Variable pitch processing method, apparatus and system based on fan, and device and medium

Abstract

Provided in the present application are a variable pitch processing method, apparatus and system based on a fan, and a device and a medium. The method comprises: acquiring wind speed data within a first preset period of time, and inputting the wind speed data into a neural network, so as to acquire one-dimensional implicit vectors corresponding to the wind speed data; converting the one-dimensional implicit vectors into a two-dimensional implicit vector matrix on the basis of the time when the wind speed data is acquired, and performing spatial pyramid convolution on the two-dimensional implicit vector matrix, so as to acquire a first feature value of the wind speed data; and acquiring discrimination complex vector containers corresponding to a fan, comparing the first feature value with a second feature value in each discrimination complex vector container, and when it is determined that the distance between the first feature value and the second feature value is less than a preset distance threshold value, sending a variable pitch instruction to a variable pitch system of the fan, such that the variable pitch system performs variable pitch processing. Therefore, the problem in the prior art of the service life of a wind turbine being shortened due to the fact that the wind turbine cannot vary a pitch in a timely manner resulting in a power system fluctuating violently is solved.

Description

A pitch control method, device, equipment, medium and system based on a wind turbine

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on December 7, 2023, with application number 202311680062.7 and application name “A method, device, equipment, medium and system for variable pitch processing based on a wind turbine”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of data processing technology, and in particular to a pitch control method, device, equipment, medium and system based on a wind turbine. Background Art

[0003] In a wind power generation system, the speed of a wind turbine is primarily determined by the wind speed. When the current wind speed is higher than the rated wind speed, variable pitch technology is used. A proportional integral (PI) controller is used to adjust the pitch angle to control the speed of the wind turbine. This ensures that the speed of the wind turbine is as close as possible to the target speed corresponding to the rated wind speed, thus preventing damage to the wind turbine caused by excessive speed.

[0004] However, wind speed fluctuations are instantaneous, sudden, and uncertain. By the time the PI controller determines that the wind turbine's actual speed deviates from the target speed and issues a pitch change command, the wind speed may have already changed several times. The wind turbine's actual speed may have mismatched the current wind speed, preventing the pitch from being adjusted to the appropriate angle in a timely manner. If the wind turbine fails to adjust its pitch to the appropriate angle in a timely manner when the current wind speed exceeds the rated speed, it will cause severe fluctuations in the power system and shorten the wind turbine's lifespan.

[0005] Summary of the Invention

[0006] The present application provides a wind turbine-based pitch change processing method, device, equipment, medium and system for solving the problem in the prior art that wind turbines cannot change pitch in time, resulting in severe fluctuations in the power system and shortened wind turbine life.

[0007] In a first aspect, the present application provides a pitch processing method based on a wind turbine, comprising: obtaining wind speed data within a first preset time, and inputting the wind speed data into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data; according to the acquisition time of the wind speed data, transforming the one-dimensional latent vector into a two-dimensional latent vector matrix, and performing spatial pyramid convolution on the two-dimensional latent vector matrix to obtain a first eigenvalue of the wind speed data; obtaining a discriminant complex vector container corresponding to the wind turbine, and comparing the first eigenvalue with the second eigenvalue in the discriminant complex vector container, and when it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, sending a pitch instruction to the pitch system of the wind turbine to enable the pitch system to perform pitch processing.

[0008] In a specific embodiment, the transforming of the one-dimensional latent vector into a two-dimensional latent vector matrix based on the acquisition time of the wind speed data includes: classifying the one-dimensional latent vector based on the acquisition time of the wind speed data to obtain a one-dimensional latent vector corresponding to daytime and a one-dimensional latent vector corresponding to nighttime; and combining the one-dimensional latent vector corresponding to daytime and the one-dimensional latent vector corresponding to nighttime to obtain a two-dimensional latent vector matrix.

[0009] In a specific embodiment, performing spatial pyramid convolution on the two-dimensional latent vector matrix to obtain the first eigenvalue of the wind speed data includes: obtaining multiple convolution kernel templates, and sequentially convolving the two-dimensional latent vector matrix with the multiple convolution kernel templates to obtain the first eigenvalue of the wind speed data; wherein the sizes of the multiple convolution kernel templates are sequentially reduced.

[0010] In a specific embodiment, obtaining the discriminant complex vector container corresponding to the wind turbine and comparing the first eigenvalue with the second eigenvalue in the discriminant complex vector container include: obtaining multiple discriminant complex vector containers corresponding to the wind turbine, and the multiple discriminant complex vector containers correspond one-to-one to multiple pitch angles respectively; comparing the first eigenvalue with the second eigenvalue in each of the discriminant complex vector containers respectively; then, when it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, sending a pitch instruction to the pitch system of the wind turbine so that the pitch system performs pitch processing, including: for each of the discriminant complex vector containers, when it is determined that the distance between the first eigenvalue and the second eigenvalue in the discriminant complex vector container is less than a preset distance threshold, sending a pitch instruction corresponding to the discriminant complex vector container to the pitch system of the wind turbine, so that the pitch system performs pitch processing according to the pitch angle corresponding to the discriminant complex vector container.

