Method and device for detecting damage to a fan blade

Abstract

The application discloses a kind of detection methods of fan blade damage, including dividing detection area on fan blade, using microwave sensor to detect detection area and obtain scattering parameter, using scattering parameter to monitor and analyze fan blade and obtain the damage condition of fan blade.The application also discloses a kind of detection device of fan blade damage.The application can more accurately obtain the damage condition of fan blade by using microwave sensor to analyze the data received by microwave sensor.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and in particular to a method and apparatus for detecting damage to wind turbine blades. Background Technology

[0002] Wind energy harvesting has garnered increasing attention due to growing public awareness of impending climate change and improved wind power system efficiency. Wind turbines are the primary equipment in wind power generation. Blade failure is a major cause of operational accidents and turbine damage; therefore, monitoring the structural health of blades and diagnosing faults are critical issues in operation and maintenance. Currently, blade fault detection often employs non-destructive testing techniques to identify damage in composite blades, such as traditional ultrasonic testing and fiber optic sensors.

[0003] Current traditional nondestructive testing methods are not applicable to all types of defects in all materials. They are easily affected by environmental factors and different material properties, resulting in large errors and inaccurate detection.

[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this invention is to provide a method and device for detecting damage to wind turbine blades, aiming to solve the technical problem of large errors in the detection results in the prior art.

[0006] In a first aspect, the present invention provides a method for detecting damage to wind turbine blades, comprising:

[0007] A microwave sensor with a preset frequency is fixed at a preset distance above a preset area of ​​a wind turbine blade moving at a preset speed, and measurement is started to obtain detection data of the preset area.

[0008] Based on the amplitude of the scattering parameter of the leaf damage state in the detection data, the amplitude of the scattering parameter of the leaf damage state in the reference data is extracted, and the amplitude of the scattering parameter of the undamaged leaf state in the reference data is obtained, so as to obtain the exponent of the average absolute amplitude difference between the leaf damage state and the undamaged leaf state in the preset area.

[0009] Based on the phase of the scattering parameters of the leaf damage state in the detection data, the phase of the scattering parameters of the leaf damage state in the reference data is extracted, and the phase of the scattering parameters of the undamaged leaf state in the reference data is obtained, so as to obtain the exponent of the average absolute phase difference between the leaf damage state and the undamaged leaf state in the preset region.

[0010] The damage gap range of the preset region of the blade is obtained based on the exponent of the mean absolute amplitude difference and the exponent of the mean absolute phase difference.

[0011] The detection data is analyzed using a preset classification regression algorithm to determine the type and number of damage gaps in the preset area of ​​the blade.

[0012] In one embodiment, the step of fixing a preset microwave sensor of a preset frequency at a preset distance above a preset area of ​​a wind turbine blade moving at a preset rotational speed and starting measurement to obtain detection data of the preset area includes:

[0013] In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at a preset distance above the preset area of ​​the wind turbine blade moving at the preset rotation speed, and the measurement is started to obtain the simulated data of the preset area;

[0014] The preset distance is changed, the preset frequency is changed under each different preset distance condition, and the amplitude of the scattering parameter in the simulation data is extracted. The relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameter in the simulation data is compared to obtain the optimal preset frequency and the optimal preset distance.

[0015] A microwave sensor with the optimal preset frequency is fixed at an optimal distance above a preset area of ​​a wind turbine blade moving at a preset speed, and measurement is started to obtain detection data of the preset area.

[0016] In one embodiment, the step of changing the preset distance, changing the preset frequency under each different preset distance condition, and extracting the amplitude of the scattering parameters in the simulation data, comparing the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulation data, yields the optimal preset frequency and optimal preset distance, which is replaced by:

[0017] The preset distance is changed, the preset frequency is changed under each different preset distance condition, and the phase of the scattering parameters in the simulation data is extracted. The relationship between each preset frequency under each different preset distance condition and the phase of the scattering parameters in the simulation data is compared to obtain the optimal preset frequency and the optimal preset distance.

[0018] In one embodiment, the step of changing the preset distance, changing the preset frequency under each different preset distance condition, and extracting the amplitude of the scattering parameters in the simulation data, comparing the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulation data, yields the optimal preset frequency and optimal preset distance, which is replaced by:

[0019] By changing the preset rotation speed, the scattering parameters in the simulation data at different preset rotation speeds are extracted and compared to obtain the optimal preset rotation speed.

[0020] In one embodiment, the step of extracting the amplitude of the scattering parameter of the leaf damage state in the reference data based on the amplitude of the scattering parameter of the leaf damage state in the detection data, and obtaining the amplitude of the scattering parameter of the undamaged leaf state in the reference data, and obtaining the exponent of the average absolute amplitude difference between the leaf damage state and the undamaged leaf state in the preset region, includes:

[0021] In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at a preset distance above the preset area of ​​the wind turbine blade moving at the preset speed and the measurement is started. The damage condition of the preset area is changed, and the preset frequency is changed within the preset frequency range to obtain the resonant frequency and scattering parameters of the preset area as reference data.

[0022] Based on the amplitude of the scattering parameters of the damaged state of the blade in the detection data, the amplitude of the scattering parameters at multiple frequency points in the reference data that are the same as the damaged state of the blade is extracted, and the amplitude of the scattering parameters at multiple frequency points in the undamaged state of the reference data is extracted, so as to obtain the exponent of the average absolute amplitude difference between the damaged state of the blade and the undamaged state of the blade in the preset region.

