A data-driven wind turbine variable pitch bearing maintenance decision method and system

Abstract

This invention belongs to the field of wind turbine operation and maintenance technology, and provides a data-driven method and system for wind turbine pitch bearing maintenance decision-making. The method includes generating predicted values ​​for the wind turbine pitch bearing's condition based on its condition characteristic data and the ambient wind speed in the wind turbine's location; generating the urgency of maintenance personnel's scheduling tasks based on their task characteristic data; generating the urgency of wind turbine pitch bearing maintenance based on the predicted bearing condition values; and generating the optimal maintenance time for the wind turbine pitch bearing based on both the urgency of maintenance personnel's scheduling tasks and the urgency of wind turbine pitch bearing maintenance. This allows for decision-making regarding the maintenance of the wind turbine pitch bearing. This invention can more accurately predict the condition of wind turbine pitch bearings and comprehensively consider equipment maintenance needs and personnel workload to optimize maintenance time decisions, reduce unplanned downtime and maintenance costs, and improve the efficiency of wind turbine operation and maintenance.

Description

Technical Field

[0001] This invention belongs to the field of wind turbine operation and maintenance technology, and in particular relates to a data-driven wind turbine pitch bearing maintenance decision-making method and system. Background Technology

[0002] As a core component of wind turbines that withstands complex alternating loads, the health of the wind turbine pitch bearing directly affects the stability of the unit's operation and its power generation efficiency. Traditional maintenance strategies mainly rely on two modes: regular inspection or post-failure repair. Regular inspection can easily lead to resource waste or insufficient maintenance, while post-failure repair can cause unplanned downtime and high repair costs.

[0003] Current research largely focuses on fault warning and lifespan prediction based on condition monitoring data such as vibration and temperature. However, existing methods generally suffer from the problem of separating equipment condition degradation modeling from environmental factors, particularly lacking quantitative processing of the impact of predicted wind speed, and failing to incorporate the dynamic effects of wind speed changes on future loads and degradation rates into the condition prediction model. In the maintenance scheduling stage, existing technologies often separate equipment maintenance needs from personnel workload, or schedule based solely on simple priority rules, resulting in the inability to achieve coordinated optimization of equipment risk and labor costs over time. Specifically, the condition prediction model ignores the real-time impact mechanism of environmental wind speed, causing prediction results to deviate from actual operating conditions; the maintenance decision-making process does not fully consider the current workload of maintenance personnel, including the number of existing tasks and total working hours, which can easily lead to task allocation imbalances, resulting in excessive concentration or idleness of personnel; at the same time, there is a lack of a unified framework to simultaneously assess the urgency of equipment maintenance and personnel scheduling, making it difficult to determine the maintenance timing with the optimal overall cost within a feasible time window. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a data-driven method and system for wind turbine pitch bearing maintenance decision-making, which solves the aforementioned problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a data-driven wind turbine pitch bearing maintenance decision-making method, which specifically includes:

[0006] Acquire the status characteristic data of the wind turbine pitch bearing, the ambient wind speed in the area where the wind turbine is located, and the task characteristic data of the maintenance personnel; among which, the task characteristic data of the maintenance personnel includes the number of existing tasks and the total man-hours of existing tasks;

[0007] A state prediction model is established based on the state characteristic data of the wind turbine pitch bearing and the ambient wind speed in the area where the wind turbine is located, and the state prediction value of the wind turbine pitch bearing is generated.

[0008] Based on the task characteristic data of maintenance personnel, the urgency of maintenance personnel scheduling tasks is generated;

[0009] Based on the predicted condition of the wind turbine pitch bearing, the maintenance urgency of the wind turbine pitch bearing is generated.

[0010] The optimal maintenance time for the wind turbine pitch bearing is generated based on the urgency of the maintenance personnel scheduling tasks and the urgency of the wind turbine pitch bearing maintenance.

[0011] The maintenance of wind turbine pitch bearings is decided based on the optimal maintenance time.

