A method and system for visualizing construction data of a power project

Abstract

The present application belongs to the technical field of new energy engineering construction informatization management, and particularly relates to a power project construction data visual presentation method and system. Through high integration of real-time collection and fusion of multi-source construction data, state mapping of digital twin model, progress prediction and quality identification of artificial intelligence algorithm, trusted evidence and task automatic transfer of blockchain mechanism, and data collaborative display of multi-terminal, a smart construction platform with dynamic sensing capability, precise control capability and visual interaction capability is constructed. A three-dimensional full-closed loop structure of collection-modeling-analysis-presentation-feedback is constituted, which significantly improves the management transparency, response efficiency and risk controllability of new energy power construction projects in the construction stage, and has good practical value and promotion prospect.

Description

A method and system for visualizing construction data in power projects Technical Field

[0001] This invention belongs to the field of information management technology for new energy engineering construction, specifically relating to a method and system for visualizing construction data of power projects. Background Technology

[0002] In the construction of new energy power projects, especially wind power and photovoltaic projects which are widely distributed, have long cycles, and involve many stages, the ability to manage information during the construction phase has become a crucial factor affecting project quality, schedule, and cost. Construction sites involve a large number of personnel, equipment, materials, and work nodes. How to achieve real-time perception, status tracking, and process recording of these heterogeneous elements is a key foundation for realizing smart construction and refined management. With the continuous evolution of technologies such as Building Information Modeling (BIM), the Internet of Things (IoT), and artificial intelligence (AI), building a data-driven engineering construction management system is gradually becoming an industry trend.

[0003] Currently, the construction process of new energy engineering projects still largely relies on traditional methods such as manual recording, scattered forms, and offline reporting. Data acquisition is lagging, management methods lack integration, and it is difficult to achieve real-time monitoring and dynamic analysis of construction milestones. While some systems have introduced 3D models linked to schedules, most are limited to static displays or drawing annotations, lacking effective interaction with on-site data. Furthermore, the identification of construction quality problems largely depends on manual inspections, and early warning mechanisms are lagging. Project managers struggle to promptly grasp risk points, leading to potential problems such as opaque task flows and delays in problem handling.

[0004] Therefore, there is an urgent need to provide a comprehensive data visualization method applicable to the infrastructure construction phase of new energy power projects, in order to solve problems such as the inability to collect and feedback construction status information in a timely manner, insufficient risk prediction capabilities, and discontinuous task node control, so as to improve the perception capabilities, visibility, and decision support efficiency of the construction process. Summary of the Invention

[0005] This invention provides a method and system for visualizing construction data of power projects, in order to solve the technical problems in the prior art, such as the inability to collect and feedback construction status information in a timely manner, insufficient risk prediction capabilities, and discontinuous control of task nodes.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for visualizing construction data of power projects, comprising the following steps: acquiring multi-source data from the construction site; constructing a three-dimensional component model based on the multi-source data, simulating the construction state to obtain component state data; using a supervised learning method based on long short-term memory networks to predict the progress based on the multi-source data and component state data, and using an image recognition model based on convolutional neural networks to perform quality identification and safety scoring analysis on the image data in the multi-source data; and graphically displaying the progress prediction results, quality identification results, and safety scoring analysis results.

[0007] The multi-source data from the construction site includes operating parameters of construction equipment, environmental parameters and structural stress information, images and videos of construction components, as well as historical completion rates of construction tasks, number of construction personnel, operating time of machinery and equipment, and environmental monitoring data.

[0008] Based on BIM (Building Information Modeling), a three-dimensional component model is established for the construction objects at the construction site. Multi-source data from the construction site is mapped to each component in the three-dimensional component model, establishing a semantic binding relationship between components and task nodes. Different status labels are applied to the corresponding components according to different construction states.

[0009] The status indicators for components are adjusted by color rendering, material replacement, and transparency adjustment based on different construction states from multi-source data at the construction site.

[0010] Based on multi-source data and component status data, a supervised learning method using long short-term memory networks is adopted for progress prediction. Specifically, a time series model is established for key construction tasks using a long short-term memory network; historical data on task completion rate, equipment utilization rate, worker attendance rate, and weather conditions at the construction site are obtained, and the time series model is trained using historical data. Multi-source data from the construction site is used as the input features of the trained time series model to predict the expected completion time of task nodes, the number of delay days, and the risk level.

[0011] The method of using a convolutional neural network-based image recognition model to perform quality identification on image data from multiple sources is as follows: Based on the convolutional neural network image recognition algorithm, the method identifies defects such as template detachment, exposed rebar, cracks, hollow areas, and missing bolts in the images and video information of construction components from multiple sources at the construction site. For the defects identified in the images, defect information including the corresponding coordinates, type, and confidence level of the defects is generated, and the defect information is bound to the corresponding components.

