Mountainous light storage power station infrastructure construction method based on full-cycle digital twin cooperation

Abstract

This application discloses a construction method for mountain photovoltaic and energy storage power stations based on full-cycle digital twin collaboration, comprising the following steps: S1, establishing a multi-dimensional digital sensing network during the entire construction process of the mountain photovoltaic and energy storage power station to comprehensively and in real-time acquire construction sensing data during the construction process; S2, establishing a multi-resource collaborative scheduling model to dynamically optimize the allocation of construction resources based on construction resource conditions and environmental conditions; S3, constructing a full-cycle digital twin model of the mountain photovoltaic and energy storage power station construction by integrating project BIM design data, construction sensing data, and construction resource allocation data; S4, adjusting the construction plan based on the deviation between the virtual simulation results of the full-cycle digital twin model and the actual construction data. This invention, through the construction of a full-cycle digital twin model, breaks down information barriers among multiple participants and achieves closed-loop management of "virtual simulation - actual construction - deviation rectification".

Description

Technical Field

[0001] This application relates to the field of new energy building technology, and in particular to a construction method for mountain photovoltaic and energy storage power station infrastructure based on full-cycle digital twin collaboration. Background Technology

[0002] Against the backdrop of the booming development of the new energy industry, the construction of mountain photovoltaic power stations and integrated photovoltaic-energy storage projects is increasing. However, existing mountain photovoltaic infrastructure and photovoltaic-energy storage construction technologies face many challenges.

[0003] For photovoltaic infrastructure in mountainous areas, firstly, due to the complex terrain, traditional sensing equipment layouts are ill-suited to the dispersed terrain and numerous construction sites. Previous deployment methods may fail to comprehensively and accurately acquire key data during construction. For example, in pile foundation construction, it's impossible to accurately capture real-time changes in concrete pouring pressure, leading to lags and inaccuracies in construction quality monitoring. Secondly, regarding data transmission, the mountainous terrain severely obstructs signals, resulting in unstable signal transmission, significant data transmission delays, and difficulty in maintaining a stable sampling frequency. This compromises data integrity and real-time performance, affecting subsequent data processing and analysis, and failing to provide reliable data support for construction quality inspection. Furthermore, manual construction quality verification is extremely inefficient, and in the complex mountainous environment with its variable lighting conditions, manual inspection lacks sufficient accuracy to meet the requirements of high-quality construction.

[0004] In terms of multi-resource scheduling for photovoltaic and energy storage construction, existing scheduling models fail to fully consider the unique factors of mountainous construction, such as travel time and equipment adaptability to terrain. This results in low overall equipment utilization, high material delivery delays, and an inability to achieve optimal resource matching, leading to resource waste and increased construction costs. Furthermore, there is a lack of effective impact assessment and adaptive response mechanisms for extreme weather. For example, during rainy and winter construction, the inability to adjust construction schedules and resource allocation in a timely manner according to weather changes can lead to decreased construction quality and even engineering accidents. In addition, the lack of dynamic coupling between the scheduling scheme and the construction schedule means that resource allocation cannot be adjusted in real time according to the construction progress, easily resulting in idle or insufficient resources, thus affecting construction efficiency.

[0005] In the full-cycle management of integrated photovoltaic and energy storage infrastructure, traditional management methods lack precise virtual platforms to reflect the actual project situation in real time. Without effective digital twin modeling technology, it is impossible to comprehensively and in real-time simulate and monitor the construction process. Collaborative management among multiple stakeholders also faces numerous problems: information transmission is delayed, communication is poor, and issues arising during construction cannot be resolved in a timely manner, affecting project progress and quality. Furthermore, the lack of effective closed-loop management throughout the construction cycle, from preparation to acceptance, makes it difficult to achieve a close integration of virtual simulation and actual construction, hindering the timely detection and rectification of deviations. Complete data support is also lacking for project acceptance and subsequent operation and maintenance.

[0006] In summary, existing technologies for mountain photovoltaic infrastructure and photovoltaic-energy storage construction have many shortcomings in terms of sensing, scheduling, and collaborative management. There is an urgent need for a new technology to solve these problems and improve construction efficiency, quality, and management level. Summary of the Invention

[0007] The main objective of this application is to provide a construction method for mountain photovoltaic and energy storage power station infrastructure based on full-cycle digital twin collaboration, in order to solve the problems of low construction efficiency, low construction quality and insufficient management level in the existing technology of mountain photovoltaic infrastructure and photovoltaic and energy storage construction technology in terms of sensing, scheduling and collaborative management.

[0008] To achieve the above objectives, this application provides the following technical solution: A construction method for mountain photovoltaic-storage power station infrastructure based on full-cycle digital twin collaboration includes the following steps: S1. During the entire construction process of the mountain photovoltaic-storage power station, a multi-dimensional digital sensing network is established to comprehensively and in real time acquire construction sensing data during the construction process. S2. Establish a multi-resource collaborative scheduling model to dynamically optimize the allocation of construction resources based on construction resource conditions and environmental conditions; S3. By integrating the project's BIM design data, construction perception data, and construction resource allocation data, a full-cycle digital twin model of the mountain photovoltaic-storage power station construction is constructed. S4. Based on the deviation between the virtual simulation results of the full-cycle digital twin model and the actual construction data, adjust the construction plan.

[0009] As a further improvement to this application, in step S1, the construction sensing data is obtained through the following steps: For any anomalous mutations found in the collected multi-source data, the following steps should be taken: Examine the continuity of the time series of abnormal data to determine whether the abnormal changes are within a reasonable time range; By combining the construction process at the location of the abnormal data collection point, analyze whether the abnormal changes were caused by normal construction operations; When the abnormal mutation is determined to be abnormal data, the abnormal data is corrected.

[0010] As a further improvement to this application, in step S1, the construction sensing data is obtained through the following steps: Edge computing nodes are set up near each sensing device in the sensing network to perform the following operations on the collected multi-source data: The multi-source data is classified, and corresponding algorithms are assigned to the classified data for feature extraction. Assign corresponding compression algorithms to the categorized data and compress the data accordingly; Set data priority transmission rules to transmit compressed data according to the set priority.

[0011] As a further improvement to this application, the priority transmission rule is as follows: Data that directly affects construction quality and safety and plays a crucial role in construction decisions will be prioritized for transmission. Data that has a relatively small immediate impact on construction decisions will be assigned as the second priority for transmission. A dynamic priority adjustment mechanism is set up to prioritize sudden and emergency data as the first priority data. The transmission priority of the first priority data is higher than that of the second priority data.

[0012] As a further improvement to this application, step S1 also includes a defect identification model based on an image recognition algorithm to detect the installation quality of photovoltaic modules, including the following steps: For different lighting conditions, the following steps are performed to extract features from photovoltaic module defect image data in a defect identification model built based on image recognition algorithms: An attention mechanism module is introduced, which encodes the coordinate information of the photovoltaic module image in the spatial dimension, enabling the model to better capture the spatial positional relationship of the defect features of the photovoltaic module. Establish direct skip connections between feature maps at different levels, enabling detailed information in shallow feature maps to be directly combined with semantic information in deep feature maps; As a further improvement to this application, the following steps are performed when training the defect recognition model constructed based on image recognition algorithms: Construct a dataset containing a large number of images of photovoltaic modules under complex lighting conditions; Image samples in the dataset are enhanced under various lighting conditions to increase data diversity; First, the defect recognition model is pre-trained on a large-scale public image dataset to obtain a representation method for general image features; The pre-trained model parameters are transferred to the task of feature extraction from photovoltaic module defect image data, and the defect recognition model is retrained using the dataset.

