Real-time monitoring method and device for periodic abnormal fluctuation state of wind turbine generator

Abstract

This invention discloses a method and apparatus for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine. The method includes the following steps: real-time acquisition of steady-state state signal data frames of the wind turbine; calculation of the wind turbine rotor frequency based on the real-time acquired steady-state state signal data frames; extraction of peak sequences from the current steady-state speed data frames, and extraction of fluctuation peak features from the peak sequences; if the fluctuation peak features exceed a preset threshold, extraction of the fluctuation frequency of the turbine state signal and the sliding window mean of the autocorrelation waveform peak sequence from the current steady-state speed data frames; and determination of whether there is a fluctuation of n times the rotor frequency based on the wind turbine rotor frequency, the fluctuation frequency of the turbine state signal, and the sliding window mean of the autocorrelation waveform peak sequence. This invention has the advantages of simple implementation, low cost, high monitoring efficiency, high accuracy, and strong robustness.

Description

Technical Field

[0001] This invention relates to the field of wind turbine monitoring technology, and in particular to a method and device for real-time monitoring of the periodic abnormal fluctuations in the state of wind turbines. Background Technology

[0002] During operation, wind turbine status signals may exhibit periodic abnormal fluctuations, including those of the rotor frequency and the frequencies of components such as the tower, blades, and drivetrain—a phenomenon known as periodic oscillations. For example, during wind turbine operation, status signals such as speed, torque, and nacelle vibration acceleration may exhibit periodic fluctuations of n times the rotor frequency (n = 1, 2, 3, 6, 9, etc.). Similarly, speed and torque signals may contain periodic fluctuations in the frequencies of components such as the tower, blades, and drivetrain. These periodic abnormal fluctuations can negatively impact the wind turbine's power generation performance, fatigue load, overall vibration, and operational stability. Therefore, real-time monitoring of these periodic abnormal fluctuations within the wind turbine is crucial to ensure its stable and reliable operation.

[0003] To address the periodic abnormal fluctuations within wind turbine units, existing technologies typically employ spectrum analysis methods, such as using spectrum analysis to extract the 1x rotational frequency characteristics of the wind turbine. However, the accuracy of directly using spectrum analysis methods is not high, it is difficult to ensure real-time monitoring, and it is also easily affected by environmental noise. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method and device for real-time monitoring of the periodic abnormal fluctuation state of wind turbine units, which is simple to implement, low in cost, efficient and accurate, and robust.

[0005] To solve the above-mentioned technical problems, the technical solution proposed by this invention is as follows:

[0006] A method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator, comprising the following steps:

[0007] Real-time acquisition of steady-state state signal data frames of wind turbine units, wherein the steady-state state signal data frames include any one or more combinations of the following: rotational speed, torque, nacelle vibration acceleration data, and pitch angle.

[0008] The wind turbine rotor frequency is calculated based on the real-time collected steady-state signal data frames.

[0009] Extract the peak sequence from the current steady-state speed data frame, and extract the fluctuation peak features from the peak sequence;

[0010] If the fluctuation peak characteristics exceed a preset threshold, the fluctuation frequency of the unit status signal and the sliding window mean of the autocorrelation waveform peak sequence are extracted from the current steady-state speed data frame.

[0011] The presence of wind turbine rotor frequency fluctuation (n times the rotor frequency) is determined based on the wind turbine rotor rotation frequency, the fluctuation frequency of the turbine status signal, and the sliding window mean of the autocorrelation waveform peak sequence, where n is a positive integer.

[0012] Furthermore, the periodic abnormal fluctuation state also includes any one or more of the following: tower frequency fluctuation, transmission chain frequency fluctuation, and blade characteristic frequency fluctuation.

[0013] Furthermore, before calculating the wind turbine rotor frequency based on the real-time collected steady-state speed data frame, the method further includes using a Butterworth bandpass filter to perform bandpass filtering on the real-time collected steady-state speed data frame to remove low-frequency trend terms and high-frequency noise signals from the steady-state signal data frame of the wind turbine.

