelines, underground transformer foundations, junction boxes and distribution boxes, etc. IMU sensors and positioning sensors are fixed on embedded parts to ensure that the sensors can collect data normally. The basic information of embedded parts includes the size, position, material, installation angle, etc. of the embedded parts. The basic information of embedded parts is input into the 3D modeling tool, and a three-dimensional dynamic model of the embedded parts is created in the digital twin model. A data communication architecture is established so that the data of the IMU sensors and positioning sensors can be transmitted to the digital twin model in real time, ensuring that the three-dimensional dynamic model can be updated in real time according to the sensor data to reflect the real-time position and posture of the embedded parts.

[0016] The monitoring data sets of IMU sensors and positioning sensors are read in real time, and construction control information with time series identification is obtained, the prediction fitting network in the three-dimensional dynamic model is activated, and the position and posture of embedded parts are predicted based on the prediction fitting network according to the construction control information and monitoring data sets, and the position prediction results and posture prediction results are established.

[0017] Read the monitoring data of IMU sensors in real time, including dynamic data such as speed and posture, and read the monitoring data of positioning sensors in real time, including location coordinate data. Obtain construction control information through the construction management system or on-site terminals, including task plans, operation steps, resource scheduling and other information. Each piece of construction control information is timestamped.

[0018] Activate the prediction fitting network in the three-dimensional dynamic model to predict the position and posture of the embedded parts through the state prediction formula and the covariance prediction formula, and establish the position prediction results and posture prediction results. The position prediction result is the predicted position coordinates of the embedded parts at the next moment, and the posture prediction result is the predicted posture information of the embedded parts at the next moment, including angle, direction, etc.

[0019] Activate the drone based on the model area identification, use the drone to collect image data of the embedded area at the construction site, establish the embedded part environment characteristics, and update the digital twin model in real time based on the embedded part environment characteristics.

[0020] According to the digital twin model, the specific area where image data collection is required is determined. According to the model area identification, the flight route and altitude of the drone are set to ensure that all areas where data collection is required can be covered. The drone flies according to the preset flight route and uses a high-definition camera to collect image data. The collected data includes photos, videos, etc. The collected image data is processed, including image enhancement and noise removal, to improve the quality and availability of the data. The environmental features of the embedded parts are extracted from the processed image data. These features include the surrounding environment of the embedded parts, position changes, potential obstacles, etc. The new embedded part environmental features are integrated with the existing digital twin model so that the model can accurately reflect the latest on-site conditions.

[0021] The position prediction results and posture prediction results are fed back to the updated digital twin model to construct a virtual reality display.

[0022] The prediction results are input into the updated digital twin model so that it can reflect the current prediction situation. The updated digital twin model is integrated into the virtual reality environment so that the model reflects the latest prediction results in virtual reality. A virtual reality display interface is generated to visualize the position and posture prediction results in a three-dimensional environment. Construction personnel can view the real-time status and changes of the virtual construction site through virtual reality helmets or displays, thereby effectively monitoring and adjusting.

[0023] Furthermore, the prediction fitting network is used to predict the position and posture of the embedded parts according to the construction control information and the monitoring data set, and to establish the position prediction result and the posture prediction result, and further includes: The position information, speed information and posture information in the monitoring data set are extracted, and a state vector is established according to the position information, speed information and posture information.

[0024] Determine the Sigma point and propagate the Sigma point as follows: ; ; in, Characterization Sigma points, For the previous moment The state estimate of is the dimension of the state vector, is the scaling parameter, Representing the last moment The covariance of is the Sigma point after propagation, is the state transfer function, For the previous moment Construction control input.

[0025] Call the state prediction formula of the prediction fitting network and perform state prediction as follows: ; in, Representing the current moment The predicted state of Represents the mean weight.

[0026] Call the covariance prediction formula of the prediction fitting network to perform covariance prediction as follows: ; in, Representing the current moment The predicted covariance of represents the covariance weight, is the process noise covariance matrix.