[0011] In a specific embodiment, the discriminant complex vector container is obtained as follows: historical wind speed data within a second preset time is obtained, and the historical wind speed data is divided into historical data and future data according to the acquisition time of the historical wind speed data; characteristic values ​​of the historical data and the characteristic values ​​of the future data are obtained respectively, and characteristic values ​​of excess data are selected from the characteristic values ​​of the future data; and the discriminant complex vector container corresponding to the historical wind speed data is obtained using a loss function based on the characteristic values ​​of the historical data, the characteristic values ​​of the future data, and the characteristic values ​​of the excess data.

[0012] In a specific embodiment, obtaining the eigenvalues ​​of the historical data includes: inputting the historical data into a neural network to obtain a one-dimensional latent vector corresponding to the historical data; transforming the one-dimensional latent vector corresponding to the historical data into a two-dimensional latent vector matrix corresponding to the historical data based on the acquisition time of the historical data, and performing spatial pyramid convolution on the two-dimensional latent vector matrix corresponding to the historical data to obtain the eigenvalues ​​of the historical data.

[0013] In the second aspect, the present application provides a pitch processing device based on a wind turbine, including: an acquisition module for acquiring wind speed data within a first preset time, and inputting the wind speed data into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data; a processing module for transforming the one-dimensional latent vector into a two-dimensional latent vector matrix according to the acquisition time of the wind speed data, and performing spatial pyramid convolution on the two-dimensional latent vector matrix to obtain a first eigenvalue of the wind speed data; the processing module is also used to obtain a discriminant complex vector container corresponding to the wind turbine, and compare the first eigenvalue with the second eigenvalue in the discriminant complex vector container, and when it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, a pitch instruction is sent to the pitch system of the wind turbine to enable the pitch system to perform pitch processing.

[0014] In a third aspect, the present application provides an electronic device comprising: a processor, a memory, and a communication interface; the memory is used to store executable instructions of the processor; wherein the processor is configured to execute the wind turbine-based pitch processing method described in the first aspect by executing the executable instructions.

[0015] In a fourth aspect, the present application provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the pitch processing method based on a wind turbine described in the first aspect is implemented.

[0016] In a fifth aspect, the present application provides a wind turbine-based pitch processing system, comprising: a wind turbine-based pitch processing device and a wind turbine as described in any one of the second to fourth aspects; wherein the wind turbine includes a pitch system.

[0017] The present application provides a pitch processing method, device, equipment, medium and system based on a wind turbine. The method includes: obtaining wind speed data within a first preset time, and inputting the wind speed data into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data; transforming the one-dimensional latent vector into a two-dimensional latent vector matrix according to the acquisition time of the wind speed data, and performing spatial pyramid convolution on the two-dimensional latent vector matrix to obtain a first eigenvalue of the wind speed data; obtaining a discriminant complex vector container corresponding to the wind turbine, and comparing the first eigenvalue with the second eigenvalue in the discriminant complex vector container, and when the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, sending a pitch command to the pitch system of the wind turbine to enable the pitch system to perform pitch processing. Compared with the prior art, which issues a pitch change instruction only when it is determined that the actual speed of the wind turbine deviates from the target speed, the pitch change processing method based on the wind turbine of the present application inputs the acquired wind speed data into the neural network to obtain a one-dimensional latent vector, transforms the one-dimensional latent vector into a two-dimensional latent vector matrix according to the acquisition time, and performs spatial pyramid convolution on the two-dimensional latent vector matrix. It can obtain significant eigenvalues ​​of the wind speed data, compare the eigenvalues ​​with the eigenvalues ​​in the discriminant complex vector container, and when the distance between the eigenvalues ​​is less than the distance threshold, it is determined that the wind speed is about to exceed the rated wind speed, and then the pitch change instruction is sent to the pitch change system in time for pitch change processing, which solves the problem in the prior art that the wind turbine cannot change the pitch in time, resulting in severe fluctuations in the power system and shortened life of the wind turbine. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] FIG1 is a flow chart of a first embodiment of a pitch control method for a wind turbine according to the present application;

[0020] FIG2 is a flow chart of a second embodiment of a pitch control method for a wind turbine provided by the present application;

[0021] FIG3 is a flow chart of a third embodiment of a pitch control method for a wind turbine provided by the present application;

[0022] FIG4 is a flow chart of a fourth embodiment of a pitch control method for a wind turbine provided by the present application;

[0023] FIG5 is a schematic structural diagram of an embodiment of a pitch control device based on a wind turbine provided by the present application;

[0024] FIG6 is a schematic structural diagram of an electronic device provided in this application. DETAILED DESCRIPTION

[0025] To make the purpose, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments made by ordinary technicians in this field based on the inspiration of these embodiments fall within the scope of protection of this application.

