Fan fault detection method, device, medium and equipment

Abstract

The present disclosure relates to a fan fault detection method, device, medium and equipment. The method comprises: acquiring current working condition data of the fan; determining whether the fan has a fault according to the acquired current working condition data, predetermined normal working condition data, a principal component analysis (PCA) algorithm, a Gaussian mixture clustering (GMM) algorithm and a kernel density estimation (KDE) algorithm. In this way, it can be accurately and reliably determined whether the fan has a fault, and the false alarm rate is reduced.

Description

Technical Field

[0001] This disclosure relates to the field of fault detection technology, specifically to a method, apparatus, medium, and equipment for detecting wind turbine faults. Background Technology

[0002] As the installed capacity of wind turbines continues to increase, the proportion of total wind power generation capacity in the entire power system is increasing year by year.

[0003] Wind turbines operate in outdoor environments for extended periods, exposed to harsh conditions such as thunderstorms, typhoons, and hail, making them prone to malfunctions and requiring timely maintenance. Failure to accurately determine if a turbine has malfunctioned, allowing it to continue operating despite a fault, can easily lead to damage and shutdown. Conversely, misdiagnosing a malfunction when it is not actually present disrupts normal operation and wastes manpower.

[0004] In related technologies, fault detection is usually performed on the operating parameters of each piece of equipment in the wind turbine separately. However, the accuracy of the detection results is low and the false alarm rate of faults is high. Summary of the Invention

[0005] The purpose of this disclosure is to provide a method, apparatus, medium, and equipment for detecting fan faults, which can accurately and reliably determine whether a fan has malfunctioned.

[0006] To achieve the above objectives, this disclosure provides a method for detecting wind turbine faults, the method comprising:

[0007] Obtain the current operating status data of the fan;

[0008] Based on the acquired current operating condition data, predetermined normal operating condition data, principal component analysis (PCA) algorithm, Gaussian mixture clustering (GMM) algorithm, and kernel density estimation (KDE) algorithm, it is determined whether the wind turbine has malfunctioned.

[0009] Optionally, determining whether the wind turbine has malfunctioned based on the acquired current operating condition data, predetermined normal operating condition data, principal component analysis (PCA) algorithm, Gaussian mixture clustering (GMM) algorithm, and kernel density estimation (KDE) algorithm includes:

[0010] The feature data of the current operating condition data obtained is filtered through the PCA algorithm;

[0011] The system determines whether the fan has malfunctioned based on the selected feature data, the normal operating condition data, the GMM algorithm, and the KDE algorithm.

[0012] Optionally, determining whether the wind turbine has malfunctioned based on the selected feature data, the normal operating condition data, the GMM algorithm, and the KDE algorithm includes:

[0013] The similarity between the selected feature data and the normal operating condition data is determined based on the GMM algorithm and the KDE algorithm.

[0014] The determination of whether the fan has malfunctioned is based on the determined similarity.

[0015] Optionally, determining whether the wind turbine has malfunctioned based on the determined similarity includes:

[0016] If the determined similarity is less than the predetermined similarity threshold, then the wind turbine is determined to have malfunctioned.

[0017] Optionally, the method further includes:

[0018] If it is determined that the fan has malfunctioned, the malfunction level is determined based on the determined similarity.

[0019] Optionally, determining the fault level based on the determined similarity includes:

[0020] The fault level corresponding to the determined similarity is found from the predetermined correspondence and used as the determined fault level. The correspondence includes the correspondence between the similarity and the fault level.

[0021] Optionally, the method further includes:

[0022] If a fault is detected in the fan, an alarm message is output.

[0023] This disclosure also provides a wind turbine fault detection device, the device comprising:

[0024] The acquisition module is used to acquire the current operating condition data of the wind turbine;

[0025] The first determining module is used to determine whether the wind turbine has malfunctioned based on the acquired current operating condition data, predetermined normal operating condition data, PCA algorithm, GMM algorithm, and KDE algorithm.

[0026] This disclosure also provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the steps of the wind turbine fault detection method described above.

