Photovoltaic support intelligent operation and maintenance system fusing industrial cloud computing

Abstract

The application relates to the technical field of intelligent operation and maintenance of photovoltaic power stations, and discloses a photovoltaic support intelligent operation and maintenance system fusing industrial cloud computing, which comprises a support state sensing layer, an edge gateway layer, an industrial cloud computing layer and an operation and maintenance execution layer; the support state sensing layer collects support inclination, vibration, strain, connector pre-tightening state, foundation displacement, corrosion environment and tracking driving parameters; the edge gateway layer performs time synchronization, abnormality elimination, feature extraction and breakpoint continuation on the collected data; the industrial cloud computing layer establishes a support asset model and a structure state digital twin model, calculates support health evaluation values, and identifies inclination drift, connection loosening, wind vibration abnormality, foundation settlement, corrosion risk and tracking mechanism jamming; and the operation and maintenance execution layer generates an inspection task, a repair work order or a remote reset instruction according to the risk level. The application can realize online diagnosis and closed-loop maintenance of the mechanical side state of the photovoltaic support, and improves the timeliness of fault discovery and operation and maintenance efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology for photovoltaic power plants, specifically to an intelligent operation and maintenance system for photovoltaic brackets that integrates industrial cloud computing. Background Technology

[0002] Photovoltaic support structures are crucial foundational structures in photovoltaic power plants, used to support, position, and adjust photovoltaic modules. Their stability directly impacts the module's light-receiving angle, array power generation efficiency, and the long-term operational safety of the power plant. For fixed support structures, long-term service is susceptible to factors such as wind loads, snow loads, foundation settlement, corrosion, and loose bolts, leading to tilt deviations, column deformation, or decreased reliability of connection nodes. For tracking support structures, issues such as slewing mechanism jamming, abnormal drive motor load, push rod stroke deviation, and tracking angle errors may also occur.

[0003] Existing photovoltaic power plant operation and maintenance systems primarily focus on monitoring electrical data such as inverters, combiner boxes, string current, voltage, and power generation, while paying insufficient attention to the mechanical aspects of the photovoltaic support structure, including structural condition, installation angle, vibration status, foundation settlement, and loose connections. In actual operation and maintenance, support anomalies typically rely on manual inspections, on-site spot checks, or post-fault troubleshooting. This approach suffers from delayed detection, high inspection costs, data dispersion, and difficulty in providing early warnings. This is particularly problematic in mountainous areas, tidal flats, deserts, and large-scale centralized photovoltaic power plants, where the efficiency and accuracy of manual inspections are insufficient to meet the demands of refined operation and maintenance.

[0004] With the development of the Industrial Internet and industrial cloud computing, photovoltaic power station equipment can achieve data collection, remote monitoring, and centralized management through edge gateways, sensor terminals, and cloud platforms. However, existing cloud-based operation and maintenance platforms typically only provide centralized display of on-site data and threshold alarms, lacking dedicated data models for the health status of photovoltaic support structures. They cannot integrate and analyze tilt angle, vibration, strain, bolt preload, corrosive environment, drive current, and historical maintenance records, nor can they form an integrated processing mechanism for support anomaly location, risk classification, maintenance recommendations, and work order closure.

[0005] Therefore, it is necessary to provide an intelligent operation and maintenance system for photovoltaic (PV) supports that integrates industrial cloud computing. By deploying structural status acquisition units at key parts of the supports and combining local preprocessing at the edge gateway, a digital twin model of the supports in the industrial cloud, and an operation and maintenance decision module, the system can achieve online monitoring, intelligent diagnosis, and closed-loop maintenance of problems such as tilt drift, abnormal wind vibration, foundation settlement, loose connections, corrosion risks, and tracking mechanism jamming of PV supports. This will improve the operation and maintenance efficiency of PV supports and the safety of power plant operation. Summary of the Invention