[0026] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the numbers used in this way are interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] In a wind power generation system, the speed of a wind turbine is primarily determined by the wind speed. When the current wind speed is higher than the rated wind speed, variable pitch technology is used. A proportional integral (PI) controller is used to adjust the pitch angle to control the speed of the wind turbine. This ensures that the speed of the wind turbine is as close as possible to the target speed corresponding to the rated wind speed, thus preventing damage to the wind turbine caused by excessive speed.

[0028] However, wind speed fluctuations are instantaneous, sudden, and uncertain. By the time the PI controller determines that the wind turbine's actual speed deviates from the target speed and issues a pitch change command, the wind speed may have already changed several times. The wind turbine's actual speed may have mismatched the current wind speed, preventing the pitch from being adjusted to the appropriate angle in a timely manner. If the wind turbine fails to adjust its pitch to the appropriate angle in a timely manner when the current wind speed exceeds the rated speed, it will cause severe fluctuations in the power system and shorten the wind turbine's lifespan.

[0029] Based on the above technical problems, the technical conception process of this application is as follows: How to provide a pitch change processing method to solve the problem in the prior art that wind turbines cannot change pitches in time, resulting in severe fluctuations in the power system and shortened wind turbine life.

[0030] The technical solution of the present application is described in detail below through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0031] FIG1 is a flow chart of a first embodiment of a pitch change processing method for a wind turbine provided by the present application. Referring to FIG1 , the pitch change processing method for a wind turbine specifically includes the following steps:

[0032] Step S101: Obtain wind speed data within a first preset time, and input the wind speed data into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data.

[0033] In this embodiment, wind speed data within a first preset time period may be acquired, where the wind speed data is vector data consisting of wind speed and wind direction.

[0034] The wind speed data is input into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data. Exemplarily, the neural network may be a long short-term memory network (LSTM). The wind speed data is input into the neural network to obtain a one-dimensional latent vector corresponding to the wind speed data.

[0035] Step S102: According to the acquisition time of the wind speed data, the one-dimensional latent vector is transformed into a two-dimensional latent vector matrix, and spatial pyramid convolution is performed on the two-dimensional latent vector matrix to obtain the first eigenvalue of the wind speed data.

[0036] In this embodiment, the one-dimensional latent vector can be transformed into a two-dimensional latent vector matrix based on the time when the wind speed data was acquired. For example, the one-dimensional latent vector can be classified into two groups of one-dimensional latent vectors based on the time when the wind speed data was acquired: one-dimensional latent vectors corresponding to daytime and one-dimensional latent vectors corresponding to nighttime. The two groups of one-dimensional latent vectors are then combined to transform into a two-dimensional latent vector matrix.

[0037] A spatial pyramid convolution is performed on the two-dimensional latent vector matrix to obtain the first eigenvalue of the wind speed data. Specifically, the two-dimensional latent vector matrix can be convolved with convolution kernel templates of successively smaller sizes to obtain the first eigenvalue of the wind speed data. The successively smaller sizes of the convolution kernel templates can reduce the resolution of the two-dimensional latent vector, thereby extracting more abstract and representative features.

[0038] Step S103: Obtain the discriminant complex vector container corresponding to the wind turbine, and compare the first eigenvalue with the second eigenvalue in the discriminant complex vector container. When it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, send a pitch control instruction to the pitch control system of the wind turbine so that the pitch control system performs pitch control processing.

[0039] In this embodiment, a discriminant complex vector container corresponding to the wind turbine may be obtained, where the discriminant complex vector container includes a second characteristic value used to determine that the wind speed is about to exceed the rated wind speed.

[0040] The first eigenvalue is compared with the second eigenvalue in the discriminant complex vector container. For example, the first eigenvalue and the second eigenvalue can be compared by a dictionary query. When the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, it is determined that the wind speed is about to exceed the rated wind speed, and a pitch control instruction is sent to a pitch control system of the wind turbine, causing the pitch control system to perform pitch control processing.