[0027] This disclosure also provides an apparatus, comprising:

[0028] processor;

[0029] Memory used to store the processor's executable instructions;

[0030] The processor is configured as follows:

[0031] Obtain the current operating status data of the fan;

[0032] The system determines whether the wind turbine has malfunctioned based on the acquired current operating condition data, the predetermined normal operating condition data, the PCA algorithm, the GMM algorithm, and the KDE algorithm.

[0033] The above technical solution acquires the current operating condition data of the wind turbine. Based on the acquired current operating condition data, predetermined normal operating condition data, and PCA, GMM, and KDE algorithms, it determines whether the wind turbine has malfunctioned. This allows for accurate and reliable determination of wind turbine malfunctions, reducing the false alarm rate.

[0034] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description

[0035] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings:

[0036] Figure 1 This is a flowchart of a wind turbine fault detection method provided in an exemplary embodiment.

[0037] Figure 2 This is a block diagram of a wind turbine fault detection device provided in an exemplary embodiment. Detailed Implementation

[0038] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.

[0039] It should be noted that all actions involving the acquisition of signals, information, or data in this disclosure are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with authorization from the owner of the relevant device.

[0040] Figure 1 This is a flowchart of a wind turbine fault detection method provided in an exemplary embodiment. For example... Figure 1 As shown, the method includes the following steps.

[0041] In step S101, the current operating condition data of the wind turbine is obtained.

[0042] A wind turbine includes components such as a rotor, generator, yaw mechanism, and speed limiting safety device. During operation, the operating data of each component reflects its individual status, which in turn reflects the overall operating status of the wind turbine. Current operating data can include environmental data, speed data, and temperature data. Environmental data may include wind speed, pitch control, and power. Speed ​​data may include rotor speed and generator speed. Temperature data may include main shaft temperature, ambient temperature, and nacelle temperature. Before fault detection, personnel can pre-set the types and frequency of operating data to be acquired based on the specific fault types to be detected for each inspection task.

[0043] In step S102, the wind turbine is determined to have malfunctioned based on the acquired current operating condition data, the predetermined normal operating condition data, the Principal Component Analysis (PCA) algorithm, the Gaussian Mixture Model (GMM) algorithm, and the Kernel Density Estimation (KDE) algorithm.

[0044] Normal operating condition data refers to the operating condition data when the wind turbine is running normally (without malfunctions), which can be preset by the designer.

[0045] PCA is a data preprocessing algorithm that can remove redundant data from the original data, making the data simpler and thus improving the data processing speed.

[0046] The Gaussian Memory Model (GMM) algorithm accurately quantifies data by using the Gaussian probability density function (normal distribution curve). It decomposes a data point into several Gaussian probability density functions (normal distribution curves), effectively characterizing the spatial distribution and characteristics of data in the parameter space. Furthermore, it facilitates data estimation and is widely used in pattern recognition and data analysis. The KDDE algorithm, on the other hand, generates multidimensional probability distribution estimates using multidimensional datasets. The GMM algorithm uses a weighted summation of different Gaussian distributions to represent the probability distribution estimate. GMM and KD differ in their strengths: GMM is better suited for analyzing circular data classifications, while KD is better suited for analyzing long, narrow data classifications. By combining GMM and KD algorithms, we can analyze the operating data of various types of wind turbines to determine if a fault has occurred, thus optimizing fault detection and ensuring the accuracy of the results.

[0047] The above technical solution acquires the current operating condition data of the wind turbine. Based on the acquired current operating condition data, predetermined normal operating condition data, and PCA, GMM, and KDE algorithms, it determines whether the wind turbine has malfunctioned. This allows for accurate and reliable determination of wind turbine malfunctions, reducing the false alarm rate.

[0048] In another embodiment, the determination of whether the wind turbine has malfunctioned based on the acquired current operating condition data, predetermined normal operating condition data, PCA algorithm, GMM algorithm, and KDE algorithm includes:

[0049] Feature data of the current operating condition data obtained is filtered using the PCA algorithm;

[0050] The system determines whether the fan has malfunctioned based on the selected feature data, normal operating condition data, GMM algorithm, and KDE algorithm.