[0006] To overcome the shortcomings of existing photovoltaic power plant operation and maintenance systems, such as an overemphasis on electrical monitoring, insufficient mechanical status perception of the support structure, reliance on manual experience for anomaly diagnosis, and inability to update models in a timely manner based on maintenance feedback, the present invention aims to provide an intelligent operation and maintenance system for photovoltaic supports that integrates industrial cloud computing. Through the collaboration of the support structure status perception layer, edge gateway layer, industrial cloud computing layer, and operation and maintenance execution layer, the system achieves online acquisition of the photovoltaic support structure status, edge feature processing, cloud-based health assessment, anomaly type identification, and closed-loop maintenance work orders.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a photovoltaic support intelligent operation and maintenance system integrating industrial cloud computing, comprising a support status perception layer, an edge gateway layer, an industrial cloud computing layer, and an operation and maintenance execution layer; The bracket status sensing layer includes status acquisition units installed on photovoltaic bracket columns, crossbeams, diagonal braces, connection nodes and / or tracking drive mechanisms, used to collect at least three types of data from bracket tilt angle, vibration, strain, connector pre-tightening status, foundation displacement, environmental corrosion parameters and tracking drive parameters; The edge gateway layer is communicatively connected to the status acquisition unit and is used to perform time synchronization, anomaly removal, feature extraction and partition caching on the acquired data, and generate a support unit operation feature package. The industrial cloud computing layer includes a support asset modeling module, a structural state digital twin module, an anomaly diagnosis module, and a maintenance decision module. The support asset modeling module is used to establish the correlation between the array number, structural type, installation angle, design load, foundation type, and historical maintenance records of the photovoltaic support. The structural state digital twin module is used to construct a real-time state model of the support based on the operational feature package of the support unit, and compare it with the baseline state model of the corresponding support to obtain the health evaluation value of the support. The anomaly diagnosis module identifies at least one type of anomaly based on the health evaluation value of the stent, including tilt drift, loose connection, abnormal wind vibration, foundation settlement, corrosion risk, and tracking mechanism jamming. The maintenance decision module generates maintenance strategies based on the anomaly type, anomaly location, and risk level, and generates inspection tasks, maintenance work orders, or remote reset commands through the operation and maintenance execution layer.

[0008] Furthermore, the status acquisition unit includes at least four of the following: tilt sensor, triaxial acceleration sensor, strain sensor, bolt preload sensor, displacement sensor, temperature and humidity sensor, salt spray or corrosion electrochemical sensor, motor current sensor, and push rod stroke sensor. The status acquisition units are grouped and numbered according to the bracket row, bracket column, or single bracket.

[0009] Furthermore, the edge gateway layer includes a data access module, an edge computing module, and a breakpoint resume module; The data access module is used to access wired sensor data and wireless sensor data; The edge computing module is used to calculate the tilt angle change rate, vibration frequency, vibration amplitude, strain change rate, cumulative displacement, bolt preload attenuation rate, and drive current fluctuation coefficient. The breakpoint resume module is used to cache the support unit's running feature packets locally when the network is interrupted, and upload them to the industrial cloud computing layer after the network is restored.

[0010] Furthermore, the structural state digital twin module includes a static structural parameter submodule and a dynamic state parameter submodule; The static structural parameter submodule is used to store the support material, column spacing, purlin specifications, foundation type, installation tilt angle, design wind load, and design snow load. The dynamic state parameter submodule is used to store real-time tilt angle, real-time vibration, real-time strain, real-time displacement, real-time corrosive environment, and real-time driving load. The structural state digital twin module matches and calculates the dynamic state parameters with the static structural parameters to obtain the state deviation of the support unit.

[0011] Furthermore, the stent health evaluation value H is calculated according to the following formula: H=100-(aA+bB+cC+dD+eE+fF), Wherein, A is the tilt angle deviation score, B is the vibration anomaly score, C is the strain anomaly score, D is the basic displacement score, E is the connection loosening score, and F is the corrosion risk score. a, b, c, d, e, and f are the corresponding score weights, and a + b + c + d + e + f = 1. When H is lower than the preset health threshold, the anomaly diagnosis module triggers the anomaly identification process.

[0012] Furthermore, the anomaly diagnosis module includes a rule diagnosis unit and a model diagnosis unit; The rule-based diagnostic unit is used to identify tilt angle abrupt changes, vibration exceeding limits, displacement accumulation exceeding limits, preload attenuation exceeding limits, and drive current exceeding limits based on preset thresholds. The model diagnostic unit is used to perform a horizontal comparison of the operational feature packages of adjacent stents in the same area, and to identify the anomaly type by combining historical failure samples of similar stents. When the outputs of the rule-based diagnostic unit and the model-based diagnostic unit are inconsistent, the anomaly diagnostic module will raise the anomaly level by one level and generate a review task.

