Vibration suppression method and device for wind turbine

Abstract

Disclosed are a vibration suppression method and device for a wind turbine. The vibration suppression method comprises: obtaining a vibration acceleration signal associated with the wind turbine (including a vibration acceleration signal at a historical moment and a vibration acceleration signal at a current moment), operating data of the wind turbine associated with a preset operating condition, and operating environment data; performing feature extraction on the vibration acceleration signal at a historical moment and the vibration acceleration signal at a current moment to obtain a vibration characteristic signal at a historical moment and a vibration characteristic signal at a current moment associated with the wind turbine; predicting a vibration characteristic signal at a future moment at the current moment based on the vibration characteristic signal at a historical moment and the vibration characteristic signal at a current moment as well as the operating data and the operating environment data; determining whether a vibration suppression condition associated with the preset operating condition is satisfied based on the vibration characteristic signal at a current moment and the vibration characteristic signal at a future moment; and executing a vibration suppression action associated with the preset operating condition in response to determining that the vibration suppression condition is satisfied.

Description

Technical Field

[0001] The present disclosure generally relates to the field of wind power generation technology, and more particularly, to a method and apparatus for suppressing vibration of a wind turbine. Background Art

[0002] In the wind power industry, with the development of wind power technology, numerous wind turbines ("turbines") are being deployed in diverse environments to generate wind power. Abnormal vibration in wind turbines ("fans") can cause serious problems and even affect the operation of the entire unit. Therefore, controlling fan vibration has become a major concern within the industry.

[0003] There are many factors that can cause abnormal vibration of wind turbines. For example, different climatic conditions may cause abnormal vibration of wind turbines. For example, multiple factors such as multi-system coupled vibration of the unit / wind turbine, insufficient blade aeroelastic damping, negative shear, large turbulence, icing, etc. can cause abnormal vibrations in different modes. In related fields, the general protection method for abnormal vibration is to detect the vibration amplitude related to the wind turbine and directly trigger a fault shutdown when the vibration amplitude reaches a specific threshold. However, existing vibration protection methods may have the disadvantages of causing large power generation losses and may lead to an increase in the failure rate of the unit.

[0004] Therefore, in order to solve the above problems, it is necessary to provide an improved protection method for fan vibration. Summary of the Invention

[0005] The embodiments of the present disclosure provide a method and apparatus for suppressing vibration of a wind turbine, thereby effectively suppressing the vibration of the wind turbine while reducing power generation loss, and significantly improving the operational safety and stability of the wind turbine.

[0006] In one general aspect, a vibration suppression method for a wind turbine is provided, the vibration suppression method comprising: acquiring a vibration acceleration signal associated with the wind turbine and operating data and operating environment data of the wind turbine associated with a preset operating condition, the vibration acceleration signal comprising a vibration acceleration signal at a historical moment and a vibration acceleration signal at a current moment; performing feature extraction on the vibration acceleration signal at a historical moment and the vibration acceleration signal at a current moment to obtain a vibration characteristic signal at a historical moment and a vibration characteristic signal at a current moment associated with the wind turbine; predicting a vibration characteristic signal at at least one future moment of the current moment based on the vibration characteristic signal at a historical moment and the vibration characteristic signal at a current moment as well as the operating data and the operating environment data; determining whether a vibration suppression condition associated with the preset operating condition is satisfied based on the vibration characteristic signal at the current moment and the vibration characteristic signal at at least one future moment; and executing a vibration suppression action associated with the preset operating condition in response to determining that the vibration suppression condition is satisfied.

[0007] Optionally, the step of determining whether the vibration suppression condition associated with the preset operating condition is met based on the vibration characteristic signal at the current moment and the vibration characteristic signal at at least one future moment may include: determining whether the vibration suppression condition associated with the preset operating condition is met by comparing the amplitude of the vibration characteristic signal at the current moment with a first preset threshold and comparing the amplitude of the vibration characteristic signal at at least one future moment with a second preset threshold.

[0008] Optionally, the step of executing the vibration suppression action associated with the preset operating condition in response to determining that the vibration suppression condition is met may include: executing the vibration suppression action associated with the preset operating condition in response to the amplitude of the vibration characteristic signal at the current moment being less than the first preset threshold and the amplitude of the vibration characteristic signal at at least one future moment being greater than or equal to the second preset threshold.

[0009] Optionally, the step of predicting the vibration characteristic signal of at least one future moment of the current moment based on the vibration characteristic signal of the historical moment, the vibration characteristic signal of the current moment, the operating data and the operating environment data may include: using the vibration characteristic signal of the historical moment, the operating data and the operating environment data to train a feature prediction model based on a neural network, and inputting the vibration characteristic signal of the current moment into the trained feature prediction model based on the neural network to obtain the vibration characteristic signal of the at least one future moment.

[0010] Optionally, the step of using the vibration characteristic signal at the historical moment, the operating data and the operating environment data to train the neural network-based feature prediction model may include: inputting the vibration characteristic signal at the historical moment, the operating data and the operating environment data into the neural network-based feature prediction model to obtain a predicted amplitude corresponding to the actual amplitude of the vibration characteristic signal at the historical moment; calculating a prediction model loss function based on the actual amplitude and the predicted amplitude of the vibration characteristic signal at the historical moment; and adjusting the model parameters of the neural network-based feature prediction model according to the prediction model loss function to obtain a trained neural network-based feature prediction model.

[0011] Optionally, the type of vibration suppressed by the vibration suppression action may include at least one of the following items: resonant vibration of the impeller speed of the wind turbine when it is close to a preset speed and the natural frequency of the wind turbine, negative damping vibration of the blades of the wind turbine when the blades are within a preset pitch angle range, generator vibration of the wind turbine when the wind turbine is within a preset power range, first-order vibration of the tower of the wind turbine under a first preset wind speed condition, and first-order vortex-induced vibration of the blades of the wind turbine under a second preset wind speed condition.

[0012] In another general aspect, a vibration suppression device for a wind turbine is provided, the vibration suppression device comprising: a data acquisition module configured to acquire a vibration acceleration signal associated with the wind turbine and operating data and operating environment data of the wind turbine associated with a preset operating condition, the vibration acceleration signal comprising a vibration acceleration signal at a historical moment and a vibration acceleration signal at a current moment; a feature extraction module configured to perform feature extraction on the vibration acceleration signal at a historical moment and the vibration acceleration signal at a current moment to obtain a vibration characteristic signal at a historical moment and a vibration characteristic signal at a current moment associated with the wind turbine; a data prediction module configured to predict a vibration characteristic signal at at least one future moment of the current moment based on the vibration characteristic signal at a historical moment and the vibration characteristic signal at a current moment as well as the operating data and the operating environment data; and a vibration suppression determination module configured to determine whether a vibration suppression condition associated with the preset operating condition is satisfied based on the vibration characteristic signal at the current moment and the vibration characteristic signal at at least one future moment, and, in response to determining that the vibration suppression condition is satisfied, execute a vibration suppression action associated with the preset operating condition.