[0012] Based on the above technical solutions, the present invention also provides the following optional technical solutions:

[0013] Further technical solution: The method for generating the predicted state value of the wind turbine pitch bearing specifically includes:

[0014] Based on the condition characteristic data of wind turbine pitch bearings, basic predicted values ​​are generated;

[0015] Based on the ambient wind speed in the area where the wind turbine is located, environmental impact factors are generated.

[0016] A state prediction model is established based on the basic predicted values ​​and environmental impact factors to generate the state prediction values ​​of the wind turbine pitch bearing.

[0017] Further technical solutions: The specific methods for generating the basic predicted values ​​include:

[0018] Through the formula: ;

[0019] Generate wind turbine pitch bearing state values ;

[0020] In the formula, This represents the normalized value of the i-th state characteristic data of the wind turbine pitch bearing. This represents the weight coefficient of the i-th state feature of the wind turbine pitch bearing, and n represents the number of state feature data of the wind turbine pitch bearing.

[0021] Through the formula: ;

[0022] Generate basic forecast values ;

[0023] In the formula, the base forecast value It refers to The basic predicted value of the wind turbine pitch bearing at any given time. This represents the wind turbine pitch bearing state value at time t. This represents the wind turbine pitch bearing state value at time t-1. This represents the wind turbine pitch bearing state value at time t-p+1, and AR represents the autoregressive model.

[0024] Further technical solutions: The specific methods for generating the environmental impact factors include:

[0025] Through the formula: ;

[0026] Generate environmental impact factors ;

[0027] In the formula, environmental impact factors It refers to Environmental influencing factors at any given time It means At any given time, the ambient wind speed This represents the wind speed threshold. This represents the wind speed sensitivity coefficient, and the calculation formula is subject to the following conditions: Otherwise, environmental impact factors The value is 1.

[0028] A further technical solution: The expression for the state prediction model is specifically as follows: ;

[0029] In the expression, It means Predicted values ​​of wind turbine pitch bearing condition at any given time. It means The basic predicted value of the wind turbine pitch bearing at any given time. It means Environmental influencing factors at any given time.

[0030] Further technical solution: The method for generating the urgency of maintenance personnel scheduling tasks specifically includes:

[0031] Through the formula: ;

[0032] Generate maintenance personnel scheduling task urgency ;

[0033] In the formula, It means The number of existing tasks for maintenance personnel. It means Total existing task man-hours for maintenance personnel This indicates that maintenance personnel are... The maximum number of tasks that can be tolerated at any given time. This indicates that maintenance personnel are... The maximum working hours that can be tolerated at any time. The weighting coefficients for the number of existing tasks. The weighting coefficients for existing task hours are as follows: .

[0034] Further technical solution: The method for generating the maintenance urgency of the wind turbine pitch bearing specifically includes:

[0035] Through the formula: ;

[0036] The urgency of generating wind turbine pitch bearing maintenance ;

[0037] In the formula, the urgency of wind turbine pitch bearing maintenance It refers to The urgency of wind turbine pitch bearing maintenance at that time It means Predicted values ​​of wind turbine pitch bearing condition at any given time. This represents the state threshold of the wind turbine pitch bearing.

[0038] Further technical solution: The method for generating the optimal maintenance time for the wind turbine pitch bearing specifically includes:

[0039] The maintenance time of wind turbine pitch bearings was taken as the independent variable, and the urgency of wind turbine pitch bearing maintenance and the urgency of maintenance personnel scheduling tasks were taken as the dependent variables.

[0040] Through the formula: ;

[0041] Optimal maintenance time for wind turbine pitch bearings ;

[0042] In the formula, This indicates the maintenance time for the wind turbine pitch bearing. The urgency of wind turbine pitch bearing maintenance at that time This indicates the maintenance time for the wind turbine pitch bearing. The urgency of the maintenance personnel scheduling task at that time This represents the weighting coefficient for the urgency of wind turbine pitch bearing maintenance. This represents the set of time windows that maintenance personnel can schedule.