[0041] In this embodiment, wind speed data within a first preset time is obtained, and the wind speed data is input into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data; according to the acquisition time of the wind speed data, the one-dimensional latent vector is transformed into a two-dimensional latent vector matrix, and spatial pyramid convolution is performed on the two-dimensional latent vector matrix to obtain a first eigenvalue of the wind speed data; a discriminant complex vector container corresponding to the wind turbine is obtained, and the first eigenvalue is compared with the second eigenvalue in the discriminant complex vector container. When it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, a pitch control instruction is sent to the pitch control system of the wind turbine to enable the pitch control system to perform pitch control processing. Compared with the prior art, which issues a pitch change instruction only when it is determined that the actual speed of the wind turbine deviates from the target speed, the pitch change processing method based on the wind turbine of the present application inputs the acquired wind speed data into the neural network to obtain a one-dimensional latent vector, transforms the one-dimensional latent vector into a two-dimensional latent vector matrix according to the acquisition time, and performs spatial pyramid convolution on the two-dimensional latent vector matrix. It can obtain significant eigenvalues ​​of the wind speed data, compare the eigenvalues ​​with the eigenvalues ​​in the discriminant complex vector container, and when the distance between the eigenvalues ​​is less than the distance threshold, it is determined that the wind speed is about to exceed the rated wind speed, and then the pitch change instruction is sent to the pitch change system in time for pitch change processing, which solves the problem in the prior art that the wind turbine cannot change the pitch in time, resulting in severe fluctuations in the power system and shortened life of the wind turbine.

[0042] FIG2 is a flow chart of a second embodiment of a pitch control method for a wind turbine provided by the present application. Based on the embodiment shown in FIG1 , step S102 specifically includes the following steps:

[0043] Step S201: Classify the one-dimensional latent vector according to the acquisition time of the wind speed data to obtain the one-dimensional latent vector corresponding to the day and the one-dimensional latent vector corresponding to the night.

[0044] Step S202: Combine the one-dimensional latent vector corresponding to daytime and the one-dimensional latent vector corresponding to nighttime to obtain a two-dimensional latent vector matrix.

[0045] Step S203: Obtain multiple convolution kernel templates, and convolve the two-dimensional latent vector matrix with the multiple convolution kernel templates in sequence to obtain the first eigenvalue of the wind speed data.

[0046] Among them, the sizes of multiple convolution kernel templates are reduced in sequence.

[0047] In this embodiment, the one-dimensional latent vector can be classified according to the time when the wind speed data was acquired. This can reflect the interaction between weather and meteorological conditions in the same time period of each cycle, which is conducive to exploring the periodic changes in wind speed. Specifically, based on whether the wind speed data was acquired during the day or at night, the one-dimensional latent vector is divided into a one-dimensional latent vector corresponding to the day and a one-dimensional latent vector corresponding to the night. For example, the one-dimensional latent vector can be "Day 1, Night 1, Day 2, Night 2, ... ", and based on the time when the wind speed data was acquired, the one-dimensional latent vector is divided into a one-dimensional latent vector corresponding to the day "Day 1, Day 2, ... " and a one-dimensional latent vector corresponding to the night "Night 1, Night 2, ... ".

[0048] For example, the one-dimensional latent vector corresponding to daytime and the one-dimensional latent vector corresponding to nighttime can be combined to form a two-dimensional latent vector matrix as shown below, which facilitates aggregation of the same features of different periods:

[0049] After forming the two-dimensional latent vector matrix, multiple convolution kernel templates can be obtained, and the two-dimensional latent vector matrix is ​​convolved with the multiple convolution kernel templates in turn to obtain the first eigenvalue of the wind speed data.

[0050] Specifically, the sizes of these multiple convolution kernel templates decrease sequentially. For example, there may be three convolution kernel templates, with sizes of 7*7, 5*5, and 3*3, respectively. By convolving the two-dimensional latent vector matrix with successively smaller convolution kernel templates, the resolution of the two-dimensional latent vector decreases. This allows for the extraction of more abstract and representative features while effectively blurring and removing period-independent redundant information.

[0051] In this embodiment, the one-dimensional latent vectors are classified and processed according to the acquisition time of the wind speed data, which can effectively reflect the interaction relationship between weather and meteorology in the same time period of each cycle, which is conducive to exploring the periodic changes of wind speed; the classified one-dimensional latent vectors are combined into a two-dimensional latent vector matrix, which can aggregate the same features of different cycles; the two-dimensional latent vector matrix is ​​convolved with multiple convolution kernel templates of successively decreasing sizes to obtain the first eigenvalue of the wind speed data, which can extract more abstract and representative features and effectively blur and eliminate redundant information that is irrelevant to the cycle. This provides a prerequisite for accurately predicting that the wind speed is about to exceed the rated wind speed, and then sending a pitch change instruction to the pitch change system for pitch change processing in a timely manner. It further solves the problem in the prior art that the wind turbine cannot change the pitch in time, resulting in severe fluctuations in the power system and shortened wind turbine life.