[0012] Multi-source data from the construction site is acquired, including environmental parameters, personnel status parameters, and equipment operating parameters. Environmental risk factors are obtained by normalizing wind speed, rainfall, humidity, and temperature among the environmental parameters. Personnel risk factors are obtained by normalizing personnel number, personnel distribution density, and personnel working hours among the personnel status parameters. Equipment risk factors are obtained by normalizing operating load rate, continuous operating time, and maintenance cycle among the equipment operating parameters. The environmental risk factors, personnel risk factors, and equipment risk factors are then weighted according to preset weighting coefficients to obtain a comprehensive safety risk index. Based on the magnitude of the comprehensive safety risk index, the risk levels are classified as low, medium, and high. When the comprehensive safety risk index reaches the high-risk level, an early warning is automatically triggered, and the warning information is pushed to the visual interface and multi-terminal interaction module.

[0013] The SHA256 hash algorithm is used to encrypt the multi-source data from the construction site, and a structured index is built based on the Merkle tree structure for the multi-source data from the construction site.

[0014] Based on the WebGL graphics rendering framework, a visual interface is created to display system results to managers and frontline staff. Based on the analysis results of progress prediction, quality identification, and safety scores, visual rendering of color, lighting, and dynamic textures is performed. For progress prediction results, a timeline function is integrated, allowing users to drag or click to view the status of any point in the project progress, achieving spatiotemporal backtracking. For quality identification results, when a task is predicted to be at risk of delay or fail to meet quality standards, the system automatically flashes a prompt and displays it centrally in the risk panel.

[0015] A power project construction data visualization system includes a data acquisition unit, a 3D simulation unit, a prediction and analysis unit, and a visualization unit. The data acquisition unit acquires multi-source data from the construction site. The 3D simulation unit constructs 3D component models based on the multi-source data to simulate construction conditions and obtain component status data. The prediction and analysis unit uses a supervised learning method based on long short-term memory networks to predict progress based on the multi-source data and component status data, employs an image recognition model based on convolutional neural networks to perform quality identification and safety scoring analysis on image data from the multi-source data, and displays the progress prediction results, quality identification results, and safety scoring analysis results graphically.

[0016] Compared with existing technologies, the present invention has the following beneficial effects: The present invention clearly acquires multi-source data from the construction site, covering construction equipment operating parameters, environmental parameters, structural stress information, construction component images and videos, historical completion rate of construction tasks, number of construction personnel, mechanical equipment operating time, and environmental monitoring data, achieving full-dimensional data coverage of equipment, environment, personnel, components, and tasks; at the same time, through data preprocessing, it ensures that the data can be quickly uploaded to the central server via industrial Ethernet or 4G / 5G networks without secondary manual processing. Compared with the traditional daily data collection cycle of manual recording, the data collection cycle is shortened to minutes or even seconds, solving the problem of scattered and untimely data collection from the source.

[0017] Furthermore, this invention constructs a 3D component model based on BIM (Building Information Modeling), maps multi-source data to each component, and establishes a binding relationship between components and task nodes. Different construction states are identified through color rendering, material changes, and transparency adjustments. Combined with a WebGL graphics rendering framework, a visual interface is formed, supporting multi-level view switching between project overview, sections, components, and details. Clicking on a component allows users to view information such as the construction unit, planned and actual construction periods, and quality inspection results. Compared to the hour-long feedback delay of traditional manual inspections followed by verbal reports, this method achieves synchronous linkage between changes in on-site status and updates to the visual interface. Abnormal states are immediately alerted through dynamic indicators, reducing feedback lag time to seconds. Managers can intuitively grasp on-site dynamics without having to sift through multiple reports, avoiding the expansion of construction deviations due to information lag.

[0018] Furthermore, a supervised learning method using Long Short-Term Memory (LSTM) networks is employed for progress prediction. A time-series model is trained using historical data on construction task completion rates, equipment utilization rates, worker attendance rates, and weather conditions. Real-time multi-source data serves as input features, outputting the estimated completion time of task nodes, the predicted number of delay days, and the risk level. Compared to the limitations of traditional static Gantt charts, which only display planned progress, the LSTM model, trained on multi-dimensional historical data, can identify high-risk delay nodes in advance and push warning information to the management end. Managers can then adjust resource allocation in advance, transforming post-delay remediation into proactive avoidance, reducing delay-related costs such as equipment rental delays and idle labor.