[0013] As a further improvement to this application, step S2 establishes a multi-resource collaborative scheduling model, including the following steps: Obtain relevant data on various resources required for construction, including: Data acquisition devices are installed on key operating components of various construction equipment to obtain the operating parameters of the construction equipment and transmit the data to the data processing platform in real time based on the Internet of Things; The project-based personnel management system collects data on workers' skill levels and attendance rates, and synchronizes the data to the data processing platform. The project's materials management system integrates inventory information, transportation progress, and demand plan information, and aggregates the data to a data processing platform. The particle swarm optimization algorithm is used to optimize the allocation of construction resources based on relevant data from various resources, including: In view of the special characteristics of mountain construction, a mountain traffic time coefficient is introduced into the particle swarm algorithm to calculate the accurate transportation time of the transportation equipment based on the transportation distance and actual driving speed. A terrain adaptability weight for construction equipment is introduced. Based on the working efficiency and applicability of construction equipment under different terrain conditions, a corresponding adaptability weight is set for each piece of construction equipment. Treating equipment resources, personnel resources, and material resources as particles in a particle swarm, with each particle representing a resource allocation scheme, and performing multiple iterative calculations to optimize the resource allocation scheme, the following operations are performed: In each iteration, the fitness value of each particle is calculated by considering the operating status of the equipment, the skill matching of the personnel, the supply of materials, and the factors of mountain transportation and terrain. Resource allocation schemes with high fitness values ​​are retained. After multiple iterations, the particle swarm algorithm converges to the optimal solution, generating a resource allocation scheme that includes equipment deployment time, personnel work arrangements, and material transportation plans.

[0014] As a further improvement to this application, step S2 involves dynamically optimizing the allocation of construction resources based on environmental conditions, including the following steps: For construction during the rainy season, please perform the following operations: Through in-depth analysis of historical rainfall data and construction records, rainfall thresholds that would significantly impact construction procedures were determined. When the rainfall exceeds the rainfall threshold, a construction suspension command is automatically triggered, and a notification containing the reason for the suspension, the estimated suspension time, and subsequent work arrangements is sent to the construction personnel. According to the preset rainproof material dispatch plan, rainproof materials and drainage materials are automatically dispatched to the construction site; For construction work in low temperatures during winter, please perform the following operations: Determine the mathematical relationship between temperature and concrete strength development, and set a low-temperature construction temperature threshold. When the ambient temperature is detected to be lower than the low-temperature construction temperature threshold, concrete transport trucks equipped with insulation devices and electric heating equipment are prioritized to be dispatched to the pile foundation pouring area. Electric heating equipment heats concrete by embedding heating wires in the concrete, thereby increasing the early strength growth rate of the concrete. Temperature sensors are embedded inside the concrete to monitor its temperature in real time, allowing for timely adjustments to the power of the heating equipment and ensuring that the concrete solidifies at a suitable temperature.

[0015] As a further improvement to this application, step S3 involves constructing a full-cycle digital twin model, including the following steps: Full-element digital twin modeling of photovoltaic-storage power stations, including: Import the project's BIM design data into a Unity3D engine-based modeling environment; The construction perception data is integrated into the modeling environment in real time to achieve real-time association between the construction perception data and the digital twin model; The equipment ledger information is associated with the modeling environment via a database connection; The construction process simulation algorithm is embedded into the modeling environment to simulate the construction process of each process; The development of a multi-party collaborative management and control module includes: The design interface is developed, allowing designers to connect to the full-cycle digital twin model through the design software and upload and update BIM drawings. Develop a construction-side interface, allowing the construction party to connect to the full-cycle digital twin model via construction-side software, and upload and update construction data. Develop a supervision terminal interface, allowing the supervision party to connect to the full-cycle digital twin model through the supervision terminal software, and conduct comprehensive supervision and management of the construction process; Develop an interface for the owner, allowing the owner to connect to the full-cycle digital twin model through the owner's software and gain a comprehensive understanding of the project's progress.

[0016] As a further improvement to this application, the adjustment of the construction plan in step S4 includes: During the construction preparation phase, the following operations shall be performed: The construction process under different construction schemes is simulated using the full-cycle digital twin model. In response to potential problems and risks identified during the simulation, the construction plan was adjusted in a timely manner. The construction personnel are trained using the full-cycle digital twin model to familiarize themselves with the construction process and key operational points. During the construction phase, the following operations shall be performed: The simulation data of the full-cycle digital twin model is compared with the actual construction data in real time to accurately monitor the construction process; For any discrepancies found in the comparison results, send a notification to the construction personnel containing detailed information about the discrepancies, rectification suggestions, and rectification deadlines; By analyzing real-time collected quality data and combining it with historical data and quality standards, potential construction quality risks can be predicted, and early warning signals can be issued to notify construction personnel and management personnel to take corresponding preventive measures.

[0017] During the construction acceptance phase, perform the following operations: Based on preset acceptance standards and quality specifications, the data in the full-cycle digital twin model is automatically analyzed and evaluated to generate an acceptance report; By using the historical data backtracking function of the full-cycle digital twin model, key events, quality issues and their handling during the construction process can be reviewed, and lessons learned can be summarized.

[0018] The beneficial effects of this application are as follows: By establishing a multi-dimensional digital sensing network throughout the entire construction process of a mountain photovoltaic and energy storage power station, a comprehensive "point-line-surface" sensing hardware system is formed to ensure the accurate and comprehensive collection of construction data for mountain photovoltaic infrastructure. This provides a reliable guarantee for subsequent data processing and analysis, placing the entire construction process under precise monitoring. Data preprocessing algorithms remove outliers, ensuring data accuracy and making quality inspection and construction decisions based on this data more reliable. Edge computing nodes perform preliminary screening and compression of data, reducing transmission bandwidth consumption, improving data transmission efficiency, and minimizing latency and congestion during data transmission. Priority-based transmission protocols ensure that critical data is transmitted first, enabling construction managers to obtain important information immediately and make timely decisions.

[0019] By constructing a defect identification model based on image recognition algorithms, automatic identification and location of defects in photovoltaic modules were achieved, significantly improving detection efficiency and accuracy compared to manual inspection. It maintains high robustness even under complex lighting conditions in mountainous terrain, overcoming the problem of manual inspection being greatly affected by lighting conditions, and ensuring accurate quality inspection of photovoltaic modules in various environments.

[0020] By establishing a multi-resource collaborative scheduling model and fully considering the unique factors of mountain construction, optimal resource matching was achieved. This reduced equipment idle time, improved equipment utilization efficiency, and lowered equipment rental costs. It also ensured the timely supply of construction materials, avoiding delays caused by material shortages, effectively shortening the construction cycle and improving the overall economic benefits of the project.

[0021] By establishing an extreme weather impact assessment and adaptive response mechanism, construction schedules and resource allocation can be adjusted promptly according to weather changes. During the rainy season, this avoids quality damage to cable laying and concrete pouring caused by rainfall; in winter, it ensures the quality of concrete construction, reduces rework and quality issues due to weather, and improves the stability and reliability of construction. The dynamic coupling technology between the scheduling plan and construction progress allows resource allocation to be adjusted in real time according to the construction schedule, avoiding resource idleness or shortages and improving construction efficiency. The visualization module allows construction managers to intuitively adjust resource allocation, further optimizing resource allocation and ensuring that construction progresses according to plan.

