A wind turbine transmission chain health assessment method based on wind speed and vibration coupling

Abstract

The present invention discloses a method for evaluating the health status of a wind turbine transmission chain based on wind speed vibration coupling, which specifically includes: installing vibration sensors on measuring points of the transmission chain to collect vibration data of each measuring point in real time; repairing the collected vibration data using a missing data reconstruction algorithm; performing bandpass filtering on the repaired vibration data of each measuring point according to a custom set standard; performing feature extraction on the filtered vibration data, calculating relevant vibration feature indicators, and obtaining feature values; judging the current operating status of the equipment according to the interval information of the speed, and if the speed is in a specific interval, amplifying the calculated feature value according to a preset amplification factor; judging the RMS value of each measuring point after amplification according to a reference value in the custom set standard, and if the calculated value is higher than the warning value and the danger value in the standard, performing time domain analysis and frequency domain analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of vibration signal testing and analysis, and mainly relates to a method for evaluating the health status of a wind turbine transmission chain based on wind speed vibration coupling. Background Art

[0002] Mechanical equipment vibration standards are primarily used for product quality assessment, operating status monitoring, and fault diagnosis. Vibration standards applicable to wind turbine drive train quality assessment and operating status monitoring are still under development worldwide. The standard established by VDI-3834 uses the "equivalent energy average" of vibration values ​​as a characteristic value for evaluation. This standard divides the vibration status of wind turbine drive train equipment into three zones:

[0003] "I" indicates that the equipment is normal and can operate for a long time;

[0004] "Zone II" indicates that the vibration is somewhat large and is not suitable for long-term operation. It is recommended to investigate and analyze the vibration source;

[0005] "Zone III" means the vibration is in a dangerous state and may cause damage to the unit or components;

[0006] Increasing the fan speed will result in more energy being input into the unit's transmission system, which in turn causes an increase in vibration amplitude. Therefore, the same mechanical component will exhibit different root mean square characteristic values ​​in different speed ranges. During operation, the data acquisition system will indiscriminately record vibration data in each speed range. However, under low wind speed conditions affected by seasonal factors, the system is more inclined to collect low-speed operation data. It should be pointed out that although mechanical failures may objectively exist, under low-speed conditions, the RMS value of the fault characteristic signal will show an amplitude attenuation phenomenon. This attenuation may cause it to fall within the normal threshold range specified by the VDI3834 standard, resulting in misjudgment or missed diagnosis of the fault, posing a potential safety hazard. Summary of the Invention

[0007] In response to the above shortcomings, the present invention proposes a method for evaluating the health status of a wind turbine transmission chain based on wind speed-vibration coupling. A data reconstruction algorithm is introduced during data preprocessing, and missing data is filled using cubic spline interpolation and LSTM time series prediction methods. Data is classified according to the speed range, and vibration data in different ranges are amplified by the corresponding speed factor. The custom-set standard is based on the vibration alarm value and danger value of VDI3834, and the RMS value of each measuring point after amplification is judged to evaluate the health status of the mechanical components of the wind turbine transmission chain. The specific steps are as follows:

[0008] S1. Install vibration sensors at the measuring points of the transmission chain to collect vibration data of each measuring point in real time;

[0009] S2. Repairing the collected vibration data using a missing data reconstruction algorithm;

[0010] S3. Perform bandpass filtering on the repaired vibration data of each measuring point according to the custom set standard;

[0011] S4. Extract features from the filtered vibration data, calculate relevant vibration feature indices, and obtain feature values;

[0012] S5. Determine the current operating status of the device based on the speed interval information. If the speed is within a specific interval, amplify the calculated characteristic value according to a preset amplification factor.

[0013] S6. Based on the reference value in the custom set standard, judge the RMS value of each measuring point after amplification. If the calculated value is higher than the warning value and danger value in the standard, perform time domain analysis and frequency domain analysis.

