Fan blade vibration real-time monitoring system and method based on edge calculation

Abstract

The invention belongs to the technical field of thermal power plant monitoring, and discloses a fan blade vibration real-time monitoring system and method based on edge computing, which are deployed in a fan site through edge computing nodes and integrate multiple modules to realize data localization real-time processing, greatly reduce data transmission delay, ensure that hidden dangers such as abnormal vibration can be diagnosed in time, and improve the monitoring efficiency. Abnormal vibration and even fracture accidents caused by fatigue load, corrosion and the like of the blade are avoided, and the timeliness and reliability of monitoring are improved. The multi-dimensional sensor network collects vibration signals and auxiliary monitoring signals, and cooperates with the data fusion unit to integrate multi-source data, so that the signal quality is effectively improved, and data support is provided for accurate abnormality diagnosis. The collaborative architecture of the edge computing node and the cloud processing platform enables the edge side to undertake preprocessing, feature extraction and abnormality diagnosis tasks with high real-time requirements, and the cloud end focuses on deep model training, historical data mining and long-term trend analysis, thereby reducing the remote data transmission amount, reducing the bandwidth occupation and transmission cost, and improving the data transmission efficiency. And real-time monitoring and long-term trend control are considered, and the operation performance of the whole monitoring system is obviously optimized.

Description

Technical Field

[0001] This invention belongs to the field of thermal power plant monitoring technology, specifically a real-time monitoring system and method for wind turbine blade vibration based on edge computing. Background Technology

[0002] Thermal power plants use combustibles as fuel to produce electricity through a conversion process of "chemical energy - thermal energy - mechanical energy - electrical energy". Prime movers include steam engines, gas turbines, etc., and fan blades are key components in their operation. Related technologies have an important application position in the power generation industry.

[0003] During long-term operation, the blades of wind turbines in thermal power plants are susceptible to fatigue loads, corrosion, and other factors, which can lead to abnormal vibrations or even breakage and safety accidents.

[0004] To address the issue of abnormal wind turbine blade operation, the current technical solution involves monitoring the blade's operational status through a monitoring system to provide early warnings and identify potential hazards such as abnormal vibrations. The core processing mode of this monitoring system is centralized cloud processing. All wind turbine blade operation-related data collected at the front end is transmitted to a cloud platform, where the cloud performs core processing tasks such as data analysis and judgment to derive the monitoring results of the blade's operational status.

[0005] Current monitoring solutions relying on centralized cloud processing have significant technical shortcomings. The core issue is their inability to meet the real-time and high-efficiency requirements of wind turbine blade monitoring. Specifically, there is a high transmission latency during the data transfer from the front-end acquisition equipment to the cloud, which prevents timely feedback on abnormal blade operating conditions. This could lead to delayed hazard identification and potentially cause safety accidents. Furthermore, the remote transmission of large amounts of monitoring data consumes significant network bandwidth, increasing data transmission costs. Bandwidth limitations may also affect the stability of data transmission, further reducing the operational efficiency of the monitoring system. Summary of the Invention

[0006] This invention provides a real-time monitoring system and method for wind turbine blade vibration based on edge computing, which solves the problem that existing technologies cannot meet the real-time and high-efficiency requirements of wind turbine blade monitoring.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A real-time monitoring system for wind turbine blade vibration based on edge computing includes: Edge computing nodes deployed at wind turbine sites integrate sensor interfaces, data processing units, and communication modules to enable localized real-time data processing. A multi-dimensional sensor network is used to collect vibration signals and auxiliary monitoring signals from wind turbine blades; The data fusion unit is used to integrate multi-source data collected by the multi-dimensional sensor network to improve signal quality; The cloud processing platform interacts with the edge computing nodes via a communication network; The edge computing nodes perform data preprocessing, feature extraction, and real-time anomaly diagnosis, while the cloud processing platform performs deep model training, historical data mining, and long-term trend analysis.

[0008] Preferably, the edge computing node is an ARM architecture processor or an embedded computing module with a computing power of not less than 20 TOPS.

