Power grid engineering project progress monitoring method

Abstract

The invention relates to a power grid engineering project progress monitoring method. An existing monitoring system can only confirm project progress and cannot automatically recognize whether project installation operation is standard or not. The method comprises the following steps: constructing a BIM model according to image data and project drawings of a project site; according to the BIM model, intelligent sensors are arranged on site, and online construction information is collected; online construction information is extracted, three-dimensional environment scene reconstruction is carried out according to the surrounding environment, and construction progress data is obtained; calculating an environment fluctuation value according to the construction progress monitoring condition and the online construction information, and setting a construction adjustment strategy; constructing a gray scale prediction model, and predicting a progress critical value result according to the construction progress data and the construction period after construction adjustment; and visualizing the construction progress data, the three-dimensional environment scene and the progress critical value prediction result to generate a BIM simulation large screen, thereby realizing scientific scheduling of the construction process and real-time monitoring of safe construction.

Description

A method for monitoring the progress of power grid engineering projects Technical Field

[0001] The technical problem solved by this invention is that it does not consider the impact of the construction environment on the construction progress, fails to achieve dynamic prediction and compensation of the construction progress, and lacks adaptive optimization configuration for construction adjustment. Background Technology

[0002] As power grid engineering projects expand in scale and become more complex, traditional quality and schedule management methods face challenges. Driven by artificial intelligence and the Internet of Things, power grid engineering management is shifting towards a new intelligent and real-time model, urgently requiring the development of corresponding monitoring systems to promote the industry's digital transformation.

[0003] Existing methods have shortcomings: they do not fully consider the impact of the construction environment on the schedule, lack dynamic prediction and compensation mechanisms for the construction schedule, and the construction adjustment strategies also lack adaptive optimization capabilities. Summary of the Invention

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0005] Firstly, a method for monitoring the construction quality and progress of power grid engineering projects includes:

[0006] Build a BIM model based on on-site image data and project drawings;

[0007] Based on the BIM model, intelligent sensors are deployed on-site to collect online construction information;

[0008] Extract online construction information, reconstruct a 3D environment scene based on the surrounding environment, and obtain construction progress data;

[0009] Calculate environmental fluctuation values ​​based on construction progress monitoring and online construction information, and set construction adjustment strategies;

[0010] A gray-scale prediction model is constructed based on the construction progress data and the critical value of the construction period prediction after construction adjustment.

[0011] The construction progress data, 3D environment scene, and progress critical value prediction results are visualized to generate a BIM simulation screen.

[0012] Preferably, constructing a BIM model based on on-site image data and project drawings specifically includes:

[0013] Using drone photography technology, image data of the engineering project site can be collected from five angles (front, back, left, right, and bottom) within a unit of time.

[0014] Obtain project drawings, draw a grid on the project drawings, mark grid points, and use the SIFT method to match the image data with the key point features of the project drawings to generate a 3D BIM model;

[0015] The BIM model is subjected to three-dimensional deviation verification, and deviation analysis is performed on axis positioning, construction dimensions and elevation, and allowable error range is set.

[0016] Preferably, based on the BIM model, deploying smart sensors on-site to collect online construction information specifically includes:

[0017] Based on the grid points marked in the BIM model, smart sensors are deployed at fixed points on the construction site corresponding to the grid points to collect online construction information. The smart sensors include temperature sensors, humidity sensors, motion sensors, smoke sensors, energy monitoring sensors, light sensors, infrared cameras, and pressure sensors. The online construction information includes environmental monitoring data, structural and gas data, energy equipment data, real-time captured images, and worker data. The environmental monitoring data includes concrete core temperature, steel structure welding area temperature, construction ambient air temperature, concrete surface humidity, material storage humidity, and wind speed. The structural and gas data includes carbon dioxide, methane, and PM2.5. The energy equipment data includes material inventory data, power monitoring data, and machinery operation data. The worker data includes personnel scheduling, personnel positioning, and fence rules.

[0018] A data communication network is constructed, consisting of a terminal layer, a convergence layer, and a core layer. The terminal layer is used to transmit environmental monitoring data and structural and gas data. The convergence layer is used to transmit energy equipment data, real-time captured images, and worker data. The core layer is used to update BIM model data.

[0019] Configure a communication protocol to transmit environmental monitoring data collected by smart sensors to the terminal layer, and aggregate the data into n node data packets according to the data volume. Send the real-time captured images to the aggregation layer. Update the corresponding nodes of the BIM model online according to the terminal layer and the aggregation layer. The communication protocol includes IEEE, HTTP, MQTT, JSONAPI, UDP, RTP and TCP.

[0020] Preferably, online construction information is extracted, and a 3D environmental scene is reconstructed based on the surrounding environment to obtain construction progress data, specifically including:

[0021] Key feature points are extracted from the real-time captured images. Based on visual SLAM stitching technology, the real-time captured images from multiple perspectives are continuously modeled to generate a dynamic three-dimensional point cloud map that reflects the real geometric changes of the construction site, thereby realizing real-time perception of the spatial form of the construction process.