[0027] The position and posture prediction of embedded parts is completed based on the state prediction results and covariance prediction results.

[0028] Extract position information based on the monitoring data of the positioning sensor, including the coordinate information (x, y, z) of the embedded part in three-dimensional space; extract speed information based on the monitoring data of the IMU sensor, indicating the movement speed of the embedded part in three-dimensional space , extract attitude information based on the monitoring data of the IMU sensor, including the tilt of the embedded part ( , and the tilt angle ( , the extracted information is integrated into a state vector, which is used to describe the current state of the embedded part. The state vector is as follows: ; The UKF (Unscented Kalman Filter) algorithm is used to determine the Sigma point. The Sigma point is a sampling method in the UKF that is used to capture the statistical characteristics of the state distribution. By generating a set of Sigma points near the state estimate, the UKF can better handle nonlinear transformations. The formula is as follows: ; ; Among them, the scaling parameter Used to control the distribution range of Sigma points. Represents the square root of the covariance matrix, which is used to generate Sigma points. represents the i-th column of the square root of the covariance matrix, through the state transfer function The Sigma point is propagated from the previous moment to the current moment. The state transfer function describes the dynamic changes of the system within a time step, taking into account the dynamic characteristics of the system.

[0029] The prediction fitting network is used to capture the dynamic characteristics of the system and predict the state of the Sigma point after propagation to obtain the predicted state at the current moment. The state prediction formula is as follows: ; This formula is used to calculate the current state prediction value. The mean weight is used to perform weighted averaging on the propagated Sigma points. The weight calculation takes into account the distribution characteristics of the Sigma points and the uncertainty of the system. By using the mean weight to perform weighted averaging on the propagated Sigma points, the predicted state at the current moment is obtained. This weighted average takes into account the mean weight of each Sigma point to ensure the accuracy of the prediction value. Substitute the state prediction formula and calculate the current moment The predicted state is completed.

[0030] Covariance prediction is used to evaluate the uncertainty of the system at the current moment. By calculating the predicted covariance matrix, the distribution characteristics of the Sigma points after propagation are reflected. The covariance prediction formula is as follows: ; Covariance weights It is used to perform weighted averaging on the propagated Sigma points. The calculation of the weights takes into account the distribution characteristics of the Sigma points and the uncertainty of the system. The covariance weights are used to calculate the weighted covariance of the propagated Sigma points to obtain the predicted covariance at the current moment. In order to consider the process noise of the system, the process noise covariance matrix is ​​added to the calculated covariance matrix. It reflects the random disturbance and uncertainty in the system process. The covariance prediction is performed through this formula, and the current moment is calculated. The predicted covariance of .

[0031] The state prediction result indicates the predicted state at the current moment, including the position information, speed information and attitude information of the embedded parts. The covariance prediction result indicates the predicted covariance at the current moment, reflecting the confidence interval and uncertainty of the predicted state. The position information and attitude information in the state prediction result are combined with the uncertainty in the covariance prediction result to complete the prediction of the final position and attitude of the embedded parts. The final prediction result not only includes the position information and attitude information, but also includes the predicted confidence interval and error range, which improves the reliability and accuracy of the prediction.

[0032] Furthermore, after building a virtual reality display, it also includes: An adaptive compensation network is configured to perform control compensation based on the state prediction result and the covariance prediction result as follows: Calculate the observed Sigma points: ,in, is the observed Sigma point, is the observation function.

[0033] Compute the observed mean and observed covariance: ; ; ; in, is the observed mean, is the observation covariance, is the observation noise covariance, is the cross covariance between states and observations.

[0034] The Kalman gain and state are updated according to the observed mean, the observed covariance, the state prediction result and the covariance prediction result, and the parameters are dynamically adjusted based on the state update result through the compensation unit of the adaptive compensation network.

[0035] Specifically, the adaptive compensation network is used to perform control compensation based on the state prediction results and covariance prediction results, thereby further improving the accuracy of embedded part position and posture prediction.