[0043] A data-driven wind turbine pitch bearing maintenance decision-making system, which executes the aforementioned data-driven wind turbine pitch bearing maintenance decision-making method, specifically includes:

[0044] The data acquisition unit is used to acquire the status characteristic data of the wind turbine pitch bearing, the ambient wind speed in the area where the wind turbine is located, and the task characteristic data of the maintenance personnel; among which, the task characteristic data of the maintenance personnel includes the number of existing tasks and the total working hours of existing tasks;

[0045] The state prediction unit is used to establish a state prediction model based on the state characteristic data of the wind turbine pitch bearing and the ambient wind speed in the area where the wind turbine is located, and generate the state prediction value of the wind turbine pitch bearing.

[0046] The personnel scheduling urgency analysis unit is used to generate the urgency of maintenance personnel scheduling tasks based on the task characteristic data of maintenance personnel;

[0047] The maintenance urgency analysis unit is used to generate the maintenance urgency of the wind turbine pitch bearing based on the predicted value of the wind turbine pitch bearing condition.

[0048] The optimal maintenance time analysis unit is used to generate the optimal maintenance time for the wind turbine pitch bearing based on the urgency of the maintenance personnel scheduling tasks and the urgency of the wind turbine pitch bearing maintenance.

[0049] The decision-making unit is used to make decisions on the maintenance of wind turbine pitch bearings based on the optimal maintenance time.

[0050] Further technical solution: The state prediction unit specifically includes:

[0051] The basic prediction module is used to generate basic prediction values ​​based on the state characteristic data of wind turbine pitch bearings.

[0052] The environmental analysis module is used to generate environmental impact factors based on the ambient wind speed in the area where the wind turbine is located.

[0053] The comprehensive analysis module is used to establish a state prediction model based on the basic predicted values ​​and environmental impact factors, and generate the state prediction values ​​of the wind turbine pitch bearing.

[0054] This invention provides a data-driven method and system for wind turbine pitch bearing maintenance decision-making, which has the following advantages compared with the prior art:

[0055] This invention acquires state characteristic data of wind turbine pitch bearings, ambient wind speed, and maintenance personnel task characteristic data to establish a state prediction model, generating state prediction values, maintenance urgency, and scheduling task urgency. This, in turn, determines the optimal maintenance time, thus solving the problems of existing technologies that ignore the impact of environmental factors on state prediction and fail to consider equipment maintenance needs and personnel workload. It has the advantages of more accurately predicting the state of wind turbine pitch bearings, comprehensively considering equipment maintenance needs and personnel workload, optimizing maintenance time decisions, reducing unplanned downtime and maintenance costs, and improving the operation and maintenance efficiency of wind turbine units. Attached Figure Description

[0056] Figure 1 This is a flowchart illustrating a data-driven wind turbine pitch bearing maintenance decision-making method provided by the present invention.

[0057] Figure 2 This is a flowchart illustrating step S20 of the present invention.

[0058] Figure 3 This is a schematic diagram of a data-driven wind turbine pitch bearing maintenance decision system provided by the present invention.

[0059] Figure 4 This is a schematic diagram of the state prediction unit provided by the present invention. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0061] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0062] Please see Figure 1 The present invention provides a data-driven wind turbine pitch bearing maintenance decision-making method according to one embodiment, comprising the following steps:

[0063] Step S10: Obtain the status characteristic data of the wind turbine pitch bearing, the ambient wind speed in the area where the wind turbine is located, and the task characteristic data of the maintenance personnel; wherein, the task characteristic data of the maintenance personnel includes the number of existing tasks and the total working hours of existing tasks;

[0064] Step S20: Establish a state prediction model based on the state characteristic data of the wind turbine pitch bearing and the ambient wind speed in the area where the wind turbine is located, and generate the state prediction value of the wind turbine pitch bearing.

