Intelligent anomaly detection platform for wind generating set

Abstract

The invention discloses an intelligent anomaly detection platform for a wind generating set, and relates to the field of generator detection. The intelligent anomaly detection platform for the wind generating set comprises an edge end, a server end and a client end, the client end achieves an interaction interface with a user, the server end is composed of a data middle table and a service processing module, and the data middle table manages sensor data and analysis result data in a unified mode with a draught fan ID as an identity label. And the business processing module is realized by adopting big data, machine learning and deep learning technologies. According to the intelligent anomaly detection platform for the wind generating set, an intelligent equipment monitoring platform is built when anomaly detection is carried out on brake pads of a fan yaw system, anomaly detection on the brake pads of the fan yaw system is achieved on the platform, and the platform has the advantages of being easy to expand and capable of managing data in a unified mode. And the maintenance operation of the wind generating set is convenient.

Description

Technical Field

[0001] This invention relates to the field of generator testing technology, specifically to an intelligent anomaly detection platform for wind turbine generator sets. Background Technology

[0002] Under normal circumstances, the mechanical noise emitted by equipment during normal operation is stable and regular. However, when equipment encounters aging, abnormal conditions, or other malfunctions during operation, it will emit significantly different operating noises. Therefore, equipment operating noise can be used as one of the bases for judging the operating status of equipment.

[0003] Traditional wind turbine generator sets rely primarily on manual methods for detecting brake pad anomalies in the yaw system, using sound to identify their status. This method is experience-dependent, inefficient, and inconvenient for unified data management across multiple wind turbine generator sets. It also fails to achieve multi-dimensional synchronous monitoring of the turbines, making maintenance operations extremely inconvenient. To address the shortcomings of existing technologies, this invention provides an intelligent anomaly detection platform for wind turbine generator sets to solve the aforementioned problems. Summary of the Invention

[0005] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides an intelligent anomaly detection platform for wind turbine generator sets. This platform solves the problem that traditional wind turbine generator sets rely primarily on manual methods for detecting anomalies in the yaw system brake pads, using sound to identify their status. This method of anomaly detection depends on work experience, is inefficient, and is not conducive to unified data management of multiple wind turbine generator sets. It also fails to achieve multi-dimensional synchronous monitoring of wind turbines, making maintenance operations extremely inconvenient.

[0006] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: an intelligent anomaly detection platform for wind turbine generators, comprising an edge terminal, a server terminal, and a client terminal, wherein the client terminal provides an interactive interface with the user. The server side consists of a data platform and a business processing module. The data platform will use the wind turbine ID as an identifier to uniformly manage sensor data and analysis results. The business processing module is implemented using big data, machine learning, and deep learning technologies, providing analysis and prediction functions for specific data and saving the results to the results database. The edge is a collective term for various sensors and data transmission units. Sensors collect data and upload it to the data platform through the transmission unit.

[0007] Preferably, the client adopts a custom structure, allowing for the addition or reduction of monitoring content and screen layout as needed.

[0008] Preferably, the results database is the basis and origin of the client-side display.

[0009] Preferably, another relatively independent module on the server side is the access module, which is mainly responsible for the interface for various sensor data. It can flexibly access sensors through parameter settings, receive data, and issue instructions.

[0010] Preferably, the edge device also includes a synchronization device that controls the edge device to associate with the action of the target it monitors.

[0011] Preferably, the client implements a multi-layered display structure centered on the wind turbine.

[0012] Preferably, the top layer displays the overall status statistics of a specific customer group, which expands to show a list of wind turbines, allowing users to select a specific device to view the status of a specific module.

[0013] Preferably, the client also provides an entry point for directly accessing a specific monitoring service, displaying statistical data within that service's scope.

[0014] Preferably, the anomaly detection based on deep learning specifically employs deep learning technology. By sampling normal sounds, a quantitative model of sound features is trained. This model projects the features of abnormal and normal sounds into different regions of a high-dimensional space. Based on this, each time a sound is detected, it is compared with the pre-saved features of normal sounds to determine whether it is abnormal.