[0022] The Iterative Closest Point (ICP) algorithm is used to spatially register the key feature points in the dynamic 3D point cloud map with the corresponding components in the BIM model, so that the dynamic point cloud data and the static BIM model are unified under the same coordinate system, thereby improving the spatial matching accuracy between the construction scene data and the design model.

[0023] Based on the completion of spatial registration, the BIM model is divided into m construction zones according to the construction type. The volume overlap between the dynamic three-dimensional point cloud in each construction zone and the corresponding BIM model zone is calculated, and the volume overlap is used as a quantitative indicator of the actual completion volume of the zone.

[0024] When the volume overlap reaches the preset completion threshold, it is determined that the corresponding construction zone has been completed, and the zone is rendered as green in the BIM model.

[0025] When the volume overlap is lower than the completion threshold but greater than zero, the corresponding construction zone is determined to be under construction and the zone is rendered in red in the BIM model; at the same time, the actual progress percentage of the construction zone is calculated based on the volume overlap, and the expected progress, actual progress and the progress percentage are saved as the construction progress data of the zone.

[0026] When no corresponding dynamic 3D point cloud feature point is detected in the construction zone of the BIM model, the corresponding spatial coordinates in the dynamic 3D point cloud are marked as redundant point clouds and visualized in yellow in the BIM model to indicate possible construction deviations or temporary storage areas.

[0027] Based on the spatial registration results, personnel positioning data, environmental sensor data, and equipment status data are fused and mapped to the corresponding construction zones and component locations in the digital twin model according to their spatial location and timestamp information, forming a comprehensive situational layer that is updated synchronously with the construction progress. This enables unified display and collaborative analysis of construction progress status, personnel activities, environmental conditions, and equipment operating status.

[0028] Preferably, the environmental fluctuation value is calculated based on the construction progress monitoring and online construction information, and the construction adjustment strategy is set, specifically including:

[0029] Based on the environmental monitoring data and energy equipment data of the online construction information, the construction progress of the zone is dynamically compensated, and the construction adjustment strategy includes a first adjustment, a second adjustment and a third adjustment;

[0030] The construction environment temperature and wind speed data are extracted from the environmental monitoring data. The standard deviation and mean of each data point within a unit of time are calculated. The sum of the ratios of the standard deviations and means of the construction environment temperature and wind speed is taken as the environmental fluctuation value. The number of days of inventory available in the material inventory data is determined based on the material usage of the next construction zone. K-means density clustering is used to dynamically classify the environmental fluctuation value to obtain a first classification interval, a second classification interval, and a third classification interval. The first classification interval < the second classification interval < the third classification interval. When the environmental fluctuation value is within the first classification interval and the number of days of inventory available is greater than three days, the first adjustment is triggered.

[0031] When the environmental fluctuation value is within the second category range and one day < number of days of inventory availability < three days, the second adjustment is triggered;

[0032] When the environmental fluctuation value falls within the third category range and the inventory availability days are less than one day, the third adjustment is triggered.

[0033] The first adjustment is used to extend the construction period and reduce the load on high-altitude operations;

[0034] The second adjustment is used to adjust the normal construction time and increase worker scheduling;

[0035] The third adjustment is used to stop outdoor operations and extend the construction period.

[0036] Preferably, the gray-scale prediction model is constructed based on the construction progress data and the adjusted construction period prediction critical value results, specifically including:

[0037] The deviation data of the construction progress of the partition is calculated, where the deviation data = expected progress - actual progress. The deviation data is stored in sequence. The least squares solution of the gray differential equation is solved using the gray prediction model GM to obtain the prediction critical value. When the deviation data is ≥ 0.7 times the prediction critical value, the dynamic compensation strategy library is called to record the actual deviation reduction and compensation time. The compensation efficiency is calculated based on the actual deviation reduction and compensation time, where the compensation efficiency = actual deviation reduction / compensation time. The gray prediction model GM is dynamically updated based on the prediction critical value and the compensation efficiency to generate a new round of prediction critical values.

[0038] Preferably, the visualization of construction progress data, 3D environment scene, and progress critical value prediction results into a BIM simulation screen specifically includes:

[0039] A BIM simulation screen is built on the cloud platform, overlaying the 3D environment scene and the BIM model, and marking the corresponding sensor points. A query interface, a grayscale prediction interface, and a visualization rendering interface are set up, and an online construction information interface is introduced. By inputting data parameters, specific element attributes are queried, and the current stage of construction completion of the 3D environment scene is modeled and rendered onto the BIM simulation screen. The completed parts and progress of different zones can be queried.

[0040] Preferably, the query interface is used to query real-time data from different sensors and update online construction information;

[0041] The grayscale prediction interface is used to deploy the grayscale prediction model GM, generate prediction thresholds, and the visualization rendering interface is used to visualize and render the 3D environment scene.

