Variable pitch motor life prediction method, device, equipment, medium and product

Abstract

The invention discloses a life prediction method, device and equipment of a variable-pitch motor, a medium and a product, and relates to the technical field of wind power generation. According to the embodiment of the invention, relevant parameters associated with the service life of the variable-pitch motor under the current time sequence and the designed service life value of the variable-pitch motor are obtained, and active power, environmental parameters and operating parameters are input into a service life loss prediction model to obtain a service life loss prediction value of the variable-pitch motor in the next time sequence of the current time sequence. According to the method, the influence of the running state of the variable-pitch motor on the service life is considered, the influence of damage of other related parts on the variable-pitch motor is also considered, and the considered factors are more comprehensive, so that when the trained service life loss prediction model is used for carrying out service life prediction, the prediction efficiency is improved. The life loss value of the variable-pitch motor in the next time sequence can be predicted more accurately, and then the residual life value of the variable-pitch motor in the next time sequence can be predicted more accurately.

Description

Technical Field

[0001] This application relates to the field of wind power generation technology, and in particular to a method, apparatus, equipment, medium and product for predicting the lifespan of a pitch motor. Background Technology

[0002] Pitch motors are the core support for the stable operation of the entire wind power generation system, playing a crucial role. By adjusting the pitch angle of the turbine blades, they effectively capture and utilize wind energy, thereby achieving precise control of the turbine's power. If a pitch motor malfunctions, it will impact the entire system's operation, causing severe economic losses and energy supply disruptions. Therefore, predicting the lifespan of pitch motors is of great significance.

[0003] Current life prediction schemes cannot accurately predict the remaining lifespan of a pitch motor at a future point in time. Summary of the Invention

[0004] This application provides a method, apparatus, device, medium, and product for predicting the lifespan of a pitch motor, which can accurately predict the remaining lifespan of the pitch motor at a future point in time.

[0005] In a first aspect, embodiments of this application provide a method for predicting the lifespan of a pitch motor, including:

[0006] Obtain relevant parameters associated with the lifespan of the pitch motor and the design lifespan value of the pitch motor under the current time series. The relevant parameters include the active power of the wind turbine generator, environmental parameters, and the operating parameters of the pitch system of the wind turbine generator.

[0007] Active power, environmental parameters, and operating parameters are input into the life loss prediction model to obtain the predicted life loss value of the pitch motor in the next time series of the current time series. The life loss prediction model is trained by taking the relevant parameters of the samples in the historical time series as input and the sample state matrix and the sample life loss value of the pitch motor in the historical time series as output. The sample state matrix includes the sample operating loss information of the pitch motor in the historical time series, as well as the sample correlation loss information of related components to the pitch motor in the historical time series. The related components are the components associated with the pitch motor.

[0008] Based on the difference between the designed life value and the predicted life loss value, the remaining life value of the pitch motor in the next time series is predicted.

[0009] Secondly, embodiments of this application provide a life prediction device for a pitch motor, comprising:

[0010] The acquisition module is used to acquire relevant parameters related to the lifespan of the pitch motor and the design lifespan value of the pitch motor under the current time series. The relevant parameters include the active power of the wind turbine generator, environmental parameters, and the operating parameters of the pitch system of the wind turbine generator.

[0011] The prediction module is used to input active power, environmental parameters, and operating parameters into the life loss prediction model to obtain the predicted life loss value of the pitch motor in the next time series of the current time series. The life loss prediction model is trained by taking the relevant parameters of the samples in the historical time series as input and the sample state matrix and the sample life loss value of the pitch motor in the historical time series as output. The sample state matrix includes the sample operating loss information of the pitch motor in the historical time series, as well as the sample correlation loss information of related components to the pitch motor in the historical time series. The related components are the components associated with the pitch motor. Based on the difference between the design life value and the predicted life loss value, the remaining life value of the pitch motor in the next time series is predicted.

[0012] Thirdly, embodiments of this application provide an electronic device, including:

[0013] processor;

[0014] Memory is used to store computer program instructions;

[0015] When the computer program instructions are executed by the processor, the method described in the first aspect is implemented.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the method described in the first aspect.

[0017] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the method described in the first aspect.

[0018] In this embodiment, relevant parameters associated with the lifespan of the pitch motor and its design lifespan value are obtained in the current time series. These relevant parameters include the active power of the wind turbine generator, environmental parameters, and operating parameters of the pitch system of the wind turbine generator. The active power, environmental parameters, and operating parameters are input into the lifespan loss prediction model to obtain the predicted lifespan loss value of the pitch motor in the next time series of the current time series. The lifespan loss prediction model is trained by taking sample relevant parameters from historical time series as input and taking a sample state matrix and sample lifespan loss values ​​of the pitch motor in historical time series as output. The sample state matrix includes sample operating loss information of the pitch motor in historical time series and sample correlation loss information of related components to the pitch motor in historical time series. Based on the difference between the design lifespan value and the predicted lifespan loss value, the remaining lifespan value of the pitch motor in the next time series is predicted. During the model training phase, this embodiment trains the model based on relevant parameters associated with the pitch motor's lifespan, the pitch motor's own operating loss information, and the correlation loss information of related components. This means that it considers not only the impact of the pitch motor's own operating state on its lifespan but also the impact of damage to other related components on the pitch motor, taking into account a more comprehensive range of factors. Therefore, when using the trained lifespan loss prediction model for lifespan prediction, it can more accurately predict the lifespan loss value of the pitch motor in the next time series, and thus more accurately predict the remaining lifespan value of the pitch motor in the next time series. Attached Figure Description

[0019] The features, advantages, and technical effects of exemplary embodiments of this application will now be described with reference to the accompanying drawings.