[0013] Optionally, the size relationship between the first task cycle of the data acquisition module performing data acquisition operations, the second task cycle of storing the acquired data, the third task cycle of the feature extraction module performing feature extraction, the fourth task cycle of the data prediction module performing prediction operations, and the fifth task cycle of the vibration suppression determination module determining to perform vibration suppression actions may be as follows: the third task cycle is equal to the fourth task cycle, the first task cycle is equal to the fifth task cycle, the first task cycle is smaller than the third task cycle, and the third task cycle is smaller than the second task cycle.

[0014] Optionally, the vibration suppression device may include a first data processing routine and a second data processing routine, the first data processing routine being used to execute data processing tasks associated with the first task cycle and the fifth task cycle, the second data processing routine being used to execute data processing tasks associated with the third task cycle and the fourth task cycle, and the first data processing routine and the second data processing routine may be respectively deployed in different controllers or may be deployed in the same controller.

[0015] Optionally, the operation of the vibration suppression determination module determining whether the vibration suppression condition associated with the preset operating condition is met based on the vibration characteristic signal at the current moment and the vibration characteristic signal at at least one future moment may include: determining whether the vibration suppression condition associated with the preset operating condition is met by comparing the amplitude of the vibration characteristic signal at the current moment with a first preset threshold and comparing the amplitude of the vibration characteristic signal at at least one future moment with a second preset threshold.

[0016] Optionally, the vibration suppression determination module, in response to determining that the vibration suppression condition is met, performs the vibration suppression action associated with the preset operating condition, which may include: in response to the amplitude of the vibration characteristic signal at the current moment being less than the first preset threshold and the amplitude of the vibration characteristic signal at at least one future moment being greater than or equal to the second preset threshold, performing the vibration suppression action associated with the preset operating condition.

[0017] Optionally, the data prediction module may include an operation of predicting the vibration characteristic signal of at least one future moment of the current moment based on the vibration characteristic signal of the historical moment, the vibration characteristic signal of the current moment, the operating data, and the operating environment data: using the vibration characteristic signal of the historical moment, the operating data, and the operating environment data to train a feature prediction model based on a neural network, and inputting the vibration characteristic signal of the current moment into the trained feature prediction model based on the neural network to obtain the vibration characteristic signal of the at least one future moment.

[0018] Optionally, the operation of the data prediction module using the vibration characteristic signal at the historical moment, the operating data and the operating environment data to train the feature prediction model based on the neural network may include: inputting the vibration characteristic signal at the historical moment, the operating data and the operating environment data into the feature prediction model based on the neural network to obtain a predicted amplitude corresponding to the actual amplitude of the vibration characteristic signal at the historical moment; calculating a prediction model loss function based on the actual amplitude and the predicted amplitude of the vibration characteristic signal at the historical moment; and adjusting the model parameters of the feature prediction model based on the neural network according to the prediction model loss function to obtain a trained feature prediction model based on the neural network.

[0019] Optionally, the type of vibration suppressed by the vibration suppression action may include at least one of the following items: resonant vibration of the impeller speed of the wind turbine when it is close to a preset speed and the natural frequency of the wind turbine, negative damping vibration of the blades of the wind turbine when the blades are within a preset pitch angle range, generator vibration of the wind turbine when the wind turbine is within a preset power range, first-order vibration of the tower of the wind turbine under a first preset wind speed condition, and first-order vortex-induced vibration of the blades of the wind turbine under a second preset wind speed condition.

[0020] In another general aspect, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by at least one processor, the at least one processor is prompted to perform the vibration suppression method for a wind turbine as described above.

[0021] In another general aspect, a computer program product is provided, comprising computer instructions, which, when executed by at least one processor, cause the at least one processor to perform the vibration suppression method for a wind turbine as described above.

[0022] In another general aspect, a computing device is provided, comprising: at least one processor; and at least one memory storing computer-executable instructions, wherein the computer-executable instructions, when executed by the at least one processor, cause the at least one processor to perform the vibration suppression method for a wind turbine as described above.

[0023] In another general aspect, a wind turbine generator set is provided, comprising: the vibration suppression device for a wind turbine generator as described above, or the computing device as described above.

[0024] According to the wind turbine vibration suppression method and device of the embodiments of the present disclosure, precise control of wind turbine vibration is achieved through vibration feature prediction and optimized control strategies, effectively suppressing wind turbine vibration while reducing power generation losses, and significantly improving the operational safety and stability of the wind turbine. In addition, by deeply integrating artificial intelligence methods, wind turbine vibration stability mechanisms, and real-time control decision-making methods at the wind turbine edge, the wind turbine vibration suppression method and device according to the present disclosure can be used to suppress wind turbine vibration related to operating conditions, thereby making the wind turbine vibration suppression method and device highly scalable. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and other objects and features of the embodiments of the present disclosure will become more apparent through the following description in conjunction with the accompanying drawings showing the embodiments, in which:

[0026] Figure 1 is a flowchart illustrating a vibration suppression method for a wind turbine generator according to an embodiment of the present disclosure;

[0027] Figure 2 is a flowchart illustrating an example of a vibration suppression method for a wind turbine generator according to an embodiment of the present disclosure;

[0028] Figure 3 is a schematic diagram illustrating abnormal vibration of an exemplary wind turbine;

[0029] Figure 4 is a timing diagram illustrating a feature extraction process according to an embodiment of the present disclosure;

[0030] Figure 5 is a block diagram illustrating a vibration suppression apparatus for a wind turbine generator according to an embodiment of the present disclosure;

[0031] Figure 6 is a block diagram illustrating a vibration suppression system for a wind turbine according to one embodiment of the present disclosure;

[0032] Figure 7 is a block diagram illustrating a vibration suppression system for a wind turbine generator according to another embodiment of the present disclosure;

[0033] Figure 8 is a block diagram illustrating a computing device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0034] The following detailed description is provided to help the reader gain a comprehensive understanding of the methods, devices and / or systems described herein. However, various changes, modifications and equivalents of the methods, devices and / or systems described herein will be clear after understanding the disclosure of the present application. For example, the order of operations described herein is merely an example and is not limited to those orders set forth herein, but can be changed as will be clear after understanding the disclosure of the present application, except for operations that must occur in a specific order. In addition, for greater clarity and conciseness, descriptions of features known in the art may be omitted.

[0035] The features described herein can be implemented in different forms and should not be construed as limited to the examples described herein. Rather, the examples described herein are provided to illustrate only some of the many possible ways to implement the methods, devices, and / or systems described herein, which will become clear after understanding the disclosure of this application.

[0036] As used herein, the term "and / or" includes any one of the associated listed items and any combination of any two or more.

[0037] The terms used herein are only used to describe various examples and are not intended to limit the disclosure. Unless the context clearly indicates otherwise, the singular is intended to include the plural. The terms "comprise," "include," and "have" indicate the presence of the recited features, quantities, operations, components, elements, and / or combinations thereof, but do not preclude the presence or addition of one or more other features, quantities, operations, components, elements, and / or combinations thereof.

[0038] Unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the present disclosure pertains after understanding the present disclosure. Unless expressly defined otherwise herein, terms (such as those defined in general dictionaries) should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and the present disclosure, and should not be interpreted in an idealized or overly formal manner.