[0014] Furthermore, the fluctuation peak characteristic is the proportion of peaks exceeding a preset threshold in the peak sequence. When the peak proportion is less than the preset proportion threshold, the current monitoring ends and the execution of the real-time acquisition of the steady-state state signal data frame of the wind turbine is returned. Otherwise, the fluctuation frequency of the unit state signal and the sliding window mean of the autocorrelation waveform peak sequence are extracted from the current steady-state speed data frame.

[0015] Furthermore, the process of extracting the fluctuation frequency of the unit status signal from the current steady-state speed data frame includes:

[0016] Calculate the autocorrelation sequence of the current steady-state speed data frame;

[0017] Extract the peak sequence of the autocorrelation sequence;

[0018] The fluctuation frequency of the unit status signal is obtained by calculating the mean time difference of the peak sequence of the autocorrelation sequence.

[0019] Further, determining whether there is a wind turbine frequency fluctuation of n times based on the wind turbine rotor frequency, the fluctuation frequency of the unit status signal, and the sliding window mean of the autocorrelation waveform peak sequence includes:

[0020] Determine whether the fluctuation frequency of the unit status signal is within a specified frequency range, wherein the specified frequency range is determined based on n times the rotational frequency of the unit's wind turbine;

[0021] If it is determined that the fluctuation frequency of the unit status signal is not within the specified frequency range, then the current monitoring ends and the process returns to executing the real-time acquisition of the wind turbine's steady-state status signal data frame.

[0022] If the fluctuation frequency of the unit status signal is determined to be within the specified frequency range, the sliding window mean of the autocorrelation waveform peak sequence is determined. If the sliding window mean of the autocorrelation waveform peak sequence is greater than the preset sliding window mean threshold, it is determined that there is a wind turbine frequency fluctuation of n times, and the corresponding early warning is triggered.

[0023] Furthermore, the specified frequency range is f p ·(1-k)~f p ·(1+k), f p The frequency is n times the rotational frequency of the wind turbine, and k is the set value of the relative frequency deviation, where 1>k>0.

[0024] A real-time monitoring device for periodic abnormal fluctuations in the state of a wind turbine generator includes:

[0025] The data acquisition module is used to acquire steady-state state signal data frames of wind turbines in real time.

[0026] The frequency parameter calculation module is used to calculate the wind turbine rotor frequency based on the real-time collected steady-state signal data frames.

[0027] The fluctuation peak feature extraction module is used to extract the peak sequence from the current steady-state speed data frame and extract the fluctuation peak features from the peak sequence;

[0028] The fluctuation frequency feature extraction module is used to extract the fluctuation frequency of the unit status signal and the sliding window mean of the autocorrelation waveform peak sequence from the current steady-state speed data frame if the fluctuation peak feature exceeds a preset threshold.

[0029] The anomaly detection and early warning module is used to determine whether an early warning is needed for the frequency fluctuation of the current component under test based on the wind turbine rotor frequency, the fluctuation frequency of the unit status signal, and the sliding window mean of the autocorrelation waveform peak sequence.

[0030] An electronic device includes a processor and a memory, the memory being used to store a computer program, and the processor being used to execute the computer program to perform the method described above.

[0031] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0032] Compared with existing technologies, the advantages of this invention are as follows: This invention acquires steady-state state signal data frames of wind turbine units in real time, extracts fluctuation peak features from the steady-state speed data frames, extracts the fluctuation frequency of the unit's state signal and the sliding window mean of the autocorrelation waveform peak sequence based on the fluctuation peak features, and determines whether there is a wind turbine frequency fluctuation of n times the rotational frequency based on the wind turbine rotor frequency, the fluctuation frequency of the unit's state signal, and the sliding window mean of the autocorrelation waveform peak sequence. It can comprehensively mine the characteristics of periodic abnormal fluctuations within the wind turbine unit by combining fluctuation peak features, state signal fluctuation frequency, and the sliding window mean of the state signal autocorrelation waveform peak sequence. It can accurately identify state signals with specified fluctuation frequencies and specific fluctuation amplitude ranges, thereby quickly and accurately realizing real-time monitoring of periodic abnormal fluctuation states such as wind turbine frequency fluctuation of n times the rotational frequency, effectively improving the monitoring accuracy and robustness, and also avoiding complex calculations, reducing data processing volume, and improving monitoring and processing efficiency. Attached Figure Description

[0033] Figure 1 This is a schematic diagram illustrating the implementation process of the real-time monitoring method for periodic abnormal fluctuations in the wind turbine generator state in this embodiment.