[0065] Step S30: Generate the urgency of maintenance personnel scheduling tasks based on the task characteristic data of maintenance personnel;

[0066] Step S40: Generate the maintenance urgency of the wind turbine pitch bearing based on the predicted value of the wind turbine pitch bearing condition;

[0067] Step S50: Generate the optimal maintenance time for the wind turbine pitch bearing based on the urgency of the maintenance personnel's scheduling tasks and the urgency of the wind turbine pitch bearing maintenance.

[0068] Step S60: Make a decision on the maintenance of the wind turbine pitch bearing based on the optimal maintenance time of the wind turbine pitch bearing;

[0069] Among them, the condition characteristic data of wind turbine pitch bearings refers to various monitoring data used to describe the current operating status and health level of wind turbine pitch bearings, such as vibration signals, temperature, lubricating oil analysis results, current, voltage, etc. These data are collected in real time by sensors and are used to assess the degree of bearing degradation after preprocessing.

[0070] Ambient wind speed refers to the real-time or predicted wind speed in the area where the wind turbine is located; changes in wind speed directly affect the load on the wind turbine rotor, which in turn affects the operating status and degradation rate of the wind turbine pitch bearing.

[0071] The task characteristic data of maintenance personnel refers to information describing the current workload and availability of maintenance personnel, including the number of existing tasks and the total working hours of existing tasks; the number of existing tasks indicates the number of tasks that maintenance personnel have been assigned or are currently executing, while the total working hours of existing tasks indicates the total working time that these tasks are expected to consume; these data reflect the scheduling flexibility and resource constraints of maintenance personnel.

[0072] Specifically, in step S10, it is necessary to acquire the status characteristic data of the wind turbine pitch bearing, the ambient wind speed in the area where the wind turbine is located, and the task characteristic data of the maintenance personnel. The status characteristic data can be monitored and collected in real time by installing various sensors on the wind turbine pitch bearing, such as vibration sensors, temperature sensors, and oil analysis sensors. These sensors transmit the raw data to the data processing unit. Ambient wind speed data can be obtained from the wind turbine's built-in anemometer or weather station, or from predicted wind speeds for a future period obtained through an external weather forecasting system. The task characteristic data of the maintenance personnel, including the number of existing tasks and the total working hours of existing tasks, can be collected through the maintenance management system or manually entered. For example, the daily work logs or shift schedules of the maintenance personnel can serve as data sources.

[0073] In step S20, based on the acquired state characteristic data of the wind turbine pitch bearing and the ambient wind speed in the area where the wind turbine is located, a state prediction model is established, and a predicted state value for the wind turbine pitch bearing is generated. The state prediction model can be established using various machine learning or statistical methods. For example, it can be based on historical state data for trend extrapolation, or a simple linear regression model can be used to predict future health indices. The predicted state value of the wind turbine pitch bearing can be a quantitative indicator of the bearing's health status at a future point in time.

[0074] In step S30, the urgency of scheduled maintenance tasks is generated based on the task characteristic data of the maintenance personnel. This urgency aims to quantify the impact of scheduling additional maintenance tasks within a specific time window on the workload of the maintenance personnel. For example, the urgency can be calculated by setting a simple threshold or proportional relationship based on the number of tasks currently assigned to the maintenance personnel and the total time expected to complete these tasks. When the workload of the maintenance personnel is high, the urgency of scheduled tasks increases accordingly.

[0075] In step S40, a maintenance urgency level for the wind turbine pitch bearing is generated based on the predicted condition value of the wind turbine pitch bearing. This urgency level is used to assess the urgency of the wind turbine pitch bearing needing maintenance at a future point in time. For example, the predicted condition value of the wind turbine pitch bearing can be compared with a preset health threshold. When the predicted value is lower than the threshold, the maintenance urgency level increases, indicating that the bearing has a high risk of failure.

[0076] In step S50, the optimal maintenance time for the wind turbine pitch bearing is generated based on the urgency of the maintenance personnel scheduling task and the urgency of the wind turbine pitch bearing maintenance. This step aims to find a balance point that maximizes the optimization of maintenance personnel scheduling while meeting equipment maintenance needs. For example, a simple weighted average function can be set to linearly combine the two urgency levels, and then within a preset time window, the time point that minimizes the combined value can be selected as the optimal maintenance time.