[0019] Furthermore, an image recognition model based on convolutional neural networks is employed to analyze images and video information of construction components from multi-source data. This allows for real-time identification of quality defects such as formwork gaps, exposed rebar, cracks, hollow areas, and missing bolts. The model generates the corresponding coordinates, type, and confidence level of each defect and assigns them to the relevant component. Compared to traditional manual inspections, which rely on personnel experience, have long inspection cycles, and are prone to missed defects, the convolutional neural network image recognition model can analyze the collected image data in real time. Defect identification response time is reduced to seconds, and the model can accurately locate and quantify the severity of defects, preventing small defects from escalating into major problems and increasing rework costs. Defect information is synchronized to a visual interface and relevant terminals, allowing supervisors to quickly arrange rectification. The quality risk handling cycle is shortened from days to hours.

[0020] Furthermore, this invention employs the SHA256 hash algorithm to encrypt multi-source data from the construction site and constructs a structured index based on a Merkle tree structure. It focuses on on-chain notarization of key data such as task completion records, quality inspection results, and safety scores. The SHA256 hash algorithm ensures that data cannot be tampered with once it is on the chain; tampering will change the hash value, allowing for rapid verification of data authenticity and preventing node gaps caused by forged task completion records. The Merkle tree uses the unique component identifier as the core association identifier, enabling rapid tracing of historical data for any task node and solving the problem of difficulty in tracing historical states.

[0021] Furthermore, this invention maps multi-source data onto a 3D component model, establishing a binding relationship between components and task nodes. Combined with the timeline function of the visualization interface, the status of components and task nodes at any point in the project progress can be viewed by dragging or clicking, completely restoring the evolution process of task nodes and avoiding control gaps caused by missing node information. Simultaneously, smart contracts can automatically trigger subsequent task flows based on node status without manual intervention, solving the discontinuity problem of node flows relying on manual triggering. Attached Figure Description

[0022] Figure 1 is a schematic diagram of the structure of a power project construction data visualization system according to an embodiment of the present invention. To further understand the content of this invention, the following detailed description is provided in conjunction with the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not intended to limit the scope of the invention.

[0025] Example 1 This example proposes a method for visualizing construction data of power projects, including the following steps: acquiring multi-source data from the construction site; constructing a three-dimensional component model based on the multi-source data to simulate the construction state and obtain component status data; using a supervised learning method based on long short-term memory networks to predict the progress based on the multi-source data and component status data; using an image recognition model based on convolutional neural networks to perform quality identification and safety scoring analysis on the image data in the multi-source data; and graphically displaying the progress prediction results, quality identification results, and safety scoring analysis results.

[0026] Based on the above methods and steps, this embodiment further proposes a power project construction data visualization system, including a data acquisition unit, a 3D simulation unit, a predictive analysis unit, and a visualization unit. This power project construction data visualization system is suitable for information management during the construction phase of new energy power projects, especially for on-site application scenarios with large project volumes, complex nodes, and long construction periods, such as wind farms and photovoltaic power stations. By integrating the data acquisition unit, 3D simulation unit, predictive analysis unit, and visualization unit into the perception layer, modeling layer, analysis layer, display layer, and interaction layer, the system achieves digital mapping, status analysis, and multi-dimensional display of the construction process, forming a data-driven intelligent construction support platform.

[0027] The data acquisition unit is used to acquire multi-source data from the construction site; the three-dimensional simulation unit is used to construct a three-dimensional component model based on the multi-source data and simulate the construction state to obtain component state data; the prediction and analysis unit is used to perform progress prediction using a supervised learning method based on long short-term memory networks, and to perform quality identification and safety scoring analysis on image data in the multi-source data using an image recognition model based on convolutional neural networks, based on the multi-source data and component state data; the visualization unit is used to graphically display the progress prediction results, quality identification results, and safety scoring analysis results.