[0013] Furthermore, the maintenance decision module generates corresponding maintenance strategies based on the anomaly type: When tilt drift is identified, an angle calibration task is generated; When a loose connection is detected, a bolt retightening task is generated; When wind-induced vibration anomalies are identified, a reinforcement inspection task is generated. When foundation settlement is identified, foundation re-measurement and support leveling tasks are generated; When a corrosion risk is identified, an anti-corrosion treatment task is generated; When a tracking mechanism jamming is detected, a drive mechanism check or remote reset command is generated.

[0014] Furthermore, the industrial cloud computing layer also includes a multi-power station collaborative analysis module, which is used to cluster the support operation data of different photovoltaic power stations according to support model, geographical region, foundation type, service life and environmental level, and update the abnormal patterns obtained from the clustering to the anomaly diagnosis module.

[0015] Furthermore, the operation and maintenance execution layer includes a mobile inspection terminal, a work order management module, and a maintenance result feedback module; The mobile inspection terminal is used to receive the array location, anomaly type, risk level, and maintenance recommendations of the abnormal support. The work order management module is used to record the dispatch time, the personnel handling the work, the handling measures, spare parts information, and the handling results. The maintenance result feedback module is used to upload the retest data after maintenance to the industrial cloud computing layer and update the historical maintenance records in the support asset modeling module.

[0016] Furthermore, the system also includes a safety early warning module, which is used to send early warning information to maintenance personnel when the detected deviation of the support tilt angle, vibration amplitude, foundation displacement, or drive current fluctuation coefficient exceeds the corresponding safety threshold, and to issue control commands to the tracking photovoltaic support to stop tracking, return to a safe angle, or reduce the tracking frequency.

[0017] Compared with the prior art, the present invention has the following beneficial technical effects: 1. This invention deploys the support status sensing layer in key parts such as columns, beams, diagonal braces, connection nodes, and tracking drive mechanisms, which can directly collect the mechanical side operating status of the support. This makes up for the shortcomings of traditional photovoltaic operation and maintenance systems that mainly focus on electrical data such as inverters, combiner boxes, and power generation, and helps to detect potential problems such as support tilt drift, loose connections, and foundation settlement in advance.

[0018] 2. This invention completes time synchronization, anomaly removal, feature extraction, and breakpoint resumption on the field side through the edge gateway layer, and only uploads the support unit operation feature packet to the industrial cloud computing layer, reducing the pressure of original data transmission and improving data continuity and system stability in weak network environments.

[0019] 3. This invention establishes a support asset model and a digital twin model of structural status at the industrial cloud computing layer. It correlates and calculates the support design parameters, installation parameters, real-time operating parameters and historical maintenance records to obtain the support health evaluation value and status deviation, thus transforming the support anomaly judgment from a single threshold alarm to a multi-factor fusion diagnosis.

[0020] 4. The anomaly diagnosis module of the present invention is equipped with both rule diagnosis unit and model diagnosis unit. When the output results of the two are inconsistent, the anomaly level is increased and a review task is generated, which can reduce the risk of missed reports and improve the reliability of identifying problems such as wind vibration anomaly, corrosion risk and tracking mechanism jamming in complex environments.

[0021] 5. This invention forms a closed loop of inspection tasks, maintenance work orders, maintenance result feedback and model updates through the maintenance decision module and the operation and maintenance execution layer. This enables the retest data after maintenance to update the support asset model and fault sample library in reverse, thereby continuously improving the accuracy of subsequent diagnosis and the efficiency of operation and maintenance management. Attached Figure Description

[0022] Figure 1 A process flow diagram for an intelligent operation and maintenance system for photovoltaic brackets that integrates industrial cloud computing. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are described clearly and completely. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Example 1 This embodiment provides an intelligent operation and maintenance system for fixed photovoltaic (PV) brackets. The fixed PV bracket includes columns, beams, diagonal braces, and purlins. Tilt sensors are installed on the columns at both ends of each row of brackets, triaxial accelerometers are installed on the beams in the middle of every three brackets, strain sensors are attached to the diagonal braces subjected to greater stress, bolt preload sensors are installed at typical connection nodes, displacement sensors are installed on the column foundation side, and temperature and humidity sensors and corrosion environment sensors are installed at typical locations within the station area.