[0042] Preferably, the dynamic compensation strategy library includes a first strategy, a second strategy, and a third strategy; the first strategy is used to increase labor to shorten the construction period, with a delay of less than two days; the second strategy is used to extend man-hours to shorten the construction period, with a delay of more than two days; and the third strategy is used to call on backup equipment to shorten the construction period, with a delay of less than one day.

[0043] Secondly, a method for monitoring the progress of power grid engineering projects includes: a project drawing construction module, a sensor deployment module, a construction progress module, a progress prediction module, and a visualization module;

[0044] The project drawing construction module is used to construct a BIM model based on the image data of the engineering site and the project drawings;

[0045] The sensor deployment module is used to deploy smart sensors on-site according to the BIM model and collect online construction information.

[0046] The construction progress module is used to extract online construction information, reconstruct a three-dimensional environmental scene based on the surrounding environment, obtain construction progress data, calculate environmental fluctuation values ​​based on construction progress monitoring and online construction information, and set construction adjustment strategies.

[0047] The progress prediction module is used to construct a gray-scale prediction model and predict the critical value of the progress based on the construction progress data and the adjusted construction period.

[0048] The visualization module is used to visualize construction progress data, three-dimensional environmental scenes, and progress critical value prediction results to generate a BIM simulation screen.

[0049] New technical feature: Intelligent safety early warning module. To further improve the safety management level of construction sites, this system also includes an intelligent safety early warning module. This module, based on multi-sensor data fusion and behavior recognition technology, enables real-time monitoring and early warning of the safety status of construction personnel.

[0050] The intelligent safety early warning module specifically includes the following steps:

[0051] The posture, position, and physiological data of construction workers are collected in real time using infrared cameras and wearable sensors.

[0052] A lightweight convolutional neural network is used to identify the behavior of construction workers and determine whether there are any violations or dangerous behaviors.

[0053] By combining hazardous area markers in the BIM model with real-time environmental data, the safety risk level can be dynamically assessed.

[0054] When high-risk behaviors or environmental anomalies are detected, a three-level early warning system is implemented through audible and visual alarms, smart bracelet vibration, and BIM large screen pop-up windows.

[0055] Preferably, the intelligent safety early warning module also supports retrospective analysis of historical accident data, constructs a safety event knowledge graph, and provides risk avoidance suggestions for subsequent construction.

[0056] New system module: Adaptive learning optimization mechanism

[0057] To improve the adaptability and accuracy of the grayscale prediction model, this system introduces an adaptive learning optimization mechanism to achieve dynamic adjustment of model parameters and self-optimization of prediction strategies.

[0058] The adaptive learning optimization mechanism includes:

[0059] Collect historical construction data and environmental fluctuation records to construct a multi-dimensional training dataset;

[0060] The time series cross-validation method was used to evaluate the prediction performance of the gray-scale prediction model at different construction stages;

[0061] The model parameters are dynamically adjusted based on the gradient descent algorithm, and the selection logic of the construction adjustment strategy is optimized by combining the reinforcement learning mechanism.

[0062] The optimized models and strategies are updated to the cloud system in real time, supporting parallel learning and knowledge transfer across multiple projects.

[0063] Preferably, the mechanism also supports user feedback annotation, allowing managers to evaluate the early warning results and adjustment strategies, and the system to further optimize the decision-making logic accordingly.

[0064] The beneficial effects of this invention are as follows: By combining BIM modeling of image data and drawings with 3D environment reconstruction, a high-precision digital twin is constructed, and intelligent sensors are deployed to accurately collect environmental data such as temperature, displacement, and wind speed. Furthermore, by constructing a grayscale prediction model, progress thresholds are predicted based on historical data, and different dynamic compensation strategies are triggered according to the prediction results, achieving scientific scheduling of the construction process. This invention also integrates 4D progress, environmental heat maps, and early warning information through a BIM simulation screen, allowing for online data querying and realizing digital management and control of the construction process. The newly added intelligent safety early warning module significantly improves on-site safety management, while the adaptive learning optimization mechanism enables the system to continuously evolve, further enhancing the system's intelligence and practicality. Attached Figure Description

[0065] Figure 1 is a schematic diagram of the basic process of a power grid engineering project progress monitoring method provided in an embodiment of the present invention. Detailed Implementation

[0066] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0067] Example 1, referring to Figure 1, is an embodiment of the present invention, providing a method for monitoring the progress of a power grid engineering project, including:

[0068] Build a BIM model based on on-site image data and project drawings;

[0069] Based on the BIM model, intelligent sensors are deployed on-site to collect online construction information;

[0070] Extract online construction information, reconstruct a 3D environment scene based on the surrounding environment, and obtain construction progress data;

[0071] Calculate environmental fluctuation values ​​based on construction progress monitoring and online construction information, and set construction adjustment strategies;

[0072] A gray-scale prediction model is constructed based on the construction progress data and the critical value of the construction period prediction after construction adjustment.