[0036] The calculation formula for the observed Sigma point is as follows: ; This formula is used to calculate the observed Sigma point. is the Sigma point after propagation The result obtained after applying the observation function, which maps the Sigma points in the state space to the observation space, thereby representing the distribution of the state in the observation space.

[0037] Calculate the observed mean and observed covariance using the following formula: ; ; ; Observed mean is the weighted average of all observed Sigma points, the observed covariance is the weighted covariance of the observation error, and the cross covariance between the state and the observation is the weighted covariance of the state error and the observation error, which provides the basis for the subsequent Kalman gain calculation and state update.

[0038] The Kalman gain is used to balance the weights of the predicted value and the observed value. The predicted result and the actual observed result are weighted and synthesized through the Kalman gain to optimize the state estimation. Specifically, the cross covariance between the state and the observation and the observation covariance are used to calculate the Kalman gain, the state estimation is updated based on the actual observation value and the observation mean at the current moment, and the state covariance matrix is ​​updated using the Kalman gain and the observation covariance. The compensation unit using the adaptive compensation network dynamically adjusts the system control parameters based on the updated state estimate and state covariance to cope with possible system changes or errors.

[0039] Furthermore, the Kalman gain and state update according to the observed mean, the observed covariance, the state prediction result and the covariance prediction result also include: ; ; ; in, is the Kalman gain, is the updated state estimate, For the current moment The observation data, is the updated state covariance.

[0040] Specifically, the Kalman gain is used to balance the weights of the predicted values ​​and the observed values ​​to minimize the estimation error. It is calculated based on the predicted covariance matrix and the observed covariance matrix, and the formula is as follows: ; The obtained Kalman gain It is an adjustment coefficient used to balance the predicted value and the observed value. Through the Kalman gain, the dependence on the prediction and observation can be adaptively adjusted to minimize the estimation error.

[0041] The Kalman gain is used to update the state estimate to combine the prediction and actual observations, as follows: ; The updated state estimate obtained It is a state estimate that combines the predicted value and the actual observed value, providing the best estimate of the current state of the system.

[0042] Update the state covariance matrix to reflect the change in uncertainty of the state estimate, as follows: ; The updated state covariance obtained It represents the uncertainty of the updated state estimate, and a smaller covariance indicates a higher estimation accuracy.

[0043] Furthermore, the method further comprises: The target position and target attitude at the current moment are established according to the analysis results; error calculation is performed based on the target position, target attitude and state update results, and adaptive compensation is performed based on the error calculation results through the compensation unit.

[0044] Through the analysis results of the construction task, the specific requirements and goals of the current construction are determined, including the setting of the target position and posture, and the target position and target posture at the current moment are established. The target position refers to the ideal position set in the construction task, given in the form of coordinates, and the target posture refers to the ideal posture set in the construction task, given in the form of angles.

[0045] The predicted position and predicted attitude are obtained according to the state update results, and the deviations between the predicted position and the target position, and between the predicted attitude and the target attitude are calculated respectively to obtain the error calculation results. The compensation unit adjusts the control parameters according to the error calculation results, such as increasing or decreasing the output of the driver, to correct the deviation and achieve more precise control to ensure that the target position and attitude of the construction task are achieved.

[0046] Furthermore, if Figure 2 As shown, the method of collecting image data of the embedded parts area of ​​the construction site by using a drone, establishing embedded parts environment characteristics, and updating the digital twin model in real time based on the embedded parts environment characteristics also includes: Configure a timing authentication network, perform timing authentication identification of image data acquisition results based on the timing authentication network, and establish static environment identification and dynamic environment identification; generate a basic attention factor based on the static environment identification and dynamic environment identification, perform acquisition control feedback of the drone through the basic attention factor, and establish embedded environment characteristics with the image data re-collected based on the acquisition control feedback results.