[0025] The status acquisition units are grouped and numbered according to the format of "power station number - array number - row number - bracket number - sensor number". The sampling period for tilt angle data is 10 minutes. The sampling frequency for vibration data is increased when the wind speed exceeds the preset wind speed threshold. The sampling period for strain, displacement, and preload data is 30 minutes, and the sampling period for environmental corrosion parameters is 1 hour. The edge gateway layer adds a uniform timestamp to each type of data and uses median filtering and a three-standard-deviation rule to remove obvious outliers.

[0026] The edge computing module calculates the tilt angle change rate, vibration dominant frequency, vibration amplitude, strain change rate, cumulative foundation displacement, and bolt preload attenuation rate. For fixed supports, the weights in the support health evaluation value H are set as a=0.25, b=0.18, c=0.18, d=0.20, e=0.12, and f=0.07. The tilt angle deviation score A is determined based on the difference between the current tilt angle and the designed installation tilt angle; the vibration anomaly score B is determined based on the vibration dominant frequency offset and vibration amplitude; the strain anomaly score C is determined based on the deviation of the measured strain from the average strain of normal supports in the same area; the foundation displacement score D is determined based on the cumulative displacement; the connector loosening score E is determined based on the preload attenuation rate; and the corrosion risk score F is determined based on temperature and humidity, salt spray, or corrosion electrochemical parameters.

[0027] When H ≥ 85, it is considered normal; when 70 ≤ H < 85, it is considered a state of concern and a key item for the next inspection is generated; when 50 ≤ H < 70, it is considered a maintenance state and a field inspection work order is generated; when H < 50, it is considered a high-risk state, and the system immediately sends an early warning message to the maintenance personnel. Through the above configuration, this embodiment can perform online identification of tilt drift, foundation settlement, and loose connections of fixed photovoltaic supports.

[0028] Example 2 This embodiment provides an intelligent operation and maintenance system for a single-axis tracking photovoltaic bracket. Compared with Embodiment 1, this embodiment adds a motor current sensor, a push rod stroke sensor, and a rotation angle sensor to the tracking drive mechanism. The edge computing module further calculates the drive current fluctuation coefficient, push rod stroke deviation, and tracking angle error.

[0029] The structural status digital twin module in the industrial cloud computing layer constructs a real-time status model of the tracking support based on the support structure parameters, solar tracking target angle, real-time rotation angle, and drive load data. The anomaly diagnosis module determines that the tracking mechanism is stuck when it detects a continuous increase in the drive current fluctuation coefficient, a push rod stroke deviation exceeding a threshold, and no identical deviation observed in adjacent supports; it determines that the angle calibration is abnormal when it detects a long-term deviation between the actual rotation angle and the target angle but no significant increase in the drive current.

[0030] When the system identifies a tracking mechanism jamming, the maintenance decision module generates a drive mechanism inspection task and issues a control command to the tracking controller to stop tracking or return to a safe angle; when it identifies an angle calibration abnormality, the maintenance decision module generates an angle calibration task and pushes the array position of the abnormal support, the abnormal period, the angle deviation curve, and the suggested handling measures to the mobile inspection terminal.

[0031] Example 3 This embodiment provides an industrial cloud-based intelligent operation and maintenance system for multi-power station collaboration. The industrial cloud computing layer accesses the support structure operation feature packages of multiple photovoltaic power stations and performs cluster analysis according to support structure model, geographical region, foundation type, service life, and environmental level. For the same category of support structures, the system extracts the normal operation feature range and the fault sample feature range to form a corresponding anomaly pattern library.

[0032] When a power station experiences a decline in the health assessment value of its support structure but lacks sufficient historical samples for a single station, the multi-power station collaborative analysis module retrieves anomaly patterns from other power stations with the same support structure model, similar environmental level, and similar foundation type for auxiliary judgment. If the combined characteristics of the target support structure's tilt deviation, abnormal vibration, and preload attenuation match the connection loosening pattern in the anomaly pattern library, the anomaly diagnosis module outputs the connection loosening risk and generates bolt retightening and node re-inspection work orders.

[0033] After completing the repairs, maintenance personnel upload photos before and after the repairs, along with remeasured tilt angles, remeasured preload, and processing results via a mobile inspection terminal. The repair result feedback module then transmits the remeasured data back to the industrial cloud computing layer to update the support asset record and fault sample database.