[0014] Furthermore, the missing data reconstruction algorithm specifically includes:

[0015] For the case of ≤5 consecutive missing sampling points, cubic spline interpolation was used;

[0016] In the case of missing more than 5 consecutive sampling points, time series prediction compensation based on long short-term memory network is activated.

[0017] Furthermore, the specific process of the bandpass filtering includes:

[0018] Use the designed bandpass filter to process the original signal through convolution operation or digital filtering algorithm;

[0019] In signal processing, discrete-time signal processing algorithms are used to filter signals; when applying bandpass filters, discrete Fourier transforms are used to convert signals from the time domain to the frequency domain, perform frequency limiting, and then convert them back to the time domain.

[0020] The specific design process of the designed bandpass filter is as follows:

[0021] Determine the low-frequency and high-frequency cutoff points of the bandpass filter based on the device's operating frequency and custom-set standards;

[0022] An infinite impulse response type filter is selected, wherein the infinite impulse response type filter adopts a Butterworth filter.

[0023] The transfer function H(s) of the Butterworth filter is expressed as:

[0024] ,

[0025] Where K represents the gain constant, s represents the complex frequency domain variable, Indicates the cutoff frequency of the low-pass filter.

[0026] Furthermore, the signal is filtered, and the specific filtering process is as follows:

[0027] Performing a fast Fourier transform on the original vibration signal to convert the original vibration signal into a frequency domain;

[0028] In the frequency domain, intercepting a signal between a low-frequency cutoff point and a high-frequency cutoff point to obtain a filtered frequency domain signal;

[0029] The filtered frequency domain signal is subjected to inverse fast Fourier transform to obtain the filtered signal in the time domain.

[0030] Furthermore, the step S5 specifically includes:

[0031] When the generator input shaft speed is greater than or equal to 1600 RPM, it is determined to be a high speed range, and the RMS value of the high speed range is multiplied by the high speed factor;

[0032] When the generator input shaft speed is greater than or equal to 1200 RPM and less than 1600 RPM, it is determined to be in the medium speed range, and the RMS value of the medium speed range is multiplied by the medium speed factor;

[0033] When the generator input shaft speed is lower than 1200 RPM, it is determined to be a low speed interval, and the RMS value of the low speed interval is multiplied by a low speed factor.

[0034] Furthermore, the time domain analysis observes the waveform characteristics of the vibration signal to determine whether there are obvious abnormal fluctuations or mutations, and then confirm potential problems with the equipment operation status; the frequency domain analysis converts the time domain signal into a spectrum through Fourier transform, analyzes the amplitude of each frequency component, and identifies whether there is an abnormal increase in a specific frequency.

[0035] According to a second aspect of the present invention, a computer program product is provided, on which one or more computer programs are stored. When the one or more computer programs are executed by a computer processor, the above method is implemented.

[0036] The above one or more technical solutions in the embodiments of the present application have at least one of the following technical effects:

[0037] 1) Accurate Speed ​​Adaptation Assessment: Breaking through the traditional single-threshold judgment model, differentiated amplification factors (1 / 1.2 / 1.6) are set for different speed ranges (high / medium / low). This design scientifically reflects the nonlinear characteristics of the impact of speed changes on vibration, particularly enhancing fault sensitivity in low-speed conditions (<1200 RPM). By dynamically adjusting the alarm threshold, it effectively addresses the lack of sensitivity of the general standard under variable speed conditions, ensuring that the assessment results are more consistent with the actual operating characteristics of the drive train.

[0038] 2) Adaptability to Variable Operating Conditions: In response to the frequent starts and stops and speed fluctuations of wind turbines, an innovative speed-vibration correlation model has been established. When the turbine is operating in the low to medium speed range, the detection sensitivity of the vibration signal is automatically increased (amplified by 20%-60%), effectively detecting early abnormalities in components such as gearboxes under non-rated operating conditions. This dynamic adjustment mechanism is more adaptable to complex environmental load variations than fixed threshold standards.