[0039] Furthermore, in describing the examples, when it is deemed that a detailed description of well-known related structures or functions would cause ambiguous interpretation of the present disclosure, such detailed description will be omitted.

[0040] Reference will now be made in detail to the embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to like parts throughout. The embodiments are described below with reference to the drawings in order to explain the present disclosure.

[0041] In the related art, as mentioned in the background, the current general protection method for abnormal vibration is to detect the vibration amplitude of the wind turbine (i.e., the amplitude that represents the vibration level of large components or the entire machine) and directly trigger a fault shutdown when the vibration amplitude reaches a specific threshold. This vibration protection method is executed only based on the detected vibration amplitude, and triggers a fault shutdown of the wind turbine when the amplitude reaches a specific threshold. However, this fault shutdown will result in significant power generation losses, increase the failure rate of the wind turbine, and thus reduce the wind turbine's availability.

[0042] In order to solve the above problems, the present disclosure proposes a vibration suppression method and device for a wind turbine. Specifically, in the present disclosure, a vibration characteristic value prediction model is established with the wind turbine vibration signal and operating conditions (for example, working conditions) as input; the output of the prediction model and the real-time vibration characteristic value are both used as inputs of the vibration suppression decision model to determine the control decision; and vibration suppression is ultimately achieved by adjusting the operating state of the wind turbine. The present disclosure achieves fine control of wind turbine vibration (for example, by controlling and adjusting in advance to avoid vibration-sensitive working conditions) through vibration characteristic value prediction and optimized control decisions, thereby effectively suppressing wind turbine vibration with minimal power generation loss and significantly improving wind turbine operation safety and stability.

[0043] Refer to the following Figures 1 to 8 A vibration suppression method and apparatus for a wind turbine according to an embodiment of the present disclosure are described in detail.

[0044] Figure 1 1 is a flowchart illustrating a vibration suppression method 100 for a wind turbine according to an embodiment of the present disclosure.

[0045] Reference Figure 1 In step S101, a vibration acceleration signal associated with a wind turbine and operating data and operating environment data of the wind turbine associated with a preset operating condition are obtained.

[0046] Here, the vibration acceleration signal includes historical vibration acceleration signals and current vibration acceleration signals. Further, the vibration acceleration signal can be a signal representing the vibration acceleration level of a large component or the entire machine in the acceleration signal of the large component or the nacelle of the wind turbine.

[0047] According to an embodiment of the present disclosure, in step S102 , feature extraction is performed on the historical vibration acceleration signal and the current vibration acceleration signal to obtain the historical vibration feature signal and the current vibration feature signal associated with the wind turbine.

[0048] According to an embodiment of the present disclosure, in step S103 , based on the vibration characteristic signals at historical moments and the vibration characteristic signals at the current moment as well as the operation data and the operation environment data, the vibration characteristic signals at at least one future moment of the current moment are predicted.

[0049] Here, the at least one future time instant may be, for example, at least one future adjacent time instant, and further may be, for example, the next time instant or a combination of future adjacent or non-adjacent time instants.

[0050] As an example, step S103 may further include: using vibration characteristic signals, operating data and operating environment data at historical moments to train a feature prediction model based on a neural network, and inputting the vibration characteristic signal at the current moment into the trained feature prediction model based on a neural network to obtain a vibration characteristic signal at at least one future moment.

[0051] According to an embodiment of the present disclosure, by using a feature prediction model, a method for predicting vibration characteristic values ​​by utilizing the correlation between vibration characteristic signals and wind turbine operating conditions is implemented, thereby realizing the prediction of vibration characteristic signals at at least one future moment.

[0052] Here, the method for training a feature prediction model based on a neural network may include the following steps S31 to S33:

[0053] In step S31, the vibration characteristic signal at the historical moment, the operating data and the operating environment data are input into a feature prediction model based on a neural network to obtain a predicted amplitude corresponding to the actual amplitude of the vibration characteristic signal at the historical moment.

[0054] In step S32, a prediction model loss function is calculated based on the actual amplitude and predicted amplitude of the vibration characteristic signal at the historical moment.

[0055] In step S33, the model parameters of the neural network-based feature prediction model are adjusted according to the prediction model loss function to obtain a trained neural network-based feature prediction model.

[0056] According to an embodiment of the present disclosure, a feature prediction model based on a neural network is adopted to predict the amplitude of the vibration characteristic signal. From the perspective of the control framework, the comparison between the predicted value and the actual value is considered to reflect the consideration of the safety of the fan operation, thereby ensuring the safety of the fan operation.

[0057] According to an embodiment of the present disclosure, in step S104 , based on the vibration characteristic signal at the current moment and at least one vibration characteristic signal at a future moment, it is determined whether a vibration suppression condition associated with a preset operating condition is satisfied.

[0058] As an example, step S104 may further include: determining whether a vibration suppression condition associated with a preset operating condition is met by comparing the amplitude of the vibration characteristic signal at the current moment with a first preset threshold and comparing the amplitude of the vibration characteristic signal at at least one future moment with a second preset threshold.

[0059] According to an embodiment of the present disclosure, the fan vibration suppression condition is determined by comparing the amplitude of the vibration characteristic signal at the current moment and the amplitude of the vibration characteristic signal at at least one future moment with corresponding conditions.

[0060] According to an embodiment of the present disclosure, in step S105 , in response to determining that the vibration suppression condition is satisfied, a vibration suppression action associated with a preset operating condition is executed.

[0061] As an example, step S105 may further include: in response to the amplitude of the vibration characteristic signal at the current moment being less than a first preset threshold and the amplitude of the vibration characteristic signal at at least one future moment being greater than or equal to a second preset threshold, performing a vibration suppression action associated with a preset operating condition.

[0062] According to an embodiment of the present disclosure, the corresponding control decision is executed by respectively satisfying the corresponding conditions of the amplitude of the vibration characteristic signal at the current moment and the amplitude of the vibration characteristic signal at at least one future moment, thereby realizing the determination of the fan vibration suppression action that needs to be executed, thereby facilitating the realization of proactive vibration suppression processing in advance.

[0063] For example, the types of vibrations suppressed by the vibration suppression action may include at least one of the following items: resonant vibration of the impeller speed of the wind turbine when it is close to a preset speed and the natural frequency of the wind turbine, negative damping vibration of the blades of the wind turbine when the blades are within a preset pitch angle range, generator vibration of the wind turbine when the wind turbine is within a preset power range, first-order vibration of the tower of the wind turbine under a first preset wind speed condition, and first-order vortex-induced vibration of the blades of the wind turbine under a second preset wind speed condition.

[0064] It should be noted that the vibration types that can be suppressed by the method disclosed herein are not limited to the above types, but can also include other vibration types that can demonstrate a specific degree of correlation between the vibration amplitude of the entire wind turbine or specific components and specific conditions (e.g., operating conditions). In other words, the method disclosed herein can suppress abnormal vibrations in wind turbines / wind turbines of different modes.