[0034] Comparative Example 1 This comparative example uses a traditional photovoltaic power plant monitoring system, which only collects data on inverter, combiner box, string voltage, current, and power generation, but does not collect data on support tilt angle, vibration, strain, connector preload, and foundation displacement. When the support experiences slight tilt drift or loose connections, the system can only detect the anomaly during a decrease in power generation or on-site manual inspection, and cannot directly pinpoint the cause of the mechanical abnormality on the support.

[0035] Comparative Example 2 This comparative example deploys tilt and vibration sensors on a support frame, but does not include an edge gateway layer; it only uploads raw data to the cloud for threshold alarms. When the network is unstable or the data volume is large, the upload delay of raw data is significant, and false alarms are easily generated during momentary sensor interference, short-term wind vibration, or communication jitter. Compared to this invention, this comparative example lacks edge feature extraction, anomaly removal, and breakpoint resume processing.

[0036] Comparative Example 3 This comparative example uses a cloud-based threshold alarm method, without establishing a support asset model or a digital twin model of the structural state, nor does it feed maintenance results back to the fault sample database. Although this method can trigger an alarm for a single parameter exceeding the limit, it is difficult to comprehensively analyze the correlation between tilt angle, vibration, strain, foundation displacement, and preload, and cannot form a closed loop for anomaly type identification, risk classification, and maintenance strategy.

[0037] Performance verification To verify the operation and maintenance effectiveness of this invention, a fixed support area with the same structural form in the same photovoltaic power station was selected for simulated operation testing. Example 1 uses the system of this invention, Comparative Example 1 uses a traditional electrical side monitoring system, Comparative Example 2 uses a sensor cloud monitoring method without edge preprocessing, and Comparative Example 3 uses a single threshold alarm method. During the test, the following were simulated: support tilt angle deviation, connector preload attenuation, foundation displacement accumulation, and short-term wind vibration interference.

[0038] Test results show that Example 1 can generate a state of concern in the early stages of bracket tilt deviation and preload decay, and automatically generate on-site inspection work orders when the anomaly continues to expand; Comparative Example 1 can only detect problems after abnormal power generation or manual inspection; Comparative Example 2 has a high number of false alarms under short-term wind vibration interference; Comparative Example 3 can identify single parameter exceeding limits, but its accuracy in identifying complex anomalies such as foundation settlement and loose connections is insufficient. This demonstrates that the present invention, through the combination of state perception, edge computing, industrial cloud digital twins, and work order feedback closed loop, can improve the timeliness of photovoltaic bracket anomaly detection and the accuracy of maintenance decisions.

[0039] Table 1: Comparison of the operating effects of the examples and the comparative examples.

[0040]

[0041] It should be noted that the sensor type, sampling period, health threshold and weight parameters in the above embodiments can be adjusted according to the power plant scale, support type, geographical environment and operation and maintenance requirements. As long as the support status perception, edge feature processing, industrial cloud health evaluation, anomaly diagnosis and operation and maintenance closed-loop mechanism described in this invention are adopted, they all fall within the protection scope of this invention.

[0042] In this document, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0043] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A photovoltaic support intelligent operation and maintenance system integrating industrial cloud computing, characterized in that, It includes a support status awareness layer, an edge gateway layer, an industrial cloud computing layer, and an operation and maintenance execution layer; The bracket status sensing layer includes status acquisition units installed on photovoltaic bracket columns, crossbeams, diagonal braces, connection nodes and / or tracking drive mechanisms, used to collect at least three types of data from bracket tilt angle, vibration, strain, connector pre-tightening status, foundation displacement, environmental corrosion parameters and tracking drive parameters; The edge gateway layer is communicatively connected to the status acquisition unit and is used to perform time synchronization, anomaly removal, feature extraction and partition caching on the acquired data, and generate a support unit operation feature package. The industrial cloud computing layer includes a support asset modeling module, a structural state digital twin module, an anomaly diagnosis module, and a maintenance decision module. The support asset modeling module is used to establish the correlation between the array number, structural type, installation angle, design load, foundation type, and historical maintenance records of the photovoltaic support. The structural state digital twin module is used to construct a real-time state model of the support based on the operational feature package of the support unit, and compare it with the baseline state model of the corresponding support to obtain the health evaluation value of the support. The anomaly diagnosis module identifies at least one type of anomaly based on the health evaluation value of the stent, including tilt drift, loose connection, abnormal wind vibration, foundation settlement, corrosion risk, and tracking mechanism jamming. The maintenance decision module generates maintenance strategies based on the anomaly type, anomaly location, and risk level, and generates inspection tasks, maintenance work orders, or remote reset commands through the operation and maintenance execution layer.

2. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The status acquisition unit includes at least four of the following: tilt sensor, triaxial acceleration sensor, strain sensor, bolt preload sensor, displacement sensor, temperature and humidity sensor, salt spray or corrosion electrochemical sensor, motor current sensor, and push rod stroke sensor. The status acquisition units are grouped and numbered according to the bracket row, bracket column, or single bracket.

3. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The edge gateway layer includes a data access module, an edge computing module, and a breakpoint resume module; The data access module is used to access wired sensor data and wireless sensor data; The edge computing module is used to calculate the tilt angle change rate, vibration frequency, vibration amplitude, strain change rate, cumulative displacement, bolt preload attenuation rate, and drive current fluctuation coefficient. The breakpoint resume module is used to cache the support unit's running feature packets locally when the network is interrupted, and upload them to the industrial cloud computing layer after the network is restored.

4. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The structural state digital twin module includes a static structural parameter submodule and a dynamic state parameter submodule; The static structural parameter submodule is used to store the support material, column spacing, purlin specifications, foundation type, installation tilt angle, design wind load, and design snow load. The dynamic state parameter submodule is used to store real-time tilt angle, real-time vibration, real-time strain, real-time displacement, real-time corrosive environment, and real-time driving load. The structural state digital twin module matches and calculates the dynamic state parameters with the static structural parameters to obtain the state deviation of the support unit.

5. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The stent health evaluation value H is calculated according to the following formula: H=100-(aA+bB+cC+dD+eE+fF), Wherein, A is the tilt angle deviation score, B is the vibration anomaly score, C is the strain anomaly score, D is the basic displacement score, E is the connection loosening score, and F is the corrosion risk score. a, b, c, d, e, and f are the corresponding score weights, and a + b + c + d + e + f = 1. When H is lower than the preset health threshold, the anomaly diagnosis module triggers the anomaly identification process.

6. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The anomaly diagnosis module includes a rule diagnosis unit and a model diagnosis unit; The rule-based diagnostic unit is used to identify tilt angle abrupt changes, vibration exceeding limits, displacement accumulation exceeding limits, preload attenuation exceeding limits, and drive current exceeding limits based on preset thresholds. The model diagnostic unit is used to perform a horizontal comparison of the operational feature packages of adjacent stents in the same area, and to identify the anomaly type by combining historical failure samples of similar stents. When the outputs of the rule-based diagnostic unit and the model-based diagnostic unit are inconsistent, the anomaly diagnostic module will raise the anomaly level by one level and generate a review task.

7. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The maintenance decision module generates corresponding maintenance strategies based on the anomaly type: When tilt drift is identified, an angle calibration task is generated; When a loose connection is detected, a bolt retightening task is generated; When wind-induced vibration anomalies are identified, a reinforcement inspection task is generated. When foundation settlement is identified, foundation re-measurement and support leveling tasks are generated; When a corrosion risk is identified, an anti-corrosion treatment task is generated; When a tracking mechanism jamming is detected, a drive mechanism check or remote reset command is generated.

8. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The industrial cloud computing layer also includes a multi-power station collaborative analysis module. This module is used to cluster the support operation data of different photovoltaic power stations according to support model, geographical region, foundation type, service life and environmental level, and update the abnormal patterns obtained from the clustering to the anomaly diagnosis module.

9. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The operation and maintenance execution layer includes a mobile inspection terminal, a work order management module, and a maintenance result feedback module. The mobile inspection terminal is used to receive the array location, anomaly type, risk level, and maintenance recommendations of the abnormal support. The work order management module is used to record the dispatch time, the personnel handling the work, the handling measures, spare parts information, and the handling results. The maintenance result feedback module is used to upload the retest data after maintenance to the industrial cloud computing layer and update the historical maintenance records in the support asset modeling module.

10. The intelligent operation and maintenance system for photovoltaic brackets integrating industrial cloud computing as described in claim 1, characterized in that, The system also includes a safety early warning module, which is used to send early warning information to maintenance personnel when the deviation of the support tilt angle, vibration amplitude, foundation displacement or driving current fluctuation coefficient exceeds the corresponding safety threshold, and to issue control commands to the tracking photovoltaic support to stop tracking, return to a safe angle or reduce the tracking frequency.

Images