[0009] Preferably, the multi-dimensional sensor network includes an accelerometer and a temperature sensor; the accelerometer is arranged at the root, middle and tip of the blade to capture the spatiotemporal characteristics of the vibration signal; the temperature sensor is used to monitor the temperature change of the blade surface to assist in the diagnosis of vibration anomalies; the accelerometer is an industrial-grade triaxial accelerometer with a range of not less than ±500g and a sampling rate of not less than 10kHz.

[0010] Preferably, the data fusion unit uses Kalman filtering or an improved algorithm thereof to integrate multi-source data.

[0011] Preferably, the edge computing node is configured with a noise suppression module, which uses a Savitzky-Golay filter or a signal smoothing algorithm to eliminate environmental noise interference.

[0012] Preferably, the edge computing node deploys a machine learning model to achieve real-time judgment of vibration exceeding limits, with a diagnostic delay of no more than 50ms and triggering a primary alarm. The cloud processing platform adopts a time-series deep learning model to identify typical faults such as blade cracks and corrosion.

[0013] Preferably, the machine learning model is a random forest model, and the time-series deep learning model is a Bi-LSTM model.

[0014] Preferably, the edge computing node is built on the ROS2 operating system and supports mixed Python / C++ programming, and the cloud processing platform adopts a containerized fault diagnosis microservice that supports horizontal scaling.

[0015] Preferably, it also includes a report generation module that supports the generation of customized report templates, automatic push and version management, and has the ability to adapt templates to different scenarios, adjusting monitoring parameters and report dimensions according to different operating environments.

[0016] A real-time monitoring method for wind turbine blade vibration based on edge computing includes the following steps: S1: Acquire blade vibration signals and auxiliary monitoring signals through a multi-dimensional sensor network; S2: Use edge computing nodes to perform noise suppression and data integration on the collected data; S3: Extract key features of the preprocessed data using time-frequency domain analysis methods; S4: Real-time anomaly detection is performed on the extracted features based on a lightweight model, triggering a primary alarm. S5: Upload local diagnostic results and key feature data to the cloud, use deep models for accurate fault identification, and generate cloud analysis results; S6: Generate operation and maintenance suggestions based on cloud-based analysis results.

[0017] Compared with existing technologies, this invention has the following advantages: This invention provides a real-time monitoring system for wind turbine blade vibration based on edge computing. By deploying edge computing nodes at the wind turbine site and integrating multiple modules, it achieves localized real-time data processing, significantly reducing data transmission latency. This ensures that potential problems such as vibration anomalies can be diagnosed promptly, preventing vibration anomalies or even blade breakage caused by fatigue loads, corrosion, etc., thus improving the timeliness and reliability of monitoring. A multi-dimensional sensor network collects vibration signals and auxiliary monitoring signals, and a data fusion unit integrates multi-source data, effectively improving signal quality and providing data support for accurate anomaly diagnosis. The collaborative architecture of edge computing nodes and the cloud processing platform allows the edge to handle high-real-time preprocessing, feature extraction, and anomaly diagnosis tasks, while the cloud focuses on deep model training, historical data mining, and long-term trend analysis. This reduces the amount of remote data transmission, lowers bandwidth usage and transmission costs, and achieves a balance between real-time monitoring and long-term trend control, significantly optimizing the overall performance of the monitoring system. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of the real-time monitoring method for wind turbine blade vibration based on edge computing according to the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0021] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0022] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0023] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0024] like Figure 1 As shown, this invention provides a real-time monitoring system for wind turbine blade vibration based on edge computing, comprising: Edge computing nodes deployed at wind turbine sites integrate sensor interfaces, data processing units, and communication modules to enable localized real-time data processing. A multi-dimensional sensor network is used to collect vibration signals and auxiliary monitoring signals from wind turbine blades; The data fusion unit is used to integrate multi-source data collected by the multi-dimensional sensor network to improve signal quality; The cloud processing platform interacts with the edge computing nodes via a communication network; The edge computing nodes perform data preprocessing, feature extraction, and real-time anomaly diagnosis, while the cloud processing platform performs deep model training, historical data mining, and long-term trend analysis.