[0052] FIG3 is a flow chart of a third embodiment of a pitch control method for a wind turbine provided by the present application. Based on the embodiments shown in FIG1 and FIG2 , referring to FIG3 , the step S103 specifically includes the following steps:

[0053] Step S301: obtaining a plurality of discriminant complex vector containers corresponding to the wind turbine, wherein the plurality of discriminant complex vector containers correspond one-to-one to a plurality of pitch angles respectively.

[0054] In this embodiment, the discriminant complex vector container includes a second characteristic value for determining whether the wind speed is about to exceed the rated wind speed. The wind turbine may correspond to multiple discriminant complex vector containers, each corresponding to a different excess target. Different excess targets also correspond to different pitch angles.

[0055] For example, a wind speed exceeding the rated wind speed by 10% to 20% can be set as the first excess target, a wind speed exceeding the rated wind speed by 20% to 30% can be set as the second excess target, a wind speed exceeding the rated wind speed by 30% to 40% can be set as the third excess target, a wind speed exceeding the rated wind speed by 40% to 50% can be set as the fourth excess target, and a wind speed exceeding the rated wind speed by 50% can be set as the fifth excess target.

[0056] A plurality of discriminant complex vector containers corresponding to the wind turbine may be obtained, and each discriminant complex vector container corresponds to a different pitch angle.

[0057] Step S302: Compare the first eigenvalue with the second eigenvalue in each discriminant complex vector container respectively.

[0058] Step S303: For each discriminant complex vector container, when it is determined that the distance between the first eigenvalue and the second eigenvalue in the discriminant complex vector container is less than a preset distance threshold, a pitch instruction corresponding to the discriminant complex vector container is sent to the pitch system of the wind turbine, so that the pitch system performs pitch processing according to the pitch angle corresponding to the discriminant complex vector container.

[0059] In this embodiment, the first eigenvalue of the wind speed data is compared with the second eigenvalue in each discriminant complex vector container. For example, the first eigenvalue and the second eigenvalue can be compared by dictionary query.

[0060] For each discriminant complex vector container, when it is determined that the distance between the first eigenvalue and the second eigenvalue in the discriminant complex vector container is less than a preset distance threshold, a pitch control instruction corresponding to the discriminant complex vector container is sent to a pitch control system of the wind turbine. The pitch control instruction may include a pitch angle corresponding to the discriminant complex vector container. The pitch control system may perform pitch control processing based on the pitch angle corresponding to the discriminant complex vector container.

[0061] In this embodiment, there can be multiple discriminant complex vector containers, each corresponding to a different pitch angle. Thus, the first eigenvalue of the wind speed data can be compared with the second eigenvalue in each discriminant complex vector container to determine the excess target that the current wind speed meets, and a pitch instruction corresponding to the excess target is sent, so that the pitch system performs pitch processing according to the corresponding pitch angle. In this way, on the basis of timely pitching when it is determined that the wind speed is about to exceed the rated wind speed, more targeted pitch processing can be performed. This further solves the problem in the prior art that the wind turbine cannot change the pitch in time, resulting in severe fluctuations in the power system and shortened life of the wind turbine.

[0062] FIG4 is a flow chart of a fourth embodiment of a pitch control method for a wind turbine provided by the present application. Based on the embodiments shown in FIG1 to FIG3 above, referring to FIG4 , a method for obtaining a complex vector container includes the following steps:

[0063] Step S401: Obtain historical wind speed data within a second preset time, and divide the historical wind speed data into historical data and future data based on the time when the historical wind speed data was obtained. In this embodiment, the historical wind speed data within the second preset time can be obtained, and the historical wind speed data is vector data consisting of wind speed and wind direction.

[0064] The historical wind speed data can be divided into historical data and future data according to the acquisition time of the historical wind speed data. For example, the second preset time is 10 days, and the historical wind speed data of the first 8 days can be used as historical data, and the historical wind speed data of the last 2 days can be used as future data.

[0065] Step S402: acquiring the characteristic values ​​of the historical data and the characteristic values ​​of the future data respectively, and selecting the characteristic values ​​of the excess data from the characteristic values ​​of the future data.

[0066] Step S403: according to the characteristic value of the historical data, the characteristic value of the future data and the characteristic value of the excess data, a loss function is used to obtain a discriminant complex vector container corresponding to the historical wind speed data.

[0067] In this embodiment, the characteristic values ​​of the historical data and the characteristic values ​​of the future data may be obtained separately.

[0068] Specifically, the historical data can be input into a neural network to obtain a one-dimensional latent vector corresponding to the historical data. Exemplarily, the neural network can be an LSTM. By inputting the historical data into the neural network, a one-dimensional latent vector corresponding to the historical data can be obtained.