[0073] The construction progress data, 3D environment scene, and progress critical value prediction results are visualized to generate a BIM simulation screen.

[0074] In this embodiment, by constructing a closed-loop intelligent construction system, digital control and dynamic optimization of the entire construction process are achieved, improving the accuracy of compensation measures for engineering construction.

[0075] Building a BIM model based on on-site image data and project drawings specifically includes:

[0076] Using drone photography technology, image data of the engineering project site can be collected from five angles (front, back, left, right, and bottom) within a unit of time.

[0077] Obtain project drawings, draw a grid on the project drawings, mark grid points, and use the SIFT method to match the image data with the key point features of the project drawings to generate a 3D BIM model;

[0078] The BIM model is subjected to three-dimensional deviation verification, and deviation analysis is performed on axis positioning, construction dimensions and elevation, and allowable error range is set.

[0079] In this embodiment, environmental and image data are collected from multiple angles, and the images are accurately located and feature points are extracted to achieve efficient matching between two-dimensional images and three-dimensional real scenes. A five-directional drone shooting scheme is adopted to break through the field of view limitations of traditional single-angle shooting, realize the three-dimensional capture of engineering real scenes, and provide a solid foundation for subsequent modeling.

[0080] Based on the BIM model, intelligent sensors are deployed on-site to collect online construction information, specifically including:

[0081] Based on the grid points marked in the BIM model, smart sensors are deployed at fixed points on the construction site corresponding to the grid points to collect online construction information. The smart sensors include temperature sensors, humidity sensors, motion sensors, smoke sensors, energy monitoring sensors, light sensors, infrared cameras, and pressure sensors. The online construction information includes environmental monitoring data, structural and gas data, energy equipment data, real-time captured images, and worker data. The environmental monitoring data includes concrete core temperature, steel structure welding area temperature, construction ambient air temperature, concrete surface humidity, material storage humidity, and wind speed. The structural and gas data includes carbon dioxide, methane, and PM2.5. The energy equipment data includes material inventory data, power monitoring data, and machinery operation data. The worker data includes personnel scheduling, personnel positioning, and fence rules.

[0082] A data communication network is constructed, consisting of a terminal layer, a convergence layer, and a core layer. The terminal layer is used to transmit environmental monitoring data and structural and gas data. The convergence layer is used to transmit energy equipment data, real-time captured images, and worker data. The core layer is used to update BIM model data.

[0083] Configure a communication protocol to transmit environmental monitoring data collected by smart sensors to the terminal layer, and aggregate the data into n node data packets according to the data volume. Send the real-time captured images to the aggregation layer. Update the corresponding nodes of the BIM model online according to the terminal layer and the aggregation layer. The communication protocol includes IEEE, HTTP, MQTT, JSONAPI, UDP, RTP and TCP.

[0084] In this embodiment, a gridded sensor deployment scheme driven by the BIM model is used to achieve full-area coverage monitoring of construction elements. Multi-protocol converged communication is used to build a protocol adaptive system. Real-time video uses the RTP protocol, equipment status data uses the MQTT protocol, and a three-layer network design of terminal layer, aggregation layer and core layer is set up to successfully build an intelligent data communication architecture.

[0085] Extracting online construction information and reconstructing a 3D environment scene based on the surrounding environment to obtain construction progress data specifically includes:

[0086] Key feature points are extracted from the real-time captured images, and visual SLAM stitching technology is used to stitch the real-time captured images together to generate a dynamic three-dimensional point cloud map.

[0087] The key feature points in the dynamic 3D point cloud map are aligned with the BIM model using the iterative nearest point algorithm and matched to the same coordinate system.

[0088] The BIM model is divided into m partitions according to construction type. When a partition of the BIM model completely overlaps with the dynamic 3D point cloud map, the partition is considered completed and rendered in green. When a partition of the BIM model partially overlaps with the dynamic 3D point cloud map, the partition is considered incomplete and rendered in red. The percentage of the volume of the overlapping part of the dynamic 3D point cloud map in the BIM model partition is calculated. The expected progress, actual progress, and the percentage of the volume are saved as the partition construction progress. When there are no key feature points in the dynamic 3D point cloud map in the partition of the BIM model, the coordinate content in the 3D point cloud map is marked as redundant point cloud and rendered in yellow in the BIM model.

[0089] In this embodiment, to achieve automatic quantitative determination of construction progress, volume overlap is introduced as an evaluation index for the actual completion of components.

[0090] The formula for calculating the volume overlap ratio R is as follows:

[0091]

[0092] in, This refers to the overlapping volume of dynamic 3D point clouds and BIM model components in the same coordinate system. This corresponds to the theoretical volume of the BIM component model.