[0025] By deploying edge computing nodes at wind turbine sites and integrating multiple modules to achieve localized real-time data processing, data transmission latency is significantly reduced. This ensures that potential hazards such as vibration anomalies can be diagnosed promptly, preventing vibration anomalies or even blade breakage caused by fatigue loads and corrosion, thus improving the timeliness and reliability of monitoring. A multi-dimensional sensor network collects vibration signals and auxiliary monitoring signals, which, combined with a data fusion unit, integrates multi-source data, effectively improving signal quality and providing data support for accurate anomaly diagnosis. The collaborative architecture of edge computing nodes and the cloud processing platform allows the edge to handle real-time preprocessing, feature extraction, and anomaly diagnosis tasks, while the cloud focuses on deep model training, historical data mining, and long-term trend analysis. This reduces remote data transmission volume, lowers bandwidth usage and transmission costs, and achieves a balance between real-time monitoring and long-term trend control, significantly optimizing the overall performance of the monitoring system.

[0026] The detailed design is as follows: The wind turbine blade vibration real-time monitoring system based on edge computing includes: deploying an edge computing workstation at the wind turbine site, integrating sensor interfaces, data processing units and communication modules; The workstation uses an ARM architecture processor, which supports real-time data preprocessing and feature extraction, reducing the amount of raw data transmission by more than 90%. By using localized computing, millisecond-level response to vibration signals is achieved, avoiding cloud processing delays. Accelerometer: Triaxial accelerometers are arranged at the root, middle and tip of the blade to capture the spatiotemporal characteristics of vibration signals; Temperature sensor: Monitors changes in blade surface temperature to aid in the diagnosis of abnormal vibration; Data fusion module: It uses the Kalman filter algorithm to integrate multi-source data, improve the signal-to-noise ratio, and reduce noise interference.

[0027] Edge nodes handle real-time alarms and basic analysis, while the cloud performs deep model training and historical data mining, forming an "edge-cloud" collaborative architecture. The edge is responsible for rapid response, while the cloud provides long-term trend analysis and optimization suggestions.

[0028] Real-time vibration signal processing: Noise suppression: Applying Savitzky-Golay filters eliminates environmental noise interference and improves signal purity; Feature extraction: Time-frequency domain features of vibration signals are extracted based on wavelet packet decomposition to generate multi-dimensional feature vectors, providing a basis for fault diagnosis; Anomaly detection: Deploy a lightweight random forest model to achieve real-time judgment of vibration exceeding limits (delay <50ms) and trigger a primary alarm.

[0029] The primary diagnostics involves edge nodes quickly identifying vibration anomalies through threshold comparison, reducing false alarm rates. The secondary diagnostics utilizes a Bi-LSTM model in the cloud to analyze time-series characteristics, identifying typical faults such as cracks and corrosion, thus improving diagnostic accuracy. The edge workstation is an NVIDIA Jetson Xavier NX module with a computing power of 21 TOPS, supporting complex algorithm execution. The sensor network consists of industrial-grade accelerometers (range ±500g, sampling rate 10kHz) to ensure data acquisition accuracy. The communication protocol uses MQTT to achieve edge-cloud data synchronization, ensuring transmission efficiency. The edge device is a ROS2-based real-time operating system, supporting Python / C++ hybrid programming to optimize resource utilization. The cloud provides a containerized fault diagnosis microservice, supporting horizontal scaling to adapt to large-scale data demands.

[0030] This invention also provides a method for real-time monitoring of wind turbine blade vibration based on edge computing, the specific steps of which are as follows: S1: Acquire blade vibration signals and auxiliary monitoring signals through a multi-dimensional sensor network; S2: Use edge computing nodes to perform noise suppression and data integration on the collected data; S3: Extract key features of the preprocessed data using time-frequency domain analysis methods; S4: Real-time anomaly detection is performed on the extracted features based on a lightweight model, triggering a primary alarm. S5: Upload local diagnostic results and key feature data to the cloud, use deep models for accurate fault identification, and generate cloud analysis results; S6: Generate operation and maintenance suggestions based on cloud-based analysis results.

[0031] Working principle: When this invention was used, a six-month field test was conducted at a 600MW thermal power plant: Data acquisition: A total of 1.2TB of vibration data has been acquired, covering normal, abnormal and fault conditions.

[0032] Performance metrics: Real-time performance: The average processing latency at the edge is 38ms, which is 87% better than the cloud solution.