[0023] In one embodiment, the step of extracting the phase of the scattering parameter of the leaf damage state in the reference data based on the phase of the scattering parameter of the leaf damage state in the detection data, and obtaining the phase of the scattering parameter of the undamaged leaf state in the reference data, to obtain the exponent of the average absolute phase difference between the leaf damage state and the undamaged leaf state in the preset region, includes:

[0024] In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at a preset distance above the preset area of ​​the wind turbine blade moving at the preset speed and the measurement is started. The damage condition of the preset area is changed, and the preset frequency is changed within the preset frequency range to obtain the resonant frequency and scattering parameters of the preset area as reference data.

[0025] Based on the phase of the scattering parameters of the leaf damage state in the detection data, the phase of the scattering parameters at multiple frequency points in the reference data that are the same as the leaf damage state is extracted, and the phase of the scattering parameters at multiple frequency points in the undamaged state in the reference data is extracted, so as to obtain the exponent of the average absolute phase difference between the leaf damage state and the undamaged state in the preset region.

[0026] In one embodiment, the step of using a preset classification regression algorithm to analyze the detection data to determine the type and number of damage gaps in the preset region of the blade includes:

[0027] Train the preset classification and regression algorithm to obtain the trained preset classification and regression algorithm;

[0028] Input the scattering parameters of the leaf damage state in the detection data, and obtain the type and number of damage gaps in the preset region of the leaf according to the trained preset classification and regression algorithm.

[0029] In one embodiment, the step of training the preset classification and regression algorithm to obtain the trained preset classification and regression algorithm includes:

[0030] The size of the scattering parameters of a first preset number of different known leaf damage states is set as the training dataset, and the size of the scattering parameters of each known leaf damage state is a training sample. Each training sample is labeled with a category label and stored in the training database.

[0031] The size of the scattering parameters of a second preset number of unknown leaf damage states is set as the test dataset, and the size of the scattering parameters of each unknown leaf damage state is a test sample.

[0032] Each time, a test sample is selected from the test dataset, the Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated, and the third preset number of training samples that are closest to the test sample in Euclidean distance from all training samples are selected as nearest neighbor samples.

[0033] Based on the query pattern of the training samples, the category labels of the third preset number of nearest neighbor samples are voted on to determine the category label of the test sample. The category label of the test sample is labeled and the query pattern of the test sample is provided. The labeled test sample and the query pattern of the test sample are stored in the training database and the next test sample in the test dataset is selected until the test dataset is selected, and the trained preset classification and regression algorithm is obtained.

[0034] In one embodiment, the step of selecting a test sample from the test dataset each time, calculating the Euclidean distance and query pattern between the test sample and all training samples in the training dataset, and selecting a third preset number of training samples from all training samples that are closest to the test sample in Euclidean distance as nearest neighbor samples includes:

[0035] Each time, a test sample is selected from the test dataset, and the Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated.

[0036] Based on the increasing Euclidean distance between the test sample and all training samples in the training dataset, all training samples in the training dataset are sorted to obtain a list of the training dataset, and the third preset number of training samples at the beginning of the list are selected as nearest neighbor samples.

[0037] Secondly, the present invention also provides a device for detecting damage to wind turbine blades, comprising:

[0038] The sensing module is used to divide a preset area on the wind turbine blade, emit microwave incident waves to the preset area and receive microwave reflected waves from the preset area, set the microwave incident waves and the microwave reflected waves as detection data and transmit them to the first analysis module, the second analysis module and the third calculation module.

[0039] The first analysis module is used to analyze the detection data, obtain the index of the average absolute amplitude difference of the preset area, and transmit it to the first calculation module.

[0040] The second analysis module is used to analyze the detection data, obtain the exponent of the average absolute phase difference of the preset area, and transmit it to the first calculation module.

[0041] The first calculation module is used to calculate the range of the damage gap of the wind turbine blades in the preset area based on the exponent of the average absolute amplitude difference and the exponent of the average absolute phase difference.

[0042] The second calculation module is used to calculate the type and number of damaged gaps in the wind turbine blades based on the preset classification regression calculation and the detection data.

[0043] The aforementioned method for detecting wind turbine blade damage involves fixing a preset microwave sensor of a preset frequency at a preset distance above a preset area of ​​a wind turbine blade moving at a preset rotational speed and starting measurement to acquire detection data for the preset area. Based on the amplitude of the scattering parameters of the blade damage state in the detection data, the amplitude of the scattering parameters of the blade damage state in the reference data is extracted, and the amplitude of the scattering parameters of the blade without damage in the reference data is obtained, resulting in an exponent of the average absolute amplitude difference between the blade damage state and the blade without damage in the preset area. Based on the phase of the scattering parameters of the blade damage state in the detection data, the phase of the scattering parameters of the blade damage state in the reference data is extracted, and the phase of the scattering parameters of the blade without damage in the reference data is obtained, resulting in an exponent of the average absolute phase difference between the blade damage state and the blade without damage in the preset area. Based on the exponent of the average absolute amplitude difference and the exponent of the average absolute phase difference, the damage gap range of the preset area of ​​the blade is obtained. Finally, a preset classification regression algorithm is used to analyze the detection data to determine the type and number of damage gaps in the preset area of ​​the blade. Through the above method, this application uses a microwave sensor for detection, utilizes the amplitude and phase in the scattering parameters and uses a classification regression algorithm to detect and calculate the damage to the wind turbine blades, so that the damage detection results of the wind turbine blades are not affected by other factors, and the detection results are more accurate. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the hardware structure of the terminal involved in the embodiments of this application;

[0045] Figure 2 This is a flowchart illustrating one embodiment of the wind turbine blade damage detection method in this application.