[0047] The time series authentication network is used to authenticate the image data in time series to ensure the consistency of the data. For example, a long short-term memory (LSTM) network is used to process the continuous image frames taken by the drone, and the number of layers, number of units, activation function and other parameters of the LSTM network are configured, and the network is trained to identify the time series features in the image sequence. The time series authentication identification refers to the verification of the time series information of the image data through the network. The image data is input into the time series authentication network, and the network analyzes each frame of the image to identify the time sequence of the data.

[0048] Static environment identification and dynamic environment identification are established from the image data after time-series authentication identification. Specifically, by comparing consecutive image frames, the static and changing parts of the environment are identified, and the static environment identification is generated based on the extracted static features. The static environment identification refers to the identification of fixed and unchanging environmental features in the image, such as the structure of the building, the ground type, etc. The dynamic environment identification is generated based on the extracted changing features. The dynamic environment identification refers to the identification of changing environmental features in the image, such as moving vehicles, activities on the construction site, etc.

[0049] Basic attention factors refer to the key parameters used to guide drone data collection. These factors integrate static and dynamic environmental characteristics to optimize the data collection process. Specifically, the information in the static and dynamic environmental identification is integrated to generate attention factors, including regional priority, collection frequency, key monitoring areas, etc. If an area contains static buildings and dynamic worker activities, the generated basic attention factors include detailed collection of the areas around the buildings and high-frequency collection of dynamic worker activity areas.

[0050] According to the basic attention factors, feedback adjustments are made to the drone's collection tasks, including adjusting the drone's flight path, camera direction, collection frequency, etc., to ensure that key areas are fully photographed. According to the collection control feedback results, the drone re-executes the image data collection task to obtain updated image data to update the embedded environment characteristics.

[0051] Furthermore, the method further comprises: The target feature is determined based on the dynamic environment identifier, and the moving frequency of the target feature is authenticated to generate a first authentication result; the moving amplitude of the target feature is authenticated to generate a second authentication result; the distance of the embedded part of the target feature is authenticated to generate a third authentication result; an additional attention factor is established based on the first authentication result, the second authentication result, and the third authentication result, and after the basic attention factor is corrected by the additional attention factor, the collection control feedback of the UAV is executed.

[0052] Specific dynamic objects, including construction equipment and porters, are extracted from dynamic environmental identification, and target features are determined. The position changes of the target features in the time series are analyzed, and the movement frequency is calculated. For example, the position changes of the target features are tracked through the image data of continuous frames, and it is determined whether the movement of the target features exceeds a preset frequency threshold. A first authentication result is generated, indicating whether the movement frequency of the target features is higher than the normal level. If the result shows that the frequency is too high, it may cause focus on the construction area.

[0053] The movement amplitude is the maximum displacement distance of the target feature within a period of time. Continuous frame image data is used to track the position change of the target feature, calculate the displacement amplitude of the target feature within a specified time period, and determine whether the actual movement amplitude of the target feature exceeds the preset safety amplitude threshold. A second authentication result is generated, indicating whether the movement amplitude of the target feature exceeds the normal range.

[0054] The distance of the embedded part is the straight-line distance from the target feature to the embedded part. The actual distance between the target feature and the embedded part is measured using image data or sensor data. The actual measured distance is compared with the preset safety distance threshold to generate a third authentication result. If the distance between the target feature and the embedded part is less than the safety distance, measures need to be taken to avoid possible interference.

[0055] A comprehensive factor is generated based on the comprehensive analysis of the first authentication result, the second authentication result, and the third authentication result, which reflects the overall impact of the target feature. This factor takes into account the situation of all authentication results. The original basic attention factor is adjusted using the additional attention factor to improve the accuracy and effectiveness of UAV collection. For example, if the movement frequency authentication and the movement amplitude authentication both show abnormalities, but the distance authentication is normal, the additional attention factor prompts that it is necessary to focus on the movement frequency and amplitude of the target feature. The UAV's collection control strategy is optimized according to the corrected attention factor to improve the effect and accuracy of data collection.