[0020] Figure 1 A flowchart illustrating a method for predicting the lifespan of a pitch motor, as provided in an embodiment of this application;

[0021] Figure 2 A flowchart of another method for predicting the lifespan of a pitch motor provided in an embodiment of this application;

[0022] Figure 3 A structural diagram of a converter model provided in an embodiment of this application;

[0023] Figure 4 A flowchart of another method for predicting the lifespan of a pitch motor provided in an embodiment of this application;

[0024] Figure 5 A schematic diagram of a remaining lifetime trend curve provided for an embodiment of this application;

[0025] Figure 6 A schematic diagram illustrating another remaining lifetime trend provided in an embodiment of this application;

[0026] Figure 7 A structural diagram of a pitch motor life prediction device provided in an embodiment of this application;

[0027] Figure 8 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0028] The features and exemplary embodiments of various aspects of this application will now be described in detail. Numerous specific details are set forth in the following detailed description to provide a comprehensive understanding of this application. However, it will be apparent to those skilled in the art that this application can be implemented without requiring some of these specific details. The following description of embodiments is merely intended to provide a better understanding of this application by illustrating examples. In the accompanying drawings and the following description, at least some well-known structures and techniques are not shown to avoid unnecessarily obscuring the application; and, for clarity, the dimensions of some structures may be exaggerated. Furthermore, the features, structures, or characteristics described below can be combined in any suitable manner in one or more embodiments.

[0029] The directional terms used in the following description refer to the directions shown in the figures and are not intended to limit the specific structure of the cable-stayed tower and wind turbine generator set of this application. It should also be noted in the description of this application that, unless otherwise explicitly specified and limited, the terms "installation" and "connection" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to direct connections or indirect connections. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0030] As mentioned above, the pitch motor is the core support for the stable operation of the entire wind power generation system and plays a crucial role. Once the pitch motor malfunctions, it will affect the operation of the entire system, causing serious economic losses and energy supply interruptions. Therefore, predicting the lifespan of the pitch motor is of great significance.

[0031] Current life prediction schemes ignore the mutual influence between wind turbine components, resulting in poor accuracy of prediction results and low guidance for on-site operation and maintenance.

[0032] The inventors discovered during the operation of wind turbines that the operating environment is a crucial factor affecting the lifespan of pitch motors. For example, pitch motors are prone to failure under harsh weather conditions such as high operating temperatures, strong winds, and heavy rain. Furthermore, the way the motor is used also impacts its lifespan. Improper usage, such as frequent start-stop cycles and overload operation, accelerates wear and tear, leading to a shorter lifespan.

[0033] Based on the electrical principles of the pitch system, the inventors also discovered that damage and failure of related components such as relay failure, cooling fan jamming, pitch reducer oil leakage, and pitch drive failure can also affect the lifespan of the pitch motor.

[0034] Based on this, the embodiments of this application provide a method, apparatus, equipment, medium and product for predicting the lifespan of a pitch motor. By combining the operating environment of the wind turbine generator set, the operating losses of the pitch motor itself and the related losses of the components of the pitch motor, the remaining lifespan of the pitch motor at a future point in time can be accurately predicted.

[0035] The following describes the method, apparatus, equipment, medium, and product for predicting the lifespan of a pitch motor provided in this application, with reference to specific embodiments.

[0036] Figure 1 The flowchart illustrates a method for predicting the lifespan of a pitch motor, as provided in this application embodiment. This method can be applied to electronic devices, which can be a part of a wind turbine, such as a whole-machine controller, or devices that are independent of the wind turbine but communicate with it, such as laptops, desktops, servers, etc.

[0037] like Figure 1 As shown, the life prediction method for the pitch motor may include the following steps:

[0038] S110. Obtain the relevant parameters associated with the lifespan of the pitch motor and the design lifespan value of the pitch motor in the current time series.

[0039] The relevant parameters include the active power of the wind turbine generator set, environmental parameters, and the operating parameters of the wind turbine generator set's pitch system.

[0040] S120. Input the active power, environmental parameters and operating parameters into the life loss prediction model to obtain the predicted life loss value of the pitch motor in the next time series of the current time series.

[0041] The life loss prediction model is trained by taking the relevant parameters of the samples in the historical time series as input and the sample state matrix and the sample life loss value of the pitch motor in the historical time series as output. The sample state matrix includes the sample operating loss information of the pitch motor in the historical time series, as well as the sample correlation loss information of the related components to the pitch motor in the historical time series. The related components are the components associated with the pitch motor.

[0042] S130. Based on the difference between the design life value and the predicted life loss value, predict the remaining life value of the pitch motor in the next time series.

[0043] During the model training phase, this embodiment trains the model based on relevant parameters associated with the pitch motor's lifespan, the pitch motor's own operating loss information, and the correlation loss information of related components. This means that it considers not only the impact of the pitch motor's own operating state on its lifespan but also the impact of damage to other related components on the pitch motor, taking into account a more comprehensive range of factors. Therefore, when using the trained lifespan loss prediction model for lifespan prediction, it can more accurately predict the lifespan loss value of the pitch motor in the next time series, and thus more accurately predict the remaining lifespan value of the pitch motor in the next time series.

[0044] The above steps are explained in detail below:

[0045] In S110, the time series here can be a sequence containing multiple sampling time points. The intervals between the sampling time points can be the same or different. In order to more accurately predict the lifespan of the pitch motor, the intervals between adjacent sampling time points can be the same. The current time series can correspond to the current time period, which can be, for example, the current day (24 hours). That is, lifespan prediction can be performed on a daily basis.

[0046] For example, the current time period can be divided into one-hour intervals, resulting in 24 sampling time points. In other words, the current time series can include 24 sampling time points. In practical applications, the length of the current time period and the division method can be varied, thus improving the flexibility and controllability of the prediction method.

[0047] The relevant parameters associated with the lifespan of the pitch motor may include, but are not limited to, the active power of the wind turbine generator, environmental parameters, and the operating parameters of the pitch system. Among them, environmental parameters can be parameters of the wind turbine generator's operating environment, such as wind speed, wind direction, temperature, and heavy rainfall.

[0048] The operating parameters of the pitch system may include, for example, the operating parameters of the pitch motor and the operating parameters of the components related to the pitch motor. The operating parameters of the pitch motor may include, for example, the pitch angle, the temperature, current, and speed of the pitch motor. The operating parameters of the components related to the pitch motor may include, for example, the temperature of the pitch inverter, the temperature of the pitch capacitor, and the voltage of the pitch capacitor.

[0049] The parameters related to the lifespan of the pitch motor can be obtained by a Supervisory Control and Data Acquisition (SCADA) system, which is part of the wind turbine generator set. That is, the embodiments of this application can use the existing data acquisition devices of the wind turbine generator set to obtain the above parameters without adding additional data acquisition devices, thereby reducing costs.