[0065] According to the embodiments of the present disclosure, by establishing a vibration characteristic value prediction model with the fan vibration signal and operating conditions as input, the above-mentioned various types of vibrations are suppressed, the vibration prediction and control related to the fan operating conditions are effectively solved, and the wide applicability of the solution of the present disclosure is improved.

[0066] Below, by reference Figures 2 to 4 The vibration suppression method 100 of the wind turbine generator described above will be described as an example. Figure 2 is a flowchart illustrating an example of a vibration suppression method for a wind turbine generator according to an embodiment of the present disclosure, Figure 3 is a schematic diagram illustrating abnormal vibration of an exemplary wind turbine, and Figure 4 FIG. 1 is a timing diagram illustrating a feature extraction process according to an embodiment of the present disclosure.

[0067] As an example, see Figure 2 An example of a vibration suppression method for a wind turbine may include the following processes: signal reading 101, signal processing 102, model M0 processing 103, model M1 processing 104, and control decision 105, which may be referred to as modules 101 to 105 below.

[0068] Specifically, module 101 can realize the acquisition of the fan vibration acceleration signal and the acquisition of the fan operation data and operating environment data associated with the preset operating conditions, modules 102 to 104 can realize the signal processing and predictive analysis of the fan vibration acceleration signal, and module 105 can realize the decision control for determining the need to perform vibration suppression action.

[0069] Here, in Figure 2 Modules 101 and 105 shown in FIG. 1 perform data processing with a task period of T0, and modules 102 to 104 perform data processing with a task period of T1. For example, task period T1 may be greater than task period T0. For example, T0 may be 10 ms or 20 ms, but the present disclosure is not limited thereto. Furthermore, for example, T1 may be 2 s or 4 s, but the present disclosure is not limited thereto.

[0070] Furthermore, regarding the task cycles of various processing processes, as an example, the task cycle of the data reading process of the signal reading module may be equal to the task cycle of the decision-making process executed by the control decision module. Furthermore, for example, the task cycle of the feature extraction process may be greater than the task cycle of the data storage process. Furthermore, for example, the task cycle of the data prediction process may be equal to the task cycle of the feature extraction process. Furthermore, for example, the task cycle of the data reading process may be less than the task cycle of the feature extraction process.

[0071] By limiting these task cycles, stable, reliable and real-time response effects can be achieved.

[0072] The vibration suppression method for a wind turbine generator according to the above-mentioned example of the present disclosure can suppress abnormal vibrations of wind turbines / units in different modes. For example, typical vibration modes include, but are not limited to, at least one of the following: resonance between the frequency multiples of the rotational speed near a specific rotational speed and the natural frequency of the generator, negatively damped vibration of blades that may occur when the blades are in a specific pitch angle range, generator vibration within a specific power range, first-order vibration of the tower within a specific wind speed range, etc. Since the common feature of these vibrations is that there is a predetermined degree of correlation between the vibration amplitude of the entire machine or large component and the specific operating conditions, the vibration suppression method of the present disclosure can predict and control vibrations.

[0073] The following examples illustrate vibration suppression methods under different vibration situations.

[0074] As an example, this example addresses the vibration problem where, at a specific speed, the impeller's rotational frequency approaches the generator's natural frequency, causing the generator to resonate. Here, rotational frequency refers to the unit / fan's impeller speed or the reciprocal of the generator's rotation time.

[0075] Here, refer to Figure 3 To explain the abnormal vibration in this example. Figure 3 As shown, the current moment is shown as k, the moment of the previous unit time (for example, its typical value can be 2s) of the current moment is shown as k-1, the moment of the next unit time of the current moment is shown as k+1, the sensitive speed is shown as n0, and the actual speeds of the current moment, the previous moment and the next moment are shown as n respectively. k 、n k-1 and n k+1 The area difference between the actual speed and the sensitive speed is shown as d k The acceleration frequency domain amplitudes at the current moment, the previous moment, and the next moment are shown as x k 、x k-1 and x k+1 .

[0076] Specifically, the sensitive speed is the speed at which resonance is most likely to occur. When the impeller's actual speed approaches the sensitive speed, the acceleration frequency domain amplitude, which represents the magnitude of the vibration, continues to increase. In other words, the acceleration frequency domain amplitude is proportional to the area difference between the actual speed and the sensitive speed per unit time. Here, the acceleration frequency domain amplitude represents the amplitude of the maximum point within a specific frequency range in the acceleration spectrum.

[0077] Refer again Figure 2, first, regarding module 101, module 101 reads a vibration acceleration signal at a predetermined task period (eg, T0), for example, a vibration acceleration signal v(t) as an analog signal read from a vibration acceleration sensor installed in a generator.

[0078] It should be noted that while the vibration acceleration sensor is shown as being installed on the generator, the present disclosure is not limited thereto and may be installed in other locations as long as it can detect relevant vibration signals. For example, it may be installed in other locations such as the center of the nacelle base. Similarly, vibration sensing sensors used to detect other vibrations may be installed in corresponding locations as long as they can detect the corresponding vibration signals used to identify vibration patterns.

[0079] Then, module 102 performs feature extraction on the vibration acceleration signal v(t) at a predetermined duty cycle (e.g., T1). For example, the extracted feature may be at least one of the following: acceleration frequency domain amplitude, acceleration RMS value, standard deviation of the acceleration raw value, standard deviation of the acceleration RMS value, and frequency domain amplitude standard deviation.

[0080] In addition, it should be noted that when extracting features from vibration signals of other vibration modes / situations, other feature extraction methods may be used, not limited to Fourier transform, and the extracted features may be features of other dimensions in addition to the above-mentioned feature examples, which is not particularly limited in this disclosure.

[0081] Here, refer to Figure 4 The timing of executing the feature extraction process is briefly described. Module 102 stores the acceleration value v(t) array of the time period T2 (for example, but not limited to 5.12s) in real time. At the same time, feature extraction processing is performed once every T1 task cycle. Here, regarding the storage of the acceleration values ​​received from module 101 by module 102, a real-time storage method can be adopted, for example, for receiving data from module 101 in real time, that is, whenever module 101 reads data at a predetermined time interval, the read data is sent to module 102 for real-time storage. In addition, module 102 stores the received data in the order in which it is received. For example, the data stored in each T2 time period can be recorded so that relevant data processing according to the data of the T2 time period can be performed when executing feature extraction.

[0082] That is, feature extraction is to perform corresponding processing on at least one set of data stored in the most recent T2 time period before the current T1. For example, the data to be processed can be a single set of data, or it can be multiple sets of data, depending on whether the features processed by several sets of data can represent the target vibration mode. For example, in the case where the cabin acceleration direction includes two directions, x and y, if the x direction itself is sufficient to effectively represent the target vibration mode, a single set of data can be used; otherwise, it is necessary to extract feature data of acceleration in the x and y directions, in which case multiple sets of data can be used. In addition, for example, in the case where the direction of the blade vibration or load signal includes two directions, edgewise (swing) and flapwise (flap), if it is necessary to represent the vibration within the impeller surface, a single set of data indicating the edgewise direction can be used.