[0034] Figure 2 This is a flowchart illustrating the process of real-time monitoring of the periodic abnormal fluctuations in the state of wind turbine generators in a specific application embodiment of the present invention.

[0035] Figure 3 This is a schematic diagram of the structural principle of the real-time monitoring device for periodic abnormal fluctuations in the wind turbine generator set in this embodiment. Detailed Implementation

[0036] The present invention will be further described below with reference to the accompanying drawings and specific preferred embodiments, but this does not limit the scope of protection of the present invention.

[0037] like Figure 1 As shown, the steps of the real-time monitoring method for periodic abnormal fluctuations in the wind turbine state in this embodiment include:

[0038] Step S01. Real-time acquisition of steady-state state signal data frames of wind turbine units, including steady-state speed, torque, nacelle vibration acceleration data, and pitch angle, etc.

[0039] The steady-state signal is the state signal of the wind turbine when it reaches a steady state of operation. This state includes parameters such as rotational speed, torque, nacelle vibration acceleration, and pitch angle. This embodiment collects real-time data from the wind turbine when it is in a steady state. Specifically, this data can be any one or a combination of two or more of the following: rotational speed, torque, nacelle vibration acceleration, and pitch angle. Data is then extracted into data frames with a specified window width, for example, a window width of W. This allows for monitoring and early warning based on data frames with a window width of W.

[0040] In a specific application embodiment, the data frame window width W required for a single early warning can be configured to be greater than 10 / f. p This allows for shorter data duration and less computational load required for a single early warning, while also enabling real-time early warning.

[0041] Step S02. Calculate the wind turbine rotor frequency based on the real-time collected steady-state signal data frames.

[0042] In this embodiment, the wind turbine's n-fold rotational frequency is calculated based on steady-state rotational speed data, where n is a positive integer, such as 1, 2, 3, 6, 9, etc., enabling real-time monitoring of the wind turbine's n-fold rotational frequency fluctuation. For example, the wind turbine's n-fold rotational frequency f can be calculated based on the generator speed data and gearbox speed ratio in the data frame. p Then, through subsequent steps S03 to S05, an early warning of abnormal fluctuations in the wind turbine's n-fold rotation frequency is achieved.

[0043] Furthermore, it can acquire frequency values ​​such as tower frequency, transmission chain frequency, and blade characteristic frequency to monitor periodic abnormal fluctuations of specific frequencies such as tower frequency fluctuations, transmission chain frequency fluctuations, and blade characteristic frequency fluctuations.

[0044] Furthermore, before calculating the wind turbine rotor frequency based on the real-time acquired steady-state speed data frames, this embodiment also includes bandpass filtering of the real-time acquired steady-state speed data frames using a Butterworth bandpass filter to remove low-frequency trend terms and high-frequency noise signals from the steady-state state signal data frames of the wind turbine. Specifically, for a data frame with a window width of W, the Butterworth bandpass filter removes the low-frequency trend terms and high-frequency noise from the wind turbine state signal to obtain the filtered signal yy_detrend. By using a Butterworth filter, the signal amplitude within the passband can be better restored, reducing the impact of the filter on the amplitude distortion of the filtered signal.

[0045] The transfer function of a Butterworth filter can be expressed as:

[0046]

[0047] In the formula,

[0048] It is understandable that, in addition to Butterworth filters, other types of filters, such as Bessel filters which have good signal amplitude reproduction in the passband, can also be used in the bandpass filtering process of wind turbine status signals, depending on actual needs.