[0050] For example, the SCADA system can obtain the parameter values ​​of the above parameters at each sampling time point in the current time series, providing a data basis for subsequent lifetime prediction.

[0051] The design life of a pitch motor is the life value at the time of manufacture. Different pitch motors may have different design life values.

[0052] In S120, the lifetime loss prediction model is a trained model used to predict the lifetime loss value of the pitch motor in the future time series. In this embodiment of the application, the lifetime loss prediction model is trained by taking the relevant parameters of the samples in the historical time series as input and taking the sample state matrix and the sample lifetime loss value of the pitch motor in the historical time series as output. The specific training process can be found in the following embodiment.

[0053] The historical time series can be one or more time series preceding the current time series. To improve the training effect of the lifetime loss prediction model, there can be multiple historical time series. The duration of each historical time series can be the same as that of the current time series. The number of sampling time points and the interval between adjacent sampling time points in each historical time series can also be the same as those in the current time series, which can further improve the training effect of the model.

[0054] The relevant parameters of the sample are the relevant parameters of the wind turbine generator set in the historical time series. For details, please refer to the above embodiment.

[0055] The state matrix here is used to characterize the operating loss information of the pitch motor and the correlation loss information of related components. The dimensions of the state matrix can be customized.

[0056] The operating loss information of the pitch motor can include at least one of power-limited pitch and feathering, and the associated loss information of related components can include at least one of electromagnetic brake failure, cooling fan failure, pitch drive failure, pitch reducer failure, and relay failure. In practical applications, the operating loss information of the pitch motor can also include pitch motor failure, and the associated loss information can also include encoder failure, pitch motor cable failure, etc.

[0057] For example, the state matrix can be represented in the following form:

[0058]

[0059] The above state matrix, using six state code bits (0-5), represents a single column. For example, each column corresponds to a loss information point. The first column (state code bit 0) corresponds to invalid data, indicating a fault shutdown; the second column (state code bit 1) corresponds to power-limited pitch; the third column (state code bit 2) corresponds to feathering; the fourth column (state code bit 3) corresponds to the electromagnetic brake not engaging; the fifth column (state code bit 4) corresponds to a damaged cooling fan; and the sixth column (state code bit 5) corresponds to a damaged drive. In practical applications, this state matrix can also include other operating states besides those mentioned above.

[0060] Each row of the state matrix corresponds to a sampling time point, meaning the number of rows in the state matrix is ​​the same as the number of sampling time points in the time series. a1-f5 in the state matrix each correspond to a state value, which is either 0 or 1. 0 indicates the absence of loss information, and 1 indicates the presence of loss information.

[0061] Taking the example of a pitch motor operating under limited power pitch control in a certain historical time series without electromagnetic brake activation, the above state matrix can be represented as follows:

[0062]

[0063] The matrix contains values ​​in state code bits 1 and 3, indicating that the pitch motor performed power-limited pitch control at the first three sampling points of the historical time series, and that the electromagnetic brake was not engaged at the last two sampling points. Based on the interval between sampling points, the duration of power-limited pitch control operation of the pitch motor in the historical time series, as well as the number of times the electromagnetic brake was not engaged, can be further determined.

[0064] The state value of each state encoding bit in the sample state matrix can be determined based on the actual operating state of the pitch motor and related components in the historical time series.

[0065] This application's embodiments predict the lifespan loss value of the pitch motor from two perspectives: the pitch motor itself and related components. This makes the prediction results more objective and accurate. At the same time, a state matrix is ​​introduced, which covers the operating loss information of the pitch motor and the correlation loss information caused by the damage of related components. This matrix is ​​more interpretable and has greater guidance significance for on-site operation and maintenance.

[0066] By inputting the active power, environmental parameters, and operating parameters of the current time series obtained above into the life loss prediction model, the life loss value of the pitch motor in the next time series can be predicted. Simultaneously, the state matrix corresponding to the next time series can also be obtained, providing a basis for subsequent statistical analysis of the operating status of each state.

[0067] In S130, the remaining life of the pitch motor in the next time series can be predicted based on the difference between the design life value and the predicted life loss value.

[0068] For example, the difference between the design life value and the predicted life loss value can be determined as the remaining life value of the pitch motor in the next time series.

[0069] For example, the difference between the design life value and the predicted life loss value can be adjusted, and the adjustment result can be used as the remaining life value of the pitch motor in the next time series. For instance, a certain value can be added or subtracted from the above difference to obtain a range, and this range can be determined as the remaining life value of the pitch motor in the next time series.

[0070] The following is combined Figure 2 The training process of the lifetime loss prediction model in the above embodiments is explained. Figure 2 and Figure 1 The difference is that, Figure 2 exist Figure 1 Before S110, there are also S210-S220.

[0071] S210. Obtain the training sample set.

[0072] The training sample set includes multiple training samples. Each training sample includes sample-related parameters under the corresponding historical time series, as well as the sample state matrix and sample lifetime loss value corresponding to the sample-related parameters.

[0073] Each historical time series can contain multiple sampling time points and multiple related parameters. Therefore, each data sample is an m*n matrix, where m represents the number of sampling time points in the historical time series and n represents the categories of related parameters. For example, if each historical time series contains 24 sampling time points and the related parameters include 9 categories, then each data sample is a 24*9 matrix.

[0074] Taking the relevant parameters of the sample, including the active power of the wind turbine generator, wind speed, pitch angle, pitch motor temperature, pitch motor current, pitch motor speed, pitch inverter temperature, pitch capacitor temperature, and pitch capacitor voltage, as an example, the lifetime loss value of each data sample can be labeled to obtain the corresponding training sample. Each training sample can be shown in Table 1. However, the state matrix corresponding to each data sample is not shown in Table 1.

[0075] Table 1 Examples of training samples

[0076]

[0077]

[0078] As shown in the table above, each lifetime loss value corresponds to one data sample.

[0079] S220. Train the transformer model using training samples until the training stopping condition is met, and obtain the lifetime loss prediction model.

[0080] The transformer model here can be a deep learning model based on the transformer model. This transformer model is trained by taking sample-related parameters as input and outputting the sample state matrix and sample lifetime loss value. Training stopping conditions could be, for example, reaching a preset number of iterations or the transformer model's loss value converging.