[0093] when When the value is ≥0.95, it is determined that the construction zone corresponding to the component has been completed, recorded as "completely overlapping", and rendered as green in the BIM model;

[0094] When 0 < When the value is less than 0.95, it is judged as "partially overlapping", indicating that the component is under construction and is rendered in red in the BIM model;

[0095] When there are point cloud voxels in the dynamic 3D point cloud that fail to match any BIM component, that part is identified as redundant point cloud, and its corresponding area is rendered in yellow in the BIM model to indicate possible construction deviations or temporary stacking behavior.

[0096] In this embodiment, improved visual SLAM technology is used to realize real-time 3D reconstruction of the construction scene. Intelligent progress recognition and scheduling are adopted to improve the accuracy of coordinate system matching. The accuracy of key point partition recognition reduces the calculation error of volume fraction ratio. In addition, red, green and yellow three-color visualization coding is used to accurately render the online status of construction progress, providing data support for subsequent construction period prediction.

[0097] Based on construction progress monitoring and online construction information, environmental fluctuation values ​​are calculated, and construction adjustment strategies are set, including:

[0098] Based on the environmental monitoring data and energy equipment data of the online construction information, the construction progress of the zone is dynamically compensated, and the construction adjustment strategy includes a first adjustment, a second adjustment and a third adjustment;

[0099] The construction environment temperature and wind speed data are extracted from the environmental monitoring data. The standard deviation and mean of each data point within a unit of time are calculated. The sum of the ratios of the standard deviations and means of the construction environment temperature and wind speed is taken as the environmental fluctuation value. The number of days of inventory available in the material inventory data is determined based on the material usage of the next construction zone. K-means density clustering is used to dynamically classify the environmental fluctuation value to obtain a first classification interval, a second classification interval, and a third classification interval. The first classification interval < the second classification interval < the third classification interval. When the environmental fluctuation value is within the first classification interval and the number of days of inventory available is greater than three days, the first adjustment is triggered.

[0100] When the environmental fluctuation value is within the second category range and one day < number of days of inventory availability < three days, the second adjustment is triggered;

[0101] When the environmental fluctuation value falls within the third category range and the inventory availability days are less than one day, the third adjustment is triggered.

[0102] The first adjustment is used to extend the construction period and reduce the load on high-altitude operations;

[0103] The second adjustment is used to adjust the normal construction time and increase worker scheduling;

[0104] The third adjustment is used to stop outdoor operations and extend the construction period.

[0105] In this embodiment, an environmental fluctuation value is introduced to quantify the impact of the construction environment on the construction progress. As an indicator for evaluating environmental stability.

[0106] The environmental fluctuation value W is calculated as follows:

[0107]

[0108] in, and These represent the standard deviation and mean of the ambient temperature during construction, respectively, within a unit of time. and These represent the standard deviation and mean of wind speed per unit time, respectively.

[0109] Environmental fluctuation value It reflects the degree of fluctuation in the construction environment; the higher the value, the more unstable the construction conditions.

[0110] After obtaining the environmental fluctuation value sample sequence, the K-means clustering algorithm was used to dynamically classify the environmental fluctuation values. The number of clusters was set to 3, corresponding to the low fluctuation range, medium fluctuation range and high fluctuation range respectively.

[0111] The following example intervals can be obtained after clustering:

[0112] First category interval: ≤0.30 (low fluctuation range);

[0113] Second category interval: 0.30< ≤0.60 (medium fluctuation range);

[0114] Third category interval: >0.60 (high volatility range).

[0115] The system combines the classification ranges with the number of days of material inventory available to trigger different levels of construction adjustment strategies, so as to achieve adaptive adjustment of construction progress to environmental changes.

[0116] In this embodiment, a dynamic compensation mechanism is set up to realize intelligent coordinated control of construction progress and resource environment. Based on the dual-dimensional evaluation model of environmental fluctuation value and inventory days, a three-level adjustment strategy is established. The environmental fluctuation value algorithm is integrated with the variation coefficients of temperature and wind speed, and K-means clustering is used to optimize the classification interval, so as to carry out digital and precise control of the level. Then, a third-order adjustment strategy is established, and the adjustment strategy is adjusted according to the magnitude of environmental fluctuation value from slow to fast, so as to realize the automatic adjustment of nonlinear data.

[0117] The construction of the gray-scale prediction model, based on the construction progress data and the adjusted construction period prediction critical value, specifically includes:

[0118] The deviation data of the construction progress of the partition is calculated, where the deviation data = expected progress - actual progress. The deviation data is stored in sequence. The least squares solution of the gray differential equation is solved using the gray prediction model GM to obtain the prediction critical value. When the deviation data is ≥ 0.7 times the prediction critical value, the dynamic compensation strategy library is called to record the actual deviation reduction and compensation time. The compensation efficiency is calculated based on the actual deviation reduction and compensation time, where the compensation efficiency = actual deviation reduction / compensation time. The gray prediction model GM is dynamically updated based on the prediction critical value and the compensation efficiency to generate a new round of prediction critical values.