[0077] In step S60, a maintenance decision is made for the wind turbine pitch bearing based on the generated optimal maintenance time. This decision may include developing a detailed maintenance plan, arranging maintenance personnel and necessary tools, and preparing spare parts. For example, if the optimal maintenance time is determined to be next Wednesday, the maintenance management system can automatically generate a maintenance work order and notify the relevant maintenance team to prepare and execute it.

[0078] This invention balances the risk of bearing failure with the cost of scheduling maintenance personnel, avoiding suboptimal results due to decisions based on a single factor. For example, even if the bearing has a high maintenance urgency on a particular day, if the scheduling of maintenance personnel on that day is also extremely urgent, the system can intelligently select a slightly later or earlier maintenance time that offers better overall efficiency. Therefore, the method in this embodiment achieves dynamic synergistic optimization of equipment risk and labor costs over time, ensuring the foresight, economy, and efficiency of maintenance decisions.

[0079] For preferred options, please refer to [link / reference]. Figure 2 The present invention further proposes a method for generating the predicted state value of the wind turbine pitch bearing, specifically including:

[0080] Step S21: Generate basic predicted values ​​based on the state characteristic data of the wind turbine pitch bearing;

[0081] Step S22: Generate environmental impact factors based on the ambient wind speed in the area where the wind turbine is located;

[0082] Step S23: Establish a state prediction model based on the basic predicted values ​​and environmental impact factors, and generate the predicted state values ​​of the wind turbine pitch bearing;

[0083] The generation of baseline predicted values ​​refers to predicting the inherent degradation trend of wind turbine pitch bearings under relatively stable environments based on their historical operating data and state characteristics. These baseline predicted values ​​reflect the evolution of the bearing's health status under normal wear, fatigue, and other mechanisms. There are various methods for generating baseline predicted values. For example, time series analysis methods, such as autoregressive (AR), moving average (MA), autoregressive moving average (ARMA), or autoregressive integral moving average (ARIMA), can be used to model and predict historical state characteristic data. Alternatively, machine learning algorithms, such as support vector machines (SVM), random forests, or neural networks, can be used to generate these values ​​by learning the relationship between state characteristic data and future states.

[0084] The generation of environmental impact factors refers to the quantification of the additional impact of ambient wind speed in the area where the wind turbine is located on the condition of the wind turbine pitch bearing. Changes in ambient wind speed may cause the bearing to experience different loads, vibrations, and temperatures, thereby accelerating or slowing down its degradation process.

[0085] Establishing a state prediction model and generating predicted values ​​for wind turbine pitch bearings involves effectively integrating the generated basic prediction values ​​with environmental impact factors to obtain a final prediction result that comprehensively considers both intrinsic degradation and external environmental impacts. Establishing a state prediction model can be done in various ways; for example, the basic prediction values ​​can be multiplicatively or additively combined with environmental impact factors to form the final state prediction value.

[0086] The proposed solution decomposes the wind turbine pitch bearing condition prediction process into two independent but interactive stages: basic prediction and environmental impact correction. First, based on the bearing's condition characteristic data, basic prediction values ​​are generated. This effectively captures the inherent degradation trend of the bearing in a relatively stable environment, reflecting its intrinsic wear and fatigue mechanisms. Second, based on the ambient wind speed in the wind turbine's location, an environmental impact factor is generated. This factor is specifically used to quantify the dynamic modulation effect of external wind speed on the bearing's condition; for example, high wind speeds may lead to additional loads and vibrations, accelerating degradation. Finally, the basic prediction values ​​and the environmental impact factor are integrated to establish a more comprehensive condition prediction model, thereby generating predicted wind turbine pitch bearing conditions. This step-by-step and integrated strategy allows the model to more precisely capture the driving factors of bearing condition changes, avoiding the limitations of a single model and ensuring the accuracy and robustness of the prediction results.