[0081] For example, such as Figure 3 As shown, the transformer model 300 may include an encoding module 301, a linear transformation module 302, and a decoding module 303. Both the encoding module 301 and the decoding module 303 consist of multiple identical stacked layers, which are functionally similar.

[0082] Each layer of the encoding module 301 contains two key sub-layers: a multi-head self-attention sub-layer 3011 and a feedforward neural network sub-layer 3012. Each sub-layer is followed by a residual connection and layer normalization 3013, which helps to alleviate the gradient vanishing problem in deep neural networks.

[0083] The output of the encoding module 301 is the state matrix. The state matrix is ​​input into the linear transformation module 302, which performs a linear transformation on the state matrix to obtain the life loss assessment value of the pitch motor in the corresponding time series.

[0084] Compared to the encoding module 301, the decoding module 303 adds a masked multi-head attention mechanism sublayer 3031. In addition, the decoding module 303 also includes a multi-head self-attention sublayer 3032 and a feedforward neural network sublayer 3033, each followed by a residual connection and layer normalization 3034. The decoding module 303 also includes a fully connected sublayer 3035, through which the predicted lifetime loss value of the pitch motor in the next time series can be obtained.

[0085] The above structure enables the Transformer model to efficiently process input and output sequences, capture long-distance dependencies, and support parallel computing, greatly improving training speed and the model's expressive power.

[0086] Based on the above model structure, S220 may include the following steps, for example:

[0087] The first sample-related parameters of the first training sample are input into the encoding module to obtain the first state matrix output by the encoding module, and the first state matrix is ​​input into the linear transformation module to obtain the first lifetime loss assessment value output by the linear transformation module.

[0088] The second training sample's relevant parameters are input into the decoding module to obtain the first lifetime loss prediction value output by the decoding module. The second training sample and the first training sample are training samples with adjacent time series.

[0089] Based on the first lifetime loss assessment value and the first sample lifetime loss value corresponding to the first training sample, a first loss function value is determined; based on the first state matrix and the sample state matrix corresponding to the first training sample, a second loss function value is determined; and based on the first lifetime loss prediction value and the sample lifetime loss value corresponding to the second training sample, a third loss function value is determined.

[0090] Based on the first loss function value, the second loss function value, and the third loss function value, the transformer model is trained until the training stopping condition is met, thus obtaining the lifetime loss prediction model.

[0091] The first and second training samples are training samples that are adjacent in time. That is, the time period corresponding to the first training sample and the time period corresponding to the second training sample are adjacent time periods. For example, each training sample corresponds to one day (24 hours). If the time period corresponding to the first training sample is August 1, 00:00:00-August 1, 23:59:59, the time period corresponding to the second training sample can be August 2, 00:00:00-August 2, 23:59:59.

[0092] The first state matrix is ​​the state matrix output by the encoding module based on the first sample related parameters of the first training sample, and the first lifetime loss assessment value is the result obtained by linear transformation of the first state matrix by the linear transformation module.

[0093] The first lifetime loss prediction value is the result obtained by the decoding module based on the second sample correlation parameters of the second training sample. The specific processing procedures of each module are not limited in the embodiments of this application.

[0094] The first sample lifetime loss value is the actual lifetime loss value generated by the first training sample. Based on the first lifetime loss assessment value output by the linear transformation module and the actual first sample lifetime loss value, the first loss function value can be determined.

[0095] The sample state matrix is ​​the actual state matrix corresponding to the first training sample. Based on the first state matrix output by the encoding module and the sample state matrix actually corresponding to the first training sample, the value of the second loss function can be determined.

[0096] The lifetime loss value corresponding to the second training sample is the actual lifetime loss value generated by the second training sample. Based on the first lifetime loss prediction value output by the decoding module and the actual lifetime loss value generated by the second training sample, the third loss function value can be determined.

[0097] The determination process for the first loss function value, the second loss function value, and the third loss function value is similar. This application does not limit the specific form of the loss function. For example, the mean squared error loss function, the cross-entropy loss function, the mean absolute error loss function, etc., can be used to calculate the above loss function values.

[0098] Based on the first loss function value, the second loss function value, and the third loss function value, the model parameters of the transformer model can be updated until the training stopping condition is met.

[0099] For example, the first, second, and third loss function values ​​can be weighted and summed to obtain the final loss function value. The model parameters of the transformer model are then updated based on this final loss function value until the training stopping condition is met. The weights of each loss function value can be set according to the scenario, requirements, experience, etc.

[0100] This application's embodiment adds a linear transformation module to the traditional Transformer model, and updates the model's parameters by combining the loss value generated from the state matrix, the output of the linear transformation module, and the output of the decoding module. This fully considers the impact of damage to the pitch motor itself and related components on the pitch motor's lifespan, thereby improving the model's training and prediction performance.

[0101] Figure 4A flowchart illustrating another method for predicting the lifespan of a pitch motor provided in this application embodiment. Figure 4 and Figure 1 The difference is that, Figure 4 Prior to S110, there was also S410, and Figure 1 S130 in the middle can be further refined into Figure 4 S420-S430.

[0102] S410. Obtain the cumulative value of the life loss assessment of the pitch motor from the start of its service to the current time series.

[0103] For example, the life loss assessment value of the pitch motor in each time series can be obtained by the linear transformation module of the life loss prediction model trained above. By accumulating the life loss assessment values ​​of the pitch motor in each time series from the start of service to the current time series, the cumulative life loss assessment value can be obtained.

[0104] For example, the above S410 may include the following steps:

[0105] The relevant parameters of each historical time series of the pitch motor from the start of its service to the current time series are input into the encoding module of the life loss prediction model to obtain the historical state matrix corresponding to each historical time series output by the encoding module.

[0106] For each historical state matrix, the historical state matrix is ​​input into the linear transformation module of the lifetime loss prediction model to obtain the lifetime loss assessment value corresponding to the historical time series output by the linear transformation module.

[0107] The cumulative lifetime loss assessment value is obtained by accumulating the lifetime loss assessment values ​​for each historical time series.

[0108] For example, the time period from the start of service of the pitch motor to the current time series can be processed according to the time series processing method described above to obtain multiple historical time series. These multiple historical time series can be partially or completely the same as the historical time series used for training the model described above.