[0028] Specifically, referring to the structural diagram of a power project construction data visualization system shown in Figure 1, the data acquisition unit includes a data acquisition and fusion module. This module is located at the forefront of the construction site and is used for the unified acquisition and preliminary processing of multi-source data, including equipment operating parameters collected by industrial sensors, photos and videos of the construction site taken by image acquisition equipment, personnel distribution and work trajectories acquired by positioning devices, and meteorological data collected by weather monitoring equipment. The acquired data is fused and processed in real time by an edge computing gateway and then uniformly uploaded to the central server. The three-dimensional simulation unit includes a digital twin modeling module. This module constructs a three-dimensional digital space that corresponds one-to-one with the construction site. Based on a BIM model, each component is assigned a unique identifier and bound to the construction task status. This module dynamically drives the construction object's state, realizing the mapping from "physical engineering" to a "computable model," and is the fundamental support unit for visualization, simulation analysis, and automatic control. The predictive analysis unit includes an intelligent analysis and early warning module and a blockchain data management module. The intelligent analysis and early warning module, based on a deep learning model, analyzes construction data in real time and identifies potential problems. This module includes a progress prediction submodule for predicting the completion time of construction nodes, and a quality identification submodule for identifying image defects (such as cracks, misassemblies, etc.). The prediction or identification results are used for front-end reminders, component status updates, and process control. The blockchain data management module encrypts and stores key node information in the construction process (such as task signature confirmation and quality inspection results) to ensure its immutability and traceability. Task flow rules are set through smart contracts, automatically triggering subsequent task startups after status change conditions are met, achieving automation of process control and ensuring data reliability. The visualization unit includes a visualization presentation module and a multi-terminal collaborative interaction module. The visualization presentation module generates a high-fidelity interactive graphical interface based on the WebGL 3D rendering engine, supporting layer-by-layer view switching from the project overview level to the individual component level. Through color changes, animation transitions, status pop-ups, and timeline functions, it achieves intuitive display of construction status and historical evolution retrospection. The multi-terminal collaborative interaction module provides three display and interaction methods: web page, mobile terminal, and information screen terminal, serving project management personnel, construction site personnel, and on-site dispatch and command scenarios, respectively. Each terminal keeps synchronized with the system's core data bus via the WebSocket protocol, enabling cross-terminal data display, integrated sharing, and feedback transmission.

[0029] The six modules mentioned above are interconnected according to data flow, logic flow, and control flow, forming a complete, well-defined, and real-time intelligent construction data visualization system platform. The data acquisition and fusion module, as the front-end perception layer, provides real-time data support to the modeling and analysis modules. The status results generated by the latter are fed back to the visualization module and terminal interaction module, creating a closed-loop flow to the control logic and process management, forming a fully intelligent collaborative chain. The specific implementation method is as follows: The data acquisition and fusion module includes various industrial sensors deployed at key locations such as construction machinery, structural nodes, and foundations; image acquisition equipment from fixed cameras, wearable camera modules, and drone platforms; environmental monitoring equipment; and personnel positioning devices. The industrial sensors collect real-time data on physical quantities such as machinery operating time, operating load, and structural stress. The image acquisition equipment periodically collects images and video data of the construction process. The environmental monitoring equipment collects environmental variables such as wind speed, rainfall, humidity, and temperature, as well as weather data, for use by the intelligent analysis module. The personnel positioning device obtains the real-time location, number of personnel, and work trajectory information of construction workers. It also obtains historical completion rate data for construction tasks from the construction information platform. The data acquisition and fusion module, through an edge computing gateway deployed in the field distribution box or central control cabinet, performs data preprocessing on the multi-source data acquired from the construction site. This preprocessing includes format standardization, anomaly removal, timestamp alignment, and local caching. The preprocessed multi-source data from the construction site is then transmitted to the central system server via industrial Ethernet or 4G / 5G networks. This module ensures the system's data real-time performance, accuracy, and front-end computing capabilities, forming the foundation for digital twin creation and analytical judgment.

[0030] The digital twin modeling module is used to construct 3D component models based on multi-source data and achieve dynamic mapping of construction status, providing a core carrier for visualization. Its specific implementation process is as follows: Based on the imported BIM building information model, 3D component models are established for construction objects such as wind turbine foundations, photovoltaic supports, transformer substation bases, cable trenches, and roads at the construction site. A unique identifier (UID) is assigned to each component, and this identifier is bound to the task number of the construction task node, forming a one-to-one correspondence between multi-source data and the 3D component model, ensuring accurate matching between data sources and model structure. The digital twin modeling module receives multi-source data from the construction site after data preprocessing from the data acquisition and fusion module. Based on different construction statuses, it performs status identification operations on the corresponding components: color rendering distinguishes construction stages, with yellow indicating construction in progress and green indicating completion; material changes reflect changes in the physical properties of the components; transparency adjustments highlight key areas of interest; and a dynamic red flashing rendering method is used to identify components in delayed or abnormal states.

[0031] Furthermore, the digital twin modeling module supports binding information such as detection results, construction logs, and image annotations from multi-source data to the corresponding 3D component models according to the component UID, realizing the associated storage of various information and components. Its stored information is linked to the timeline function of the visualization module, calling historical status data based on timeline parameters to reconstruct and display the component status at any point in the project, achieving spatiotemporal backtracking of the construction status. Through the above process, the digital twin modeling module organically integrates multi-source data with 3D component models, forming a real-time mapping of the construction status. This provides basic model support for the subsequent visualization of progress prediction, quality identification, and safety scoring, and its processing results are directly pushed to the visualization module for graphical display.