[0049] Step S03. Extract the peak sequence from the current steady-state speed data frame, and extract the fluctuation peak features from the peak sequence.

[0050] For the steady-state speed data frame signal yy_detrend after bandpass filtering, the peak sequence G of the signal yy_detrend can be extracted by peak detection. pk Then, the fluctuation peak features are extracted from the peak sequence.

[0051] In this embodiment, the peak fluctuation characteristic can be the proportion of peaks exceeding a preset threshold in the peak sequence. This peak proportion can be used to effectively characterize the state of the peak fluctuation. When the peak proportion is less than the preset proportion threshold, the current monitoring ends and the process returns to the execution of real-time acquisition of the wind turbine's steady-state state signal data frame. Otherwise, the fluctuation frequency of the unit's state signal and the sliding window mean of the autocorrelation waveform peak sequence are extracted from the current steady-state speed data frame.

[0052] Specifically, the peak sequence G is calculated. pk The value exceeds the preset threshold G pk＿thresh Peak percentage P k If the peak percentage P k Less than the preset percentage threshold P k_thres If the peak percentage P is not reached, the monitoring process ends and step S01 is executed again; k Greater than the preset percentage threshold P k_thres If so, proceed to the next step S04.

[0053] Understandably, other peak-related features can also be used for fluctuation peak characteristics according to actual needs, such as the sliding window mean of the peak sequence.

[0054] Step S04. If the fluctuation peak characteristics exceed the preset threshold, extract the fluctuation frequency of the unit status signal and the sliding window mean of the autocorrelation waveform peak sequence from the current steady-state speed data frame.

[0055] The fluctuation peak characteristics can reflect the abnormal state of the fluctuation peak. If the fluctuation peak characteristics do not exceed the preset threshold, it indicates that there is no abnormal fluctuation at present, and the current monitoring can be ended and the process can return to step S01 to continue the next monitoring. If the fluctuation peak characteristics exceed the preset threshold, it indicates that there may be abnormal fluctuation. Then, the fluctuation frequency of the unit status signal is further extracted from the current steady-state speed data frame, and the sliding window mean of the autocorrelation waveform peak sequence is calculated. The sliding window mean of the autocorrelation waveform peak sequence is the peak sequence extracted from the autocorrelation sequence of the steady-state speed data frame, and then the sliding window mean of the peak sequence is calculated.

[0056] In this embodiment, the step of extracting the fluctuation frequency of the unit status signal from the current steady-state speed data frame specifically includes:

[0057] Step S401. Calculate the autocorrelation sequence of the current steady-state speed data frame;

[0058] Step S402. Extract the peak sequence from the autocorrelation sequence;

[0059] Step S403. Calculate the mean time difference of the peak sequence of the autocorrelation sequence to obtain the fluctuation frequency of the unit status signal.

[0060] Specifically, in order to extract the fluctuation frequency of the unit status signal, this embodiment first calculates the autocorrelation sequence R of the steady-state speed data frame signal yy_detrend after bandpass filtering. corr Then, R is extracted through peak detection. corr Peak sequence R pk Finally, the peak sequence R is calculated. pk Time difference mean T t The fluctuation frequency f of the unit status signal is obtained. s Furthermore, the peak sequence R is calculated. pk sliding window mean R pk＿roling Then the fluctuation frequency f of the unit status signal can be obtained. s and peak sequence R pk sliding window mean R pk＿roling .

[0061] The expression for calculating the autocorrelation above is:

[0062]

[0063] In the formula, N is the number of data points, and τ is the time-shift variable.

[0064] Step S05. Determine whether there is a wind turbine frequency fluctuation of n times the rotational frequency based on the wind turbine rotor frequency, the fluctuation frequency of the unit status signal, and the sliding window mean of the autocorrelation waveform peak sequence, where n is a positive integer.