[0015] This invention discloses an intelligent anomaly detection platform for wind turbine generator sets, which has the following beneficial effects: 1. This intelligent anomaly detection platform for wind turbine generators replaces the traditional manual detection method for anomaly detection of brake pads in the yaw system of wind turbines. It establishes an intelligent equipment monitoring platform and realizes anomaly detection of brake pads in the yaw system of wind turbines on this platform. The platform can access various monitoring information and inspection information applications related to wind turbines. It has the characteristics of easy expansion and unified data management, which makes the maintenance of wind turbine generators convenient.

[0016] 2. This intelligent anomaly detection platform for wind turbine generators adopts a platform-based solution. The system loosely couples equipment, data, business logic, and user interface. Through the data middle platform, data from different sensors in different business processes are managed in a unified manner, enabling multi-dimensional synchronous monitoring of wind turbines. By decoupling data and applications, different application logics can share data. The addition of new business processes and data is very convenient, without affecting existing business processes or requiring cumbersome redevelopment. In the future, maintenance operations management will be incorporated into the platform, allowing for the monitoring and analysis of maintenance operation effectiveness. The platform uses deep learning methods for anomaly detection, exhibiting high robustness to noise. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the detection platform system structure of the present invention; Figure 2 This is a diagram of the sound feature quantization model of the present invention. Detailed Implementation

[0019] 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.

[0020] This application provides an intelligent anomaly detection platform for wind turbine generators. For the brake pads of the wind turbine yaw system, the platform uses sound to identify the status, detect anomalies, and provide corresponding alerts. On the other hand, intelligent monitoring and maintenance of wind turbines will rapidly develop in the future. More sensor information will be combined with big data, AI, and other technologies to intelligently monitor and analyze the operating status of various modules of the wind turbine. The types of signals and the diversity of data will greatly increase, and all this diverse and complex information is related to the wind turbine. Therefore, building a flexible and scalable platform that integrates data and business functions around the wind turbine becomes particularly important.

[0021] This solves the problem that traditional wind turbine generator sets rely primarily on manual methods for detecting brake pad anomalies in the yaw system, using sound to identify their status. This anomaly detection depends on work experience, is inefficient, and is not conducive to unified data management of multiple wind turbine generator sets, making multi-dimensional synchronous monitoring of wind turbines impossible and maintenance operations extremely inconvenient.

[0022] The goal of this project is to build an intelligent equipment monitoring platform and enable anomaly detection of brake pads in the wind turbine yaw system. This platform can integrate various monitoring and inspection information related to the wind turbine, and features easy expansion and unified data management.

[0023] According to the development goals, the main development contents include: the architecture design and implementation of the wind turbine monitoring platform, the brake pad anomaly diagnosis algorithm based on soundprint and the monitoring function based on the algorithm, and the integration of these data and functions into the platform.

[0024] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0025] This invention discloses an intelligent anomaly detection platform for wind turbine generator sets, according to the appendix. Figure 1-2 As shown, the entire system of the intelligent anomaly detection platform for wind turbine generators consists of an edge terminal, a server terminal, and a client terminal. The client terminal provides an interactive interface with the user and adopts a custom structure, allowing for the addition or reduction of monitoring content and screen layout as needed.

[0026] The server side consists of a data platform and a business processing module. The data platform will uniformly manage sensor data and analysis results data using the wind turbine ID as the identifier. The business processing module uses big data, machine learning and deep learning technologies to perform analysis and prediction functions for specific data and save the results to the results database. The results database is the basis and source of the client display.

[0027] Another relatively independent module on the server side is the access module, which is mainly responsible for the interface for various sensor data. It flexibly connects to sensors through parameter settings, receives data, and issues commands.

[0028] The edge is a collective term for various sensors and data transmission units. Sensors collect data and upload it to the data platform via transmission units.

[0029] The edge also includes a synchronization device that controls the edge devices to associate with the actions of the targets they monitor.

[0030] The client implements a multi-layered display structure centered around the wind turbines. The top layer displays overall status statistics for a specific customer group. Expanding it reveals a list of wind turbines, allowing users to select a specific device to view the status of a particular module.

[0031] The client also provides an entry point for directly accessing a specific monitoring service, displaying statistical data within that service's scope. The expanded view is similar to the overall view.