[0083] In this example, the steps of performing feature extraction processing in module 102 include:

[0084] Step 1): In the current task cycle, a fast Fourier transform is performed on the stored vibration acceleration value array to obtain an acceleration spectrum (ie, acceleration frequency domain data).

[0085] Step 2), assuming that the natural frequency of the generator at the resonance point is f, take the amplitude x(k) of the maximum point in the frequency range [f-f0, f+f0] in the spectrum.

[0086] Here, the frequency f0 used to limit the above frequency range is typically set to 1 Hz, but the present disclosure is not limited thereto.

[0087] In addition, module 102 can also output the acceleration frequency domain amplitudes x(k-1), x(k-2), ... of several cycles before the current cycle in addition to the acceleration frequency domain amplitude x(k) of the current cycle, so as to input x(k) into module 103, and input x(k-1), x(k-2), ... into module 104, so that model M1 can use the acceleration frequency domain amplitudes of the past period of time to predict the future acceleration frequency domain amplitudes (reflecting the vibration trend).

[0088] Next, refer to Figure 2 , regarding module 103, the input of model M0 is the acceleration frequency domain amplitude x(k) of the current cycle, and the output of model M0 is u0(k). The relationship between the input and output of model M0 is shown in the following equation (1):

[0089]

[0090] Where a represents the acceleration amplitude exceeding the limit shutdown threshold.

[0091] Furthermore, regarding module 104, model M1 uses a recurrent neural network prediction model to predict the acceleration frequency domain amplitude. It should be noted that the method according to the present disclosure is not limited to the recurrent neural network prediction model in terms of the selection of prediction models. Other types of prediction models can also be used, such as perceptron neural networks containing one or more hidden layers, multivariate linear regression, and other different prediction models.

[0092] Specifically, the training process of the prediction model can be as follows:

[0093] (1) The acceleration frequency domain amplitudes x(k), x(k-1)... of the past j T1 cycles and the rotation speeds n(k-1), n(k-2)... that have a strong correlation with vibration, and / or at least one of the wind turbine operating power data, wind speed data, etc. are used as inputs of the prediction model as the actual test set data.

[0094] (2) Using the test set data, a prediction model is constructed, for example, with a single hidden layer of a recurrent neural network with 256 hidden units, a batch size of 32, and 300 training iterations.

[0095] (3) Predicting the test set data using the prediction model to obtain the test set prediction value, calculating the loss function based on the predicted value and the actual value, and adjusting the parameters of the prediction model by making the loss function meet the preset conditions. For example, the root mean square error between the predicted value and the actual value can be adjusted to an acceptable level as the goal of adjusting the model parameters, but the present disclosure is not limited to this.

[0096] After the prediction model is trained, the trained model M1 is used to predict the acceleration frequency domain amplitudes x(k+1), x(k+2), ..., x(k+i) at times k+1, k+2, ..., k+i. When the predicted value x(k+1) exceeds the fault shutdown threshold a, the output u1(k) of M1 takes on the value 1. In other words, the relationship between the value of u1(k) and x(k+1) is similar to that in Equation (1).

[0097] Finally, regarding module 105, module 105 performs control decision processing at a predetermined task cycle (e.g., T0). Module 105 judges its inputs u0(k) and u1(k), and determines the corresponding wind turbine control decision action based on the judgment results. Here, the rules for module 105 to perform judgment are shown in Table 1 below. In addition, Figure 2 In , z(k) represents the preset system disturbance.

[0098] Table 1

[0099]

[0100] Referring to Table 1, when u0(k) is 1, u(k) takes the value of 1, and the corresponding control decision is: directly execute the shutdown action. Here, the shutdown state is maintained for a period of time T3, which can be, for example, 10 minutes, but the present disclosure is not limited to this. When only u1(k) is 1 (i.e., u0(k) is 0 and u1(k) is 1), u(k) takes the value of 2, and the corresponding control decision is: limit the speed and maintain the time period T4, and the speed limit value is n. Here, T4 can take the value of, for example, 10 minutes, but the present disclosure is not limited to this.

[0101] In addition, the speed limit value n can be expressed as follows:

[0102] n=no-Δn (2)

[0103] The value of Δn may be, for example, 1.5 rpm, so that after the speed limit operation, the actual speed is far away from the sensitive speed (this also reflects the effect of power reduction), thereby achieving the effect of vibration suppression.

[0104] In addition, when u0(k) and u1(k) are both 0, u(k) takes the value of 0, and the corresponding control decision is: the fan continues to run.

[0105] For the above example, a deep learning vibration frequency domain amplitude prediction model is established using the wind turbine vibration signal and operating conditions as input; the output of the prediction model and the real-time vibration amplitude are used as inputs to the vibration control decision model to make control decisions; and vibration control is achieved by adjusting the operating state of the wind turbine. According to the above example of the present disclosure, fine control of wind turbine vibration is achieved through vibration amplitude prediction and optimized control based on the control decision model. It is possible to proactively control and adjust to avoid vibration-sensitive operating conditions in advance, thereby effectively suppressing wind turbine vibration with minimal power generation loss, significantly improving the safety and stability of wind turbine operation.

[0106] In addition, as another example, the vibration situation to be addressed in this example is: negatively damped blade vibration that may occur when a turbine with a specific blade model (e.g., a large turbine) is under specific wind conditions and within a specific pitch angle range. For this vibration situation, blade vibration prediction can be performed using the pitch angle (e.g., which can be used to replace the characteristics of the rotational speed in the above example), and / or variables such as power and wind speed, as well as corresponding characteristic values ​​and blade vibration characteristic values ​​(e.g., which can be used to replace the characteristics of the acceleration frequency domain amplitude in the above example). Based on the predicted blade vibration values, the minimum pitch angle is actively increased (which also reflects the effect of power reduction), thereby achieving vibration suppression in this abnormal vibration mode.

[0107] In this example, the relevant rule that can be used to replace the execution of the speed limit judgment in the above example can be, for example: if the vibration amplitude is greater than the threshold value b, increase the minimum pitch angle by 3 degrees, otherwise do nothing.

[0108] As another example, this example addresses the vibration scenario where, under specific wind speed and wind deviation conditions, first-order tower vibration and first-order blade vortex-induced vibration may occur when the wind turbine is in a shutdown state. To address this vibration scenario, variables such as wind speed and wind deviation, along with their corresponding eigenvalues ​​and tower and blade vibration eigenvalues ​​(for example, these can be used to replace the acceleration frequency-domain amplitude features in the above example), can be used to predict tower and blade vibrations, respectively. Active minimum pitch angle adjustment or yaw action can then be performed based on the predicted tower and blade vibration values, thereby suppressing first-order tower vibration and blade vortex-induced vibration when the wind turbine is in a shutdown state.

[0109] Another example is generator vibration within a specific power range. This can occur in some units due to structural issues with the generator, causing the amplitude of the generator's natural frequency in the nacelle acceleration spectrum to continuously increase at rated power. Once this vibration condition is identified, power limiting can be implemented, and an alarm can be issued to prompt relevant operators to investigate.