[0051] Feature data refers to the data remaining after removing overlapping information from the acquired operating condition data using the PCA algorithm. The PCA algorithm can classify different types of current operating condition data (e.g., environmental data, speed data, and temperature data). By transforming high-dimensional data to low-dimensional data and extracting the largest individual differences from the principal components, the feature data of the current operating condition data is selected.

[0052] After filtering out the characteristic data of the acquired current operating condition data, it is possible to determine whether the fan has malfunctioned based on the filtered characteristic data, normal operating condition data, GMM algorithm and KDE algorithm.

[0053] In this embodiment, the PCA algorithm is used to filter out the feature data of the current operating condition data, which speeds up the data processing. Based on the filtered feature data, normal operating condition data, GMM algorithm and KDE algorithm, it is possible to determine in a timely, accurate and reliable manner whether the wind turbine has failed.

[0054] In another embodiment, the determination of whether the wind turbine has malfunctioned based on the selected feature data, normal operating condition data, GMM algorithm, and KDE algorithm includes:

[0055] The similarity between the selected feature data and the normal operating condition data is determined based on the GMM algorithm and the KDE algorithm.

[0056] The determination of whether the fan has malfunctioned is based on the established similarity.

[0057] After filtering out the feature data of the current operating condition data, the filtered feature data and the predetermined normal operating condition data can be input into the GMM algorithm and the KDE algorithm, respectively. After the GMM algorithm and KDE algorithm perform the calculations, they respectively obtain the similarity between the input feature data and the normal operating condition data. Both the GMM algorithm and the KDE algorithm can output the similarity between the feature data and the normal data. Because they are good at handling different types of data, the similarity determined by them will have a small difference. The average of the similarity determined by the GMM algorithm and the KDE algorithm can be determined as the similarity between the feature data and the normal data.

[0058] Similarity represents the repetition rate. The higher the similarity, the higher the repetition rate between the feature data and the normal operating condition data, and the more stable the operation of the wind turbine. The lower the similarity, the lower the repetition rate between the feature data and the normal operating condition data, and the more chaotic the operation of the wind turbine.

[0059] In this embodiment, compared with the related technology, which determines whether a wind turbine has failed by analyzing the operating data of each type of wind turbine individually, the GMM algorithm and KDE algorithm can determine the similarity between the feature data of multiple types of operating data and the predetermined normal operating data. Based on the determined similarity, it is possible to accurately and reliably determine whether a wind turbine has failed.

[0060] In another embodiment, the similarity between the selected feature data and the normal operating condition data can be determined using the K-means clustering algorithm. K-means clustering is an iterative clustering analysis algorithm that divides the data into K groups, randomly selects K objects as initial cluster centers, calculates the distance between each object and each seed cluster center, and assigns each object to the nearest cluster center. Each cluster center and the objects assigned to it represent a cluster. Each time a sample is assigned, the cluster centers are recalculated based on the existing objects in the cluster. This process is repeated until a termination condition is met. The termination condition may be that no (or a minimum number) objects are reassigned to different clusters, no (or a minimum number) cluster centers change, or the sum of squared errors reaches a local minimum. After selecting the feature data for the current operating condition, the selected feature data and the predetermined normal operating condition data can be input into the K-means clustering algorithm to calculate and determine the similarity between the feature data and the normal operating condition data.

[0061] In yet another embodiment, determining whether a fan has malfunctioned based on the determined similarity includes:

[0062] If the determined similarity is less than the predetermined similarity threshold, then the wind turbine is determined to have malfunctioned.

[0063] The similarity threshold can be preset by the designer, for example, 90%. When the determined similarity threshold is less than the preset threshold, the repetition rate between the feature data and normal data is low, the wind turbine's operating state is unstable, and it can be determined that the wind turbine has failed. When the determined similarity threshold is greater than the preset threshold, the repetition rate between the feature data and normal data is high, the wind turbine's operating state is stable, and it can be determined that the wind turbine has not failed.

[0064] In this embodiment, when the determined similarity is less than a predetermined similarity threshold, the fan failure can be accurately and reliably determined.

[0065] In yet another embodiment, the method further includes:

[0066] If a fan failure is determined, the failure level is determined based on the determined similarity.