[0032] The intelligent analysis and early warning module is a core functional module of the predictive analysis unit. It receives multi-source construction site data preprocessed by the data acquisition and fusion module and component status data generated by the digital twin modeling module. It achieves progress prediction based on the supervised learning method of long short-term memory network and quality identification based on the image recognition model of convolutional neural network. The output results are pushed to the digital twin modeling module, the visualization module and the blockchain data management module, respectively. It provides technical support for the predictive analysis and graphical display functions in the claims and is the key to the system's construction risk management.

[0033] The intelligent analysis and early warning module performs progress prediction by: acquiring multi-source data preprocessed by the data acquisition and fusion module, including historical completion rates of construction tasks, equipment utilization rates corresponding to the number of construction personnel, worker attendance rates, and historical weather data collected by environmental monitoring equipment; firstly, for key construction tasks, a time series model is constructed using a long short-term memory network, and then the time series model is trained using the aforementioned historical data; after training, the multi-source data from the construction site is used as the input features of the time series model, and the model calculates the estimated completion time of task nodes, the estimated number of delay days, and the risk level. After the progress prediction results are generated, they are pushed to the digital twin modeling module to update the progress status of the corresponding components, to the visualization module for subsequent graphical display, and to the blockchain data management module for data storage. High-risk and medium-risk early warning information is also pushed to relevant management and construction personnel through a multi-terminal collaborative interaction module.

[0034] The intelligent analysis and early warning module performs quality identification by: acquiring images and videos of construction components from multi-source data at the construction site through the data acquisition and fusion module; and using a convolutional neural network image recognition algorithm to identify defects in the construction components within these images and videos. Identifiable defect types include formwork gaps, missing rebar, cracks, hollow areas, and missing bolts. For each defect identified in the image, defect information is generated, including the corresponding coordinates, defect type, and confidence level, and this defect information is then linked to the corresponding construction component. The quality identification results and defect information are pushed to the digital twin modeling module to update the quality status of the corresponding component, to the visualization module for subsequent graphical display, and to the blockchain data management module for data storage. For defect identification results requiring re-inspection, the multi-terminal collaborative interaction module is also linked to push the results to the supervisor to trigger a manual review process.

[0035] The intelligent analysis and early warning module performs safety scoring analysis as follows: First, it acquires multi-source data from the construction site from the data acquisition and fusion module, including environmental parameters, personnel status parameters, and equipment operating parameters. Specifically, the environmental parameters include meteorological environmental information such as wind speed, rainfall, humidity, and temperature; the personnel status parameters include the number of construction workers, personnel distribution density, personnel working hours, and trajectory information; and the equipment operating parameters include data such as the operating load rate of mechanical equipment, continuous operating time, and equipment maintenance cycle.

[0036] Based on multi-source data of environmental parameters, personnel status parameters, and equipment operating parameters at the construction site, a safety risk assessment model is constructed. This model employs a hierarchical weighted approach for comprehensive calculation. Specifically, wind speed, rainfall, humidity, and temperature are normalized to obtain environmental risk factors; personnel number, personnel density, and personnel working hours are normalized to obtain personnel risk factors; and operating load rate, continuous operating time, and maintenance cycle are normalized to obtain equipment risk factors. Then, based on project safety management regulations and historical accident data, weighting coefficients are set for the environmental risk factors, personnel risk factors, and equipment risk factors, and a comprehensive safety risk index is calculated using the following formula:

[0037] in, To comprehensively assess the safety risk index, As an environmental risk factor, For personnel risk factors, As equipment risk factors, These are the weighting coefficients for environmental risk factors. The weighting coefficients for personnel risk factors. Weighting coefficients for equipment risk factors satisfy .

[0038] Furthermore, the safety risk index is graded according to a set threshold range, divided into low risk (0–0.4), medium risk (0.4–0.7), and high risk (0.7–1.0). The comprehensive safety risk index is calculated based on the above calculations. When the comprehensive safety risk index reaches a high risk level, the system automatically triggers a risk warning, which is highlighted in red on the 3D component model by the visualization module and pushed to relevant management personnel and on-site workers through the multi-terminal collaborative interaction module.

[0039] In addition, the system also performs trend analysis on the time series data of risk factors and calculates the rate of change of the risk index using the sliding window method. If a significant upward trend appears in a short period of time, an early warning will be triggered even if the current level is medium risk, in order to prevent safety hazards from evolving into accidents.