[0056] Furthermore, the method further comprises: Abnormal distance authentication is performed based on the position prediction results, the posture prediction results and the real-time updated digital twin model; a display mark is generated based on the abnormal distance authentication result, and the display mark includes a color mark of the warning level based on the abnormal distance authentication result; and a virtual reality display is constructed based on the display mark.

[0057] The position prediction results and posture prediction results of the embedded parts are matched with the environmental data in the real-time updated digital twin model to determine the actual distance between the embedded parts and the target features. A safety distance is defined to determine whether the actual distance between the target features and the embedded parts exceeds the range. If the actual distance is lower than the safety distance, it is marked as an abnormality, indicating that measures need to be taken.

[0058] The warning level is determined based on the severity of the abnormal distance, and different colors are usually used. For example, green indicates normal, yellow indicates warning, and red indicates serious problems. The color identification is combined with the description information of the abnormal situation, such as the specific abnormal distance, affected area, etc., to generate the final display identification.

[0059] Display logos are embedded into the virtual reality environment. Logos can be displayed on the virtual reality interface in the form of icons, annotations, color areas, etc., and relevant areas can be highlighted according to the warning level to help users quickly identify problem areas.

[0060] In summary, the construction control method based on the three-dimensional dynamic model provided in the embodiment of the present application has the following technical effects: By reading data from the construction site and establishing a digital twin model, the real-time digitization of the construction site status is achieved. The data from the IMU sensor and positioning sensor is transmitted to the three-dimensional dynamic model in real time, ensuring the instant update of the construction data, improving the real-time and accuracy of the data, and enabling the construction management to grasp the on-site situation in real time, thereby improving the efficiency of construction control and the timeliness of decision-making; by activating the prediction fitting network, the position and posture of the embedded parts are predicted based on the monitoring data of the IMU sensor and positioning sensor. By combining real-time data and prediction models, dynamic prediction of the construction status is achieved, so that the position and posture during the construction process can be predicted in advance, thereby reducing the error caused by The problems caused by this problem were solved and the construction process was optimized. Image data of the construction site was collected by drones, the environmental characteristics of embedded parts were established, and the digital twin model was updated in real time based on these characteristics, ensuring that the model can promptly reflect changes in the on-site environment, improving the monitoring capability of environmental changes, ensuring the consistency of the digital twin model with the actual construction environment, and enabling construction management to quickly respond to environmental changes. The position prediction results and posture prediction results were fed back to the updated digital twin model, and a virtual reality display was constructed. Through virtual reality technology, an intuitive construction progress and status display platform was provided, which strengthened the visual management of the construction status and improved the convenience and decision-making efficiency of construction management.

[0061] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A construction control method based on a three-dimensional dynamic model, characterized in that: The method comprises: Read data from the construction site, establish a digital twin model of the construction site, read construction tasks, and identify model areas of the digital twin model based on the construction tasks; Analyze the construction tasks, configure IMU sensors and positioning sensors on the embedded parts based on the analysis results, and establish a three-dimensional dynamic model in the digital twin model based on the basic information of the embedded parts. The three-dimensional dynamic model communicates with the IMU sensors and positioning sensors in real time digitally; Read the monitoring data sets of IMU sensors and positioning sensors in real time, obtain construction control information with time series identification, activate the prediction fitting network in the three-dimensional dynamic model, predict the position and posture of embedded parts based on the prediction fitting network according to the construction control information and monitoring data sets, and establish position prediction results and posture prediction results; Activate the drone based on the model area identification, use the drone to collect image data of the embedded area of ​​the construction site, establish the embedded environment characteristics, and update the digital twin model in real time based on the embedded environment characteristics; The position prediction results and posture prediction results are fed back to the updated digital twin model to construct a virtual reality display.