[0033] Accuracy: The fault identification accuracy rate reaches 92.3%, and the false alarm rate is less than 3%.

[0034] Typical case: Successfully provided early warning of blade root cracks, avoiding unplanned downtime losses and verifying system reliability.

[0035] This system achieves millisecond-level response time for wind turbine vibration monitoring through edge computing, and field verification has shown that it improves equipment availability. Future inventions will focus on the application of 5G+TSN networks at edge nodes to further improve data transmission efficiency.

[0036] Predictive maintenance driven by digital twin technology enables full lifecycle management of equipment.

[0037] The multi-unit collaborative monitoring algorithm has been optimized to support fault diagnosis under complex operating conditions.

[0038] This health report can intuitively present the time-domain waveform, frequency-domain characteristics and trend changes of blade vibration, and form a digital archive by combining the equipment's full life cycle data, providing data support for operation and maintenance decisions.

[0039] Another embodiment of the present invention provides a system that supports integration with the asset management system of thermal power plants, enabling automatic dispatching and progress tracking of maintenance work orders, and ensuring closed-loop management of the fault handling process.

[0040] This multi-system collaborative monitoring system can not only improve the reliability of individual wind turbines, but also provide standardized solutions for the management of equipment groups throughout the entire thermal power plant, helping enterprises achieve a comprehensive upgrade of intelligent operation and maintenance.

[0041] Digital archives can be stored immutably using blockchain technology, ensuring the authenticity and credibility of the basis for operation and maintenance decisions.

[0042] Another embodiment of the present invention provides a system that supports the generation of customized report templates to meet the needs of different management levels, such as fault detail reports for front-line maintenance personnel and equipment health trend overview reports for management.

[0043] Furthermore, by combining digital twin technology, the evolution of abnormal blade vibration can be simulated in a virtual environment, providing visual support for root cause analysis and maintenance solution verification, and further improving the scientific nature and accuracy of operation and maintenance decisions.

[0044] Users can also adjust the module layout by dragging and dropping, and set the color scheme and axis parameters of data visualization charts to make the report presentation more in line with their own usage habits.

[0045] The system provides a template saving function. Customized report templates can be saved to the template library with one click and can be directly called when generating reports later, greatly reducing the time spent on repetitive configuration.

[0046] For periodic reports that need to be generated regularly, the system supports setting up an automatic generation schedule, which pushes the reports to a designated email address or platform on a daily, weekly, or monthly basis, thereby achieving automated management of operation and maintenance reports.

[0047] The system supports template version management, allowing users to update and iterate templates while retaining historical versions, facilitating the review of report configuration logic at different stages. It also provides a template sharing mechanism, allowing users across departments to copy high-quality templates with one click, quickly promote the standardized reporting system, and further improve the standardization and consistency of the enterprise's overall operation and maintenance documents.

[0048] Another embodiment of the present invention provides a system that also supports adding update logs for each version, recording the time, personnel, and specific adjustments made to the template, such as adding a vibration characteristic parameter module and optimizing the fault risk assessment algorithm, which helps team members quickly understand the core logic of template iteration.

[0049] For templates shared across departments, the system provides hierarchical permission control, allowing users to set different permission levels such as "view only," "copyable," and "editable" to ensure the security of core template configurations while meeting the personalized needs of different departments.

[0050] The update log supports association with specific operation and maintenance scenario requirements. For example, it can be marked that a template adjustment is to adapt to new monitoring parameters of a certain type of wind turbine, or to respond to the special report format requirements of a power plant in a certain region, making subsequent template reuse more targeted.

[0051] Meanwhile, the system provides a log retrieval function, which can quickly locate historical version change records by keywords (such as "vibration characteristic parameters" and "fault algorithm"), helping newly hired maintenance personnel to quickly grasp the evolution of template configuration and reduce the cost of knowledge transfer.

[0052] For example, to address the monitoring needs of wind turbine blades in high salt spray environments at coastal thermal power plants, the system can quickly filter out suitable corrosion resistance parameter templates through scene tags associated with logs. For the analysis of low-pressure vibration characteristics of wind turbines in high-altitude areas, it can also accurately match targeted report templates, avoiding monitoring deviations of general templates in special scenarios.