[0039] 3) Intelligent Missing Data Reconstruction: A hierarchical compensation strategy based on the duration of data missing is adopted. For short-term data missing (≤5 sampling points), a cubic spline interpolation algorithm is applied to ensure smooth waveform transitions. For long-term data missing (>5 sampling points), a long short-term memory (LSTM) network is used for time series prediction compensation. This method overcomes the reconstruction distortion problem of traditional single interpolation methods when processing continuous long-term missing data. By modeling the time series characteristics of the vibration signal through the LSTM network, the reconstruction accuracy of long-term missing data is improved. This intelligent hierarchical processing mechanism maximizes the preservation of the dynamic characteristics of the original vibration signal while ensuring data integrity, providing high-quality data input for subsequent vibration analysis based on the VDI3834 standard. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated into and constitute a part of this specification. The accompanying drawings illustrate the embodiments and, together with the description, serve to explain the principles of the present invention. Other embodiments and many of the expected advantages of the embodiments will be readily apparent as they become better understood by reference to the following detailed description. The elements of the drawings are not necessarily to scale with respect to each other. Like reference numerals designate corresponding similar parts.

[0041] Figure 1 A flow chart of a method for evaluating the health status of a wind turbine transmission chain based on wind speed-vibration coupling according to an embodiment of the present invention is shown.

[0042] Figure 2 It is a structural diagram of a computer system suitable for implementing the electronic device of the embodiment of the present application. DETAILED DESCRIPTION

[0043] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the relevant invention are shown in the accompanying drawings.

[0044] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0045] Figure 1 FIG. 1 is a flow chart of a method for evaluating the health status of a wind turbine transmission chain based on wind speed-vibration coupling according to an embodiment of the present invention. Figure 1 As shown:

[0046] S1. Install vibration sensors at the measuring points of the transmission chain to collect vibration data from each measuring point in real time. The vibration data should contain sufficient time domain information for subsequent analysis and processing.

[0047] S2. Repairing the collected vibration data using a missing data reconstruction algorithm;

[0048] S2.1. For the case where there are ≤5 consecutive missing sampling points, cubic spline interpolation is used. The standard form is:

[0049] ,

[0050] in, represents a known data point, is a constant term, which represents the offset of the polynomial, which is usually determined by the starting point of the interpolation. Represents the coefficient of the first term, which determines the linear rate of change of the polynomial in the interval. Represents the quadratic term coefficient, controls the curvature of the polynomial, and affects the acceleration of the curve. Represents the cubic coefficient, affects the curvature change of the polynomial, and determines the sharp change or turning point of the interpolation curve. It represents the starting point of the interval, that is, the horizontal coordinate of the known data point, and x is the independent variable, which represents the position to be interpolated.

[0051] For all data points, solve the cubic spline interpolation equation for the coefficients. Based on the interpolation conditions, form a system of equations and solve for each coefficient.

[0052] Once we have the coefficients of the cubic polynomial, we can use these equations to interpolate the missing data points.

[0053] The result obtained by interpolation will be smooth, which can ensure the continuity and smoothness of the signal.

[0054] S2.2. For the case where there are more than 5 consecutive missing sampling points, start the time series prediction compensation based on the long short-term memory (LSTM) network.

[0055] For consecutive missing sampling points, an LSTM model is used to predict these missing values. In order to train an LSTM model, it is usually necessary to convert the original time series data into a supervised learning problem, that is, to use the historical data of the time series to predict future data values.

[0056] Before training LSTM, it is usually necessary to standardize the input data (such as normalization) to improve the convergence speed and prediction accuracy of the model.

[0057] Furthermore, a deep learning framework such as Keras is used to build an LSTM model. The input is the past time step, and the output is the predicted value of the next time step. The data can be divided into multiple time windows, and the historical data in each window is used to predict the missing data outside the window. In a preferred implementation, the specific steps are as follows:

[0058] S2.2.1. Obtain a CSV file, where the data columns are time series data.