[0119] In this embodiment, a gray-scale prediction model GM(1,1) is constructed based on historical construction progress deviation data to predict the critical value of future construction progress risk.

[0120] The construction progress deviation data is represented in time series form as follows:

[0121] , ,in As expected, This represents the actual progress.

[0122] The system is set to trigger a schedule risk warning when the real-time schedule deviation reaches 70% of the predicted critical value. When the time comes, the corresponding dynamic compensation strategy will be automatically invoked. To predict the critical value.

[0123] After the compensation is completed, the system records the change in deviation before and after compensation, as well as the compensation duration, and calculates the compensation efficiency η, which is calculated using the following formula:

[0124]

[0125] in, This refers to the actual reduction in schedule deviation after the implementation of the compensatory measures. The duration of the compensation measures.

[0126] Based on the compensation efficiency results, the system dynamically updates the gray-scale prediction model parameters and the schedule risk trigger threshold to improve the accuracy of subsequent predictions and decisions.

[0127] In this embodiment, based on the improved gray-scale prediction system, an intelligent compensation decision-making mechanism based on the progress prediction critical value is established, and a dynamic self-learning closed-loop optimization system is set up. The model parameters are automatically updated every hour to improve the accuracy of the prediction critical value, thereby realizing intelligent management and control of construction progress risks.

[0128] The visualization of construction progress data, 3D environment scenes, and progress threshold prediction results into a BIM simulation dashboard specifically includes:

[0129] A BIM simulation screen is built on the cloud platform, overlaying the 3D environment scene and the BIM model, and marking the corresponding sensor points. A query interface, a grayscale prediction interface, and a visualization rendering interface are set up, and an online construction information interface is introduced. By inputting data parameters, specific element attributes are queried, and the current stage of construction completion of the 3D environment scene is modeled and rendered onto the BIM simulation screen. The completed parts and progress of different zones can be queried.

[0130] In this embodiment, a three-dimensional scene fusion intelligent visualization management and control platform is built based on the BIM model. The three-dimensional scene is integrated and aligned with the BIM model, and multiple sensor data are mapped to the visualization screen in real time to realize multi-scale observation of data. In addition, the screen can also support the deployment of critical value prediction results, which makes it easier for staff to effectively and quickly track the real-time construction progress and provide effective solutions for potential risks in the future of the project, thereby improving the efficiency of project management and control.

[0131] The query interface is used to query real-time data from different sensors and update online construction information;

[0132] The grayscale prediction interface is used to deploy the grayscale prediction model GM and generate prediction critical values.

[0133] The visualization rendering interface is used to visualize and render 3D environment scenes.

[0134] In this embodiment, three core functions are achieved through an intelligent visualization platform, which significantly improves the efficiency of project management. The query interface for multi-source data fusion supports real-time monitoring by multiple sensors, the dynamic prediction interface integrates an improved GM(1,1) model to improve the speed of early warning response, and the 3D rendering engine realizes high-precision scene reconstruction to improve the accuracy of progress visualization.

[0135] The dynamic compensation strategy library includes a first strategy, a second strategy, and a third strategy;

[0136] The first strategy is used to increase labor to shorten the construction period, with a delay of less than two days;

[0137] The second strategy is used to shorten the construction period by extending workers' working hours, with the delay exceeding two days;

[0138] The third strategy is used to call on backup equipment to shorten the construction period, with a delay of less than one day.

[0139] In this embodiment, the dynamic compensation strategy library is divided into three levels through a hierarchical response mechanism to achieve precise schedule optimization. The first strategy is used as labor adjustment to solve short-term schedule delay recovery. The second strategy is used as working time extension adjustment to solve medium- and long-term delays. The third strategy is used as equipment emergency strategy to achieve crisis management within 24 hours and reduce schedule delay rate.

[0140] Specific implementation method of the intelligent safety early warning module:

[0141] The intelligent safety early warning module operates as follows: The system uses an infrared camera and sensor network deployed on key nodes of tower cranes, scaffolding, and workers' safety helmets to collect video streams and posture data in real time. A lightweight YOLOv5s model is used for real-time behavior detection, identifying violations such as "not wearing a safety helmet" and "working at heights without protection." Simultaneously, predefined "high-risk work areas" (such as foundation pits and hoisting areas) in the BIM model are overlaid with real-time personnel positioning data for spatial collision detection. When the system determines that the risk level exceeds a threshold, firstly, an on-site audible and visual alarm sounds at the corresponding location; secondly, a vibration alarm is sent to the smart bracelet worn by the at-risk personnel via a LoRa wireless network; finally, a flashing red pop-up window alerts management personnel on the project's BIM management screen, and a screenshot is automatically taken and archived. The implementation of this module elevates safety supervision from passive response to proactive prevention, significantly reducing the incidence of safety accidents at construction sites.