[0067] The fault level represents the severity of a wind turbine malfunction; the more severe the malfunction, the greater the potential damage. Once a wind turbine malfunction is confirmed, the fault level can be determined based on the established similarity.

[0068] In this embodiment, when a fan failure is determined, the failure level can be determined based on the determined similarity, which can reflect the severity of the fan failure and help staff take corresponding measures based on the determined failure level.

[0069] In yet another embodiment, determining the fault level based on the determined similarity includes:

[0070] The fault level corresponding to the determined similarity is found from the predetermined correspondence, and is used as the determined fault level. The correspondence includes the correspondence between similarity and fault level.

[0071] The correspondence between similarity and fault severity can be pre-defined by the designer. For example, if the determined similarity is 80%, the fan is classified as experiencing a Level 1 fault; if the determined similarity is 70%, the fan is classified as experiencing a Level 2 fault; and if the determined similarity is 60%, the fan is classified as experiencing a Level 3 fault. The severity of a Level 1 fault is lower than the severity of a Level 2 fault.

[0072] In this embodiment, a lookup table method is used to quickly determine the fault level corresponding to the similarity. The method is simple and the data processing speed is fast.

[0073] In yet another embodiment, the method further includes:

[0074] If a fan malfunction is detected, an alarm message will be output.

[0075] When a fan malfunction is detected, an alarm message can be output. For example, a pop-up message can be displayed on the screen of the display device: "The fan has experienced a level one malfunction. Please handle it promptly."

[0076] In this embodiment, when a malfunction is detected in the blower, an alarm message is output to facilitate timely action by staff, thereby improving safety.

[0077] Figure 2 This is a block diagram of a wind turbine fault detection device provided in an exemplary embodiment. Figure 2 As shown, the wind turbine fault detection device 200 includes: an acquisition module 201 and a first determination module 202.

[0078] The acquisition module 201 is used to acquire the current operating condition data of the wind turbine.

[0079] The first determining module 202 is used to determine whether the wind turbine has malfunctioned based on the acquired current operating condition data, the predetermined normal operating condition data, the PCA algorithm, the GMM algorithm, and the KDE algorithm.

[0080] Optionally, the first determining module 202 includes a filtering submodule and a determining submodule.

[0081] The filtering submodule is used to filter the feature data of the acquired current operating condition data using the PCA algorithm.

[0082] The determination submodule is used to determine whether the fan has failed based on the selected feature data, normal operating condition data, GMM algorithm and KDE algorithm.

[0083] Optionally, the determination submodule is also used to determine the similarity between the selected feature data and the normal operating condition data based on the GMM algorithm and the KDE algorithm;

[0084] The determination of whether the fan has malfunctioned is based on the established similarity.

[0085] Optionally, the determining submodule is also used to determine that the wind turbine has failed if the determined similarity is less than a predetermined similarity threshold.

[0086] Optionally, the wind turbine fault detection device 200 also includes a second determination module.

[0087] The second determination module is used to determine the fault level based on the determined similarity if it is determined that the wind turbine has failed.

[0088] Optionally, the second determining module includes a lookup submodule.

[0089] The lookup submodule is used to find the fault level corresponding to the determined similarity from the predetermined correspondence, and the correspondence includes the correspondence between similarity and fault level.

[0090] Optionally, the fan fault detection device 200 also includes an output module.

[0091] The output module is used to output an alarm message if a fan malfunction is determined.

[0092] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0093] The above technical solution acquires the current operating condition data of the wind turbine. Based on the acquired current operating condition data, predetermined normal operating condition data, and PCA, GMM, and KDE algorithms, it determines whether the wind turbine has malfunctioned. This allows for accurate and reliable determination of wind turbine malfunctions, reducing the false alarm rate.

[0094] This disclosure also provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the steps of the wind turbine fault detection method described above.

[0095] This disclosure also provides an apparatus, comprising:

[0096] processor;

[0097] Memory used to store processor-executable instructions;

[0098] The processor is configured as follows:

[0099] Obtain the current operating status data of the wind turbine;

[0100] The system determines whether the wind turbine has malfunctioned based on the acquired current operating condition data, the predetermined normal operating condition data, the PCA algorithm, the GMM algorithm, and the KDE algorithm.