[0022] By constructing a full-element digital twin model of the photovoltaic-storage power station, a precise virtual platform is provided for the construction process. Through real-time mapping of actual construction progress and quality parameters, as well as simulation and early warning functions for each process, construction and management personnel can identify problems in advance and optimize construction plans. The multi-party collaborative management module enables efficient collaboration among design, construction, supervision, and owner parties. All parties can share data and information promptly, avoiding information delays and improving communication efficiency. A closed-loop management system of "virtual simulation - actual construction - deviation rectification" is achieved. Attached Figure Description

[0023] Figure 1 This is a schematic diagram illustrating the steps of the construction method for mountain photovoltaic and energy storage power stations based on full-cycle digital twin collaboration as described in this application.

[0024] Figure 2 This is a flowchart illustrating the method steps performed in step S1 of this application for handling abnormal mutations that occur in the collected multi-source data.

[0025] Figure 3 This is a flowchart illustrating the method steps for processing multi-source data collected through edge computing nodes in step S1 of this application.

[0026] Figure 4 This is a flowchart illustrating the method steps for priority transmission rules in step S1 of this application.

[0027] Figure 5 This is a flowchart illustrating the method steps in step S1 of this application for detecting the installation quality of photovoltaic modules using a defect identification model constructed based on an image recognition algorithm.

[0028] Figure 6 This is a flowchart illustrating the method steps for training the defect recognition model constructed based on the image recognition algorithm in step S1 of this application.

[0029] Figure 7 This is a flowchart illustrating the steps of establishing a multi-resource collaborative scheduling model in step S2 of this application.

[0030] Figure 8 This is a flowchart illustrating the method for dynamically optimizing the allocation of construction resources for rainy season construction in step S2 of this application.

[0031] Figure 9 This is a flowchart illustrating the method for dynamically optimizing the allocation of construction resources for low-temperature construction in winter, as described in step S2 of this application.

[0032] Figure 10 This is a flowchart illustrating the method steps for constructing a full-cycle digital twin model in step S3 of this application.

[0033] Figure 11 This is a flowchart illustrating the steps and procedures performed during the construction preparation phase when adjusting the construction plan in step S4 of this application.

[0034] Figure 12 This is a flowchart illustrating the steps and procedures performed during the construction implementation phase when adjusting the construction plan in step S4 of this application.

[0035] Figure 13 This is a flowchart illustrating the steps and procedures performed during the construction acceptance phase when adjusting the construction plan in step S4 of this application. Detailed Implementation

[0036] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0037] like Figure 1 As shown, this application discloses a construction method for mountain photovoltaic-storage power station infrastructure based on full-cycle digital twin collaboration, including the following steps: S1. During the entire construction process of the mountain photovoltaic-storage power station, a multi-dimensional digital sensing network is established to comprehensively and in real time acquire construction sensing data during the construction process. S2. Establish a multi-resource collaborative scheduling model to dynamically optimize the allocation of construction resources based on construction resource conditions and environmental conditions; S3. By integrating the project's BIM design data, construction perception data, and construction resource allocation data, a full-cycle digital twin model of the mountain photovoltaic-storage power station construction is constructed. S4. Based on the deviation between the virtual simulation results of the full-cycle digital twin model and the actual construction data, adjust the construction plan.

[0038] This application deeply integrates distributed sensing, dynamic scheduling, and digital twin technology, avoiding the limitations of traditional infrastructure technologies such as "single links and severe data silos," and forming data interoperability covering the entire construction cycle. Specifically, the multi-dimensional digital sensing network enables "full-process-multi-parameter" monitoring of mountain photovoltaic infrastructure; the multi-resource collaborative scheduling model solves the problem of resource misallocation under extreme conditions, achieving dynamic optimization of construction resource allocation; and by constructing a full-cycle digital twin model, it breaks down information barriers among multiple participants, achieving closed-loop management of "virtual simulation-actual construction-deviation rectification." Through the coordinated operation of these three links, data-driven development of mountain photovoltaic and energy storage infrastructure is achieved.

[0039] The following will elaborate on the construction method for mountain photovoltaic and energy storage power stations based on full-cycle digital twin collaboration, as proposed in this application, through a specific implementation plan.

[0040] In step S1, a multi-dimensional digital sensing network is established, including the following steps: Stress sensors are installed in the pile foundation construction area at predetermined standard intervals. For example, after precise measurement and planning, vibrating wire stress sensors are installed at standard intervals of 5 meters. To ensure the sensors are stable and can accurately sense pressure, small installation pits are first dug at predetermined locations. The bottom of the sensor is fixed to the pre-poured concrete base with high-strength bolts. The sensor's sensing end is connected to the concrete pouring area using flexible connecting material to ensure sensitive capture of pressure changes generated during concrete pouring and to convert the pressure signal into an easily transmitted electrical signal.

[0041] Before the concrete is poured, a sonar detection module is pre-embedded in a predetermined position inside the pile foundation. For example, the sonar detection module is accurately installed at one-third of the height above the bottom of the pile foundation using a customized fixing bracket. This position can effectively reflect the compaction of the entire pile foundation concrete. The sonar detection module integrates a high-performance ultrasonic transducer, signal processing circuit, and data storage unit. When the concrete is poured and vibrated, the ultrasonic transducer emits high-frequency sound wave signals to the surrounding concrete at set time intervals, such as every 100ms. When the sound waves propagate in the concrete, they encounter defects inside the concrete, such as honeycomb, pitting, pores, or interfaces between different media, such as the interface between concrete and steel reinforcement, resulting in reflection, refraction, and scattering. Some of the sound wave signals return to the ultrasonic transducer. The transducer converts the received sound wave signals into electrical signals and transmits them to the signal processing circuit. The signal processing circuit pre-amplifies the electrical signals to increase their amplitude for subsequent processing.

[0042] Displacement monitoring devices are installed at the tower foundation and photovoltaic module installation locations. These devices are tightly connected to the tower foundation and photovoltaic module supports via customized clamps, ensuring synchronized movement between the monitoring devices and the supports. The devices employ laser rangefinders or microelectromechanical systems (MEMS) to accurately measure minute displacement changes in all directions.

[0043] Multiple industrial cameras were installed at different locations within the photovoltaic module installation area. Based on a detailed analysis of the photovoltaic module layout and lighting conditions, the installation positions and angles of the industrial cameras were rationally planned. To ensure comprehensive capture of image information from all parts of the photovoltaic modules, a multi-camera collaborative shooting scheme was adopted for large photovoltaic module arrays, ensuring seamless connection between the shooting fields of each camera and avoiding blind spots. Simultaneously, the industrial cameras were equipped with modules for automatically adjusting parameters such as exposure and white balance according to the varying lighting conditions at different times of day in the mountainous terrain. Ambient light sensors were used to monitor the ambient light intensity and color temperature in real time, and the general-purpose intelligent algorithms built into the industrial cameras automatically adjusted the shooting parameters based on sensor data to ensure clear, high-quality images of the photovoltaic modules under various complex lighting conditions.

[0044] Along the cable laying path, temperature and humidity sensors and slope displacement sensors are installed at pre-set second standard intervals. For example, along the cable laying path, temperature and humidity sensors and slope displacement sensors are installed every 10m. The temperature and humidity sensors adopt a wall-mounted design, installed above the side of the cable trench to avoid the influence of water accumulation, and are connected to the data acquisition system through waterproof cables. The slope displacement sensors are installed by embedding, digging holes of appropriate depth in the stable soil layer of the slope, burying and fixing the sensor body, and using internal high-precision tilt sensors and displacement measuring devices to monitor the displacement of the slope soil in real time, ensuring the stability of the cable laying area.

[0045] For data transmission, LoRa wireless communication modules were selected. Addressing the challenges of complex mountainous terrain and severe signal obstruction, the optimal antenna installation angle was determined through multiple field tests and electromagnetic simulation analyses. For nodes in higher elevations and open areas, the antennas were installed vertically to maximize signal coverage; while in valleys or densely vegetated areas, the antennas were tilted at a certain angle to avoid major obstructions and enhance signal propagation. To further improve communication reliability, signal boosters were installed at each communication node to amplify and relay the signal, forming a stable and reliable "point-line-surface" all-area sensing hardware system.