[0142] Specific implementation methods of the adaptive learning optimization mechanism:

[0143] The adaptive learning optimization mechanism runs as a background service. It continuously collects deviation data, environmental fluctuation values, triggered adjustment strategies, and their compensation efficiencies from the progress prediction module. Each week, the system automatically initiates an optimization cycle: First, it retrains the gray-scale prediction model (GM) using data from the past four weeks, finding the optimal background value coefficients through a grid search method; second, it constructs a reward function using compensation efficiency data, optimizing the mapping relationship from "environmental fluctuation value - inventory days" to "adjustment strategy" using a Q-learning algorithm, enabling the system to select the strategy with the highest historical compensation efficiency under similar scenarios. After the optimized model and strategy pass simulation verification, they are automatically deployed to the production environment during low-load nighttime hours. This mechanism allows the system to adapt to different project types (such as residential buildings, bridges, and factories) and the construction characteristics of different seasons, with prediction accuracy continuously improving over time.

[0144] Referring to Figure 1, which illustrates an embodiment of the present invention, a method system for monitoring the progress of power grid engineering projects is provided, comprising a project drawing construction module, a sensor deployment module, a construction progress module, a progress prediction module, and a visualization module;

[0145] The project drawing construction module is used to construct a BIM model based on the image data of the engineering site and the project drawings;

[0146] The sensor deployment module is used to deploy smart sensors on-site according to the BIM model and collect online construction information.

[0147] The construction progress module is used to extract online construction information, reconstruct a three-dimensional environmental scene based on the surrounding environment, obtain construction progress data, calculate environmental fluctuation values ​​based on construction progress monitoring and online construction information, and set construction adjustment strategies.

[0148] The progress prediction module is used to construct a gray-scale prediction model and predict the critical value of the progress based on the construction progress data and the adjusted construction period.

[0149] The visualization module is used to visualize construction progress data, three-dimensional environmental scenes, and progress critical value prediction results to generate a BIM simulation screen.

[0150] In this embodiment, a high-precision BIM model is constructed by matching multi-angle drone photography (front / back / left / right / bottom) with the gridded drawings, resolving the drawing conflict problem in traditional modeling. Furthermore, eight types of sensors are deployed using BIM grid points to achieve complete coverage monitoring of construction elements. Subsequently, SLAM technology is used to generate a dynamic point cloud map, which is compared with the BIM model to calculate volume overlap, and an adaptive compensation strategy is set to reduce resource waste. An improved grayscale prediction model, incorporating compensation feedback, predicts critical values ​​seven days in advance and dynamically optimizes and updates the model, improving the efficiency of digital collaboration in the construction process and driving the power grid project from experience-driven to data-driven. After integrating intelligent safety early warning and adaptive learning optimization mechanisms, the system forms a complete intelligent closed loop of "perception-decision-early warning-optimization".

[0151] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer application code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0152] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for monitoring the progress of a power grid engineering project, characterized in that, include: A BIM digital twin model is constructed based on image data and project drawings from the construction site. The physical space of the construction site is divided into multiple management grids based on construction procedures and safety risk levels. Intelligent sensing terminals are deployed at corresponding locations according to the attributes of each management grid to collect online construction information, and a multi-layer data communication network is established. Multi-source data collected by the intelligent sensing terminals is integrated to perform dynamic 3D scene reconstruction and real-time status rendering in the digital twin model, and the construction progress percentage of components within each grid is automatically calculated. A multi-dimensional state vector is constructed for each management grid. Based on a preset coupling analysis rule base, grayscale prediction model, and cross-grid event detection algorithm, construction progress deviations, safety risks, and abnormal cross-grid flow events are identified. Based on the identification results of the previous step, structured grid-linked alarm information containing event elements and handling suggestions is generated and published. Based on the handling feedback data of the alarm information, the coupling analysis rule base and the grayscale prediction model are iteratively optimized.

2. The method according to claim 1, characterized in that: Based on image data and project drawings from the engineering site, a BIM digital twin model is constructed. Based on construction procedures and safety risk levels, the physical space of the engineering site is divided into multiple management grids. Specifically, this includes: acquiring power grid engineering design drawings and on-site real-world data collected through UAV oblique photography to construct a 3D BIM digital twin model containing terrain, lines, towers, and substation components; dividing the physical space corresponding to the digital twin model into multiple management grids based on construction procedures, work area division, and safety risk assessment results; and configuring an attribute set for each management grid, the attribute set including at least a unique grid identifier, a list of associated BIM components, a preset safety level, and key progress nodes.