[0059] S2.2.2. Use MinMaxScaler to scale the data to the range [0, 1] to ensure that the data is suitable for LSTM model training.

[0060] S2.2.3. Use the create_dataset function to convert the raw data into a data format suitable for LSTM training by specifying the time_step. Each data sample X is the value at the past time_step, and the corresponding target y is the value at the next time step.

[0061] S2.2.4. Use the Sequential model to build the network structure: the first LSTM layer has 50 units, and return_sequences=True, indicating that the complete output sequence is returned, which is necessary for stacking multiple LSTM layers; the second LSTM layer does not return a sequence, so return_sequences=False; the last layer is a fully connected layer Dense, which outputs a value (that is, the predicted target value).

[0062] S2.2.5. Train the model using the Adam optimizer and the mean_squared_error loss function. The validation set loss is also calculated during training.

[0063] S2.2.6. Plot the loss changes during training and validation to help view the training effect of the model.

[0064] S2.2.7. After making predictions using the test set, plot a comparison chart of the predicted results and the actual results.

[0065] Furthermore, the processed time series data is input into the LSTM network for training, and the historical data and labels (the true values ​​corresponding to the missing data) are used to optimize the network weights.

[0066] After training the LSTM model, the model is used to predict the missing time steps and fill in the missing values.

[0067] S3. Perform bandpass filtering on the repaired vibration data of each measuring point according to the custom set standard;

[0068] The custom settings are based on the VDI 3834 standard, the vibration condition monitoring standard for wind turbines developed by the Verein Deutscher Ingenieure (German Engineers Association). VDI 3834 specifies the collection, analysis, and processing of vibration data, ensuring sufficient data quality and accuracy. According to VDI 3834, a bandpass filter is used to remove unnecessary high-frequency noise and low-frequency drift, retaining only the frequency information relevant to the equipment fault. This helps highlight the equipment's vibration characteristics and prevents noise or irrelevant signals from affecting the results in subsequent analysis.

[0069] Furthermore, the specific process of bandpass filtering includes:

[0070] Use the designed bandpass filter to process the original signal through convolution operation or digital filtering algorithm;

[0071] In signal processing, discrete-time signal processing algorithms are used to filter signals; when applying bandpass filters, discrete Fourier transforms are used to convert signals from the time domain to the frequency domain, perform frequency limiting, and then convert them back to the time domain.

[0072] The specific design process of the designed bandpass filter is as follows:

[0073] Determine the low-frequency and high-frequency cutoff points of the bandpass filter based on the device's operating frequency and custom-set standards;

[0074] An infinite impulse response type filter is selected, wherein the infinite impulse response type filter adopts a Butterworth filter, and the transfer function H(s) of the Butterworth filter is expressed as:

[0075] ,

[0076] Where K represents the gain constant, s represents the complex frequency domain variable, Indicates the cutoff frequency of the low-pass filter.

[0077] Common types of bandpass filters include IIR (Infinite Impulse Response) filters and FIR (Finite Impulse Response) filters. For filtering vibration signals, IIR filters are often chosen because of their high efficiency and ability to achieve good frequency selectivity.

[0078] IIR bandpass filters generally use double inversion filters, such as Butterworth filters and Chebyshev filters.

[0079] The Chebyshev filter allows a certain amount of ripple (that is, it allows some gain fluctuation within the passband) and can provide a steeper cutoff response. It is suitable for situations where higher filtering efficiency is required.

[0080] After filtering, examine the filtered signal to verify that it meets the expected frequency range and noise removal. Plot the spectrogram before and after filtering to see if sufficient signal energy is retained within the bandpass frequency band while removing noise from other frequency bands. Finally, output the bandpass-filtered vibration signal for subsequent feature extraction and analysis steps.