[0046] Figure 3 This is a schematic diagram of a preset area in one embodiment of the wind turbine blade damage detection method of this application.

[0047] Figure 4 This is a schematic diagram of the preset distance in one embodiment of the wind turbine blade damage detection method of this application.

[0048] Figure 5 This is a flowchart illustrating one embodiment of the wind turbine blade damage detection method in this application.

[0049] Figure 6 This is a schematic diagram of amplitude variation in one embodiment of the wind turbine blade damage detection method in this application.

[0050] Figure 7 This is a schematic diagram of phase changes in one embodiment of the wind turbine blade damage detection method in this application.

[0051] Figure 8 This is a flowchart illustrating one embodiment of the wind turbine blade damage detection method in this application.

[0052] Figure 9 This is a flowchart illustrating one embodiment of the wind turbine blade damage detection method in this application.

[0053] Figure 10 This is a flowchart illustrating one embodiment of the wind turbine blade damage detection method in this application.

[0054] Figure 11 This is a flowchart illustrating one embodiment of the wind turbine blade damage detection method in this application.

[0055] Figure 12 This is a flowchart illustrating one embodiment of the wind turbine blade damage detection method in this application.

[0056] Figure 13 This is a flowchart illustrating an embodiment of the wind turbine blade damage detection device in this application.

[0057] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0059] The terminal provided in this application embodiment can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, portable wearable devices, and servers. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc.

[0060] Its internal structure diagram can be as follows Figure 1As shown, the terminal includes a processor 1001, a memory 1005, a user interface 1003, and a network interface 1004 connected via a system bus. The processor 1001 provides computing and control capabilities. The memory 1005 includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The memory 1005 also stores various data such as detection data, training datasets, and test datasets. The network interface 1004 is used for communication with external terminals via a network connection. When executed by the processor, this terminal implements a method for detecting wind turbine blade damage.

[0061] Those skilled in the art will understand that Figure 1 The terminal structure shown does not constitute a limitation on the terminal and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0062] Reference Figure 2 This invention provides a method for detecting damage to wind turbine blades, the method comprising:

[0063] S110, Fix a preset microwave sensor of preset frequency at a preset distance above a preset area of ​​a wind turbine blade moving at a preset speed and start measuring to obtain detection data of the preset area;

[0064] The preset area is a cuboid with a preset length, preset width and preset thickness.

[0065] Specifically, a preset microwave sensor sends an incident wave and receives a reflected wave from a preset region, using the incident wave signal and the reflected wave signal as detection data for the preset region. For example, such as... Figure 3 As shown, select an area on the wind turbine blade that is 50mm long, 10mm wide, and 1mm thick; Figure 4 As shown, the microwave sensor is fixed at a preset distance from the wind turbine blade.

[0066] S120, based on the amplitude of the scattering parameter of the leaf damage state in the detection data, extract the amplitude of the scattering parameter of the leaf damage state in the reference data, and obtain the amplitude of the scattering parameter of the undamaged leaf state in the reference data, and obtain the index of the average absolute amplitude difference between the leaf damage state and the undamaged leaf state in the preset area.

[0067] The formula for calculating the exponent of the mean absolute amplitude difference is:

[0068]

[0069] Specifically, based on the amplitude of the scattering parameters of the leaf damage state in the detection data, N frequency points in the reference data that are identical to the leaf damage state are obtained. The amplitudes of the damaged state at these N frequency points in the reference data are then extracted and calculated along with the amplitudes of the N frequency points in the undamaged state of the leaf extracted from the reference data. This yields the average absolute amplitude difference between the damaged state and the undamaged state, where N represents the number of frequency points, and S... i (x) represents the amplitude of the scattering parameter of the blade damage state at the i-th frequency, S i (0) represents the amplitude of the scattering parameter of the blade in the undamaged state at the i-th frequency.

[0070] S130, based on the phase of the scattering parameter of the leaf damage state in the detection data, extract the phase of the scattering parameter of the leaf damage state in the reference data, and obtain the phase of the scattering parameter of the leaf undamaged state in the reference data, and obtain the exponent of the average absolute phase difference between the leaf damage state and the leaf undamaged state in the preset area.

[0071] The formula for calculating the exponent of the mean absolute amplitude difference is:

[0072]

[0073] Specifically, based on the amplitude of the scattering parameters of the damaged state of the leaf in the detection data, N frequency points in the reference data that are the same as the damaged state of the leaf are obtained. The amplitude of the damaged state at these N frequency points in the reference data and the amplitude at the N frequency points of the undamaged state of the leaf extracted from the reference data are calculated to obtain the average absolute amplitude difference between the damaged state and the undamaged state of the leaf. Here, N represents the number of frequency points, ∠Si(x) represents the phase of the scattering parameters of the damaged state of the leaf at the i-th frequency, and ∠Si(0) represents the phase of the scattering parameters of the undamaged state of the leaf at the i-th frequency.

[0074] S140, based on the exponent of the mean absolute amplitude difference and the exponent of the mean absolute phase difference, the damage gap range of the preset region of the blade is obtained;

[0075] Specifically, by comparing the exponent of the mean absolute amplitude difference and the exponent of the mean absolute phase difference, the damage gap range of the preset region of the blade is obtained.

[0076] S150, using a preset classification regression algorithm to analyze the detection data to determine the type and number of damage gaps in the preset area of ​​the blade.

[0077] The preset classification algorithm can be the K-nearest neighbor algorithm. In particular, different classification and regression algorithms can be selected for calculation as needed, and no limitation is made here.