3. The method according to claim 2, characterized in that: Based on the attributes of each management grid, intelligent sensing terminals are deployed at corresponding locations to collect online construction information and establish a multi-layer data communication network. Specifically, this includes: deploying intelligent sensing terminals, including high-definition cameras, panoramic cameras, temperature and humidity sensors, wind speed sensors, equipment status sensors, and personnel positioning tags, at key locations in each management grid according to the grid's security level and function; constructing a three-layer communication architecture consisting of a terminal layer, an edge layer, and a platform layer. The terminal layer is used for data acquisition and encapsulation, the edge layer is deployed on-site for video analysis, data aggregation, and real-time computation, and the platform layer performs data fusion and centralized analysis; configuring an adaptive communication protocol to transmit sensor data via the MQTT protocol and transmit video streams to the edge layer via the RTSP or RTP protocol.

4. The method according to claim 3, characterized in that: By integrating multi-source data collected by the intelligent sensing terminal, dynamic 3D scene reconstruction and real-time state rendering are performed in the digital twin model, and the construction progress percentage of components within each grid is automatically calculated. Specifically, this includes: using video acquisition equipment deployed within the management grid to perform real-time modeling of the construction scene based on visual SLAM technology, generating a dynamic 3D point cloud map reflecting the actual state of the construction site, thereby achieving continuous perception of the construction entity form of each management grid; and employing the Iterative Closest Point (ICP) algorithm to spatially register the dynamic 3D point cloud map with the static BIM component model of the corresponding management grid in the digital twin model, unifying them to the same coordinate system. Improve the spatial matching accuracy between dynamic construction data and design models; based on the completion of spatial registration, calculate the volume overlap between the dynamic 3D point cloud and the corresponding BIM component model, and use the volume overlap as a quantitative indicator of the actual completion of the component. Automatically calculate and update the construction progress percentage of the component based on the overlap value, thereby realizing the automated and objective measurement of construction progress; fuse and map personnel positioning data, environmental sensor data, and equipment status data into the digital twin model according to the spatial registration results to form a comprehensive situational layer that is updated synchronously with the construction progress, so as to realize the intuitive display and collaborative analysis of construction status, environmental conditions, and resource operation status.

5. The method according to claim 4, characterized in that: To identify construction progress deviations, safety risks, and abnormal cross-grid flow events, a multi-dimensional state vector is constructed for each management grid. Based on a pre-defined coupling analysis rule base, a gray-scale prediction model, and a cross-grid event detection algorithm, the following steps are taken: A multi-dimensional state vector V=(P,H,E,S) is constructed for each management grid, where P is the progress deviation index, H is the personnel activity density index, E is the environmental risk index, and S is the equipment status index. A pre-defined coupling analysis rule base is established, which includes at least: Rule A, used to trigger a resource shortage warning when progress is lagging and personnel activity is insufficient; Rule B, used to trigger a violation cross-grid flow warning when an unauthorized target is detected moving to a high-risk grid. The gray-scale prediction model GM(1,1) is used to model historical progress deviation sequences, predict future progress risk thresholds, and trigger progress risk warnings when real-time progress deviations reach a preset proportion of the predicted threshold. For cross-grid events, the existence of cross-grid flow events is determined by analyzing the spatiotemporal continuity of dynamic target activity states between adjacent grids.

6. The method according to claim 5, characterized in that: Based on the identification results from the previous step, a structured grid-linked alarm information containing event elements and handling suggestions is generated and published. Specifically, this includes: extracting event elements of the warning event, including event type, source and target grid identifiers, risk level, associated progress tasks, and occurrence time; generating handling suggestions by calling the dynamic compensation strategy library based on the event type, wherein the strategy library includes one or more of the following: increasing construction resources, adjusting work time, and activating backup plans; combining the event elements and handling suggestions to generate structured grid-linked alarm information, and pushing it through the BIM visualization screen and mobile terminals.

7. The method according to claim 6, characterized in that: The iterative optimization of the coupled analysis rule base and the gray-scale prediction model based on the alarm information handling feedback data specifically includes: recording the handling response time, the measures taken, and the subsequent progress recovery status of each alarm information; calculating the actual compensation efficiency of the invoked dynamic compensation strategy based on the handling feedback data; and periodically iteratively optimizing the parameters of the gray-scale prediction model GM(1,1) and the trigger threshold of the coupled analysis rule base using the actual compensation efficiency and result data.

8. A grid-based digital twin-based collaborative monitoring system for the progress and safety of power grid projects, used to implement the method described in any one of claims 1 to 7, characterized in that, include: The digital twin and grid management module is used to build and maintain BIM digital twin models, as well as to perform grid division and attribute configuration. The gridded intelligent perception layer consists of various sensors, video acquisition devices, and positioning terminals deployed in each management grid; the edge computing and data aggregation module is used for real-time intelligent analysis of video streams, aggregation of multi-source data, and localized scene reconstruction; the progress-safety coupled analysis engine has a built-in coupled analysis rule base, grayscale prediction model, and cross-grid event detection algorithm for comprehensive risk assessment and event judgment; and the early warning and collaborative handling platform is used to generate and publish structured alarm information and provide handling strategy suggestions. The data storage and model optimization module is used to store data throughout the entire process and to optimize the models and rules in the analysis engine using a feedback mechanism.

Images