[0065] Step S501. Determine whether the fluctuation frequency of the unit status signal is within the specified frequency range, which is determined based on the n times rotation frequency of the unit's wind turbine;

[0066] Step S502. If it is determined that the fluctuation frequency of the unit status signal is not within the specified frequency range, then end the current monitoring and return to the execution of real-time acquisition of steady-state status signal data frames of the wind turbine unit;

[0067] Step S503. If it is determined that the fluctuation frequency of the unit status signal is within the specified frequency range, determine the sliding window mean of the autocorrelation waveform peak sequence. If the sliding window mean of the autocorrelation waveform peak sequence is greater than the preset sliding window mean threshold, it is determined that there is a wind turbine frequency fluctuation of n times, and the corresponding early warning is triggered.

[0068] Taking the monitoring of abnormal fluctuations in the nth multiple of the wind turbine's rotational frequency as an example, the specified frequency range can be the key frequency range to be monitored, for example, it can be represented as f p ·(1-k)~f p ·(1+k), where f p Let f be the frequency of the wind turbine, n times its rotational speed, and k be the frequency relative deviation setpoint, where 1 > k > 0. If the fluctuation frequency f of the unit status signal... s Not in f p ·(1-k)~f p If the frequency is between (1+k), the current monitoring process ends, and step S01 is re-executed. The fluctuation frequency f of the unit status signal is... s At f p ·(1-k)~f p Between (1+k), continue to determine the autocorrelation waveform peak sequence R. pk sliding window mean R pk＿roling Size, if R pk＿roling Less than threshold R thres Then the current monitoring process ends, and step S01 is re-executed; if R pk＿roling Greater than the threshold R thres This will trigger an early warning for wind turbine frequency fluctuations up to n times.

[0069] Correspondingly, the frequency ranges for the drive train frequency and blade characteristic frequency are determined in a similar manner to those for the n-times multiple rotational frequency of the wind turbine. Taking the drive train frequency as an example, the drive train frequency f... t The range of values ​​can be represented as f t (1-k)~f t(1+k), where k is a fixed value (e.g., 5%), and can be fine-tuned according to the actual machine design parameters. If the real-time drive train frequency is not at f... t (1-k)~f t If the frequency fluctuation of the transmission chain is within (1+k), it is determined that there is a fluctuation in the transmission chain frequency.

[0070] This embodiment first calculates the wind turbine rotor frequency by real-time acquisition of steady-state state signal data frames from the wind turbine. Then, it extracts the fluctuation peak features from the steady-state speed data frames. Based on the fluctuation peak features, it further extracts the fluctuation frequency of the unit's state signal and the sliding window mean of the autocorrelation waveform peak sequence. Finally, it determines whether there is a fluctuation of n times the rotor frequency based on the wind turbine rotor frequency, the fluctuation frequency of the unit's state signal, and the sliding window mean of the autocorrelation waveform peak sequence. It can comprehensively extract the characteristics of periodic abnormal fluctuations within the wind turbine by combining the fluctuation peak features, the fluctuation frequency of the state signal, and the sliding window mean of the autocorrelation waveform peak sequence of the state signal. It can accurately identify state signals with specified fluctuation frequencies and specific fluctuation amplitude ranges, thereby quickly and accurately realizing real-time monitoring of periodic abnormal fluctuation states such as wind turbine n times frequency fluctuations. This effectively improves the monitoring accuracy and robustness, avoids complex calculations, reduces data processing volume, and improves monitoring and processing efficiency.

[0071] The following example, using the method described above to achieve real-time monitoring of wind turbine rotor frequency fluctuations, further illustrates the present invention.

[0072] Rotor imbalance is a typical anomaly in wind turbines, impacting their power generation performance, fatigue load, overall vibration, and operational stability. Rotor imbalance includes two main modes: aerodynamic imbalance and mass imbalance. Aerodynamic imbalance is primarily caused by deviations in the aerodynamic performance of the three blades, while mass imbalance is mainly caused by uneven mass distribution within the rotor (including blades) in the rotor's plane of rotation. Common causes of rotor imbalance include blade icing, external blade damage, blade tip breakage, blade installation angle deviation, and large blade mass moment deviation. For wind turbines with rotor imbalance, periodic fluctuations at one times the rotor's rotational frequency are typically observed in their speed and torque status signals. Since wind turbine speed and torque are nonlinear and non-stationary signals, extracting the rotor's one-time rotational frequency characteristic using traditional spectral analysis methods is insufficient to guarantee real-time monitoring.