[0078] Specifically, the K-nearest neighbor algorithm is used to analyze the detection data to determine the type and number of damage gaps in the preset area of ​​the blade.

[0079] In this embodiment, a preset microwave sensor of a preset frequency is fixed at a preset distance above a preset area of ​​a wind turbine blade moving at a preset rotation speed, and measurement is started to acquire detection data of the preset area. Based on the amplitude of the scattering parameters of the blade damage state in the detection data, the amplitude of the scattering parameters of the blade damage state in the reference data is extracted, and the amplitude of the scattering parameters of the blade undamaged state in the reference data is obtained, thus obtaining the exponent of the average absolute amplitude difference between the blade damage state and the blade undamaged state in the preset area. Based on the phase of the scattering parameters of the blade damage state in the detection data, the phase of the scattering parameters of the blade damage state in the reference data is extracted, and the phase of the scattering parameters of the blade undamaged state in the reference data is obtained, thus obtaining the exponent of the average absolute phase difference between the blade damage state and the blade undamaged state in the preset area. Based on the exponent of the average absolute amplitude difference and the exponent of the average absolute phase difference, the damage gap range of the preset area of ​​the blade is obtained. A preset classification regression algorithm is used to analyze the detection data to determine the type and number of damage gaps in the preset area of ​​the blade. Through the above method, this application uses a microwave sensor for detection, utilizes the amplitude and oscillation parameters in the scattering parameters and uses a classification regression algorithm to detect and calculate the damage to the wind turbine blades, so that the damage detection results of the wind turbine blades are not affected by other factors, and the detection results are more accurate.

[0080] Furthermore, refer to Figure 5 In one embodiment of the wind turbine blade damage detection method of the present invention, step S110: fixing a preset microwave sensor of a preset frequency at a preset distance above a preset area of ​​a wind turbine blade moving at a preset rotational speed and starting measurement to obtain detection data of the preset area may specifically include:

[0081] S111, In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at the preset distance above the preset area of ​​the wind turbine blade moving at the preset speed and the measurement is started to obtain the simulated data of the preset area;

[0082] Specifically, in the preset simulation software, the preset microwave sensor of the preset frequency is fixed at a preset distance above the preset area of ​​the wind turbine blade moving at the preset rotation speed, and the measurement is started to obtain the simulated data of the preset area.

[0083] S112, change the size of the preset distance, change the size of the preset frequency under each different preset distance condition and extract the amplitude of the scattering parameter in the simulation data, compare the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameter in the simulation data to obtain the optimal preset frequency and the optimal preset distance.

[0084] S113, Fix the preset microwave sensor with the optimal preset frequency at the optimal distance position above the preset area of ​​the wind turbine blade moving at the preset speed and start measuring to obtain the detection data of the preset area.

[0085] Specifically, the preset distance is changed, and the preset frequency is changed under each different preset distance condition. The amplitude of the scattering parameters in the simulation data is extracted, and the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulation data is compared to obtain the optimal preset frequency and optimal preset distance. For example, as shown... Figure 6 As shown, the preset distance is set to 1mm, 2mm, 3mm, 4mm, and 5mm. Under each preset distance condition, the magnitude of the preset frequency is changed and the amplitude of the scattering parameter in the simulation data is extracted. The relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameter in the simulation data is compared, and the optimal frequency is found to be 24GHz and the optimal preset distance is 2mm. A 24GHz microwave sensor is fixed 2mm above the preset area of ​​the wind turbine blade moving at a preset speed and measurement is started to obtain the detection data of the preset area.

[0086] In this embodiment, a microwave sensor of a preset frequency is fixed at a preset distance above a preset region of a wind turbine blade moving at a preset rotational speed in a preset simulation software, and measurement is started to acquire simulated data of the preset region. The preset distance is changed, and the preset frequency is changed under each different preset distance condition. The amplitude of the scattering parameters in the simulated data is extracted. The relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulated data is compared to obtain the optimal preset frequency and optimal preset distance. Through this method, multiple loss states are simulated and changes in scattering parameters are detected, resulting in more comprehensive processing of the detection data and improving the speed and accuracy of detection data analysis and calculation.

[0087] Furthermore, in one embodiment of the wind turbine blade damage detection method of the present invention, step S112: changing the magnitude of the preset distance, changing the magnitude of the preset frequency under each different preset distance condition and extracting the amplitude of the scattering parameters in the simulation data, comparing the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulation data to obtain the optimal preset frequency and optimal preset distance, can be replaced by:

[0088] The preset distance is changed, the preset frequency is changed under each different preset distance condition, and the phase of the scattering parameters in the simulation data is extracted. The relationship between each preset frequency under each different preset distance condition and the phase of the scattering parameters in the simulation data is compared to obtain the optimal preset frequency and the optimal preset distance.

[0089] Specifically, by changing the magnitude of the preset distance, the magnitude of the preset frequency is changed under each different preset distance condition, and the phase of the scattering parameters in the simulation data is extracted. The relationship between each preset frequency under each different preset distance condition and the phase of the scattering parameters in the simulation data is compared to obtain the optimal preset frequency and the optimal preset distance. For example, as shown... Figure 6 As shown, the preset distance was set to 1mm, 2mm, 3mm, 4mm, and 5mm. Under each preset distance condition, the preset frequency was changed, and the phase of the scattering parameters in the simulated data was extracted. The relationship between each preset frequency under each different preset distance condition and the phase of the scattering parameters in the simulated data was compared, resulting in an optimal frequency of 24GHz and an optimal preset distance of 2mm. A 24GHz microwave sensor was fixed 2mm above a preset area on a wind turbine blade moving at a preset rotation speed, and measurements were taken to acquire detection data for the preset area. By selecting the phase of the scattering parameters for comparison in this way, the analysis of the preset area became more comprehensive, and the results more accurate.