[0109] For each historical time series, the relevant parameters of the corresponding samples can be input into the encoding module of the lifetime loss prediction model, and the linear transformation module will output the lifetime loss assessment value. By accumulating the lifetime loss assessment values ​​of each historical time series, the cumulative lifetime loss assessment value can be obtained.

[0110] This application embodiment utilizes a trained life loss prediction model, combined with relevant parameters from samples of each historical time series, to obtain the life loss assessment value of the pitch motor in each historical time series. This leads to the cumulative life loss assessment value of the pitch motor from the start of its service to the current time series, providing a more accurate basis for subsequent prediction of the remaining life of the pitch motor.

[0111] S420. Based on the difference between the design life value and the cumulative life loss assessment value, predict the remaining life assessment value of the pitch motor in the current time series.

[0112] For example, Where T represents the design life value of the pitch motor, N represents the number of time series included in the pitch motor from the start of service to the current time series, t(i) represents the life loss assessment value of the pitch motor in the i-th time series, and Se represents the remaining life assessment value of the pitch motor in the current time series.

[0113] S430. Based on the difference between the remaining life assessment value and the life loss prediction value, predict the remaining life value of the pitch motor in the next time series.

[0114] For example, Sp = Se - p(j), where Sp represents the remaining lifetime value of the pitch motor in the next time series, and p(j) represents the predicted lifetime loss value of the pitch motor in the j-th time series, which is the next time series after the current time series.

[0115] Based on the design life value and the cumulative life loss assessment value, this application embodiment combines the life loss prediction value output by the life loss prediction model to predict the remaining life of the pitch motor in the next time series. It fully considers the operating loss of the pitch motor itself and the associated loss of related components, thus improving the accuracy of the prediction results.

[0116] In some embodiments, after “inputting the sample-related parameters corresponding to each historical time series of the pitch motor from the start of its service to the current time series into the encoding module of the life loss prediction model to obtain the historical state matrix corresponding to each historical time series output by the encoding module”, the life prediction method for the pitch motor may further include the following steps:

[0117] Based on the state values ​​at the corresponding state positions in each historical state matrix, the first cumulative result of the operating loss information of the pitch motor from the start of its service to the current time series and the second cumulative result of the associated loss information are determined.

[0118] Displays the first and second cumulative results.

[0119] The state values ​​here are 0 or 1. For each historical time series, a historical state matrix can be obtained. Based on the state values ​​of this historical state matrix, it can be determined whether the pitch motor has operating losses in each historical time series, and if so, the duration or frequency of those losses. Similarly, it can also be determined whether related components have correlated losses in each historical time series, and if so, the duration or frequency of those losses.

[0120] The first cumulative result here may include the cumulative duration or number of times the pitch motor has operating losses, and the second cumulative result may include the cumulative duration or number of times related components have associated losses.

[0121] For example, taking the operating loss information of the pitch motor, which includes power-limited pitch and feathering, as an example, the first cumulative result may include the cumulative duration of power-limited pitch and the cumulative number of feathering operations.

[0122] For example, taking the associated loss information of related components, including the damage to the cooling fan of the pitch motor, the failure of the electromagnetic brake to engage, and the damage to the pitch drive, as an example, the second cumulative result may include the cumulative duration of the cooling fan damage, the cumulative number of times the electromagnetic brake was not engaged, and the cumulative duration of the pitch drive damage.

[0123] Taking the historical state matrix as an example:

[0124]

[0125] As can be seen from the matrix above, the pitch motor experienced power-limited pitch and electromagnetic brake failure in this historical time series. Assuming the interval between two adjacent sampling time points is t, we can obtain that the cumulative duration of power-limited pitch of the pitch motor in this historical time series is 2t, and the cumulative number of times the electromagnetic brake failed to engage in this historical time series is 2.

[0126] The present application does not limit the display method of the first cumulative result and the second cumulative result. For example, the cumulative results of operating loss information and related operating loss in each historical time series can be displayed by means of curves or charts, as well as the total first cumulative result and the second cumulative result of each historical time series. In this way, the operation and maintenance personnel can understand the operating status of the pitch motor and related components in each historical time series, make preparations in advance, and avoid power generation loss caused by failure downtime.

[0127] For example, the final first and second cumulative results can also be displayed in text form.

[0128] This application not only includes the operating loss information of the pitch motor and the correlation loss information of related components in the state matrix, making the predicted lifespan more accurate, but also obtains the cumulative operating results of the pitch motor and related components in various historical time series and from the start of service to the current time series through the state matrix. This helps maintenance personnel to understand the operating status of the pitch motor and related components during the operation of the wind turbine, thereby enabling them to plan maintenance work scientifically in advance, rationally allocate human and material resources, greatly reduce the huge economic losses caused by long-term downtime due to sudden pitch motor failure, effectively ensure the continuity of power generation, avoid power outages due to motor failure, and ensure that electricity can be continuously and stably delivered to users.

[0129] In some embodiments, after S130, the pitch motor life prediction method may further include the following steps:

[0130] Based on the remaining lifespan of the pitch motor in a preset time series, the remaining lifespan trend curve of the pitch motor is displayed.

[0131] If the remaining lifetime value is less than the remaining lifetime threshold of the corresponding preset time series, or if the rate of change of the remaining lifetime trend curve within a preset time period is greater than the rate of change threshold, an early warning message will be output. The preset time period includes at least one preset time series.

[0132] The preset time series here can be a time series between the start of service of the pitch motor and a certain time series after the current time series. That is, the embodiments of this application can display the remaining life assessment value of the pitch motor in the historical time series and the remaining life prediction result in a certain time series in the future.

[0133] For example, the remaining lifetime corresponding to each time series can be connected to obtain the remaining lifetime trend curve of the pitch motor. The curve can also display the remaining lifetime of the pitch motor in each time series, so that relevant personnel can understand the remaining lifetime of the pitch motor in each historical time series.

[0134] The remaining lifetime threshold can be set according to the scenario, requirements, etc., and the remaining lifetime thresholds for different time series can be the same or different.

[0135] The rate of change of the remaining life trend curve within a preset time period can be the rate of change of the remaining life trend curve between two adjacent time series or multiple consecutive time series. When multiple time series are included, the rate of change of the remaining life trend curve between two adjacent time series can be determined separately, or the overall rate of change of the remaining life from the first time series to the last time series can be obtained based on the remaining life of the first and last time series among multiple time series.