[0046] like Figure 2 As shown, in step S1, the construction sensing data is obtained through the following steps: For any anomalous mutations found in the collected multi-source data, the following steps should be taken: S111. Check the continuity of the time series of abnormal data to determine whether the mutation is within a reasonable time range.

[0047] S112. Based on the construction process at the location of the abnormal data collection point, analyze whether the sudden change was caused by normal construction operations.

[0048] S113. When the abnormal mutation is determined to be abnormal data, the abnormal data is corrected.

[0049] Continuous numerical data such as stress, displacement, temperature, and humidity can all experience abrupt changes. Taking stress data as an example, when abrupt changes occur in the collected stress data, the first step is to analyze the trend of stress data changes over a past period to determine whether the abrupt change is consistent with the stress change pattern during normal construction. For instance, during concrete pouring, stress usually increases gradually as the amount of concrete poured increases. If a sudden and significant drop or increase occurs that is inconsistent with the normal pouring rhythm, it may be an outlier.

[0050] Furthermore, the installation location of the stress sensor and the construction process were considered for further confirmation. If the stress sensor was located near the concrete vibration area, and the data mutation occurred precisely during the vibration operation, it might be a temporary disturbance caused by vibration, thus identifying the data mutation as abnormal data.

[0051] Once outlier data is identified, a median filtering method based on a sliding window is used for correction. A sliding window is formed by selecting several data points within a certain time span, centered on the current data point, and the median value of the data within the window is used to replace the outlier data point.

[0052] By detecting and correcting abnormal mutations, the accuracy of the data is ensured, making quality inspection and construction decisions based on this data more reliable.

[0053] In step S1, the construction sensing data is also obtained through the following steps: Edge computing nodes are deployed near each sensing device in the sensing network. The hardware architecture of these edge computing nodes employs high-performance multi-core processors, such as ARM-based Cortex-A series processors, paired with large-capacity DDR4 memory to meet the demands of parallel processing of multi-source data. They are also equipped with high-speed storage devices, such as solid-state drives, for temporary storage and rapid data retrieval. The edge computing nodes are also configured with multiple communication interfaces, including Ethernet, Wi-Fi, and LoRa interfaces, to facilitate connectivity with different types of sensing devices and networks.

[0054] like Figure 3 As shown, the following operations are performed on the collected multi-source data through edge computing nodes: S121. Classify the multi-source data and assign corresponding algorithms to the classified data for feature extraction.

[0055] S122. Assign the corresponding compression algorithm to the classified data and compress the data.

[0056] S123. Set data priority transmission rules to transmit compressed data according to the set priority.

[0057] Image data can be assigned to a dedicated graphics processing unit (GPU) for processing, leveraging the GPU's parallel computing capabilities to accelerate image feature extraction. For example, for images of photovoltaic modules, existing convolutional neural networks can be used for rapid feature extraction to identify the module's outline, location, and potential defects.

[0058] Sensor data, such as stress, displacement, temperature, and humidity, is processed by the central processing unit (CPU). The CPU uses digital signal processing algorithms to filter and denoise this data, removing high-frequency noise and interference signals. Then, existing feature selection algorithms, such as the chi-square test and wrap-around algorithms, are used to select the most representative features from a large amount of data, reducing data dimensionality and improving subsequent processing efficiency.

[0059] For image data compression, a compression method based on discrete cosine transform (DCT) and quantization is employed. The image is divided into multiple small blocks, and a DCT transform is performed on each block to convert the spatial domain image data to the frequency domain. Then, the frequency domain coefficients are quantized to remove details in the high-frequency components. Finally, entropy coding is performed to encode the quantized coefficients into a binary data stream, thereby achieving image data compression.

[0060] For sensor data, due to its relatively small data volume and high accuracy requirements, lossless compression algorithms are used, such as the existing LZ77 algorithm. The LZ77 algorithm achieves compression by finding repeating patterns in the data and replacing the repeating data with pointers to the repeating parts.

[0061] By using edge computing nodes to perform preliminary filtering and compression of data, the transmission bandwidth utilization is reduced, the data transmission efficiency is improved, and the latency and congestion during the data transmission process are reduced.

[0062] like Figure 4 As shown, the priority transmission rule is as follows: S131. Data that directly affects construction quality and safety and plays a key role in construction decision-making is set as the first priority for transmission.

[0063] S132. Data with relatively small immediate impact on construction decisions will be designated as second-priority data for transmission.

[0064] S133. Set a dynamic priority adjustment mechanism to set sudden emergency data as the first priority data.

[0065] In the above steps, the transmission priority of first-priority data is higher than that of second-priority data. For example, key data such as pile foundation pouring pressure and photovoltaic module installation deviation are set as first-priority transmission data because these data are directly related to construction quality and safety and have a crucial guiding role in construction decisions. Environmental data such as temperature and humidity are set as second-priority data. Although these data are also important, their immediate impact on construction decisions is relatively small.

[0066] To ensure effective priority management, a priority label is assigned to each type of data. During data transmission, data is scheduled based on the priority label. Simultaneously, a dynamic priority adjustment mechanism is implemented. In emergency situations, such as a sudden increase in slope displacement in a certain area that could lead to a safety accident, the priority of displacement data related to that area and other relevant data will be temporarily elevated from second-priority data to first-priority data, ensuring that this data is transmitted to the management platform with priority.

[0067] like Figure 5 As shown, step S1 also includes a defect identification model built based on an image recognition algorithm to detect the installation quality of photovoltaic modules, including the following steps: For different lighting conditions, the following steps are performed to extract features from photovoltaic module defect image data in a defect identification model built based on image recognition algorithms: S141. An attention mechanism module is introduced, which encodes the coordinate information of the photovoltaic module image in the spatial dimension, enabling the model to better capture the spatial positional relationship of the defect features of the photovoltaic module.

[0068] S142. Establish direct jump connections between feature maps at different levels, so that the detailed information in the shallow feature map can be directly combined with the semantic information in the deep feature map.

[0069] In this embodiment, addressing the challenge of extracting defect feature images from photovoltaic (PV) modules under complex lighting conditions in mountainous terrain, an attention mechanism module is introduced into the existing YOLOv8 feature extraction network. For example, in the existing feature extraction network CSPDarknet, a CA module is introduced. This module encodes the coordinate information of the PV module image in two-dimensional space and calculates attention weights, fusing the horizontal and vertical attention weight vectors to obtain a final attention weight matrix. This matrix is ​​then element-wise multiplied with the original feature map, making the network more attentive to the spatial location and feature channels related to subtle defects in the PV module. Thus, under complex lighting conditions, local features crucial for defect identification are enhanced in the weighted feature map. The weighted feature map, as the output of the CA module, is input into the subsequent feature extraction network CSPDarknet for further feature extraction and defect identification. Because the CA module effectively captures the spatial positional relationships of PV module features and enhances the extraction capability of local PV module features, the subsequent feature extraction network CSPDarknet can more accurately learn the feature patterns of subtle defects in the PV module, thereby improving the accuracy of identifying subtle defects in PV modules under complex lighting conditions. Enhance the ability to extract local features of photovoltaic modules and improve the accuracy of identifying subtle defects in photovoltaic modules under complex lighting conditions.