[0090] Furthermore, in one embodiment of the wind turbine blade damage detection method of the present invention, step S112: changing the magnitude of the preset distance, changing the magnitude of the preset frequency under each different preset distance condition and extracting the amplitude of the scattering parameters in the simulation data, comparing the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulation data to obtain the optimal preset frequency and optimal preset distance, can be replaced by:

[0091] By changing the preset rotation speed, the scattering parameters in the simulation data at different preset rotation speeds are extracted and compared to obtain the optimal preset rotation speed.

[0092] Specifically, by changing the preset rotational speed, the scattering parameters in the simulated data at different preset rotational speeds are extracted and compared to obtain the optimal preset rotational speed. A microwave sensor with a preset frequency is fixed at a preset distance above a preset area of ​​the wind turbine blades moving at the optimal preset rotational speed, and measurement begins to acquire detection data for the preset area. This method enables the preset microwave sensor to accurately detect the preset area, improving the accuracy of the acquired detection data.

[0093] Furthermore, refer to Figure 8 In one embodiment of the wind turbine blade damage detection method of the present invention, step S120: extracting the amplitude of the scattering parameter of the blade damage state in the reference data based on the amplitude of the scattering parameter of the blade damage state in the detection data, and obtaining the amplitude of the scattering parameter of the blade undamaged state in the reference data, and obtaining the exponent of the average absolute amplitude difference between the blade damage state and the blade undamaged state in the preset region, may specifically include:

[0094] S121, In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at the preset distance above the preset area of ​​the wind turbine blade moving at the preset speed and the measurement is started. The damage condition of the preset area is changed, and the preset frequency is changed in the preset frequency range to obtain the resonant frequency and scattering parameters of the preset area as reference data.

[0095] S122, based on the amplitude of the scattering parameter of the leaf damage state in the detection data, extract the amplitude of the scattering parameter at multiple frequency points in the reference data that are the same as the leaf damage state, extract the amplitude of the scattering parameter at multiple frequency points in the undamaged state in the reference data, and obtain the exponent of the average absolute amplitude difference between the leaf damage state and the undamaged state in the preset region.

[0096] Specifically, in this embodiment, a microwave sensor of a preset frequency is fixed at a preset distance above a preset region of a wind turbine blade moving at a preset rotational speed in a preset simulation software, and measurement begins. The damage condition of the preset region is varied, and the preset frequency is changed within a preset frequency range to obtain the resonant frequency and scattering parameters of the preset region as reference data. Based on the amplitude of the scattering parameters of the blade damage state in the detection data, the amplitudes of the scattering parameters at multiple frequency points in the reference data that are the same as the blade damage state are extracted. The amplitudes of the scattering parameters at multiple frequency points in the reference data under the undamaged state are also extracted. The exponent of the average absolute amplitude difference between the blade damage state and the undamaged state in the preset region is obtained. Through this method, multiple damage conditions are simulated, and the exponent of the average absolute amplitude difference is selected, making the detection method more reasonable.

[0097] Furthermore, refer to Figure 9 In one embodiment of the wind turbine blade damage detection method of the present invention, step S130: extracting the phase of the scattering parameter of the blade damage state in the reference data based on the phase of the scattering parameter of the blade damage state in the detection data, and obtaining the phase of the scattering parameter of the blade undamaged state in the reference data, and obtaining the exponent of the average absolute phase difference between the blade damage state and the blade undamaged state in the preset region, may specifically include:

[0098] S131, In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at the preset distance above the preset area of ​​the wind turbine blade moving at the preset speed and the measurement is started. The damage condition of the preset area is changed, and the preset frequency is changed in the preset frequency range to obtain the resonant frequency and scattering parameters of the preset area as reference data.

[0099] S132, based on the phase of the scattering parameters of the leaf damage state in the detection data, extract the phase of the scattering parameters at multiple frequency points in the reference data that are the same as the leaf damage state, extract the phase of the scattering parameters at multiple frequency points in the reference data under the undamaged state, and obtain the exponent of the average absolute phase difference between the leaf damage state and the undamaged state in the preset region.

[0100] Specifically, in this embodiment, a microwave sensor of a preset frequency is fixed at a preset distance above a preset region of a wind turbine blade moving at a preset rotational speed in a preset simulation software, and measurement begins. The damage condition of the preset region is varied, and the resonant frequency and scattering parameters of the preset region are obtained as reference data by changing the preset frequency within a preset frequency range. Based on the phase of the scattering parameters of the blade damage state in the detection data, the phase of the scattering parameters at multiple frequency points in the reference data that are the same as the blade damage state is extracted. The phase of the scattering parameters at multiple frequency points in the reference data under the undamaged state is also extracted, and the exponent of the average absolute phase difference between the blade damage state and the undamaged state in the preset region is obtained. Through this method, multiple damage conditions are simulated, and the exponent of the average absolute phase difference is selected, making the detection method more reasonable.

[0101] Furthermore, refer to Figure 10 In one embodiment of the wind turbine blade damage detection method of the present invention, step S150: using a preset classification regression algorithm to analyze the detection data to determine the type and number of damage gaps in the preset region of the blade, may specifically include:

[0102] S151, Train the preset classification and regression algorithm to obtain the trained preset classification and regression algorithm;

[0103] The preset classification algorithm can be the K-nearest neighbor algorithm. In particular, different classification and regression algorithms can be selected for calculation as needed, and no limitation is made here.