2. The construction control method based on a three-dimensional dynamic model as claimed in claim 1, characterized in that: The prediction fitting network is used to predict the position and posture of the embedded parts according to the construction control information and the monitoring data set, and establish the position prediction result and the posture prediction result, and further includes: Extracting position information, speed information and attitude information from the monitoring data set, and establishing a state vector according to the position information, speed information and attitude information; Determine the Sigma point and propagate the Sigma point as follows: ; ; in, Characterization Sigma points, For the previous moment The state estimate of is the dimension of the state vector, is the scaling parameter, Representing the last moment The covariance of is the Sigma point after propagation, is the state transfer function, For the previous moment Construction control input; Call the state prediction formula of the prediction fitting network and perform state prediction as follows: ; in, Representing the current moment The predicted state of Characterize the mean weight; Call the covariance prediction formula of the prediction fitting network to perform covariance prediction as follows: ; in, Representing the current moment The predicted covariance of represents the covariance weight, is the process noise covariance matrix; The position and posture prediction of embedded parts are completed based on the state prediction results and covariance prediction results.

3. The construction control method based on a three-dimensional dynamic model as claimed in claim 2, characterized in that: After building the VR display, it also includes: An adaptive compensation network is configured to perform control compensation based on the state prediction result and the covariance prediction result as follows: Calculate the observed Sigma points: ,in, is the observed Sigma point, is the observation function; Compute the observed mean and observed covariance: ; ; ; in, is the observed mean, is the observation covariance, is the observation noise covariance, is the cross covariance between states and observations; The Kalman gain and state are updated according to the observed mean, the observed covariance, the state prediction result and the covariance prediction result, and the parameters are dynamically adjusted based on the state update result through the compensation unit of the adaptive compensation network.

4. The construction control method based on a three-dimensional dynamic model as claimed in claim 3, characterized in that: The Kalman gain and state update are performed according to the observation mean, the observation covariance, the state prediction result and the covariance prediction result, and further includes: ; ; ; in, is the Kalman gain, is the updated state estimate, For the current moment The observation data, is the updated state covariance.

5. The construction control method based on a three-dimensional dynamic model as claimed in claim 4, characterized in that: The method further comprises: Establish the target position and target posture at the current moment according to the analysis results; Error calculation is performed based on the target position, target posture and state update results, and adaptive compensation is performed based on the error calculation results through a compensation unit.

6. The construction control method based on a three-dimensional dynamic model as claimed in claim 1, characterized in that: The method of collecting image data of the embedded parts area of ​​the construction site by using a drone, establishing embedded parts environment characteristics, and updating the digital twin model in real time based on the embedded parts environment characteristics also includes: Configure a time sequence authentication network, perform time sequence authentication identification of image data acquisition results based on the time sequence authentication network, and establish static environment identification and dynamic environment identification; A basic attention factor is generated according to the static environment identifier and the dynamic environment identifier, and the collection control feedback of the UAV is performed through the basic attention factor, and the embedded part environment characteristics are established by the image data re-collected according to the collection control feedback result.

7. The construction control method based on a three-dimensional dynamic model as claimed in claim 6, characterized in that: The method further comprises: Determine the target feature based on the dynamic environment identifier, and perform mobile frequency authentication on the target feature to generate a first authentication result; Performing movement range authentication on the target feature to generate a second authentication result; Perform embedded part distance authentication on target features and generate a third authentication result; An additional attention factor is established according to the first authentication result, the second authentication result, and the third authentication result. After the basic attention factor is corrected by the additional attention factor, the collection control feedback of the drone is executed.

8. The construction control method based on three-dimensional dynamic model as claimed in claim 1, characterized in that: The method further comprises: Abnormal distance authentication is performed based on the position prediction results, posture prediction results, and the real-time updated digital twin model; Generate a display mark based on the abnormal distance authentication result, wherein the display mark includes a color mark based on the warning level of the abnormal distance authentication result; A virtual reality display is constructed based on the display identifier.

Citation Information

Patent Citations

  • Foundation pit construction risk assessment method and system based on digital twin simulation platform

    CN115936437A

  • Steel structure machining and mounting method based on oblique photography of unmanned aerial vehicle

    CN117274499A

  • Determining Terrain Model Error

    US20160170090A1

Cited By

  • Municipal road construction section scheduling optimization method and system based on artificial intelligence

    CN120124985A