[0081] In a specific embodiment, the parameters of the bandpass filter are selected as follows:

[0082] Low-frequency cutoff frequency: Set as the upper limit of the low-frequency part of the noise in the vibration signal, usually set to the lowest frequency of the equipment operation or the vibration frequency of the system.

[0083] High-frequency cutoff frequency: Set as the lower limit of the high-frequency part of the noise in the vibration signal, usually set according to the highest operating frequency of the transmission system.

[0084] Filter Order: You can select the filter order based on the desired filtering effect. A higher order filter can achieve a steeper frequency response.

[0085] S4. Extract features from the filtered vibration data, calculate relevant vibration feature indices, and obtain feature values;

[0086] Calculate characteristic indicators of the filtered vibration signal, including the root mean square value (RMS), also known as the RMS value or effective value. This value is a measure of the signal's energy level, which is also the indicator used by VDI-3834.

[0087] Feature extraction includes time domain features, frequency domain features and time-frequency domain features. Among them, the time-frequency domain features combine the advantages of the time domain and frequency domain and are used to describe the change of the signal's frequency over time. Common methods include: short-time Fourier transform: divide the signal into short time windows and calculate the spectrum of each window; wavelet transform: a multi-scale time-frequency analysis method suitable for non-stationary signals; Hilbert-Huang transform: a time-frequency analysis method based on empirical mode decomposition.

[0088] S5. According to the speed interval information, the current operating state of the device is judged. If the speed is within a specific interval, the calculated characteristic value is amplified according to a preset amplification factor.

[0089] When the generator input shaft speed is greater than or equal to 1600 RPM, it is determined to be a high speed range, and the RMS value of the high speed range is multiplied by the high speed factor;

[0090] When the generator input shaft speed is greater than or equal to 1200 RPM and less than 1600 RPM, it is determined to be in the medium speed range, and the RMS value of the medium speed range is multiplied by the medium speed factor;

[0091] When the generator input shaft speed is lower than 1200 RPM, it is determined to be a low speed interval, and the RMS value of the low speed interval is multiplied by a low speed factor.

[0092] In addition to the RMS value, similar adjustments can be made to other characteristic values ​​(such as vibration amplitude, temperature, etc.). The specific steps are as follows:

[0093] Vibration amplitude adjustment: In different speed ranges, the sensitivity of the vibration amplitude may be different, and different adjustment factors need to be multiplied according to the speed range.

[0094] Temperature adjustment: If the speed is related to the device temperature, you can also set different temperature adjustment factors.

[0095] The operating status of the device is further judged based on the adjusted characteristic values, including:

[0096] Normal state: The characteristic value is within a certain range and no further adjustment is required.

[0097] Abnormal status: The characteristic value exceeds the preset threshold and maintenance or shutdown may be required.

[0098] The equipment monitoring system enables real-time detection and automatically adjusts characteristic values ​​based on speed information. The health status of the equipment is acquired and updated in real time at different speed ranges.

[0099] S6. Based on the reference values ​​in the custom set standard, the RMS values ​​of each measurement point after amplification are determined. If the calculated value is higher than the warning value and danger value in the standard, the next step of analysis is carried out, including time domain analysis and frequency domain analysis.

[0100] The RMS (root mean square) value of each measuring point is adjusted and judged based on the reference values ​​in the VDI3834 standard. If the calculated value after speed range adjustment exceeds the warning and danger values ​​defined in the standard, further detailed analysis is required. At this point, time domain analysis is first performed to observe the waveform characteristics of the vibration signal to determine whether there are obvious abnormal fluctuations or mutations, thereby confirming potential problems with the equipment's operating status. Secondly, frequency domain analysis is performed. The time domain signal is converted into a spectrum through Fourier transform, and the amplitude of each frequency component is analyzed to identify abnormal increases in specific frequencies. This helps diagnose whether the equipment has mechanical failures or problems such as imbalance, looseness, etc. Through dual analysis in the time and frequency domains, a more comprehensive assessment of the equipment's operating health can be achieved, providing support for subsequent maintenance decisions.