[0104] Specifically, the K-nearest neighbor algorithm is trained to obtain a trained K-nearest neighbor algorithm.

[0105] S152, Input the scattering parameter magnitude of the leaf damage state in the detection data, and obtain the type and number of damage gaps in the preset region of the leaf according to the trained preset classification regression algorithm.

[0106] Specifically, in this embodiment, a trained preset classification and regression algorithm is obtained by training the preset classification and regression algorithm; the scattering parameters of the leaf damage state in the detection data are input, and the type and number of damage gaps in the preset region of the leaf are obtained according to the trained preset classification and regression algorithm. Through the above method, the leaf damage state is classified and regressed using the preset classification and regression algorithm, making the determination of the leaf damage state more accurate.

[0107] Furthermore, refer to Figure 11In one embodiment of the wind turbine blade damage detection method of the present invention, step S151: training the preset classification and regression algorithm to obtain the trained preset classification and regression algorithm may specifically include:

[0108] S1511, set the size of the scattering parameters of a first preset number of different known leaf damage states as a training dataset, the size of the scattering parameters of each known leaf damage state as a training sample, and label each training sample with a category label and store it in the training database;

[0109] Specifically, the size of the scattering parameters of n different known leaf damage states is set as the training dataset, and the size of the scattering parameters of each known leaf damage state is a training sample. Each training sample is labeled with a category label and stored in the training database.

[0110] S1512, set the size of the scattering parameters of the second preset number of unknown leaf damage states as the test dataset, and the size of the scattering parameters of each unknown leaf damage state is a test sample;

[0111] Specifically, the size of the scattering parameters of m unknown blade damage states is set as the test dataset, and the size of the scattering parameters of each unknown blade damage state is a test sample.

[0112] S1513, Each time a test sample is selected from the test dataset, the Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated, and the third preset number of training samples that are closest to the test sample in Euclidean distance from all training samples are selected as nearest neighbor samples.

[0113] Specifically, each time a test sample is selected from the test dataset, the Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated, and the K training samples that are closest to the test sample in terms of Euclidean distance from all training samples are selected as nearest neighbor samples.

[0114] S1514, based on the query pattern of the training samples, vote on the category labels of the third preset number of nearest neighbor samples to determine the category label of the test sample, label the test sample with category labels and provide the query pattern of the test sample, store the labeled test sample and the query pattern of the test sample into the training database and select the next test sample in the test dataset until the test dataset is selected, and obtain the trained preset classification and regression algorithm.

[0115] Specifically, based on the query pattern of the training samples, the category labels of the K nearest neighbor samples are voted on to determine the category label of the test sample. The category label of the test sample is labeled and the query pattern of the test sample is provided. The labeled test sample and the query pattern of the test sample are stored in the training database and the next test sample in the test dataset is selected. This process continues until the test dataset is selected, resulting in the trained preset classification and regression algorithm.

[0116] In this embodiment, a training dataset is set with the scattering parameters of a first preset number of different known leaf damage states, each known leaf damage state having its scattering parameter value as a training sample. Each training sample is labeled with a category and stored in a training database. A test dataset is set with the scattering parameters of a second preset number of unknown leaf damage states, each unknown leaf damage state having its scattering parameter value as a test sample. Each time, a test sample is selected from the test dataset. The Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated. A third preset number of training samples with the closest Euclidean distance to the test sample are selected as nearest neighbors. Based on the query pattern of the training samples, the category labels of the third preset number of nearest neighbors are voted on to determine the category label of the test sample. The test sample is then labeled with a category label and its query pattern is provided. The labeled test sample and its query pattern are stored in the training database. The next test sample is selected from the test dataset, and this process continues until the test dataset is fully selected, resulting in a trained preset classification and regression algorithm. This method improves the accuracy of the preset classification and regression algorithm.

[0117] Furthermore, refer to Figure 12 In one embodiment of the wind turbine blade damage detection method of the present invention, step S1513: selecting a test sample from the test dataset each time, calculating the Euclidean distance and query pattern between the test sample and all training samples in the training dataset, and selecting a third preset number of training samples from all training samples that are closest to the test sample in Euclidean distance as nearest neighbor samples, may specifically include:

[0118] S15131, Each time a test sample is selected from the test dataset, the Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated;

[0119] S15132, based on the increasing Euclidean distance between the test sample and all training samples in the training dataset, sort all training samples in the training dataset to obtain a list of the training dataset, and select the third preset number of training samples at the beginning of the list as nearest neighbor samples.

[0120] Specifically, in this embodiment, by selecting a test sample from the test dataset each time, calculating the Euclidean distance between the test sample and all training samples in the training dataset, and sorting all training samples in the training dataset according to the increasing Euclidean distance between the test sample and all training samples in the training dataset, a list of the training dataset is obtained. The third preset number of training samples at the top of the list are selected as nearest neighbor samples. Through this method, all training samples in the training dataset are sorted according to the increasing Euclidean distance between the test sample and all training samples in the training dataset, thereby improving the classification speed of the preset classification algorithm.

[0121] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0122] In one embodiment, refer to Figure 13 This application provides a device for detecting damage to wind turbine blades. The device for detecting damage to wind turbine blades includes:

[0123] The sensing module 1310 is used to divide a preset area on the wind turbine blade, emit microwave incident waves to the preset area and receive microwave reflected waves from the preset area, set the microwave incident waves and the microwave reflected waves as detection data and transmit them to the first analysis module, the second analysis module and the third calculation module.