[0070] To more effectively detect defects in photovoltaic modules of different sizes, the multi-scale feature fusion method of YOLOv8 can be optimized. In addition to the existing top-down and bottom-up feature fusion paths, cross-scale connection structures are added. For example, direct skip connections are established between feature maps at different levels, allowing detailed information in shallow feature maps to be more directly combined with semantic information in deeper feature maps. This improved multi-scale feature fusion method helps the feature extraction network accurately detect various defects such as photovoltaic module misalignment, loose clamping blocks, and junction box damage at different scales, especially significantly improving the detection performance of small-sized defects.

[0071] like Figure 6 As shown, the following steps are performed when training a defect recognition model built based on an image recognition algorithm: S143. Construct a dataset containing a large number of photovoltaic module images under complex lighting conditions.

[0072] S144. Perform various lighting condition enhancement processes on the image samples in the dataset to expand the diversity of the data.

[0073] S145. First, the defect recognition model is pre-trained on a large-scale public image dataset to obtain a representation method for general image features.

[0074] S146. Transfer the pre-trained model parameters to the task of feature extraction from photovoltaic module defect image data, and use the dataset to perform secondary training on the defect recognition model.

[0075] Specifically, the photovoltaic module image dataset was constructed to cover different times of day, such as early morning, noon, and evening, and different weather conditions, such as sunny, cloudy, and overcast, as well as photovoltaic module image samples under different lighting directions. Various lighting enhancement processes were applied to the images in the dataset, such as randomly adjusting brightness, contrast, and saturation, and adding Gaussian noise to simulate lighting interference, to expand the diversity of the data. During training, transfer learning techniques were employed. The YOLOv8 model was pre-trained on a large-scale public image dataset, such as the existing COCO image dataset, allowing the pre-trained model to learn a wide range of general image features on COCO, such as object edges, textures, and shapes. These general features provide good initial parameters for the model to quickly adapt to the new task when transferred to the mountain photovoltaic module detection task, reducing training time and computational resource consumption. Then, the parameters of the pre-trained model were transferred to the mountain photovoltaic module detection task and fine-tuned using a constructed mountain complex lighting dataset. During fine-tuning, hyperparameters such as the learning rate and optimizer parameters were adjusted to adapt to the specific detection task. For example, the learning rate is initially set to 0.001. After a certain number of training rounds, such as 20 rounds, the learning rate is reduced to 0.0001, and after another 20 rounds, it is further reduced to 0.00001. Simultaneously, the learning rate is dynamically adjusted based on changes in the loss function and performance metrics on the validation set during training to prevent the model from getting trapped in local optima. This enables the model to accurately identify defects in photovoltaic modules under complex lighting conditions.

[0076] like Figure 7 As shown, in step S2, establishing a multi-resource collaborative scheduling model includes the following steps: S21. Obtain relevant data on various resources required for construction, including: Data acquisition devices are installed on key operating components of various construction equipment to acquire operating parameters and transmit the data to a data processing platform in real time via the Internet of Things (IoT). For example, pressure, angle, and displacement sensors are installed on key parts of truck cranes, such as the boom, slewing mechanism, and hoisting mechanism, to acquire operating parameters such as load rate, boom angle, and hoisting height in real time. Vibration and torque sensors are installed on the drill rod and power head of drilling rigs to monitor drilling speed and torque changes. These sensors are connected to intelligent data acquisition terminals on the construction equipment via wired or wireless means. The acquisition terminals perform preliminary processing and packaging of the sensor data, and then transmit the data to the data processing platform in real time via the Industrial Internet of Things (IIoT).

[0077] The project-based personnel management system collects data on workers' skill levels and attendance rates, and synchronizes this data to a data processing platform. For example, skill levels are determined through professional skills examinations and practical work experience assessments, covering multiple trades such as electrical installation, machinery operation, and civil construction. Attendance rates are tracked in real-time using facial recognition devices or electronic check-in systems installed at construction site entrances. Employees are required to check in when arriving at or leaving get off work, or entering specific construction areas. The system automatically records employees' arrival and departure times and synchronizes this data to the data processing platform.

[0078] The project-based materials management system integrates inventory information, transportation progress, and demand planning information, and aggregates the data to a data processing platform. For example, inventory information is obtained through regular inventory checks and real-time inventory monitoring equipment installed in the warehouse. This equipment uses RFID or laser scanning technology to automatically identify the type, quantity, and location changes of materials. Transportation progress is tracked in real time by installing GPS positioning systems on transport vehicles and connecting them to a logistics information platform. The logistics information platform provides detailed information such as transportation routes and estimated arrival times. Demand planning is formulated based on the project construction schedule and engineering design requirements, specifying the required quantities and timeframes for various materials at different construction stages. All this material-related data is aggregated on the data processing platform.

[0079] S22. Using the particle swarm optimization algorithm, based on the relevant data of various resources obtained, the construction resources are optimized and allocated, including: To address the unique challenges of construction in mountainous terrain, a mountainous traffic time coefficient is incorporated into the particle swarm optimization algorithm. Based on the transport distance and actual travel speed, the accurate transport time for the equipment is calculated. Actual travel speed data for different slopes and road conditions is obtained through on-site surveys of mountain roads. For example, on sections with a slope greater than 25°, the average travel speed decreases to 60%-70% of normal speed due to increased vehicle resistance. The particle swarm optimization algorithm calculates the accurate transport time based on the transport distance and actual travel speed, serving as a crucial reference factor for resource scheduling.

[0080] A terrain adaptability weight is introduced for construction equipment. Based on the efficiency and suitability of each piece of equipment under different terrain conditions, a corresponding adaptability weight is assigned. For example, small drilling rigs have better mobility and efficiency in steep slope areas, so their priority scheduling coefficient is set to 1.2, while large equipment may have difficulty moving in the same area, so its priority scheduling coefficient is set to 0.8. By adjusting the priority scheduling coefficient parameter, the particle swarm algorithm can more accurately assess the rationality of resource allocation under different construction stages and terrain conditions, based on the actual situation of mountain construction.

[0081] S23. Treat equipment, personnel, and material resources as particles in a particle swarm, with each particle representing a resource allocation scheme. Perform multiple iterative calculations to optimize the resource allocation scheme. The particle's position indicates the allocation of resources under different construction tasks and time nodes, while its velocity indicates the direction and magnitude of adjustments to the resource allocation scheme. Through continuous iterative calculations, particles adjust their positions based on their historical optimal position and global optimal position, thus optimizing the resource allocation scheme.

[0082] In each iteration, the fitness value of each particle is calculated by considering the operating status of the equipment, the skill matching of the personnel, the supply of materials, and factors such as mountain transportation and terrain. The higher the fitness value, the better the resource allocation scheme. Resource allocation schemes with high fitness values ​​are retained.

[0083] When calculating the fitness value of each particle, scores are assigned to factors such as equipment operating status, personnel skill matching, material supply, and terrain and transportation factors, with appropriate weights assigned to reflect their relative importance in the resource allocation scheme. For example, the weight for equipment operating status is set to 0.3, personnel skill matching to 0.25, material supply to 0.25, and terrain and transportation factors to 0.2. These weights are not fixed and can be adjusted according to the characteristics and needs of the actual project.

[0084] Assume the equipment operating status score is denoted as S. 设备 Personnel skill matching score is recorded as S 人员 The score for the supply of materials is recorded as S.物资 The scores for mountainous transportation and terrain factors are denoted as S. 山地 The fitness value for each particle is calculated using the following formula: .

[0085] Through the detailed fitness value calculation process described above, the merits and demerits of the resource allocation scheme represented by each particle can be comprehensively evaluated. This guides the particle swarm algorithm to gradually search for a better resource allocation scheme during the iteration process, thereby achieving efficient scheduling of photovoltaic and energy storage construction resources.