[0069] Based on the acquisition time of the historical data, the one-dimensional latent vector corresponding to the historical data is transformed into a two-dimensional latent vector matrix corresponding to the historical data. For example, based on the acquisition time of the historical data, the one-dimensional latent vectors corresponding to the historical data can be classified into two groups of one-dimensional latent vectors: one-dimensional latent vectors corresponding to daytime and one-dimensional latent vectors corresponding to nighttime. The two groups of one-dimensional latent vectors are then combined to transform into a two-dimensional latent vector matrix corresponding to the historical data.

[0070] A spatial pyramid convolution is performed on the two-dimensional latent vector matrix corresponding to the historical data to obtain the eigenvalues ​​of the historical data. Specifically, the two-dimensional latent vector matrix can be convolved with convolution kernel templates of successively smaller sizes to obtain the eigenvalues ​​of the historical data. The successively smaller sizes of the convolution kernel templates can reduce the resolution of the two-dimensional latent vector corresponding to the historical data, thereby extracting more abstract and representative features.

[0071] Specifically, the future data can be input into a neural network to obtain a one-dimensional latent vector corresponding to the future data; according to the acquisition time of the future data, the one-dimensional latent vector corresponding to the future data is transformed into a two-dimensional latent vector matrix corresponding to the future data, and spatial pyramid convolution is performed on the two-dimensional latent vector matrix corresponding to the future data to obtain the eigenvalue of the future data.

[0072] After obtaining the characteristic values ​​of the future data, characteristic values ​​of the excess data may be selected from the characteristic values ​​of the future data. For example, characteristic values ​​of the excess data may be selected from the characteristic values ​​of the future data according to an excess target. The excess target may be, for example, that the wind speed exceeds the rated wind speed by 10% to 20%.

[0073] In this embodiment, the discriminant complex vector container corresponding to the historical wind speed data can be obtained by using a loss function according to the characteristic values ​​of the historical data, the characteristic values ​​of the future data, and the characteristic values ​​of the excess data.

[0074] Specifically, the eigenvalues ​​of the historical data form the discriminant complex vector container d g , the eigenvalues ​​of future data form the discriminant complex vector container f a , the eigenvalues ​​of the excess data form the discriminant complex vector container f g , the loss function is as follows:

[0075] Among them, the numerator is d g With f g The similarity degree measurement result, the denominator is d g With f a The similarity degree measurement result of , all is the eigenvalue of all future data, and goal is the eigenvalue of all excess data. The closer the loss function approaches 0, the closer the numerator and denominator are. The discriminant complex vector container d g The better you are at distinguishing excess targets from non-excess targets.

[0076] In this embodiment, multiple discriminant complex vector containers may be formed according to different excess targets. For example, a wind speed exceeding the rated wind speed by 10% to 20% may be set as the first excess target, a wind speed exceeding the rated wind speed by 20% to 30% may be set as the second excess target, a wind speed exceeding the rated wind speed by 30% to 40% may be set as the third excess target, a wind speed exceeding the rated wind speed by 40% to 50% may be set as the fourth excess target, and a wind speed exceeding the rated wind speed by 50% may be set as the fifth excess target.

[0077] For example, the feature values ​​of historical data can be divided into shared features and unique features based on different excess targets. Shared features are common to all excess targets, i.e., the features that determine the significance of all excess targets, and they constitute a shared feature container. Unique features are distinguishing features that discriminate between different excess targets, and they constitute multiple unique feature sub-containers. Each unique feature sub-container and the shared feature container are then combined to form multiple discriminant complex vector containers.

[0078] In this embodiment, a loss function is used to obtain a discriminant complex vector container corresponding to the historical wind speed data based on the eigenvalues ​​of historical data, the eigenvalues ​​of future data, and the eigenvalues ​​of excess data. This effectively distinguishes excess targets from non-excess targets, providing a prerequisite for accurately predicting when the wind speed is about to exceed the rated wind speed and promptly sending pitch control commands to the pitch control system for pitch control. This further solves the existing problem of wind turbines failing to change pitch in a timely manner, resulting in severe power system fluctuations and shortened wind turbine lifespans.

[0079] The following are device embodiments of the present application, which can be used to implement the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

[0080] FIG5 is a schematic structural diagram of an embodiment of a pitch processing device based on a wind turbine provided by the present application; as shown in FIG5 , the pitch processing device based on a wind turbine 50 includes: an acquisition module 51 and a processing module 52. The acquisition module 51 is used to acquire wind speed data within a first preset time, and input the wind speed data into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data; the processing module 52 is used to transform the one-dimensional latent vector into a two-dimensional latent vector matrix according to the acquisition time of the wind speed data, and perform spatial pyramid convolution on the two-dimensional latent vector matrix to obtain the first eigenvalue of the wind speed data; the processing module 52 is also used to acquire a discriminant complex vector container corresponding to the wind turbine, and compare the first eigenvalue with the second eigenvalue in the discriminant complex vector container. When it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, a pitch instruction is sent to the pitch system of the wind turbine to enable the pitch system to perform pitch processing.