[0110] Additionally, as another specific example, under conditions of deep power limiting (eg, 5% of rated power), most vibrations (eg, the examples of typical vibration types / modes described above) may be eliminated.

[0111] That is to say, in control scenarios for different vibration types / modes, different suppression methods can be adopted, such as power limiting, increasing the minimum pitch angle, torque reverse resistance (as a suppression method of "tower lateral resistance"), or a combination of different suppression methods can be adopted to achieve targeted control for specific vibration types, but the present disclosure is not limited to this.

[0112] According to an embodiment of the present disclosure, the vibration suppression method of the wind turbine disclosed in the present disclosure can be deployed in a wind turbine edge side control system, or can also be deployed in a wind farm controller with real-time computing capabilities. Here, the wind turbine edge side control system (Wind Turbine Edge Side Control System) is an on-site control system located on the edge side of the wind turbine. It is a carrier for realizing wind turbine edge computing (Wind Turbine Edge Computing, a computing architecture for data processing, control, and storage), and it has the advantages of low latency (high real-time performance), high efficiency (for example, on-site online perception, analysis, and control), data optimization, and high security. By being deployed in a system capable of realizing real-time online signal processing, the vibration suppression method disclosed in the present disclosure has the advantages of high real-time performance, high efficiency, and high security.

[0113] Figure 5 is a block diagram illustrating a vibration suppression apparatus 500 for a wind turbine generator according to an embodiment of the present disclosure.

[0114] Reference Figure 5 The vibration suppression device 500 for a wind turbine according to an embodiment of the present disclosure may include a data acquisition module 510 , a feature extraction module 520 , a data prediction module 530 and a vibration suppression determination module 540 .

[0115] According to an embodiment of the present disclosure, the data acquisition module 510 may execute: acquiring a vibration acceleration signal associated with the wind turbine and operating data and operating environment data of the wind turbine associated with a preset operating condition.

[0116] Here, the vibration acceleration signal includes a historical moment vibration acceleration signal and a current moment vibration acceleration signal.

[0117] As an example, the size relationship of the first task cycle in which the data acquisition module 510 performs the data acquisition operation, the second task cycle in which the acquired data is stored, the third task cycle in which the feature extraction module performs feature extraction, the fourth task cycle in which the data prediction module performs the prediction operation, and the fifth task cycle in which the vibration suppression determination module determines to perform the vibration suppression action can be as follows: the third task cycle is equal to the fourth task cycle, the first task cycle is equal to the fifth task cycle, the first task cycle is less than the third task cycle, and the third task cycle is less than the second task cycle.

[0118] According to the embodiments of the present disclosure, by setting the task cycles corresponding to different operations, data processing efficiency can be improved, data processing time can be saved, and the data processing process can be optimized while ensuring data processing accuracy and data processing precision.

[0119] According to an embodiment of the present disclosure, the feature extraction module 520 may perform feature extraction on the historical vibration acceleration signal and the current vibration acceleration signal to obtain the historical vibration feature signal and the current vibration feature signal associated with the wind turbine.

[0120] According to an embodiment of the present disclosure, the data prediction module 530 may execute: based on the vibration characteristic signals at historical moments and the vibration characteristic signals at the current moment, as well as the operation data and the operation environment data, predict the vibration characteristic signals at at least one future moment of the current moment.

[0121] Here, the at least one future time instant may be, for example, at least one future adjacent time instant, and further may be, for example, the next time instant or a combination of future adjacent or non-adjacent time instants.

[0122] As an example, the data prediction module 530 may include an operation of predicting the vibration characteristic signal of at least one future moment of the current moment based on the vibration characteristic signal of the historical moment, the vibration characteristic signal of the current moment, the operating data, and the operating environment data: using the vibration characteristic signal of the historical moment, the operating data, and the operating environment data to train a feature prediction model based on a neural network, and inputting the vibration characteristic signal of the current moment into the trained feature prediction model based on the neural network to obtain a vibration characteristic signal of at least one future moment.

[0123] For example, the operation of the data prediction module 530 using the historical vibration feature signals, operation data, and operation environment data to train a neural network-based feature prediction model may include the following operations 5301) to 5303):

[0124] In operation 5301), the vibration characteristic signal at the historical moment, the operation data and the operation environment data are input into the feature prediction model based on the neural network to obtain the predicted amplitude corresponding to the actual amplitude of the vibration characteristic signal at the historical moment.

[0125] In operation 5302), a prediction model loss function is calculated based on the actual amplitude and predicted amplitude of the vibration characteristic signal at the historical moment.

[0126] In operation 5303), the model parameters of the neural network-based feature prediction model are adjusted according to the prediction model loss function to obtain a trained neural network-based feature prediction model.

[0127] According to an embodiment of the present disclosure, the vibration suppression determination module 540 may execute: based on the vibration characteristic signal at the current moment and the vibration characteristic signal at at least one future moment, determining whether a vibration suppression condition associated with a preset operating condition is met, and in response to determining that the vibration suppression condition is met, executing a vibration suppression action associated with the preset operating condition.

[0128] Here, the types of vibrations suppressed by the vibration suppression action may include, for example, at least one of the following items: the resonant vibration of the impeller speed of the wind turbine when it is close to the preset speed and the natural frequency of the wind turbine, the negative damping vibration of the blades of the wind turbine when the blades are within the preset pitch angle range, the generator vibration of the wind turbine when the wind turbine is within the preset power range, the first-order vibration of the tower of the wind turbine under the first preset wind speed condition, the first-order vortex-induced vibration of the blades of the wind turbine under the second preset wind speed condition, etc.

[0129] As an example, the vibration suppression determination module 540 may include determining whether the vibration suppression condition associated with the preset operating condition is met based on the vibration characteristic signal at the current moment and the vibration characteristic signal at at least one future moment by comparing the amplitude of the vibration characteristic signal at the current moment with a first preset threshold and comparing the amplitude of the vibration characteristic signal at at least one future moment with a second preset threshold to determine whether the vibration suppression condition associated with the preset operating condition is met.

[0130] In addition, for example, the vibration suppression determination module 540, in response to determining that the vibration suppression conditions are met, performs an operation of a vibration suppression action associated with a preset operating condition, which may include: in response to the amplitude of the vibration characteristic signal at the current moment being less than a first preset threshold and the amplitude of the vibration characteristic signal at at least one future moment being greater than or equal to a second preset threshold, performing the vibration suppression action associated with the preset operating condition.

[0131] In addition, the vibration suppression device 500 may include a first data processing routine and a second data processing routine. Specifically, the first data processing routine is used to perform data processing tasks associated with the first and fifth duty cycles, and the second data processing routine is used to perform data processing tasks associated with the third and fourth duty cycles.

[0132] For example, the first data processing routine and the second data processing routine may be respectively deployed in different controllers or may be deployed in the same controller.

[0133] According to the embodiments of the present disclosure, data processing of different task cycles can be assigned to different routines according to different data processing requirements (for example, the amount of data to be processed and the real-time requirements, etc.), thereby achieving an improvement in processing speed and processing efficiency, and further ensuring the timely application of prediction results, thereby improving the effectiveness of vibration suppression.