[0073] This embodiment first monitors and identifies data of the wind turbine at steady-state speed, extracts data frames with a window width of W, and calculates the wind turbine's 1x rotational speed based on the generator speed data and gearbox ratio in the data frames. For data frames with a window width of W, a Butterworth bandpass filter is used to remove low-frequency trend terms and high-frequency noise from the wind turbine's state signal to obtain a bandpass filtered signal. Then, the peak sequence of the bandpass filtered signal is extracted by peak detection, and the proportion of peaks exceeding the peak threshold in the peak sequence is calculated. The autocorrelation waveform of the bandpass filtered signal is obtained through autocorrelation analysis. The peak sequence of the autocorrelation waveform and the fluctuation frequency of the state signal are obtained based on peak detection and differential calculation. Then, a sliding window mean filter is applied to the peak sequence of the autocorrelation waveform. Finally, the wind turbine's 1x rotational speed fluctuation early warning is determined by comprehensively considering the proportion of state signal fluctuation peaks exceeding the threshold, the state signal fluctuation frequency, and the sliding window mean of the autocorrelation waveform peak sequence, and the monitoring results are output.

[0074] like Figure 2 As shown, the detailed steps for real-time monitoring of wind turbine rotor frequency fluctuations (1x speed fluctuation) in this embodiment are as follows:

[0075] Step 1: Monitor and identify data on the wind turbine's steady-state speed, and issue early warnings in data frames with a window width of W. Calculate the wind turbine's rotational frequency f (1 times the rotational speed) based on the generator speed data and gearbox ratio in the data frames. p .

[0076] Step 2: For a data frame with a window width of W, a Butterworth bandpass filter is used to remove the low-frequency trend term and high-frequency noise from the wind turbine state signal to obtain the signal yy_detrend. The peak sequence G of yy_detrend is then extracted using peak detection. pk .

[0077] Step 3: Calculate G pk Exceeding the threshold G pk＿thresh Peak percentage P k If P k Less than threshold P k＿thres If P k Greater than threshold P k_thres If so, proceed to step 4.

[0078] Step 4: Calculate the autocorrelation sequence R of the signal yy_detrend corr R is extracted by peak detection corr Peak sequence R pk And calculate R pk Time difference mean T of the sequence t The fluctuation frequency f of the unit status signal is obtained. s .

[0079] Step 55: If the fluctuation frequency f of the unit status signal s At f p ·(1-k)~f p If the value is between (1+k), continue with step 6; otherwise, the monitoring process ends and step 1 is executed again.

[0080] Step 6: Calculate the peak sequence R pk sliding window mean R pk＿roling If R pk＿roling Less than threshold R thres The monitoring process ends, and step 1 is executed again. If R pk＿roling Greater than the threshold R thres This will trigger a warning for wind turbine frequency fluctuation of 1x.

[0081] This embodiment uses the above method to efficiently and accurately monitor the 1x frequency fluctuation of the wind turbine's status signal, thereby enabling real-time monitoring of wind turbine rotor imbalance and its development trend, and avoiding adverse effects on the unit's power generation performance, fatigue load, overall vibration, operational stability, and operational safety caused by wind turbine rotor imbalance.

[0082] like Figure 3 As shown, the real-time monitoring device for the periodic abnormal fluctuation state of wind turbine units in this embodiment includes:

[0083] The data acquisition module is used to acquire steady-state state signal data frames of the wind turbine in real time. The steady-state state signal data frames include any one or a combination of two or more of the following: speed, torque, nacelle vibration acceleration data and pitch angle under steady-state conditions.