[0086] After multiple iterations, the particle swarm algorithm converges to the optimal solution, which achieves optimal matching of resources under different construction stages and terrain conditions, and generates a resource allocation scheme that includes equipment deployment time, personnel work arrangements, and material transportation plans.

[0087] Combination Figure 8 and Figure 9 As shown, in step S2, the dynamic optimization of construction resource allocation based on environmental conditions includes the following steps: like Figure 8 As shown, for construction during the rainy season, the following operations shall be performed: S241. Through in-depth analysis of historical rainfall data and construction records, determine the rainfall threshold that will have a significant impact on construction procedures.

[0088] S242. When the rainfall is detected to exceed the rainfall threshold, a construction suspension command is automatically triggered, and a notification containing the reason for the suspension, the estimated suspension time, and subsequent work arrangements is sent to the construction personnel.

[0089] S243. Based on the preset rainproof material dispatch plan, automatically dispatch rainproof materials and drainage materials to the construction site.

[0090] During the rainy season, rainfall primarily impacts construction processes such as cable laying and concrete pouring. In-depth analysis of historical rainfall data and construction records shows that hourly rainfall exceeding 8mm significantly affects these processes. In such cases, the system automatically triggers a work stoppage command via wireless communication and sends the notification to the handheld terminals of relevant construction personnel. The rainproofing material deployment plan utilizes tarpaulins, which are quickly deployed using specialized winding equipment to cover the construction area. Drainage materials include drainage pumps, which are automatically connected to drainage pipes to promptly remove accumulated water. Furthermore, drones can be used to monitor water accumulation at the construction site in real time, providing more accurate information for drainage efforts.

[0091] like Figure 9 As shown, the following operations should be performed for construction in low temperatures during winter: S251. Determine the mathematical relationship between temperature and concrete strength development, and set the temperature threshold for low-temperature construction.

[0092] When the ambient temperature is detected to be lower than the low-temperature construction temperature threshold, concrete transport trucks equipped with insulation devices and electric heating equipment are dispatched to the pile foundation pouring area with priority.

[0093] S252. Electric heating equipment heats concrete by pre-embedded heating wires in the concrete, thereby increasing the early strength growth rate of the concrete.

[0094] S253. Install temperature sensors inside the concrete to monitor the concrete temperature in real time and adjust the power of the heating equipment in a timely manner to ensure that the concrete solidifies at a suitable temperature.

[0095] During winter construction, low temperatures primarily affect the setting and solidification of concrete. Existing experimental conditions can be used to simulate the concrete setting process under different temperature conditions, measure the strength gain of concrete at different time points, and determine the mathematical relationship between temperature and concrete strength development. Analysis revealed that when the ambient temperature is below the -3℃ threshold, low-temperature construction measures are necessary. In this case, the system automatically prioritizes dispatching insulated concrete trucks and electric heating equipment to the pile foundation pouring area. The insulated trucks utilize high-efficiency insulation materials and heating systems to ensure stable concrete temperature during transportation. The electric heating equipment heats the concrete through heating wires embedded in it, accelerating the early strength gain. Simultaneously, real-time monitoring of concrete temperature is strengthened by embedding temperature sensors within the concrete, transmitting temperature data to the monitoring system in real time to adjust the power of the heating equipment promptly, ensuring the concrete sets at a suitable temperature. Furthermore, an emergency resource reserve early warning mechanism can be established to pre-allocate antifreeze, anti-slip mats, and other materials based on weather forecasts. The material management system monitors the inventory of emergency resources in real time. When the inventory falls below the warning threshold, an alarm is automatically issued to remind managers to replenish materials in a timely manner, thereby improving the resilience and response capabilities of construction under extreme weather conditions.

[0096] like Figure 10 As shown, in step S3, constructing a full-cycle digital twin model includes the following steps: S31. Full-element digital twin modeling of photovoltaic and energy storage power stations, including: The project's BIM design data was imported into a modeling environment based on the industry-specific Unity3D engine. The BIM design data includes detailed geometric structures, dimensional parameters, and internal equipment layout information for the booster station, photovoltaic array, and energy storage compartment. First, the BIM data underwent lightweight processing, removing details not essential for building the digital twin model and optimizing the data structure to reduce data volume and improve loading speed. The lightweight BIM data was then imported into the Unity3D engine-based modeling environment through an industry-specific data conversion interface. During the import process, data format conversion and coordinate system unification were performed to ensure compatibility with the Unity3D engine's coordinate system and data format. Based on the Unity3D engine's modeling capabilities, a full-element digital twin model of the photovoltaic-energy storage power station was constructed at a 1:1 scale according to the actual dimensions.

[0097] The construction perception data is integrated into the modeling environment in real time, enabling real-time association between the construction perception data and the digital twin model. The construction perception data is acquired through the multi-dimensional digital perception network established in step S1, and then integrated into the Unity3D modeling environment in real time via a network communication interface, achieving real-time association between the construction perception data and the digital twin model.

[0098] The equipment ledger information is linked to the modeling environment via a database connection. This information includes the model, specifications, and technical parameters of equipment such as inverters and transformer substations. The equipment ledger information is organized into a structured data table and linked to the Unity3D modeling environment via a database connection. During model building, each equipment model is assigned corresponding ledger information attributes, enabling convenient querying and management of detailed equipment information within the digital twin model.

[0099] A construction process simulation algorithm is integrated into the modeling environment to simulate the construction process of each step. For the pile foundation pouring process, a concrete solidification simulation algorithm based on thermodynamics and fluid mechanics is implemented. This algorithm considers factors such as concrete mix proportions, pouring temperature, and ambient temperature and humidity, and simulates the hydration heat process and strength development of concrete after pouring using finite element analysis. In the digital twin model, the changes in the internal temperature field, moisture migration, and gradual strength development of the concrete are visualized, helping construction personnel to understand the concrete solidification state in real time and predict potential quality problems such as temperature cracks. For support installation, a three-dimensional coordinate matching function is developed. The three-dimensional coordinates of the support installation position are obtained using high-precision measuring equipment at the actual construction site and transmitted to the digital twin model in real time. The digital twin model compares the actual coordinates with the design coordinates in real time. When the deviation exceeds 2mm, the system automatically triggers an early warning mechanism, informing construction personnel of the deviation location and amount with a prominent color and prompt information on the digital twin model interface, so that timely adjustments can be made. For the hoisting of energy storage containers, Unity3D's physics simulation engine and collision detection algorithm are used to pre-simulate the hoisting path in a digital twin model. Inputting the dimensions, weight, lifting point location of the energy storage container, and the location information of surrounding facilities such as towers and cables, the system simulates the container's movement trajectory and spatial attitude during hoisting. During the simulation, the system detects collisions between the energy storage container and surrounding objects in real time. If a collision risk is detected, the system immediately issues an alarm and provides optimized hoisting path suggestions to avoid collisions during actual hoisting and ensure the safe execution of the hoisting operation.

[0100] S32. Development of a multi-participant collaborative management and control module, including: The system develops a design-side interface, allowing designers to connect to the full-lifecycle digital twin model via design-side software, enabling them to upload and update BIM drawings. Designers can directly annotate the BIM drawings using the software's rich annotation tools, such as text annotations, arrow indicators, and graphic markers, to detail the technical requirements of key processes and construction precautions. Simultaneously, designers can view construction progress and quality data uploaded by the construction team to promptly understand the implementation of design intent during construction. When discrepancies are found between construction and design, designers can promptly propose modifications and notify relevant stakeholders through the system.