[0134] It should be noted that the operations performed by the above modules can be compared with those in Figure 1 The relevant contents described are similar and will not be repeated here.

[0135] Next, refer to Figure 6 and Figure 7 The vibration suppression system of the wind turbine according to the embodiment of the present disclosure is described as an example. Figure 6 is a block diagram illustrating a vibration suppression system for a wind turbine generator according to one embodiment of the present disclosure, and Figure 7 is a block diagram illustrating a vibration suppression system for a wind turbine according to another embodiment of the present disclosure.

[0136] Reference Figure 6According to an embodiment of the present disclosure, a vibration suppression system for a wind turbine generator may include a vibration sensor 601 , a data acquisition device 602 , a controller 604 , an operating condition sensor and data acquisition device 603 , and an actuator 605 .

[0137] Specifically, the vibration sensor 601 is used to realize vibration perception of the whole machine or large components for a specific vibration mode. The data acquisition device 602 is used to collect vibration signals and transmit the signals to the controller 604. The working condition sensor and data acquisition device 603 include sensors for collecting working condition information such as rotational speed and corresponding data acquisition devices, and are used to transmit working condition signals to the controller 604. The controller 604 is used to send control decision module instructions to the actuator 605. The actuator 605 includes a wind turbine pitch system, a flow conversion system, a yaw system, etc., which are used to execute the actions determined by the control decision in the method disclosed in the present invention, and ultimately achieve active vibration suppression of the target vibration mode.

[0138] As an example, controller 604 includes two processor cores, namely core 1 and core 2. As an example, core 1 is used to perform various processes in the method of the present disclosure with a task cycle of T0, and core 2 is used to perform various processes in the method of the present disclosure with a task cycle of T1. In addition, communication between cores 1 and 2 is performed in the task cycle of T0.

[0139] Reference Figure 7 According to another embodiment of the present disclosure, a vibration suppression system for a wind turbine may include a vibration sensor 701 , a data acquisition device 702 , two controllers (shown as controller 1 and controller 2 ) 704 and 705 , an operating condition sensor and data acquisition device 703 , and an actuator 706 .

[0140] Specifically, the vibration sensor 701 is used to realize vibration perception of the whole machine or large components for a specific vibration mode. The data acquisition device 702 is used to collect vibration signals and transmit the signals to the controller 704. The working condition sensor and data acquisition device 703 include sensors and corresponding data acquisition devices for collecting working condition information such as rotational speed, and are used to transmit working condition signals to the controller 705. The controller 704 and the controller 705 can communicate in real time. The controller 705 is used to send control decision module instructions to the actuator 706. The actuator 706 includes a wind turbine pitch system, a flow conversion system, a yaw system, etc., which are used to execute the actions determined by the control decision in the method disclosed in the present invention, and ultimately achieve active vibration suppression of the target vibration mode.

[0141] As an example, the controller 704 is used to execute the processing tasks corresponding to the signal reading module, signal processing module, model M1 module, and model M0 module in the method of the present disclosure with an execution cycle of T0 task cycle, and its functions are the same as Figure 6The controller 705 is used to execute the processing tasks corresponding to the control decision module with the execution cycle of T0 task cycle in the method disclosed in the present invention, and its functions are similar to those of the core 2 in the embodiment of the present invention. Figure 6 The function of kernel 1 in is similar.

[0142] By adopting Figure 6 Kernel 2 in and Figure 7 The controller 704 in the embodiment can improve the computing power of the system and separate the operations related to the prediction model from the system input and control commands, thereby effectively improving the computing efficiency of the system while meeting the prediction method based on complex machine learning / deep learning algorithms.

[0143] Figure 8 is a block diagram illustrating a computing device 800 according to an embodiment of the present disclosure.

[0144] Reference Figure 8 According to an embodiment of the present disclosure, a computing device 800 may include at least one processor 810 and a memory 820. The at least one processor 810 may include (but is not limited to) a central processing unit (CPU), a digital signal processor (DSP), a microcomputer, a field programmable gate array (FPGA), a system on a chip (SoC), a microprocessor, an application-specific integrated circuit (ASIC), etc. The memory 820 may store computer-executable instructions to be executed by the at least one processor 810. The memory 820 includes a high-speed random access memory and / or a non-volatile computer-readable storage medium. When the at least one processor 810 executes the computer-executable instructions stored in the memory 820, the vibration suppression method for the wind turbine described above may be implemented.

[0145] The vibration suppression method of a wind turbine according to an embodiment of the present disclosure can be written as a computer program / instruction to form a computer program product and stored on a computer-readable storage medium. When the computer program / instruction is executed by a processor, the vibration suppression method of the wind turbine as described above can be implemented. When the instructions in the computer-readable storage medium are executed by at least one processor (for example, a processor of an electronic device / server), the at least one processor is enabled to execute the vibration suppression method of the wind turbine as described above. Examples of computer-readable storage media include: read-only memory (ROM), random access programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), flash memory, non-volatile memory, CD-ROM, CD-R, CD+R, CD-RW, CD+RW, DVD-ROM, DVD-R, DVD+R, DVD-RW, DVD+RW, DVD-RAM, BD-ROM, BD-R, BD-R LTH, BD-RE, Blu-ray or optical disk storage, hard disk drive (HDD), solid state drive (SSD), card storage (such as, multimedia card, secure digital (SD) card or ultra fast digital (XD) card), magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid state disk and any other device, any other device configured to store the computer program and any associated data, data files and data structures in a non-transitory manner and provide the computer program and any associated data, data files and data structures to a processor or computer so that the processor or computer can execute the computer program. In one example, the computer program and any associated data, data files and data structures are distributed on a networked computer system so that the computer program and any associated data, data files and data structures are stored, accessed and executed in a distributed manner by one or more processors or computers.

[0146] In addition, according to an embodiment of the present disclosure, a wind turbine generator set may be provided, which includes: the vibration suppression device for the wind turbine generator as described above and / or the computing device as described above.

[0147] According to the vibration suppression method and device of the wind turbine according to the embodiments of the present disclosure, fine control of wind turbine vibration is achieved through vibration characteristic prediction and optimization control strategy, which effectively suppresses the vibration of the wind turbine while reducing power generation loss and significantly improves the operating safety and stability of the wind turbine.

[0148] On the other hand, by deeply combining artificial intelligence methods, wind turbine vibration stability mechanisms and real-time control decision methods on the wind turbine edge side, the vibration suppression method and device of the wind turbine according to the present disclosure can be used to suppress wind turbine vibrations related to working conditions, thereby making the vibration suppression method and device of the wind turbine highly scalable.