[0040] Through the above safety scoring and analysis methods, the system can quantify and comprehensively evaluate environmental changes, personnel status, and equipment operation during the construction process in real time, form a dynamic safety level and provide immediate feedback, thereby enabling the prediction and pre-control of construction safety risks.

[0041] The blockchain data management module ensures the immutability of critical system data and the automatic execution of process controls, providing technical support for data trustworthiness. Its specific implementation process is as follows: It receives critical data from the data acquisition and fusion module, intelligent analysis module, and digital twin modeling module, including construction signature records, quality inspection reports, task completion logs, component status information, and AI quality inspection results. The SHA256 hash algorithm is used to encrypt the data, generating a fixed-length hash value. This hash value, along with the data index information, is stored in a private blockchain node, achieving on-chain evidence storage of critical data. A structured index is built on the on-chain data blocks based on a Merkle tree structure. Using the component UID as the core association identifier, the hash values ​​of related data for the same component are used as leaf nodes. Non-leaf nodes are generated layer by layer through hash operations, ultimately forming a root node hash bound to the component UID, improving data retrieval efficiency. Smart contracts are deployed to automate the process control. Contract triggering conditions include a completed component status pushed by the digital twin modeling module, a qualified AI quality inspection result generated by the intelligent analysis module, and continuous consistency of on-chain work log records. When all the above conditions are met simultaneously, the contract automatically activates the subsequent task status, pushes task flow information to the visualization module and related terminals, and drives the construction process forward. Data auditing functionality is also provided, allowing regulatory agencies to query the full historical information of corresponding data based on the component UID, including the time and responsible person for each signing action, the inspection results and review records of the quality inspection process, and the complete chain of component status changes, ensuring the openness, transparency, and traceability of construction process data. The blockchain data management module collaborates with other modules of the system to achieve encrypted storage of key data, automatic process driving, and data auditability, providing a reliable data foundation and process guarantee for the visualization presentation method of the entire power project construction data.

[0042] The visualization module, built on the WebGL graphics rendering framework, graphically displays progress prediction results, quality identification results, and safety score analysis results, providing managers and frontline personnel with an intuitive information interaction interface. Its specific implementation process is as follows: The visualization module receives the 3D component models and component status data output by the digital twin modeling module. Combined with the progress prediction results, quality identification results, and safety scores generated by the predictive analysis unit, it constructs a multi-level view structure: from the project overview view, users can switch layer by layer to the section area view, the overall component view, and the unit construction view. Clicking allows for view drill-down, ensuring accurate viewing of construction information at different levels. Based on the component status pushed by the digital twin modeling module and the results of the predictive analysis unit, visual rendering is performed on the 3D component models: color changes distinguish the component's construction progress (e.g., under construction, completed), quality status (e.g., qualified, defective), and safety level (e.g., low risk, high risk); lighting adjustments highlight key construction areas; and dynamic texture changes indicate real-time updates to the component status, achieving a visual mapping of the construction status. The visualization module, based on the analysis results of progress prediction, quality identification, and safety scoring, sets up interactive hotspots on the 3D component model. Clicking on a component's hotspot brings up an information panel displaying basic component information, construction unit, planned and actual construction periods, quality inspection conclusions, and defect images, ensuring quick access to key data associated with the component. It also integrates a timeline function, linking with historical status data from the digital twin modeling module. By dragging the timeline slider or clicking on historical time points, the module retrieves the corresponding component status data and progress prediction results, reconstructing and displaying the construction status at any point in the project's history, enabling spatiotemporal regression of construction progress. Furthermore, the visualization module includes a risk warning mechanism. Based on delay risk warnings and quality non-compliance information pushed by the predictive analysis unit, components with potential risks are automatically flashed, and the risk information is centrally displayed in the interface's risk panel, facilitating quick identification and handling of risk issues. The graphical display of progress prediction, quality identification, and safety scoring analysis results provides foundational data for subsequent multi-terminal collaborative interaction modules, ensuring consistency in display content and interaction logic across different terminals.