[0136] If the rate of change of the remaining life trend curve within a preset time period is greater than the rate of change threshold, it indicates that the remaining life trend curve has a sharp downward trend within the preset time period. The magnitude of the rate of change threshold can be set according to scenarios, requirements, etc., and different pitch motors can correspond to different rate of change thresholds in different time series.

[0137] For example, when the remaining lifespan of the pitch motor is less than the remaining lifespan threshold of the corresponding time series, or when the rate of change of the remaining lifespan trend curve within a preset time period is greater than the rate of change threshold, a warning message can be output. The warning message can be output in various ways, such as through voice, light, SMS, email, or telephone. The warning message may include, but is not limited to, the warning time series, remaining lifespan, and the cumulative operating results of the pitch motor and related components.

[0138] Among them, the rate of change of the life trend curve within a preset time period is greater than the rate of change threshold. This can be defined as the rate of change of the life trend curve between any two time series within the preset time period being greater than the rate of change threshold. In other words, within the preset time period, as long as the rate of change of the life trend curve between any two time series is greater than the rate of change threshold, the rate of change of the life trend curve within the preset time period is considered to be greater than the rate of change threshold.

[0139] The form of the remaining life trend curve can be found in [reference]. Figure 5 In this schematic diagram, in addition to displaying the remaining life trend curve, the cumulative operating loss information of the pitch motor and related components obtained in the above embodiment can also be displayed.

[0140] For example, the diagram may also include a remaining lifetime concern threshold line, also known as a remaining lifetime warning line, which can force the wind turbine to shut down when the remaining lifetime is lower than the remaining lifetime corresponding to the remaining lifetime concern threshold line.

[0141] For example, such as Figure 6 As shown, the remaining lifetime of the pitch motor in each time series can also be displayed as a bar chart, with the rest of the display content remaining unchanged. In this case, the predicted remaining lifetime of the pitch motor in future time series can be displayed differently from the estimated remaining lifetime in historical time series, making it easier for maintenance personnel to view. For example... Figure 6 The shaded area represents the remaining lifetime assessment value of the pitch motor in the historical time series, while the unshaded area represents the predicted remaining lifetime value of the pitch motor in the future time series.

[0142] This application embodiment can plot the remaining life trend chart based on the remaining life value of the pitch motor in each time series, and output early warning information when the remaining life is lower than the remaining life threshold or when the remaining life has a sharp downward trend, so that operation and maintenance personnel can make corresponding strategies in a timely manner to avoid power generation loss caused by failure downtime.

[0143] This application's embodiments can predict the lifespan of a pitch motor based on SCADA data without requiring additional sensors, reducing the data access cost for lifespan prediction. Simultaneously, a state matrix is ​​introduced during the modeling process. This matrix includes the operating loss characteristics of the pitch motor and the associated loss characteristics of other related components, enhancing the interpretability and objectivity of the prediction results. Furthermore, the length of the input time series of the model can be adjusted according to actual needs, thereby improving the flexibility and controllability of the prediction method.

[0144] Based on the same inventive concept, this application also provides a life prediction device for a pitch motor, which is described below in conjunction with... Figure 7 The life prediction device for a pitch motor provided in the embodiments of this application will be described in detail.

[0145] Figure 7 A structural diagram of a pitch motor life prediction device provided in an embodiment of this application is shown below. Figure 7 As shown, the life prediction device 700 for the pitch motor may include:

[0146] The acquisition module 701 is used to acquire relevant parameters related to the lifespan of the pitch motor and the design lifespan value of the pitch motor under the current time series. The relevant parameters include the active power of the wind turbine generator set, environmental parameters, and the operating parameters of the pitch system of the wind turbine generator set.

[0147] The prediction module 702 is used to input active power, environmental parameters, and operating parameters into the life loss prediction model to obtain the predicted life loss value of the pitch motor in the next time series of the current time series. The life loss prediction model is trained by taking the relevant parameters of the samples in the historical time series as input and the sample state matrix and the sample life loss value of the pitch motor in the historical time series as output. The sample state matrix includes the sample operating loss information of the pitch motor in the historical time series, as well as the sample correlation loss information of related components to the pitch motor in the historical time series. The related components are the components associated with the pitch motor. Based on the difference between the design life value and the predicted life loss value, the remaining life value of the pitch motor in the next time series is predicted.

[0148] During the model training phase, this embodiment trains the model based on relevant parameters associated with the pitch motor's lifespan, the pitch motor's own operating loss information, and the correlation loss information of related components. This means that it considers not only the impact of the pitch motor's own operating state on its lifespan but also the impact of damage to other related components on the pitch motor, taking into account a more comprehensive range of factors. Therefore, when using the trained lifespan loss prediction model for lifespan prediction, it can more accurately predict the lifespan loss value of the pitch motor in the next time series, and thus more accurately predict the remaining lifespan value of the pitch motor in the next time series.

[0149] In some embodiments, the acquisition module 701 is further configured to acquire a training sample set before acquiring the relevant parameters associated with the lifespan of the pitch motor under the current time series and the design lifespan value of the pitch motor. The training sample set includes multiple training samples, and each training sample includes the sample-related parameters under the corresponding historical time series, as well as the sample state matrix and sample lifespan loss value corresponding to the sample-related parameters.

[0150] The life prediction device 700 for the pitch motor may also include:

[0151] The training module is used to train the transformer model using training samples until the training stopping condition is met, thus obtaining the lifetime loss prediction model.

[0152] In some embodiments, the converter model includes an encoding module, a linear transformation module, and a decoding module;

[0153] The training module is specifically used for:

[0154] The first sample-related parameters of the first training sample are input into the encoding module to obtain the first state matrix output by the encoding module, and the first state matrix is ​​input into the linear transformation module to obtain the first lifetime loss assessment value output by the linear transformation module.

[0155] The second training sample's relevant parameters are input into the decoding module to obtain the first lifetime loss prediction value output by the decoding module. The second training sample and the first training sample are training samples with adjacent time series.

[0156] Based on the first lifetime loss assessment value and the first sample lifetime loss value corresponding to the first training sample, a first loss function value is determined; based on the first state matrix and the sample state matrix corresponding to the first training sample, a second loss function value is determined; and based on the first lifetime loss prediction value and the sample lifetime loss value corresponding to the second training sample, a third loss function value is determined.