[0149] Although some embodiments of the present disclosure have been disclosed and described, it will be understood by those skilled in the art that changes and modifications may be made to the embodiments without departing from the concept and spirit of the disclosure, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for suppressing vibration of a wind turbine, characterized in that: The vibration suppression method comprises: Acquire a vibration acceleration signal associated with the wind turbine and operating data and operating environment data of the wind turbine associated with a preset operating condition, wherein the vibration acceleration signal includes a historical vibration acceleration signal and a current vibration acceleration signal; Performing feature extraction on the vibration acceleration signal at the historical moment and the vibration acceleration signal at the current moment to obtain a vibration feature signal at the historical moment and a vibration feature signal at the current moment associated with the wind turbine; Predicting a vibration characteristic signal at at least one future moment of the current moment based on the vibration characteristic signal at the historical moment and the vibration characteristic signal at the current moment, the operating data, and the operating environment data; determining, based on the vibration characteristic signal at the current moment and the vibration characteristic signal at the at least one future moment, whether a vibration suppression condition associated with the preset operating condition is satisfied; In response to determining that the vibration suppression condition is satisfied, executing a vibration suppression action associated with the preset operating condition to proactively and actively control and adjust to avoid a vibration-sensitive operating condition; The step of determining whether the vibration suppression condition associated with the preset operating condition is satisfied based on the vibration characteristic signal at the current moment and the vibration characteristic signal at the at least one future moment includes: determining whether the vibration suppression condition associated with the preset operating condition is satisfied by comparing the amplitude of the vibration characteristic signal at the current moment with a first preset threshold value and comparing the amplitude of the vibration characteristic signal at the at least one future moment with a second preset threshold value. Among them, the step of executing the vibration suppression action associated with the preset operating condition in response to determining that the vibration suppression condition is met includes: executing the vibration suppression action associated with the preset operating condition in response to the amplitude of the vibration characteristic signal at the current moment being less than the first preset threshold and the amplitude of the vibration characteristic signal at at least one future moment being greater than or equal to the second preset threshold.

2. The vibration suppression method according to claim 1, wherein: The step of predicting a vibration characteristic signal at at least one future moment of the current moment based on the vibration characteristic signal at the historical moment, the vibration characteristic signal at the current moment, the operating data, and the operating environment data comprises: The vibration characteristic signal at the historical moment, the operating data and the operating environment data are used to train a feature prediction model based on a neural network, and the vibration characteristic signal at the current moment is input into the trained feature prediction model based on the neural network to obtain the vibration characteristic signal at at least one future moment.

3. The vibration suppression method according to claim 2, wherein: The step of using the vibration characteristic signal at the historical moment, the operation data and the operation environment data to train the feature prediction model based on the neural network includes: Inputting the vibration characteristic signal at the historical moment, the operating data, and the operating environment data into the neural network-based feature prediction model to obtain a predicted amplitude corresponding to the actual amplitude of the vibration characteristic signal at the historical moment; Calculating a prediction model loss function based on the actual amplitude and the predicted amplitude of the vibration characteristic signal at the historical moment; According to the prediction model loss function, the model parameters of the neural network-based feature prediction model are adjusted to obtain a trained neural network-based feature prediction model.

4. The vibration suppression method according to claim 1, wherein: The types of vibrations suppressed by the vibration suppression action include at least one of the following: resonant vibration of the impeller speed of the wind turbine when the impeller is close to the preset speed and the natural frequency of the wind turbine, negative damping vibration of the blades of the wind turbine when the blades are within the preset pitch angle range, generator vibration of the wind turbine when the wind turbine is within the preset power range, first-order vibration of the tower of the wind turbine under the first preset wind speed condition, and first-order vortex-induced vibration of the blades of the wind turbine under the second preset wind speed condition.

5. A vibration suppression device for a wind turbine, characterized in that: The vibration suppression device comprises: a data acquisition module configured to: acquire a vibration acceleration signal associated with the wind turbine and operating data and operating environment data of the wind turbine associated with a preset operating condition, wherein the vibration acceleration signal includes a historical vibration acceleration signal and a current vibration acceleration signal; a feature extraction module configured to: perform feature extraction on the vibration acceleration signal at the historical moment and the vibration acceleration signal at the current moment to obtain a vibration feature signal at the historical moment and a vibration feature signal at the current moment associated with the wind turbine; a data prediction module configured to: predict a vibration characteristic signal at at least one future moment of the current moment based on the vibration characteristic signal at the historical moment and the vibration characteristic signal at the current moment, the operation data, and the operation environment data; a vibration suppression determination module configured to: determine whether a vibration suppression condition associated with the preset operating condition is satisfied based on the vibration characteristic signal at the current moment and the vibration characteristic signal at the at least one future moment; and in response to determining that the vibration suppression condition is satisfied, execute a vibration suppression action associated with the preset operating condition to proactively and actively control and adjust to avoid a vibration-sensitive operating condition; The operation of determining whether the vibration suppression condition associated with the preset operating condition is satisfied based on the vibration characteristic signal at the current moment and the vibration characteristic signal at the at least one future moment includes: determining whether the vibration suppression condition associated with the preset operating condition is satisfied by comparing the amplitude of the vibration characteristic signal at the current moment with a first preset threshold value and comparing the amplitude of the vibration characteristic signal at the at least one future moment with a second preset threshold value. Among them, the operation of executing the vibration suppression action associated with the preset operating condition in response to determining that the vibration suppression condition is met includes: executing the vibration suppression action associated with the preset operating condition in response to the amplitude of the vibration characteristic signal at the current moment being less than the first preset threshold and the amplitude of the vibration characteristic signal at at least one future moment being greater than or equal to the second preset threshold.

6. The vibration suppression device according to claim 5, wherein: The relationship between the first task cycle in which the data acquisition module performs a data acquisition operation, the second task cycle in which the acquired data is stored, the third task cycle in which the feature extraction module performs feature extraction, the fourth task cycle in which the data prediction module performs a prediction operation, and the fifth task cycle in which the vibration suppression determination module determines to perform a vibration suppression action is as follows: The third duty cycle is equal to the fourth duty cycle, the first duty cycle is equal to the fifth duty cycle, the first duty cycle is smaller than the third duty cycle, and the third duty cycle is smaller than the second duty cycle.

7. The vibration suppression device according to claim 6, wherein: The vibration suppression device includes a first data processing routine and a second data processing routine, the first data processing routine is used to perform data processing tasks associated with the first task cycle and the fifth task cycle, and the second data processing routine is used to perform data processing tasks associated with the third task cycle and the fourth task cycle, and the first data processing routine and the second data processing routine are respectively deployed in different controllers or deployed in the same controller.

8. A computer-readable storage medium, characterized in that When the instructions in the computer-readable storage medium are executed by at least one processor, the at least one processor is prompted to perform the vibration suppression method for a wind turbine according to any one of claims 1 to 4.

9. A computer program product comprising computer instructions, characterized in that When the computer instructions are executed by at least one processor, the at least one processor is prompted to perform the vibration suppression method for a wind turbine according to any one of claims 1 to 4.

10. A computing device, characterized in that The computing device includes: at least one processor; and at least one memory storing computer-executable instructions, wherein the computer-executable instructions, when executed by the at least one processor, prompt the at least one processor to execute the vibration suppression method for a wind turbine according to any one of claims 1 to 4.

11. A wind turbine generator set, characterized in that: The wind turbine generator set comprises: The vibration suppression device for a wind turbine according to any one of claims 5 to 7, or The computing device of claim 10.

Images