[0043] The multi-terminal collaborative interaction module is designed with differentiated functions for different usage scenarios, enabling multi-terminal adaptation and user interaction of the content output by the visualization module, and ensuring the flow and synchronization of system data across different terminals. The specific implementation process is as follows: For the graphical display content output by the visualization module, the multi-terminal collaborative interaction module designs three types of terminal functions to meet different user needs: The web-based terminal is for management personnel, providing a complete information console, construction plan adjustment tools, and construction data statistical analysis panel, supporting in-depth querying and operation of progress prediction results, quality identification results, and safety scores; the mobile terminal is for construction teams and supervisors, supporting the acquisition of component association information by scanning codes, uploading construction image data, and recording voice notes, while also receiving construction task change notifications; the information screen terminal is set up in the project command center, using a combination of charts and 3D graphics to centrally display the overall construction status of the site, the risk heat map of the construction area, and the critical path of construction, adapting to multi-party consultation and dispatch command scenarios. The multi-terminal collaborative interaction module establishes a high-frequency data synchronization link with the system's main server via the WebSocket protocol. It receives updated display content from the visualization module in real time and simultaneously feeds back interactive information such as image data and voice annotations uploaded by mobile devices to the data acquisition and fusion module. This enables bidirectional flow of core data between terminals and the system, ensuring that the progress prediction results, quality identification results, and security score analysis results displayed by each terminal remain consistent. When a terminal triggers a data update, the system's main server synchronously pushes the updated content to other terminals, ensuring that the information obtained by different users is consistent in real time and supporting cross-role collaborative work.

[0044] Example 2: In a centralized photovoltaic power station construction project, the power project construction data visualization system described in this invention was deployed on-site. The following uses "photovoltaic support foundation construction" as an example to illustrate the typical operation process of the system on-site: Step 1: Front-end acquisition and fusion of construction data 1.1 Industrial sensors, image acquisition equipment and personnel positioning devices were deployed in the construction area of ​​the project site.

[0045] 1.2 Industrial sensors collect data in real time, such as the discharge information of concrete mixing plants, the operating status of construction machinery, and the running time of vibratory compaction equipment.

[0046] 1.3 The image acquisition device acquires construction images every 10 minutes and uploads them to the edge computing gateway.

[0047] 1.4 The edge computing gateway cleans the raw data, removes anomalies, and synchronizes timestamps before uploading it to the central server.

[0048] Step 2: Component Status Modeling and Binding 2.1 The system automatically reads the component model number from the BIM drawings and binds the collected construction task data with the component UID one by one.

[0049] 2.2 Each component is displayed as gray (not under construction), yellow (under construction), green (completed), or red (abnormal) according to its current construction status.

[0050] 2.3 Personnel attendance records, image acquisition records, and construction unit information are bound to the corresponding component nodes, forming a complete state mapping chain.

[0051] Step 3: Intelligent Analysis and Anomaly Prediction 3.1 The system calls the LSTM prediction model to analyze the construction data of Area C, outputs the delay risk level as "medium", and estimates the delay days as 2.3 days.

[0052] 3.2 The system synchronously performs convolutional neural network recognition on the uploaded images, identifies two instances where the reinforcing bars are not properly tied, marks them as "pending rectification", and pushes the results to the supervision terminal.

[0053] Step 4: Blockchain Data Storage and Task Control 4.1 After the on-site verification personnel confirm that the components are qualified, the system performs SHA256 hash processing on the completion confirmation information, quality inspection records, and other contents, and generates block data.

[0054] 4.2 The blockchain module indexes data through a Merkle tree structure and writes it into the private chain to achieve data immutability and subsequent traceability.

[0055] 4.3 Once the smart contract determines that the construction status is "completed", the quality status is "qualified", and the construction log is consistent, it will automatically activate the next construction task "scaffold installation" and send the task assignment information to the responsible person.

[0056] Step 5: 3D Visualization and Terminal Linkage Display 5.1 The visualization module switches the color of the support foundation components from "yellow" to "green" in real time, and pops up a window on the 3D interface to display the completion time, quality inspection report, and on-site images.

[0057] 5.2 Users can drag the timeline to any historical point in time to view the daily construction status evolution and review delays.

[0058] 5.3 The mobile application synchronously receives status change data, and construction personnel can scan the code to view component details, defect pictures and task feedback information.

[0059] 5.4 The information screen in the project department's dispatch hall displays a real-time situation map of the entire site, an overview of the construction progress, task warning areas, and a heat map of safety risks, for unified command and dispatch by leaders and managers.

[0060] This embodiment realizes full-process collaboration from front-end data collection to back-end data fusion, model mapping, AI analysis, on-chain authentication, and multi-terminal linkage display, effectively supporting the intelligent, data-driven, and transparent management of large-scale new energy power projects during the construction phase.