[0124] The first analysis module 1320 is used to analyze the detection data, obtain the index of the average absolute amplitude difference of the preset area, and transmit it to the first calculation module.

[0125] The second analysis module 1330 is used to analyze the detection data, obtain the index of the average absolute phase difference of the preset area, and transmit it to the first calculation module.

[0126] The first calculation module 1340 is used to calculate the range of the damage gap of the wind turbine blades in the preset area based on the exponent of the average absolute amplitude difference and the exponent of the average absolute phase difference.

[0127] The second calculation module 1350 is used to calculate the type and number of damaged gaps in the wind turbine blades based on the preset classification regression calculation and the detection data.

[0128] The modules in the aforementioned wind turbine blade damage detection device correspond to the steps in the aforementioned wind turbine blade damage detection method embodiment, and their functions and implementation processes will not be described in detail here.

[0129] Each module in the aforementioned wind turbine blade damage detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0130] In one embodiment, this application provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of the wind turbine blade damage detection method described above.

[0131] The method implemented when the computer program is executed can be referred to in various embodiments of the wind turbine blade damage detection method of this application, and will not be repeated here.

[0132] It should be noted that, 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 system 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 system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0133] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0134] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0135] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for detecting damage to wind turbine blades, characterized in that, The method for detecting damage to the wind turbine blades includes: A microwave sensor with a preset frequency is fixed at a preset distance above a preset area of ​​a wind turbine blade moving at a preset speed, and measurement is started to obtain detection data of the preset area. Based on the amplitude of the scattering parameter of the leaf damage state in the detection data, the amplitude of the scattering parameter of the leaf damage state in the reference data is extracted, and the amplitude of the scattering parameter of the undamaged leaf state in the reference data is obtained, so as to obtain the exponent of the average absolute amplitude difference between the leaf damage state and the undamaged leaf state in the preset area. Based on the phase of the scattering parameters of the leaf damage state in the detection data, the phase of the scattering parameters of the leaf damage state in the reference data is extracted, and the phase of the scattering parameters of the undamaged leaf state in the reference data is obtained, so as to obtain the exponent of the average absolute phase difference between the leaf damage state and the undamaged leaf state in the preset region. The damage gap range of the preset region of the blade is obtained based on the exponent of the mean absolute amplitude difference and the exponent of the mean absolute phase difference. The type and number of damage gaps in the preset region of the blade are determined by analyzing the detection data using a preset classification regression algorithm. The step of extracting the phase of the scattering parameter of the leaf damage state in the reference data based on the phase of the scattering parameter of the leaf damage state in the detection data, and obtaining the phase of the scattering parameter of the undamaged leaf state in the reference data, and obtaining the exponent of the average absolute phase difference between the leaf damage state and the undamaged leaf state in the preset region, includes: In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at a preset distance above the preset area of ​​the wind turbine blade moving at the preset speed and the measurement is started. The damage condition of the preset area is changed, and the preset frequency is changed within the preset frequency range to obtain the resonant frequency and scattering parameters of the preset area as reference data. Based on the phase of the scattering parameters of the leaf damage state in the detection data, the phase of the scattering parameters at multiple frequency points in the reference data that are the same as the leaf damage state is extracted, and the phase of the scattering parameters at multiple frequency points in the undamaged state in the reference data is extracted, so as to obtain the exponent of the average absolute phase difference between the leaf damage state and the undamaged state in the preset region. The step of using a preset classification regression algorithm to analyze the detection data to determine the type and number of damage gaps in the preset region of the blade includes: Train the preset classification and regression algorithm to obtain the trained preset classification and regression algorithm; Input the scattering parameters of the leaf damage state in the detection data, and obtain the type and number of damage gaps in the preset region of the leaf according to the trained preset classification and regression algorithm.

2. The method for detecting wind turbine blade damage as described in claim 1, characterized in that, The step of fixing a preset microwave sensor of a preset frequency at a preset distance above a preset area of ​​a wind turbine blade moving at a preset rotation speed and starting measurement to obtain detection data of the preset area includes: In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at a preset distance above the preset area of ​​the wind turbine blade moving at the preset rotation speed, and the measurement is started to obtain the simulated data of the preset area; The preset distance is changed, the preset frequency is changed under each different preset distance condition, and the amplitude of the scattering parameter in the simulation data is extracted. The relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameter in the simulation data is compared to obtain the optimal preset frequency and the optimal preset distance. A microwave sensor with the optimal preset frequency is fixed at an optimal distance above a preset area of ​​a wind turbine blade moving at a preset speed, and measurement is started to obtain detection data of the preset area.

3. The method for detecting wind turbine blade damage as described in claim 2, characterized in that, The process involves changing the preset distance, altering the preset frequency under each different preset distance condition, extracting the amplitude of the scattering parameters from the simulation data, comparing the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulation data, and obtaining the optimal preset frequency and optimal preset distance. This is then replaced with: The preset distance is changed, the preset frequency is changed under each different preset distance condition, and the phase of the scattering parameters in the simulation data is extracted. The relationship between each preset frequency under each different preset distance condition and the phase of the scattering parameters in the simulation data is compared to obtain the optimal preset frequency and the optimal preset distance.