[0084] The frequency parameter calculation module is used to calculate the wind turbine rotor frequency based on the real-time collected steady-state signal data frames.

[0085] The fluctuation peak feature extraction module is used to extract the peak sequence from the current steady-state speed data frame and extract the fluctuation peak features from the peak sequence;

[0086] The fluctuation frequency feature extraction module is used to extract the fluctuation frequency of the unit status signal and the sliding window mean of the autocorrelation waveform peak sequence from the current steady-state speed data frame if the fluctuation peak feature exceeds a preset threshold.

[0087] The anomaly detection and early warning module is used to determine whether an early warning is needed for the frequency fluctuation of the current component under test based on the wind turbine rotor frequency, the fluctuation frequency of the unit status signal, and the sliding window mean of the autocorrelation waveform peak sequence.

[0088] Specifically, the fluctuation peak feature extraction module identifies the unit's steady-state speed data, performs bandpass filtering on the state signal in the window width W data frame to remove low-frequency trend terms and high-frequency noise, and then extracts the peak sequence G of the bandpass filtered signal through peak detection. pk And calculate G pk Exceeding the threshold G pk＿thresh Peak percentage P k The fluctuation frequency feature extraction module calculates the autocorrelation waveform of the bandpass filtered signal and extracts the peak sequence of the autocorrelation waveform through peak detection. From the peak sequence, the fluctuation frequency of the state signal and the sliding window mean signal of the autocorrelation waveform peak sequence are extracted. The anomaly detection and early warning module comprehensively considers the wind turbine rotation frequency, the fluctuation frequency of the state signal, and the sliding window mean of the autocorrelation waveform peak sequence to determine a wind turbine rotation frequency fluctuation of 1x and outputs the monitoring results.

[0089] The real-time monitoring device for the periodic abnormal fluctuation state of wind turbines in this embodiment corresponds one-to-one with the above-mentioned real-time monitoring method for the periodic abnormal fluctuation state of wind turbines, and will not be described in detail here.

[0090] This embodiment further provides an electronic device, including a processor and a memory, wherein the memory is used to store a computer program and the processor is used to execute the computer program to perform the method as described above.

[0091] It is understood that the method described in this embodiment can be executed by a single device, such as a computer or server, or it can be applied to a distributed scenario where multiple devices cooperate to complete the task. In a distributed scenario, one of the multiple devices may execute only one or more steps of the method described in this embodiment, and the multiple devices interact to complete the method. The processor can be implemented using a general-purpose CPU, microprocessor, application-specific integrated circuit, or one or more integrated circuits, and is used to execute relevant programs to implement the method described in this embodiment. The memory can be implemented using read-only memory (ROM), random access memory (RAM), static storage devices, and dynamic storage devices. The memory can store the operating system and other applications. When the method described in this embodiment is implemented through software or firmware, the relevant program code is stored in the memory and called and executed by the processor.

[0092] This embodiment further provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0093] Those skilled in the art will understand that the above embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create an implementation for the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 a process or multiple processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0094] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention should fall within the protection scope of the present invention.

Claims

1. A method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator, characterized by the following steps: include: Real-time acquisition of steady-state state signal data frames of wind turbine units, wherein the steady-state state signal data frames include any one or more combinations of speed, torque, nacelle vibration acceleration data and pitch angle under steady-state conditions; The wind turbine rotor frequency is calculated based on the real-time collected steady-state signal data frames. Extract the peak sequence from the current steady-state speed data frame, and extract the fluctuation peak features from the peak sequence; If the fluctuation peak characteristics exceed a preset threshold, the fluctuation frequency of the unit status signal and the sliding window mean of the autocorrelation waveform peak sequence are extracted from the current steady-state speed data frame. The presence of wind turbine rotor frequency fluctuation (n times the rotor frequency) is determined based on the wind turbine rotor rotation frequency, the fluctuation frequency of the turbine status signal, and the sliding window mean of the autocorrelation waveform peak sequence, where n is a positive integer.