[0061] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that can be understood by those skilled in the art. The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A method for visualizing construction data of power projects, characterized in that, Includes the following steps: Acquire multi-source data from the construction site; based on the multi-source data, construct a three-dimensional component model to simulate the construction state and obtain component status data; based on the multi-source data and component status data, use a supervised learning method with long short-term memory networks for progress prediction; use an image recognition model based on convolutional neural networks to perform quality identification and safety scoring analysis on the image data in the multi-source data; and graphically display the progress prediction results, quality identification results, and safety scoring analysis results.

2. The method for visualizing construction data of a power project according to claim 1, characterized in that, The multi-source data from the construction site includes operating parameters of construction equipment, environmental parameters and structural stress information, images and videos of construction components, as well as historical completion rates of construction tasks, number of construction personnel, operating time of machinery and equipment, and environmental monitoring data.

3. The method for visualizing construction data of a power project according to claim 1, characterized in that, Based on BIM (Building Information Modeling), a three-dimensional component model is established for the construction objects at the construction site. Multi-source data from the construction site is mapped to each component in the three-dimensional component model, establishing a semantic binding relationship between components and task nodes. Different status labels are applied to the corresponding components according to different construction states.

4. The method for visualizing construction data of a power project according to claim 3, characterized in that, The status indicators for components are adjusted by color rendering, material replacement, and transparency adjustment based on different construction states from multi-source data at the construction site.

5. The method for visualizing construction data of a power project according to claim 1, characterized in that, Based on multi-source data and component status data, a supervised learning method using long short-term memory networks is adopted for progress prediction. Specifically, a time series model is established for key construction tasks using a long short-term memory network; historical data on task completion rate, equipment utilization rate, worker attendance rate, and weather conditions at the construction site are obtained, and the time series model is trained using historical data. Multi-source data from the construction site is used as the input features of the trained time series model to predict the expected completion time of task nodes, the number of delay days, and the risk level.

6. The method for visualizing construction data of a power project according to claim 1, characterized in that, The method of using a convolutional neural network-based image recognition model to perform quality identification on image data from multiple sources is as follows: Based on the convolutional neural network image recognition algorithm, the method identifies defects such as template detachment, exposed rebar, cracks, hollow areas, and missing bolts in the images and video information of construction components from multiple sources at the construction site. For the defects identified in the images, defect information including the corresponding coordinates, type, and confidence level of the defects is generated, and the defect information is bound to the corresponding components.

7. The method for visualizing construction data of a power project according to claim 1, characterized in that, Multi-source data from the construction site is acquired, including environmental parameters, personnel status parameters, and equipment operating parameters. Environmental risk factors are obtained by normalizing wind speed, rainfall, humidity, and temperature in the environmental parameters. Personnel risk factors are obtained by normalizing personnel number, personnel distribution density, and personnel working time in the personnel status parameters. Equipment risk factors are obtained by normalizing operating load rate, continuous operating time, and maintenance cycle in the equipment operating parameters. The environmental risk factor, personnel risk factor, and equipment risk factor are weighted according to preset weighting coefficients to obtain a comprehensive safety risk index, and are divided into low-risk, medium-risk, and high-risk levels based on the magnitude of the comprehensive safety risk index. When the comprehensive security risk index reaches a high risk level, an early warning will be automatically triggered and the warning information will be pushed to the visual interface and multi-terminal interaction module.

8. The method for visualizing construction data of a power project according to claim 1, characterized in that, The SHA256 hash algorithm is used to encrypt the multi-source data from the construction site, and a structured index is built based on the Merkle tree structure for the multi-source data from the construction site.

9. The method for visualizing construction data of a power project according to claim 1, characterized in that, Based on the WebGL graphics rendering framework, a visual interface is created to display system results to managers and frontline staff. Based on the analysis results of progress prediction, quality identification, and safety scores, visual rendering of color, lighting, and dynamic textures is performed. For progress prediction results, a timeline function is integrated, allowing users to drag or click to view the status of any point in the project progress, achieving spatiotemporal backtracking. For quality identification results, when a task is predicted to be at risk of delay or fail to meet quality standards, the system automatically flashes a prompt and displays it centrally in the risk panel.

10. A power project construction data visualization system, characterized in that, It includes a data acquisition unit, a 3D simulation unit, a predictive analysis unit, and a visualization unit. The data acquisition unit is used to acquire multi-source data from the construction site. The 3D simulation unit is used to construct a 3D component model based on the multi-source data and simulate the construction state to obtain component state data. The predictive analysis unit is used to predict the progress using a supervised learning method based on long short-term memory networks, and to perform quality identification and safety scoring analysis on image data from the multi-source data using an image recognition model based on convolutional neural networks, based on the multi-source data and component state data. The visualization unit is used to graphically display the progress prediction results, quality identification results, and safety scoring analysis results.

Images