[0101] A construction-side interface has been developed, allowing the construction team to connect to the full-cycle digital twin model via construction-side software, enabling them to upload and update construction data. The software features convenient data upload capabilities, allowing construction personnel to upload real-time images and videos of concealed works, such as photos and videos of pile foundation reinforcement binding, and records of cable trench excavation and cable laying. These images and videos provide a clear picture of the construction process and quality status. Simultaneously, quality inspection data can be uploaded, such as concrete test block strength reports, photovoltaic module performance test data, and electrical equipment insulation resistance test results. Furthermore, the construction team can view the technical requirements marked by the designer and the rectification suggestions proposed by the supervisor, allowing for timely adjustments to construction plans and techniques to ensure that construction quality meets requirements. The construction team can also submit construction progress reports within the software, detailing the completion status of each construction stage, encountered problems, and the anticipated next steps.

[0102] A monitoring interface has been developed, allowing the monitoring party to connect to the full-cycle digital twin model via monitoring software for comprehensive supervision and management of the construction process. The monitoring software provides document review functionality, enabling monitoring personnel to review construction plans submitted by the construction party online and assess their rationality and feasibility. For problems or quality concerns in the construction plans, the monitoring party can use annotation tools to make detailed annotations and propose specific rectification suggestions. Simultaneously, the monitoring party can view hidden works images and quality inspection data uploaded by the construction party, enabling real-time monitoring of construction quality. When quality issues are discovered, the monitoring party can promptly issue rectification notices to the construction party through the system and track the rectification process to ensure proper resolution. Furthermore, the monitoring party can record a monitoring log within the software, detailing daily monitoring work content, problems discovered, and their handling.

[0103] A client-side interface is developed, allowing the client to connect to the full-cycle digital twin model via client-side software and gain a comprehensive understanding of project progress. The client-side software visually displays construction progress, quality data, and communication information from all stakeholders. Clients can intuitively understand the overall project progress through charts and reports, such as the percentage of work completed and comparisons of actual completion times with planned times for key milestones. They can also view statistical analysis results of construction quality data to understand the achievement of quality standards. Clients can also review communication records between the design, construction, and supervision teams to promptly understand existing problems and solutions. Furthermore, clients can propose overall project management requirements and objectives within the software, approve major project decisions, and ensure the project progresses as expected.

[0104] Combination Figures 11 to 13 As shown, in step S4 of this application, the construction plan is adjusted based on the deviation between the virtual simulation results of the full-cycle digital twin model and the actual construction data, including: like Figure 11 As shown, during the construction preparation phase, the following operations shall be performed: S411. Simulate the construction process under different construction schemes using the full-cycle digital twin model.

[0105] S412. Adjust the construction plan in a timely manner in response to potential problems and risks discovered during the simulation process.

[0106] S413. Train construction personnel using the full-cycle digital twin model to familiarize them with the construction process and key operational points.

[0107] During the construction preparation phase, the main approach is to simulate the construction process, including pile foundation construction and photovoltaic module installation, using a full-cycle digital twin model. For example, when simulating pile foundation construction, parameters such as the pile driving sequence and concrete pouring speed can be adjusted to observe potential problems during construction, such as pile tilting and insufficient concrete compaction. For photovoltaic module installation, different installation sequences and personnel configurations are simulated to analyze their impact on construction efficiency and quality.

[0108] By simulating different construction scenarios, potential problems and risks were identified, and construction plans were adjusted accordingly. For example, model simulations revealed that following the conventional photovoltaic module installation sequence in areas with steep slopes could lead to uneven spacing between modules, affecting power generation efficiency. Based on the cause of the problem, the construction team adjusted the installation sequence, installing the modules on the steeper slopes first and then gradually installing those on the steeper slopes downwards, effectively resolving the issue.

[0109] Meanwhile, by training construction workers through a full-cycle digital twin model, and through virtual environment demonstrations, construction workers can more intuitively understand the specific operation methods, precautions, possible problems and countermeasures for each construction step, thereby improving the skill level and construction efficiency of construction workers.

[0110] like Figure 12 As shown, during the construction phase, the following operations shall be performed: S421. Real-time comparison of the simulation data of the full-cycle digital twin model with the actual construction data to accurately monitor the construction process.

[0111] S422. For any discrepancies found in the comparison results, send a notification to the construction personnel containing detailed information about the discrepancies, rectification suggestions, and rectification deadlines.

[0112] S423. By analyzing the quality data collected in real time, and combining it with historical data and quality standards, we can predict potential construction quality risks and issue early warning signals to notify construction personnel and management personnel to take corresponding preventive measures.

[0113] The system automatically compares the simulated data from the full-cycle digital twin model with the actual construction data acquired in real time by the multi-dimensional sensing network in step S1. For example, for the installation of substation equipment, it automatically detects whether the equipment spacing and terminal block positions meet the design requirements; for the installation of photovoltaic modules, it detects whether parameters such as the installation angle and flatness of the photovoltaic modules meet the standards.

[0114] When a deviation exceeding a preset range is detected, the location and amount of the deviation are marked on the digital twin model interface. A notification, including detailed information about the deviation, rectification suggestions, and a rectification deadline, is sent to the construction personnel's handheld terminals to remind them to make timely adjustments. Simultaneously, the time, location, and rectification process of the deviation are recorded, forming a complete construction quality traceability record.

[0115] like Figure 13 As shown, during the construction and acceptance phase, the following operations shall be performed: S431. Based on the preset acceptance standards and quality specifications, automatically analyze and evaluate the data in the full-cycle digital twin model to generate an acceptance report.

[0116] S432. Through the historical data backtracking function of the full-cycle digital twin model, review key events, quality issues and their handling during the construction process, and summarize construction experience and lessons learned.

[0117] Because the digital twin model integrates quality data from the entire construction process, including various testing data for pile foundation construction, deviation records for photovoltaic module installation, and test results for electrical equipment, acceptance personnel can intuitively view the quality status of each construction section and quickly obtain the necessary quality information by operating the digital twin model. The system automatically analyzes and evaluates the data in the model based on preset acceptance standards and quality specifications, generating an acceptance report. The acceptance report includes a project overview, an overall evaluation of construction quality, quality acceptance results for each construction section, existing problems, and rectification suggestions. The acceptance report is presented in a visually appealing format, utilizing visualization elements in the model such as images, charts, and animations to intuitively display the construction quality, enabling acceptance personnel to more clearly and accurately understand the project's quality status and improving the efficiency and accuracy of the acceptance work.

[0118] Through the historical data retrospective function of the digital twin model, acceptance personnel can query the construction status and quality parameters at any point in time. By reviewing key events, quality issues, and their handling during the construction process, they can summarize lessons learned. This retrospective analysis of historical data provides a reference for the construction of similar projects in the future, preventing similar problems from recurring. At the same time, historical data also provides important information for subsequent operation and maintenance of the project. Maintenance personnel can understand the equipment's operating history, maintenance records, and other information by querying historical data, enabling them to better formulate operation and maintenance plans and conduct fault diagnosis.

[0119] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or photovoltaic modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0120] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0121] The specific embodiments of the invention have been described in detail above, but they are only examples, and this application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of this application. Therefore, all equivalent changes, modifications, and improvements made without departing from the spirit and principles of this application should be covered within the scope of this application.

Claims

1. A mountainous light storage power station infrastructure construction method based on full-cycle digital twin cooperation, characterized in that, Includes the following steps: S1. During the entire construction process of the mountain photovoltaic-storage power station, a multi-dimensional digital sensing network is established to comprehensively and in real time acquire construction sensing data during the construction process. S2. Establish a multi-resource collaborative scheduling model to dynamically optimize the allocation of construction resources based on construction resource conditions and environmental conditions; S3. By integrating the project's BIM design data, construction perception data, and construction resource allocation data, a full-cycle digital twin model of the mountain photovoltaic-storage power station construction is constructed. S4. Based on the deviation between the virtual simulation results of the full-cycle digital twin model and the actual construction data, adjust the construction plan.

2. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 1, characterized in that, In step S1, the construction sensing data is obtained through the following steps: For any anomalous mutations found in the collected multi-source data, the following steps should be taken: Examine the continuity of the time series of abnormal data to determine whether the abnormal changes are within a reasonable time range; By combining the construction process at the location of the abnormal data collection point, analyze whether the abnormal changes were caused by normal construction operations; When the abnormal mutation is determined to be abnormal data, the abnormal data is corrected.

3. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 1, characterized in that, In step S1, the construction sensing data is obtained through the following steps: Edge computing nodes are set up near each sensing device in the sensing network to perform the following operations on the collected multi-source data: The multi-source data is classified, and corresponding algorithms are assigned to the classified data for feature extraction. Assign corresponding compression algorithms to the categorized data and compress the data accordingly; Set data priority transmission rules to transmit compressed data according to the set priority.

4. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 3, characterized in that, The priority transmission rule is as follows: Data that directly affects construction quality and safety and plays a crucial role in construction decisions will be prioritized for transmission. Data that has a relatively small immediate impact on construction decisions will be assigned as the second priority for transmission. A dynamic priority adjustment mechanism is set up to prioritize sudden and emergency data as the first priority data. The transmission priority of the first priority data is higher than that of the second priority data.

5. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 1, characterized in that, Step S1 also includes a defect identification model built based on image recognition algorithms to detect the installation quality of photovoltaic modules, including the following steps: For different lighting conditions, the following steps are performed to extract features from photovoltaic module defect image data in a defect identification model built based on image recognition algorithms: An attention mechanism module is introduced, which encodes the coordinate information of the photovoltaic module image in the spatial dimension, enabling the model to better capture the spatial positional relationship of the defect features of the photovoltaic module. Establish direct jump connections between feature maps at different levels, enabling detailed information in shallow feature maps to be directly combined with semantic information in deep feature maps.

6. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 5, characterized in that, When training a defect identification model built on image recognition algorithms, the following steps are performed: Construct a dataset containing a large number of images of photovoltaic modules under complex lighting conditions; Image samples in the dataset are enhanced under various lighting conditions to increase data diversity; First, the defect recognition model is pre-trained on a large-scale public image dataset to obtain a representation method for general image features; The pre-trained model parameters are transferred to the task of feature extraction from photovoltaic module defect image data, and the defect recognition model is retrained using the dataset.

7. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 1, characterized in that, In step S2, a multi-resource collaborative scheduling model is established, including the following steps: Obtain relevant data on various resources required for construction, including: Data acquisition devices are installed on key operating components of various construction equipment to obtain the operating parameters of the construction equipment and transmit the data to the data processing platform in real time based on the Internet of Things; The project-based personnel management system collects data on workers' skill levels and attendance rates, and synchronizes the data to the data processing platform. The project's materials management system integrates inventory information, transportation progress, and demand plan information, and aggregates the data to a data processing platform. The particle swarm optimization algorithm is used to optimize the allocation of construction resources based on relevant data from various resources, including: In view of the special characteristics of mountain construction, a mountain traffic time coefficient is introduced into the particle swarm algorithm to calculate the accurate transportation time of the transportation equipment based on the transportation distance and actual driving speed. A terrain adaptability weight for construction equipment is introduced. Based on the working efficiency and applicability of construction equipment under different terrain conditions, a corresponding adaptability weight is set for each piece of construction equipment. Treating equipment resources, personnel resources, and material resources as particles in a particle swarm, with each particle representing a resource allocation scheme, and performing multiple iterative calculations to optimize the resource allocation scheme, the following operations are performed: In each iteration, the fitness value of each particle is calculated by considering the operating status of the equipment, the skill matching of the personnel, the supply of materials, and the factors of mountain transportation and terrain. Resource allocation schemes with high fitness values ​​are retained. After multiple iterations, the particle swarm algorithm converges to the optimal solution, generating a resource allocation scheme that includes equipment deployment time, personnel work arrangements, and material transportation plans.

8. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 1, characterized in that, In step S2, the allocation of construction resources is dynamically optimized based on environmental conditions, including the following steps: For construction during the rainy season, please perform the following operations: Through in-depth analysis of historical rainfall data and construction records, rainfall thresholds that would significantly impact construction procedures were determined. When the rainfall exceeds the rainfall threshold, a construction suspension command is automatically triggered, and a notification containing the reason for the suspension, the estimated suspension time, and subsequent work arrangements is sent to the construction personnel. According to the preset rainproof material dispatch plan, rainproof materials and drainage materials are automatically dispatched to the construction site; For construction work in low temperatures during winter, please perform the following operations: Determine the mathematical relationship between temperature and concrete strength development, and set a low-temperature construction temperature threshold. When the ambient temperature is detected to be lower than the low-temperature construction temperature threshold, concrete transport trucks equipped with insulation devices and electric heating equipment are prioritized to be dispatched to the pile foundation pouring area. Electric heating equipment heats concrete by embedding heating wires in the concrete, thereby increasing the early strength growth rate of the concrete. Temperature sensors are embedded inside the concrete to monitor its temperature in real time, allowing for timely adjustments to the power of the heating equipment and ensuring that the concrete solidifies at a suitable temperature.

9. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration of claim 1, characterized in that, Step S3 involves constructing a full-cycle digital twin model, including the following steps: Full-element digital twin modeling of photovoltaic-storage power stations, including: Import the project's BIM design data into a Unity3D engine-based modeling environment; The construction perception data is integrated into the modeling environment in real time to achieve real-time association between the construction perception data and the digital twin model; The equipment ledger information is associated with the modeling environment via a database connection; The construction process simulation algorithm is embedded into the modeling environment to simulate the construction process of each process; The development of a multi-party collaborative management and control module includes: The design interface is developed, allowing designers to connect to the full-cycle digital twin model through the design software and upload and update BIM drawings. Develop a construction-side interface, allowing the construction party to connect to the full-cycle digital twin model via construction-side software, and upload and update construction data. Develop a supervision terminal interface, allowing the supervision party to connect to the full-cycle digital twin model through the supervision terminal software, and conduct comprehensive supervision and management of the construction process; Develop an interface for the owner, allowing the owner to connect to the full-cycle digital twin model through the owner's software and gain a comprehensive understanding of the project's progress.

10. The mountainous light storage power station infrastructure construction method based on full-cycle digital twin collaboration according to claim 1 or 9, characterized in that, In step S4, the adjustments to the construction plan include: During the construction preparation phase, the following operations shall be performed: The construction process under different construction schemes is simulated using the full-cycle digital twin model. In response to potential problems and risks identified during the simulation, the construction plan was adjusted in a timely manner. The construction personnel are trained using the full-cycle digital twin model to familiarize themselves with the construction process and key operational points. During the construction phase, the following operations shall be performed: The simulation data of the full-cycle digital twin model is compared with the actual construction data in real time to accurately monitor the construction process; For any discrepancies found in the comparison results, send a notification to the construction personnel containing detailed information about the discrepancies, rectification suggestions, and rectification deadlines; By analyzing the quality data collected in real time, and combining it with historical data and quality standards, we can predict potential construction quality risks and issue early warning signals to notify construction personnel and managers to take corresponding preventive measures. During the construction acceptance phase, perform the following operations: Based on preset acceptance standards and quality specifications, the data in the full-cycle digital twin model is automatically analyzed and evaluated to generate an acceptance report; By using the historical data backtracking function of the full-cycle digital twin model, key events, quality issues and their handling during the construction process can be reviewed, and lessons learned can be summarized.

Images