[0032] In deep learning-based anomaly detection, due to the large variety of on-site anomaly features, it is difficult to obtain sufficient data in advance. Therefore, it is almost impossible to manually construct features of anomalies, and only anomaly detection strategies based on normal samples can be adopted.

[0033] In deep learning-based anomaly detection, deep learning technology is used to train a quantitative model of sound features by sampling normal sounds.

[0034] This model projects the features of abnormal and normal sounds into different regions of a high-dimensional space, such as... Figure 2 Based on this, each time a sound is detected, it is compared with the pre-saved normal sound characteristics to determine whether it is abnormal.

[0035] Basic plan: (1) Train a deep learning model to reduce normal business to an ideal point (or small area); (2) If the current input sound is found to deviate significantly from the ideal point during the detection process, it can be assumed that an abnormality has occurred.

[0036] (3) As the usage process progresses, any anomalies encountered (that have been manually confirmed) are saved as anomaly samples and used for identification.

[0037] The system adopts a platform-based approach, loosely coupling devices, data, business logic, and the user interface. Its main features are: (1) By using the data platform, data from different sensors in different businesses can be managed in a unified manner, enabling multi-dimensional synchronous monitoring of wind turbines. (2) By decoupling data and applications, different application logics can share data. (3) The addition of new business and data is very convenient, will not affect the original business, and does not require heavy redevelopment. (4) In the future, the management of maintenance operations will be incorporated into the platform, which will enable the monitoring and analysis of the effectiveness of maintenance operations. (5) The use of deep learning methods for anomaly detection has high robustness to noise. This intelligent anomaly detection platform for wind turbine generators replaces the traditional manual detection method for anomaly detection of brake pads in the yaw system of wind turbines. It establishes an intelligent equipment monitoring platform and realizes anomaly detection of brake pads in the yaw system of wind turbines on this platform. The platform can access various monitoring information and inspection information applications related to wind turbines. It has the characteristics of easy expansion and unified data management, which makes the maintenance of wind turbine generators convenient.

[0038] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0039] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent anomaly detection platform for wind turbine generator sets, comprising an edge terminal, a server terminal, and a client terminal, characterized in that: The client implements the interactive interface with the user; The server side consists of a data platform and a business processing module. The data platform will use the wind turbine ID as an identifier to uniformly manage sensor data and analysis results. The business processing module is implemented using big data, machine learning, and deep learning technologies, providing analysis and prediction functions for specific data and saving the results to the results database. The edge is a collective term for various sensors and data transmission units. Sensors collect data and upload it to the data platform through the transmission unit.

2. The intelligent anomaly detection platform for wind turbine generator sets according to claim 1, characterized in that: The client uses a custom structure, allowing users to add or remove monitoring content and screen layout as needed.

3. The intelligent anomaly detection platform for wind turbine generator sets according to claim 1, characterized in that: The results database is the foundation and origin of the client-side display.

4. The intelligent anomaly detection platform for wind turbine generator sets according to claim 1, characterized in that: Another relatively independent module on the server side is the access module, which is mainly responsible for the interface for various sensor data. It flexibly connects to sensors through parameter settings, receives data, and issues commands.

5. The intelligent anomaly detection platform for wind turbine generator sets according to claim 1, characterized in that: The edge also includes a synchronization device that controls the edge devices to associate with the actions of the targets they monitor.

6. The intelligent anomaly detection platform for wind turbine generator sets according to claim 1, characterized in that: The client implements a multi-layered display structure centered around the wind turbine.

7. The intelligent anomaly detection platform for wind turbine generator sets according to claim 6, characterized in that: The top layer displays the overall status statistics for a specific customer group. Expanding it reveals a list of wind turbines, allowing users to select a specific device to view the status of a particular module.

8. The intelligent anomaly detection platform for wind turbine generator sets according to claim 1, characterized in that: The client also provides an entry point for directly accessing a specific monitoring service, displaying statistical data within that service's scope.

9. The intelligent anomaly detection platform for wind turbine generator sets according to claim 1, characterized in that: Anomaly detection based on deep learning specifically employs deep learning technology. By sampling normal sounds, a quantification model of sound features is trained. This model projects the features of abnormal and normal sounds into different regions of a high-dimensional space. Based on this, each time a sound is detected, it is compared with the pre-saved features of normal sounds to determine whether it is abnormal.

Images