2. The method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator as described in claim 1, characterized in that, The periodic abnormal fluctuation state also includes any one or more of the following: tower frequency fluctuation, transmission chain frequency fluctuation, and blade characteristic frequency fluctuation.

3. The method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator as described in claim 1, characterized in that, Before calculating the wind turbine rotor frequency based on the real-time collected steady-state speed data frame, the method further includes using a Butterworth bandpass filter to perform bandpass filtering on the real-time collected steady-state speed data frame to remove low-frequency trend terms and high-frequency noise signals from the steady-state signal data frame of the wind turbine.

4. The method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator as described in claim 1, characterized in that, The fluctuation peak characteristic is the proportion of peaks exceeding a preset threshold in the peak sequence. When the peak proportion is less than the preset proportion threshold, the current monitoring ends and the execution of the real-time acquisition of the steady-state state signal data frame of the wind turbine is returned. Otherwise, the fluctuation frequency of the unit state signal and the sliding window mean of the autocorrelation waveform peak sequence are extracted from the current steady-state speed data frame.

5. The method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator according to claim 1, characterized in that, The fluctuation frequency of the unit status signal extracted from the current steady-state speed data frame includes: Calculate the autocorrelation sequence of the current steady-state speed data frame; Extract the peak sequence of the autocorrelation sequence; The fluctuation frequency of the unit status signal is obtained by calculating the mean time difference of the peak sequence of the autocorrelation sequence.

6. The method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator according to any one of claims 1 to 5, characterized in that, The step of determining whether there is a wind turbine frequency fluctuation of n times based on the wind turbine rotor frequency, the fluctuation frequency of the unit status signal, and the sliding window mean of the autocorrelation waveform peak sequence includes: Determine whether the fluctuation frequency of the unit status signal is within a specified frequency range, wherein the specified frequency range is determined based on n times the rotor speed. If it is determined that the fluctuation frequency of the unit status signal is not within the specified frequency range, then the current monitoring ends and the process returns to executing the real-time acquisition of the wind turbine's steady-state status signal data frame. If the fluctuation frequency of the unit status signal is determined to be within the specified frequency range, the sliding window mean of the autocorrelation waveform peak sequence is determined. If the sliding window mean of the autocorrelation waveform peak sequence is greater than the preset sliding window mean threshold, it is determined that there is a wind turbine frequency fluctuation of n times, and the corresponding early warning is triggered.

7. The method for real-time monitoring of periodic abnormal fluctuations in the state of a wind turbine generator according to claim 6, characterized in that, The specified frequency range is f p ·(1-k)~f p ·(1+k), f p The frequency is n times the rotational frequency of the wind turbine, and k is the set value of the relative frequency deviation, where 1>k>0.

8. A real-time monitoring device for periodic abnormal fluctuations in the state of a wind turbine generator, characterized in that, include: The data acquisition module is used to acquire steady-state state signal data frames of the wind turbine in real time. The steady-state state signal data frames include any one or a combination of two or more of the following: speed, torque, nacelle vibration acceleration data and pitch angle under steady-state conditions. The frequency parameter calculation module is used to calculate the wind turbine rotor frequency based on the real-time collected steady-state signal data frames. The fluctuation peak feature extraction module is used to extract the peak sequence from the current steady-state speed data frame and extract the fluctuation peak features from the peak sequence; The fluctuation frequency feature extraction module is used to extract the fluctuation frequency of the unit status signal and the sliding window mean of the autocorrelation waveform peak sequence from the current steady-state speed data frame if the fluctuation peak feature exceeds a preset threshold. The abnormal fluctuation determination module is used to determine whether there is a wind turbine frequency fluctuation of n times the rotation frequency based on the wind turbine rotor frequency, the fluctuation frequency of the unit status signal, and the sliding window mean of the autocorrelation waveform peak sequence, where n is a positive integer.

9. An electronic device comprising a processor and a memory, the memory being used to store a computer program, characterized in that, The processor is used to execute the computer program to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method as described in any one of claims 1 to 7.

Images