[0101] This process not only helps accurately locate faults but also provides early warning of potential risks, avoiding downtime and production losses caused by equipment failure. Furthermore, the analysis results can guide maintenance personnel to take appropriate repair measures, thereby improving equipment reliability and service life.

[0102] The present invention overcomes the difficulty of traditional vibration monitoring's insufficient sensitivity under variable speed conditions through a hierarchical amplification mechanism for speed intervals. This method sets differentiated amplification factors (1, 1.2, and 1.6) for three speed intervals: greater than or equal to 1600 RPM, greater than or equal to 1200 RPM but less than 1600 RPM, and less than 1200 RPM. In particular, by enhancing the detection sensitivity of low-speed conditions (amplification by 60%), it improves the ability to identify early faults in components such as gearboxes at non-rated speeds. This dynamic assessment strategy based on speed adaptation is more consistent with the actual operating characteristics of the transmission chain than the fixed threshold method.

[0103] Reference below Figure 2 , which shows a structural diagram of a computer system 200 suitable for implementing an electronic device of an embodiment of the present application. Figure 2 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0104] like Figure 2As shown, the computer system 200 includes a central processing unit (CPU) 201, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 202 or a program loaded from a storage unit 208 into a random access memory (RAM) 203. Various programs and data required for the operation of the system 200 are also stored in the RAM 203. The CPU 201, the ROM 202, and the RAM 203 are connected to each other via a bus 204. An input / output (I / O) interface 205 is also connected to the bus 204.

[0105] The following components are connected to the I / O interface 205: an input section 206 including a keyboard, a mouse, and the like; an output section 207 including a liquid crystal display (LCD) and speakers; a storage section 208 including a hard disk; and a communication section 209 including a network interface card such as a LAN card or a modem. The communication section 209 performs communication processing via a network such as the Internet. A drive 210 is also connected to the I / O interface 205 as needed. A removable medium 211, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 210 as needed, so that computer programs read therefrom can be installed into the storage section 208 as needed.

[0106] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 209, and / or installed from the removable medium 211. When the computer program is executed by the central processing unit (CPU) 201, the above-mentioned functions defined in the method of the present application are executed. It should be noted that the computer-readable storage medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0107] Computer program code for performing the operations of the present application can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0108] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0109] The modules described in the embodiments of the present application may be implemented by software or hardware.

[0110] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or it may exist independently and not be assembled into the electronic device. The computer-readable storage medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device: S1, installs a vibration sensor at a measuring point of the transmission chain to collect vibration data of each measuring point in real time; S2, repairs the collected vibration data using a missing data reconstruction algorithm; S3, performs bandpass filtering on the repaired vibration data of each measuring point according to a custom set standard; S4, performs feature extraction on the filtered vibration data, calculates relevant vibration feature indicators, and obtains a feature value; S5, judges the current operating status of the device based on the speed interval information, and if the speed is in a specific interval, amplifies the calculated feature value according to a preset amplification factor; S6, judges the RMS value of each measuring point after amplification based on the reference value in the custom set standard, and if the calculated value is higher than the warning value and danger value in the standard, performs time domain analysis and frequency domain analysis.