[0157] Based on the first loss function value, the second loss function value, and the third loss function value, the transformer model is trained until the training stopping condition is met, thus obtaining the lifetime loss prediction model.

[0158] In some embodiments, the acquisition module 701 is further configured to acquire the cumulative value of the life loss assessment of the pitch motor from the start of service to the current time series before acquiring the relevant parameters associated with the life of the pitch motor in the current time series and the design life value of the pitch motor.

[0159] Prediction module 702 is specifically used for:

[0160] Based on the difference between the design life value and the cumulative life loss assessment value, the remaining life assessment value of the pitch motor in the current time series is predicted.

[0161] The remaining life of the pitch motor in the next time series is predicted based on the difference between the remaining life assessment value and the life loss prediction value.

[0162] In some embodiments, the acquisition module 701 is specifically used for:

[0163] The relevant parameters of each historical time series of the pitch motor from the start of its service to the current time series are input into the encoding module of the life loss prediction model to obtain the historical state matrix corresponding to each historical time series output by the encoding module.

[0164] For each historical state matrix, the historical state matrix is ​​input into the linear transformation module of the lifetime loss prediction model to obtain the lifetime loss assessment value corresponding to the historical time series output by the linear transformation module.

[0165] The cumulative lifetime loss assessment value is obtained by accumulating the lifetime loss assessment values ​​for each historical time series.

[0166] In some embodiments, the pitch motor life prediction device 700 may further include:

[0167] The determination module is used to input the sample-related parameters corresponding to each historical time series between the start of service and the current time series of the pitch motor into the encoding module of the life loss prediction model after the acquisition module 701 inputs them respectively. After obtaining the historical state matrix corresponding to each historical time series output by the encoding module, the module determines the first cumulative result of the operating loss information and the second cumulative result of the correlation loss information of the pitch motor from the start of service to the current time series based on the state value of the corresponding state position in each historical state matrix.

[0168] The display module is used to display the first cumulative result and the second cumulative result.

[0169] In some embodiments, the display module is further configured to display the remaining life trend curve of the pitch motor based on the remaining life value of the pitch motor in a preset time series after the prediction module 702 predicts the remaining life value of the pitch motor in the next time series based on the difference between the design life value and the life loss prediction value.

[0170] In some embodiments, the pitch motor life prediction device 700 may further include:

[0171] The output module is used to output warning information when the remaining lifetime value is less than the remaining lifetime threshold of the corresponding preset time series, or when the rate of change of the remaining lifetime trend curve within a preset time period is greater than the rate of change threshold. The preset time period includes at least one preset time series.

[0172] In some embodiments, the operating loss information of the pitch motor includes at least one of power-limited pitch and feathering;

[0173] The associated loss information of related components includes at least one of the following: damage to the cooling fan of the pitch motor, failure of the electromagnetic brake to engage, and failure of the pitch drive.

[0174] This application's embodiments can predict the lifespan of a pitch motor based on SCADA data without requiring additional sensors, reducing the data access cost for lifespan prediction. Simultaneously, a state matrix is ​​introduced during the modeling process. This matrix includes the operating loss characteristics of the pitch motor and the associated loss characteristics of other related components, enhancing the interpretability and objectivity of the prediction results. Furthermore, the length of the input time series of the model can be adjusted according to actual needs, thereby improving the flexibility and controllability of the prediction method.

[0175] The pitch motor life prediction device provided in this application embodiment can achieve... Figures 1-4 The various processes in the embodiment of the pitch motor life prediction method shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0176] Based on the same inventive concept, embodiments of this application also provide an electronic device, such as a tablet computer, a laptop computer, a handheld computer, etc. The following, in conjunction with... Figure 8 The electronic devices provided in the embodiments of this application will be described in detail.

[0177] like Figure 8 As shown, the electronic device 800 may include a processor 801 and a memory 802 for storing computer program instructions.

[0178] The processor 801 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0179] Memory 802 may include mass storage for data or instructions. For example, and not limitingly, memory 802 may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. In one instance, memory 802 may include removable or non-removable (or fixed) media, or memory 802 may be non-volatile solid-state memory. In one instance, memory 802 may be read-only memory (ROM). In one instance, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0180] The processor 801 reads and executes computer program instructions stored in the memory 802 to achieve... Figures 1-4 The method in the illustrated embodiment achieves... Figures 1-4 The corresponding technical effects achieved by the methods in the illustrated embodiments are described briefly and will not be elaborated further here.

[0181] In one example, the electronic device 800 may also include a communication interface 803 and a bus 804. For example, Figure 8 As shown, the processor 801, memory 802, and communication interface 803 are connected through bus 804 and complete communication with each other.

[0182] The communication interface 803 is mainly used to realize communication between various modules, devices and / or equipment in the embodiments of this application.

[0183] Bus 804 includes hardware, software, or both, that couples the various components of electronic device 800 together. For example, and not as a limitation, bus 804 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 804 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0184] After acquiring the relevant parameters associated with the lifespan of the pitch motor in the current time series and the design lifespan value of the pitch motor, the electronic device 800 can execute the pitch motor lifespan prediction method in this embodiment of the application, thereby achieving a combination of... Figures 1-4 The described method for predicting the lifespan of a pitch motor and Figure 7 The described life prediction device for a pitch motor.

[0185] Furthermore, in conjunction with the pitch motor life prediction method in the above embodiments, this application embodiment can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the pitch motor life prediction methods in the above embodiments.

[0186] Furthermore, in conjunction with the pitch motor life prediction method in the above embodiments, this application embodiment can provide a computer program product to implement it. This computer program product includes a computer program that, when executed by a processor, implements any of the pitch motor life prediction methods in the above embodiments.