[0081] The wind turbine-based pitch processing device provided in the embodiment of the present application can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar and will not be repeated here.

[0082] In one possible implementation, the processing module 52 is specifically configured to classify the one-dimensional latent vector according to the acquisition time of the wind speed data to obtain a one-dimensional latent vector corresponding to the day and a one-dimensional latent vector corresponding to the night; and combine the one-dimensional latent vector corresponding to the day and the one-dimensional latent vector corresponding to the night to obtain a two-dimensional latent vector matrix.

[0083] In a possible implementation scheme, the processing module 52 is specifically used to obtain multiple convolution kernel templates, and convolve the two-dimensional latent vector matrix with the multiple convolution kernel templates in turn to obtain the first eigenvalue of the wind speed data; wherein the sizes of the multiple convolution kernel templates are reduced in turn.

[0084] The wind turbine-based pitch processing device provided in the embodiment of the present application can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar and will not be repeated here.

[0085] In a possible implementation scheme, the processing module 52 is specifically used to obtain multiple discriminant complex vector containers corresponding to the wind turbine, and the multiple discriminant complex vector containers correspond one-to-one to multiple pitch angles respectively; the first eigenvalue is compared with the second eigenvalue in each discriminant complex vector container respectively; for each discriminant complex vector container, when it is determined that the distance between the first eigenvalue and the second eigenvalue in the discriminant complex vector container is less than a preset distance threshold, a pitch instruction corresponding to the discriminant complex vector container is sent to the pitch system of the wind turbine, so that the pitch system performs pitch processing according to the pitch angle corresponding to the discriminant complex vector container.

[0086] The wind turbine-based pitch processing device provided in the embodiment of the present application can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar and will not be repeated here.

[0087] In a possible implementation scheme, the processing module 52 is also used to obtain historical wind speed data within a second preset time, and divide the historical wind speed data into historical data and future data according to the acquisition time of the historical wind speed data; obtain the characteristic values ​​of the historical data and the characteristic values ​​of the future data respectively, and select the characteristic values ​​of the excess data from the characteristic values ​​of the future data; and use the loss function to obtain the discriminant complex vector container corresponding to the historical wind speed data based on the characteristic values ​​of the historical data, the characteristic values ​​of the future data and the characteristic values ​​of the excess data.

[0088] In one possible implementation, the processing module 52 is specifically used to input the historical data into a neural network to obtain a one-dimensional latent vector corresponding to the historical data; based on the acquisition time of the historical data, the one-dimensional latent vector corresponding to the historical data is transformed into a two-dimensional latent vector matrix corresponding to the historical data; and spatial pyramid convolution is performed on the two-dimensional latent vector matrix corresponding to the historical data to obtain the eigenvalues ​​of the historical data.

[0089] The wind turbine-based pitch processing device provided in the embodiment of the present application can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar and will not be repeated here.

[0090] Figure 6 is a schematic diagram of the structure of an electronic device provided in this application. As shown in Figure 6, the electronic device 60 includes: a processor 61, a memory 62, and a communication interface 63; wherein the memory 62 is used to store executable instructions of the processor 61; the processor 61 is configured to execute the technical solution of any of the aforementioned method embodiments by executing the executable instructions.

[0091] Optionally, the memory 62 can be independent or integrated with the processor 61.

[0092] Optionally, when the memory 62 is a device independent of the processor 61 , the electronic device 60 may further include: a bus 64 for connecting the above devices.

[0093] The electronic device is used to execute the technical solution in any of the aforementioned method embodiments, and its implementation principles and technical effects are similar and will not be repeated here.

[0094] An embodiment of the present application further provides a readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the technical solution provided by any of the aforementioned embodiments is implemented.

[0095] An embodiment of the present application further provides a wind turbine-based pitch processing system, comprising a wind turbine-based pitch processing device and a wind turbine as provided in any of the aforementioned embodiments; wherein the wind turbine includes a pitch system.

[0096] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A pitch control method based on a wind turbine, characterized in that: include: Obtaining wind speed data within a first preset time, and inputting the wind speed data into a neural network to obtain a one-dimensional latent vector corresponding to the wind speed data; According to the acquisition time of the wind speed data, transform the one-dimensional latent vector into a two-dimensional latent vector matrix, and perform spatial pyramid convolution on the two-dimensional latent vector matrix to obtain a first eigenvalue of the wind speed data; Obtain a discriminant complex vector container corresponding to the wind turbine, and compare the first eigenvalue with the second eigenvalue in the discriminant complex vector container; when it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, send a pitch control instruction to the pitch control system of the wind turbine so that the pitch control system performs pitch control processing.