[0111] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned inventive concept. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for evaluating the health status of a wind turbine transmission chain based on wind speed and vibration coupling, characterized in that: include: S1. Install vibration sensors at the measuring points of the transmission chain to collect vibration data of each measuring point in real time; S2. Repair the collected vibration data using a missing data reconstruction algorithm. In the missing data reconstruction algorithm, if there are more than 5 consecutive missing sampling points, a time series prediction compensation based on a long short-term memory network is started. The specific construction steps include: Get a CSV file where the data columns are time series data; Use MinMaxScaler to scale the data to the range [0,1]; Use the create_dataset function to convert the original data into a data format suitable for long short-term memory network training by specifying time_step; each data sample X is the value at the past time_step moment, and the corresponding target y is the value at the next moment; Use the Sequential model to build the network structure: the first layer of the long short-term memory network has 50 units, and return_sequences = True, indicating that the complete output sequence is returned; the second layer of the long short-term memory network does not return a sequence, return_sequences = False; the last layer is a fully connected layer Dense, which outputs a predicted target value; The model is trained using the Adam optimizer and the mean_squared_error loss function, and the loss of the validation set is calculated during training. S3. According to the custom set standard, the vibration data after repair at each measuring point is subjected to band-pass filtering. The band-pass filtering adopts a Butterworth filter, and the transfer function H(s) is expressed as: Where K represents the gain constant, s represents the complex frequency domain variable, ω c Indicates the cutoff frequency of the low-pass filter; S4. Extract features from the filtered vibration data, calculate relevant vibration feature indices, and obtain feature values; S5. Judging the current operating state of the device based on the speed range information. If the speed is within a specific range, amplifying the calculated characteristic value according to a preset amplification factor. Specifically, the following steps are included: When the generator input shaft speed is greater than or equal to 1600 RPM, it is determined to be a high speed range, and the RMS value of the high speed range is multiplied by the high speed factor; When the generator input shaft speed is greater than or equal to 1200 RPM and less than 1600 RPM, it is determined to be in the medium speed range, and the RMS value of the medium speed range is multiplied by the medium speed factor; When the generator input shaft speed is lower than 1200 RPM, it is determined to be in the low speed range, and the RMS value of the low speed range is multiplied by the low speed factor; S6. Based on the reference value in the custom set standard, judge the RMS value of each measuring point after amplification. If the calculated value is higher than the warning value and danger value in the standard, perform time domain analysis and frequency domain analysis.

2. The method according to claim 1, characterized in that The missing data reconstruction algorithm specifically includes: For the case of ≤5 consecutive missing sampling points, cubic spline interpolation was used; In the case of missing more than 5 consecutive sampling points, time series prediction compensation based on long short-term memory network is activated.

3. The method according to claim 1, characterized in that The specific process of the bandpass filtering process includes: Use the designed bandpass filter to process the original signal through convolution operation or digital filtering algorithm; In signal processing, discrete-time signal processing algorithms are used to filter signals; when applying bandpass filters, discrete Fourier transforms are used to convert signals from the time domain to the frequency domain, perform frequency limiting, and then convert them back to the time domain.

4. The method according to claim 3, characterized in that The specific design process of the designed bandpass filter is as follows: Determine the low-frequency and high-frequency cutoff points of the bandpass filter based on the device's operating frequency and custom-set standards; An infinite impulse response type filter is selected, wherein the infinite impulse response type filter adopts a Butterworth filter.

5. The method according to claim 3, characterized in that The signal is filtered, and the specific filtering process is as follows: Performing a fast Fourier transform on the original vibration signal to convert the original vibration signal into a frequency domain; In the frequency domain, intercepting a signal between a low-frequency cutoff point and a high-frequency cutoff point to obtain a filtered frequency domain signal; The filtered frequency domain signal is subjected to inverse fast Fourier transform to obtain the filtered signal in the time domain.

6. The method according to claim 1, wherein The time domain analysis observes the waveform characteristics of the vibration signal to determine whether there are abnormal fluctuations or mutations, and then confirms potential problems with the equipment's operating status. The frequency domain analysis converts the time domain signal into a spectrum through Fourier transform, analyzes the amplitude of each frequency component, and identifies whether there is an abnormal increase in a specific frequency.

7. A computer program product, characterized in that A computer program is stored thereon, which implements the method according to any one of claims 1 to 6 when executed by a processor.

8. A computing system, characterized in that: The method comprises a processor and a memory, wherein the processor is configured to execute the method according to any one of claims 1 to 6.

Images