[0053] In addition, the log retrieval function also supports classification and filtering by scenario type (such as "coastal high salt fog" and "high altitude low air pressure"), and forms a scenario-based knowledge base by combining historical operation and maintenance data, providing mature template references for monitoring new equipment in similar scenarios, and further improving the system's scenario adaptability and operation and maintenance efficiency.

[0054] For monitoring the low-temperature brittle vibration of wind turbine blades in the frigid winter regions of northern China, the system can match scene tags with a template specific to low-temperature environments to enhance the early identification parameters of material fatigue cracks. For wind turbines in windy and sandy areas, the template will focus on the blade surface wear and vibration coupling analysis module to accurately capture vibration anomalies caused by sand and dust erosion.

[0055] The scenario-based knowledge base also supports users to upload new scenario cases, which are then incorporated into the tag system after system verification, continuously enriching the coverage dimensions of the template library and ensuring the reliability of monitoring in extreme climates and complex geographical environments.

[0056] For power plants in the southern coastal areas with high temperature and humidity, a dedicated template will be added to analyze the correlation between humidity and vibration, focusing on monitoring the corrosion fatigue vibration characteristics of blade materials under hot and humid conditions. For wind turbines in areas prone to thunderstorms, the template enhances the monitoring parameters for sudden vibration changes after lightning strikes, enabling timely detection of blade structural damage caused by lightning strikes.

[0057] The system also supports automatically generating new scenario tags based on user-uploaded scenario cases, such as "high temperature and humidity corrosion" and "lightning impact damage". After verification by the technical team, these tags are updated to the tag system to ensure that the template library can quickly respond to emerging and complex scenario needs, providing accurate and efficient technical support for wind turbine blade monitoring in various extreme environments.

[0058] This module collects environmental humidity data and blade vibration signals in real time, establishes a nonlinear correlation model, accurately identifies vibration frequency shifts and amplitude anomalies caused by material corrosion under humid and hot conditions, and provides early warning of blade corrosion fatigue risks 7-14 days in advance. Simultaneously, it integrates electrochemical corrosion sensor data to achieve coupled analysis of vibration characteristics and corrosion rate, providing maintenance personnel with a quantitative assessment report of "corrosion degree - vibration risk" to assist in the formulation of targeted anti-corrosion maintenance plans.

[0059] For example, when the vibration frequency of a blade is detected to deviate by more than 5% continuously within 24 hours and the humidity value is higher than 85%, the system will automatically trigger a level 2 warning and push a maintenance plan containing specific content such as "blade leading edge corrosion risk level B" and "re-inspection of surface anti-corrosion coating recommended within 3 days". This helps the maintenance team to accurately locate the problem area and take preventive measures to avoid corrosion fatigue from further developing into structural damage.

[0060] The system will also synchronize the blade's historical maintenance records. For example, if the blade has undergone two anti-corrosion coating repairs within six months, the warning message will additionally indicate "risk of insufficient coating durability" and recommend upgrading to a moisture- and heat-resistant anti-corrosion material.

[0061] Meanwhile, the early warning information can directly trigger the automatic creation of maintenance work orders. The work orders are pre-filled with core information such as blade number, risk location, and suggested maintenance measures. Maintenance personnel only need to confirm the work order priority to dispatch work, which greatly simplifies the fault handling process.

[0062] In addition, the system will automatically calculate the theoretical remaining life of the anti-corrosion coating based on data such as the cumulative running time of the blades and the exposure period to the corrosive environment. It will also include time nodes such as "suggested coating upgrade window (next 15 days)" in the warning information to help the operation and maintenance team to reasonably arrange the shutdown and maintenance plan and avoid conflicts with the high load power generation period of the unit.

[0063] For blades that have been upgraded with moisture- and heat-resistant materials, the system will continuously track changes in their vibration characteristics, verify the effect of material performance improvement, and feed the data back to the template library to optimize the material selection recommendation model for similar scenarios.

[0064] The system will also combine the power plant’s annual maintenance plan with grid load forecast data to automatically generate the optimal timing for coating upgrades. For example, it will avoid peak summer electricity demand and winter heating guarantee periods, and prioritize maintenance during low load periods in spring and autumn to minimize the impact of downtime on power generation efficiency.