2. The pitch processing method based on a wind turbine according to claim 1, characterized in that: The step of transforming the one-dimensional latent vector into a two-dimensional latent vector matrix according to the acquisition time of the wind speed data includes: Classifying the one-dimensional latent vector according to the acquisition time of the wind speed data to obtain a one-dimensional latent vector corresponding to the day and a one-dimensional latent vector corresponding to the night; The one-dimensional latent vector corresponding to the daytime and the one-dimensional latent vector corresponding to the nighttime are combined to obtain a two-dimensional latent vector matrix.

3. The pitch processing method based on a wind turbine according to claim 1 or 2, characterized in that: The performing spatial pyramid convolution on the two-dimensional latent vector matrix to obtain the first eigenvalue of the wind speed data includes: Acquire multiple convolution kernel templates, and sequentially convolve the two-dimensional latent vector matrix with the multiple convolution kernel templates to obtain a first eigenvalue of the wind speed data; Among them, the sizes of the multiple convolution kernel templates decrease successively.

4. The pitch processing method based on a wind turbine according to claim 1 or 2, characterized in that: The obtaining the discriminant complex vector container corresponding to the wind turbine and comparing the first eigenvalue with the second eigenvalue in the discriminant complex vector container includes: Acquire a plurality of discriminant complex vector containers corresponding to the wind turbine, wherein the plurality of discriminant complex vector containers correspond one-to-one to a plurality of variable pitch angles respectively; respectively comparing the first eigenvalue with the second eigenvalue in each of the discriminant complex vector containers; When it is determined that the distance between the first characteristic value and the second characteristic value is less than a preset distance threshold, sending a pitch change instruction to the pitch change system of the wind turbine so that the pitch change system performs pitch change processing, including: For each of the discriminant complex vector containers, when it is determined that the distance between the first eigenvalue and the second eigenvalue in the discriminant complex vector container is less than a preset distance threshold, a pitch instruction corresponding to the discriminant complex vector container is sent to the pitch system of the wind turbine, so that the pitch system performs pitch processing according to the pitch angle corresponding to the discriminant complex vector container.

5. The pitch processing method based on a wind turbine according to claim 1 or 2, characterized in that: The method for obtaining the discriminant complex vector container is: Acquire historical wind speed data within a second preset time, and divide the historical wind speed data into historical data and future data according to the acquisition time of the historical wind speed data; respectively obtaining the characteristic values ​​of the historical data and the characteristic values ​​of the future data, and selecting the characteristic values ​​of the excess data from the characteristic values ​​of the future data; According to the characteristic values ​​of the historical data, the characteristic values ​​of the future data and the characteristic values ​​of the excess data, a discriminant complex vector container corresponding to the historical wind speed data is obtained by using a loss function.

6. The pitch processing method based on a wind turbine according to claim 5, characterized in that: The obtaining of the characteristic value of the historical data includes: Inputting the historical data into a neural network to obtain a one-dimensional latent vector corresponding to the historical data; According to the acquisition time of the historical data, the one-dimensional latent vector corresponding to the historical data is transformed into a two-dimensional latent vector matrix corresponding to the historical data, and spatial pyramid convolution is performed on the two-dimensional latent vector matrix corresponding to the historical data to obtain the eigenvalue of the historical data.

7. A variable pitch processing device based on a wind turbine, characterized in that: include: An acquisition module, used to acquire wind speed data within a first preset time, and input the wind speed data into a neural network to acquire a one-dimensional latent vector corresponding to the wind speed data; A processing module, configured to transform the one-dimensional latent vector into a two-dimensional latent vector matrix according to the acquisition time of the wind speed data, and perform spatial pyramid convolution on the two-dimensional latent vector matrix to obtain a first eigenvalue of the wind speed data; The processing module is also used to obtain the discriminant complex vector container corresponding to the wind turbine, and compare the first eigenvalue with the second eigenvalue in the discriminant complex vector container. When it is determined that the distance between the first eigenvalue and the second eigenvalue is less than a preset distance threshold, a pitch control instruction is sent to the pitch control system of the wind turbine so that the pitch control system performs pitch control processing.

8. An electronic device, characterized in that: include: Processor, memory, communication interface; The memory is used to store executable instructions of the processor; The processor is configured to execute the wind turbine-based pitch processing method according to any one of claims 1 to 6 by executing the executable instructions.

9. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the pitch processing method based on a wind turbine according to any one of claims 1 to 6 is implemented.

10. A variable pitch processing system based on a wind turbine, characterized in that: include: A wind turbine-based pitch processing device and a wind turbine as described in any one of claims 7 to 9; wherein the wind turbine comprises a pitch system.

Images