[0187] Although this application has been described with reference to preferred embodiments, various modifications can be made thereto and components can be replaced with equivalents without departing from the scope of this application. In particular, the technical features mentioned in the various embodiments can be combined in any manner, provided there is no structural conflict. This application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for predicting the lifespan of a pitch motor, characterized in that, include: Obtain relevant parameters associated with the lifespan of the pitch motor under the current time series, as well as the design lifespan value of the pitch motor. The relevant parameters include the active power of the wind turbine generator set, environmental parameters, and the operating parameters of the pitch system of the wind turbine generator set. The active power, environmental parameters, and operating parameters are input into the lifespan loss prediction model to obtain the predicted lifespan loss value of the pitch motor in the next time series of the current time series. The lifespan loss prediction model is trained using sample-related parameters from historical time series as input and a sample state matrix and the sample lifespan loss value of the pitch motor from the historical time series as output. The sample state matrix includes sample operating loss information of the pitch motor from the historical time series, and sample correlation loss information of associated components to the pitch motor from the historical time series. The associated components are those related to the pitch motor. Based on the difference between the designed lifespan value and the predicted lifespan loss value, the remaining lifespan value of the pitch motor in the next time series is predicted.

2. The method according to claim 1, characterized in that, Before obtaining the relevant parameters associated with the lifespan of the pitch motor in the current time series and the design lifespan value of the pitch motor, the method further includes: Obtain a training sample set, which includes multiple training samples. Each training sample includes sample-related parameters under the corresponding historical time series, as well as a sample state matrix and sample lifetime loss value corresponding to the sample-related parameters. The transformer model is trained using the training samples until the training stopping condition is met, thus obtaining the lifetime loss prediction model.

3. The method according to claim 2, characterized in that, The converter model includes an encoding module, a linear transformation module, and a decoding module; The step of training the transformer model using the training samples until the training stopping condition is met to obtain the lifetime loss prediction model includes: The first sample-related parameters of the first training sample are input into the encoding module to obtain the first state matrix output by the encoding module, and the first state matrix is ​​input into the linear transformation module to obtain the first lifetime loss assessment value output by the linear transformation module. The second sample related parameters of the second training sample are input into the decoding module to obtain the first lifetime loss prediction value output by the decoding module. The second training sample and the first training sample are training samples that are adjacent in time series. Based on the first lifetime loss assessment value and the first sample lifetime loss value corresponding to the first training sample, a first loss function value is determined; based on the first state matrix and the sample state matrix corresponding to the first training sample, a second loss function value is determined; and based on the first lifetime loss prediction value and the sample lifetime loss value corresponding to the second training sample, a third loss function value is determined. The converter model is trained based on the first loss function value, the second loss function value, and the third loss function value until the training stopping condition is met, thereby obtaining the lifetime loss prediction model.

4. The method according to claim 2, characterized in that, Before obtaining the relevant parameters associated with the lifespan of the pitch motor in the current time series and the design lifespan value of the pitch motor, the method further includes: Obtain the cumulative lifespan loss assessment value of the pitch motor from the start of its service to the current time series; The step of predicting the remaining lifespan of the pitch motor in the next time series based on the difference between the designed lifespan value and the predicted lifespan loss value includes: Based on the difference between the design life value and the cumulative life loss assessment value, the remaining life assessment value of the pitch motor in the current time series is predicted. The remaining life of the pitch motor in the next time series is predicted based on the difference between the remaining life assessment value and the life loss prediction value.

5. The method according to claim 4, characterized in that, The process of obtaining the cumulative lifespan loss assessment value of the pitch motor from the start of its service life to the present time series includes: The relevant parameters of each historical time series of the pitch motor from the start of its service to the current time series are input into the encoding module of the life loss prediction model to obtain the historical state matrix corresponding to each historical time series output by the encoding module. For each historical state matrix, the historical state matrix is ​​input into the linear transformation module of the lifetime loss prediction model to obtain the lifetime loss assessment value corresponding to the historical time series output by the linear transformation module. The cumulative lifetime loss assessment value is obtained by summing the lifetime loss assessment values ​​of each of the aforementioned historical time series.

6. The method according to claim 5, characterized in that, After inputting the sample correlation parameters corresponding to each historical time series of the pitch motor from the start of its service to the current time series into the encoding module of the life loss prediction model to obtain the historical state matrix corresponding to each historical time series output by the encoding module, the method further includes: Based on the state values ​​of the corresponding state positions in each of the historical state matrices, the first cumulative result of the operating loss information of the pitch motor from the start of its service to the current time series and the second cumulative result of the associated loss information are determined. Display the first cumulative result and the second cumulative result.

7. The method according to any one of claims 1-6, characterized in that, After predicting the remaining lifespan of the pitch motor in the next time series based on the difference between the design lifespan value and the predicted lifespan loss value, the method further includes: Based on the remaining lifespan value of the pitch motor in a preset time series, display the remaining lifespan trend curve of the pitch motor; If the remaining lifetime value is less than the remaining lifetime threshold of the corresponding preset time series, or if the rate of change of the remaining lifetime trend curve within a preset time period is greater than the rate of change threshold, an early warning message is output, wherein the preset time period includes at least one preset time series.

8. The method according to any one of claims 1-6, characterized in that, The operating loss information of the pitch motor includes at least one of power-limited pitch and feathering; The associated loss information of the associated components includes at least one of the following: damage to the cooling fan of the pitch motor, failure of the electromagnetic brake to engage, and failure of the pitch drive.

9. A life prediction device for a pitch motor, characterized in that, include: The acquisition module is used to acquire relevant parameters related to the lifespan of the pitch motor under the current time series and the design lifespan value of the pitch motor. The relevant parameters include the active power of the wind turbine generator set, environmental parameters, and the operating parameters of the pitch system of the wind turbine generator set. The prediction module is used to input the active power, environmental parameters, and operating parameters into a lifespan loss prediction model to obtain the predicted lifespan loss value of the pitch motor in the next time series of the current time series. The lifespan loss prediction model is trained using sample-related parameters from historical time series as input and a sample state matrix and the sample lifespan loss value of the pitch motor in the historical time series as output. The sample state matrix includes sample operating loss information of the pitch motor in the historical time series, and sample correlation loss information of associated components to the pitch motor in the historical time series. The associated components are those related to the pitch motor. Based on the difference between the design lifespan value and the predicted lifespan loss value, the remaining lifespan value of the pitch motor in the next time series is predicted.

10. An electronic device, characterized in that, include: processor; Memory is used to store computer program instructions; When the computer program instructions are executed by the processor, the method as described in any one of claims 1-8 is implemented.

11. A computer-readable storage medium storing computer program instructions thereon, characterized in that, When the computer program instructions are executed by the processor, the method as described in any one of claims 1-8 is implemented.

12. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, implements the method as described in any one of claims 1-8.

Images