[0101] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0102] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0103] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

Claims

1. A method for detecting fan faults, characterized in that, The method includes: Obtain the current operating status data of the fan; Based on the acquired current operating condition data, the predetermined normal operating condition data, the principal component analysis (PCA) algorithm, the Gaussian mixture clustering (GMM) algorithm, and the kernel density estimation (KDE) algorithm, it is determined whether the wind turbine has malfunctioned. The step of determining whether the wind turbine has malfunctioned based on the acquired current operating condition data, predetermined normal operating condition data, principal component analysis (PCA) algorithm, Gaussian mixture clustering (GMM) algorithm, and kernel density estimation (KDE) algorithm includes: The feature data of the current operating condition data obtained is filtered through the PCA algorithm; Based on the selected feature data, the normal operating condition data, the GMM algorithm, and the KDE algorithm, it is determined whether the wind turbine has malfunctioned. The step of determining whether the wind turbine has malfunctioned based on the selected feature data, the normal operating condition data, the GMM algorithm, and the KDE algorithm includes: The similarity between the selected feature data and the normal operating condition data is determined based on the GMM algorithm and the KDE algorithm. The determination of whether the fan has malfunctioned is based on the determined similarity. Determining whether the wind turbine has malfunctioned based on the determined similarity includes: If the determined similarity is less than the predetermined similarity threshold, then the wind turbine is determined to have malfunctioned.

2. The method according to claim 1, characterized in that, The method further includes: If it is determined that the fan has malfunctioned, the malfunction level is determined based on the determined similarity.

3. The method according to claim 2, characterized in that, The step of determining the fault level based on the determined similarity includes: The fault level corresponding to the determined similarity is found from the predetermined correspondence and used as the determined fault level. The correspondence includes the correspondence between the similarity and the fault level.

4. The method according to claim 1, characterized in that, The method further includes: If a fault is detected in the fan, an alarm message is output.

5. A fan fault detection device, characterized in that, The device includes: The acquisition module is used to acquire the current operating condition data of the wind turbine; The first determining module is used to determine whether the wind turbine has malfunctioned based on the acquired current operating condition data, the predetermined normal operating condition data, the PCA algorithm, the GMM algorithm, and the KDE algorithm. The first determining module includes a filtering submodule and a determining submodule; The filtering submodule is used to filter the feature data of the acquired current working condition data through the PCA algorithm. The determination submodule is used to determine whether the fan has malfunctioned based on the selected feature data, the normal operating condition data, the GMM algorithm, and the KDE algorithm. The determining submodule is further configured to determine the similarity between the screened feature data and the normal operating condition data based on the GMM algorithm and the KDE algorithm; The determination of whether the fan has malfunctioned is based on the determined similarity. The determining submodule is further configured to determine that the wind turbine has malfunctioned if the determined similarity is less than a predetermined similarity threshold.

6. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When executed by a processor, the program instructions implement the steps of the method described in any one of claims 1 to 4.

7. A device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured as follows: Obtain the current operating status data of the wind turbine; The system determines whether the wind turbine has malfunctioned based on the acquired current operating condition data, the predetermined normal operating condition data, the PCA algorithm, the GMM algorithm, and the KDE algorithm. The step of determining whether the wind turbine has malfunctioned based on the acquired current operating condition data, predetermined normal operating condition data, principal component analysis (PCA) algorithm, Gaussian mixture clustering (GMM) algorithm, and kernel density estimation (KDE) algorithm includes: The feature data of the current operating condition data obtained is filtered through the PCA algorithm; Based on the selected feature data, the normal operating condition data, the GMM algorithm, and the KDE algorithm, it is determined whether the wind turbine has malfunctioned. The step of determining whether the wind turbine has malfunctioned based on the selected feature data, the normal operating condition data, the GMM algorithm, and the KDE algorithm includes: The similarity between the selected feature data and the normal operating condition data is determined based on the GMM algorithm and the KDE algorithm. The determination of whether the fan has malfunctioned is based on the determined similarity. Determining whether the wind turbine has malfunctioned based on the determined similarity includes: If the determined similarity is less than the predetermined similarity threshold, then the wind turbine is determined to have malfunctioned.

Images