[0065] Meanwhile, the theoretical remaining life calculation model is regularly updated dynamically based on the latest environmental monitoring data (such as salt spray concentration and humidity fluctuations) to ensure that the accuracy of life prediction remains above 90%, providing a reliable time dimension reference for operation and maintenance decisions.

[0066] For example, when the system identifies a 15-day standby window for generating units in the annual maintenance plan of a coastal power plant in Q2, and the power grid predicts that the regional load rate during this period will be only 58%, it will automatically mark this period as a "priority maintenance window" and simultaneously push a solution report containing benefit analysis such as "3 days required for coating upgrade" and "expected reduction of power generation loss of approximately 210,000 kWh", to help the power plant management quickly evaluate and make decisions.

[0067] In addition, the system supports integration with the power plant's material management system, automatically checking the inventory of anti-corrosion materials when generating maintenance plans. If the inventory is insufficient, a procurement warning is triggered to ensure that the materials required for maintenance arrive in advance and to avoid maintenance delays due to material shortages.

[0068] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the scope of the invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0069] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that can be understood by those skilled in the art. The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A real-time monitoring system for wind turbine blade vibration based on edge computing, characterized in that, include: Edge computing nodes deployed at wind turbine sites integrate sensor interfaces, data processing units, and communication modules to enable localized real-time data processing. A multi-dimensional sensor network is used to collect vibration signals and auxiliary monitoring signals from wind turbine blades; The data fusion unit is used to integrate multi-source data collected by the multi-dimensional sensor network to improve signal quality; The cloud processing platform interacts with the edge computing nodes via a communication network; The edge computing nodes perform data preprocessing, feature extraction, and real-time anomaly diagnosis, while the cloud processing platform performs deep model training, historical data mining, and long-term trend analysis.

2. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 1, characterized in that, The edge computing node is an ARM architecture processor or an embedded computing module with a computing power of not less than 20 TOPS.

3. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 1, characterized in that, The multi-dimensional sensor network includes an accelerometer and a temperature sensor. The accelerometer is arranged at the root, middle and tip of the blade to capture the spatiotemporal characteristics of the vibration signal. The temperature sensor is used to monitor the temperature change of the blade surface to assist in the diagnosis of vibration anomalies. The accelerometer is an industrial-grade triaxial accelerometer with a range of not less than ±500g and a sampling rate of not less than 10kHz.

4. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 1, characterized in that, The data fusion unit uses Kalman filtering or its improved algorithm to integrate multi-source data.

5. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 1, characterized in that, The edge computing node is equipped with a noise suppression module, which uses a Savitzky-Golay filter or a signal smoothing algorithm to eliminate environmental noise interference.

6. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 1, characterized in that, The edge computing nodes deploy machine learning models to achieve real-time judgment of vibration exceeding limits, with a diagnostic delay of no more than 50ms and triggering a primary alarm. The cloud processing platform adopts a time-series deep learning model to identify typical faults such as blade cracks and corrosion.

7. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 6, characterized in that, The machine learning model is a random forest model, and the temporal deep learning model is a Bi-LSTM model.

8. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 1, characterized in that, The edge computing nodes are built on the ROS2 operating system and support mixed Python / C++ programming. The cloud processing platform adopts containerized deployment of fault diagnosis microservices and supports horizontal scaling.

9. The real-time monitoring system for wind turbine blade vibration based on edge computing according to claim 1, characterized in that, It also includes a report generation module that supports the generation of customized report templates, automatic push and version management, and has the ability to adapt templates to different scenarios, adjusting monitoring parameters and report dimensions according to different operating environments.

10. A method for real-time monitoring of wind turbine blade vibration based on edge computing, characterized in that, Includes the following steps: S1: Acquire blade vibration signals and auxiliary monitoring signals through a multi-dimensional sensor network; S2: Use edge computing nodes to perform noise suppression and data integration on the collected data; S3: Extract key features of the preprocessed data using time-frequency domain analysis methods; S4: Real-time anomaly detection is performed on the extracted features based on a lightweight model, triggering a primary alarm. S5: Upload local diagnostic results and key feature data to the cloud, use deep models for accurate fault identification, and generate cloud analysis results; S6: Generate operation and maintenance suggestions based on cloud-based analysis results.

Images