4. The method for detecting wind turbine blade damage as described in claim 2, characterized in that, The process involves changing the preset distance, altering the preset frequency under each different preset distance condition, extracting the amplitude of the scattering parameters from the simulation data, comparing the relationship between each preset frequency under each different preset distance condition and the amplitude of the scattering parameters in the simulation data, and obtaining the optimal preset frequency and optimal preset distance. This is then replaced with: By changing the preset rotation speed, the scattering parameters in the simulation data at different preset rotation speeds are extracted and compared to obtain the optimal preset rotation speed.

5. The method for detecting wind turbine blade damage as described in claim 1, characterized in that, The step of extracting the amplitude of the scattering parameter of the leaf damage state in the reference data based on the amplitude of the scattering parameter of the leaf damage state in the detection data, and obtaining the amplitude of the scattering parameter of the undamaged leaf state in the reference data, and obtaining the exponent of the average absolute amplitude difference between the leaf damage state and the undamaged leaf state in the preset region, includes: In the preset simulation software, the preset microwave sensor of the preset frequency is fixed at a preset distance above the preset area of ​​the wind turbine blade moving at the preset speed and the measurement is started. The damage condition of the preset area is changed, and the preset frequency is changed within the preset frequency range to obtain the resonant frequency and scattering parameters of the preset area as reference data. Based on the amplitude of the scattering parameters of the damaged state of the blade in the detection data, the amplitude of the scattering parameters at multiple frequency points in the reference data that are the same as the damaged state of the blade is extracted, and the amplitude of the scattering parameters at multiple frequency points in the undamaged state of the reference data is extracted, so as to obtain the exponent of the average absolute amplitude difference between the damaged state of the blade and the undamaged state of the blade in the preset region.

6. The method for detecting wind turbine blade damage as described in claim 1, characterized in that, The step of training the preset classification and regression algorithm to obtain the trained preset classification and regression algorithm includes: The size of the scattering parameters of a first preset number of different known leaf damage states is set as the training dataset, and the size of the scattering parameters of each known leaf damage state is a training sample. Each training sample is labeled with a category label and stored in the training database. The size of the scattering parameters of a second preset number of unknown leaf damage states is set as the test dataset, and the size of the scattering parameters of each unknown leaf damage state is a test sample. Each time, a test sample is selected from the test dataset, the Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated, and the third preset number of training samples that are closest to the test sample in Euclidean distance from all training samples are selected as nearest neighbor samples. Based on the query pattern of the training samples, the category labels of the third preset number of nearest neighbor samples are voted on to determine the category label of the test sample. The category label of the test sample is labeled and the query pattern of the test sample is provided. The labeled test sample and the query pattern of the test sample are stored in the training database and the next test sample in the test dataset is selected until the test dataset is selected, and the trained preset classification and regression algorithm is obtained.

7. The method for detecting wind turbine blade damage as described in claim 6, characterized in that, The step of selecting a test sample from the test dataset each time, calculating the Euclidean distance and query pattern between the test sample and all training samples in the training dataset, and selecting a third preset number of training samples from all training samples that are closest to the test sample in Euclidean distance as nearest neighbor samples includes: Each time, a test sample is selected from the test dataset, and the Euclidean distance and query pattern between the test sample and all training samples in the training dataset are calculated. Based on the increasing Euclidean distance between the test sample and all training samples in the training dataset, all training samples in the training dataset are sorted to obtain a list of the training dataset, and the third preset number of training samples at the beginning of the list are selected as nearest neighbor samples.

8. A device for detecting damage to wind turbine blades, characterized in that, The wind turbine blade damage detection device includes: The sensing module is used to divide a preset area on the wind turbine blade, emit microwave incident waves to the preset area and receive microwave reflected waves from the preset area, set the microwave incident waves and the microwave reflected waves as detection data and transmit them to the first analysis module, the second analysis module and the third calculation module. The first analysis module is used to analyze the detection data, and obtain the exponent of the average absolute amplitude difference of the preset area based on the amplitude of the scattering parameters of the leaf damage state in the detection data, the amplitude of the scattering parameters of the leaf damage state in the reference data, and the amplitude of the scattering parameters of the undamaged leaf state in the reference data, and transmit it to the first calculation module. The second analysis module is used to analyze the detection data, and based on the phase of the scattering parameters of the leaf damage state in the detection data, the phase of the scattering parameters of the leaf damage state in the reference data, and the phase of the scattering parameters of the undamaged state of the leaf in the reference data, obtain the exponent of the average absolute phase difference of the preset area and transmit it to the first calculation module. The first calculation module is used to calculate the range of the damage gap of the wind turbine blades in the preset area based on the exponent of the average absolute amplitude difference and the exponent of the average absolute phase difference. The second calculation module is used to calculate the type and number of damaged gaps in the wind turbine blades based on a preset classification regression algorithm and the detection data. The second analysis module is further configured to: simulate in preset simulation software fixing the sensing module of preset frequency at a preset distance above the preset area of ​​a wind turbine blade moving at a preset rotational speed and start measurement, change the damage condition of the preset area, change the preset frequency within a preset frequency range to obtain the resonant frequency and scattering parameters of the preset area as reference data; extract the phase of the scattering parameters of the blade damage state in the reference data at multiple frequency points with the same blade damage state according to the phase of the scattering parameters of the blade damage state in the detection data, extract the phase of the scattering parameters of the blade damage state at multiple frequency points in the reference data under the undamaged state, and obtain the exponent of the average absolute phase difference between the blade damage state and the blade undamaged state in the preset area; The second calculation module is also used to train the preset classification and regression algorithm to obtain the trained preset classification and regression algorithm; input the scattering parameter of the leaf damage state in the detection data, and obtain the type and number of damage gaps in the preset region of the leaf according to the trained preset classification and regression algorithm.

Images