[0087] The above technical solution decomposes the wind turbine pitch bearing condition prediction process into two independent but interacting stages: basic prediction and environmental impact correction. This decomposition allows the model to more precisely capture the inherent degradation patterns of the bearing and dynamically consider the accelerating or mitigating effects of external environmental wind speed on its degradation process. Therefore, the generated wind turbine pitch bearing condition prediction values ​​can more accurately reflect the true health condition of the bearing, improving the accuracy and reliability of the prediction. This provides a more solid data foundation for subsequent maintenance decisions and effectively avoids maintenance delays or over-maintenance caused by inaccurate predictions.

[0088] Preferably, the present invention further proposes a method for generating the basic predicted value, specifically including:

[0089] Through the formula: ;

[0090] Generate wind turbine pitch bearing state values ;

[0091] In the formula, This represents the normalized value of the i-th state characteristic data of the wind turbine pitch bearing. This represents the weight coefficient of the i-th state feature of the wind turbine pitch bearing, and n represents the number of state feature data of the wind turbine pitch bearing.

[0092] Through the formula: ;

[0093] Generate basic forecast values ;

[0094] In the formula, the base forecast value It refers to The basic predicted value of the wind turbine pitch bearing at any given time. This represents the wind turbine pitch bearing state value at time t. This represents the wind turbine pitch bearing state value at time t-1. This represents the wind turbine pitch bearing state value at time t-p+1, and AR represents the autoregressive model.

[0095] The generation of wind turbine pitch bearing state values ​​aims to integrate multiple discrete wind turbine pitch bearing state characteristic data into a unified comprehensive index that reflects the overall health of the bearing. This index is achieved by weighted summation of the normalized values ​​of the i-th state characteristic data of the wind turbine pitch bearing, where... Let represent the weight coefficient of the i-th state feature, and n represent the number of state feature data. This method can effectively integrate heterogeneous data from different sensors or monitoring points, providing a quantitative and easily understandable measure of bearing health. In addition to weighted summation, dimensionality reduction techniques such as principal component analysis (PCA) can be used to extract the main health features, or a fuzzy logic system can be used to classify the health status based on the feature values.

[0096] The generation of baseline forecasts is based on historical wind turbine pitch bearing condition values, using an autoregressive (AR) model to predict future bearing health trends. The AR model predicts future values ​​by analyzing the linear relationship between the current and previous wind turbine pitch bearing condition values. This method captures the inherent dependencies and trends in time-series data, thus predicting the natural degradation path of bearings in the absence of external environmental disturbances. Besides AR models, other models such as Autoregressive Moving Average (ARMA), Autoregressive Moving Average Difference (ARIMA), or more complex deep learning models like Recurrent Neural Networks (RNNs) or Long Short-Term Memory Networks (LSTMs) can be used to adapt to more complex time-series data patterns.

[0097] This application's solution generates basic predicted values ​​through a phased, data-driven approach. First, for the original wind turbine pitch bearing condition characteristic data, a weighted summation method is used to integrate multiple normalized condition characteristic data and their corresponding weight coefficients, thereby calculating a comprehensive wind turbine pitch bearing condition value. This step effectively unifies multi-dimensional and potentially heterogeneous sensor data into a single, quantifiable health indicator, laying the foundation for subsequent predictions. Subsequently, historical wind turbine pitch bearing condition values ​​are used as input, and an autoregressive (AR) model is employed to model these time-series data. The AR model can identify and learn the inherent correlations and evolution patterns between historical condition values, thereby predicting the basic predicted values ​​of wind turbine pitch bearings at future times. This two-stage processing approach ensures that the basic predicted values ​​not only comprehensively reflect the current health status of the bearings but also accurately capture their inherent degradation trends. This provides a solid and accurate foundation for subsequently combining environmental impact factors to generate the final wind turbine pitch bearing condition predicted values, significantly improving the prediction accuracy and reliability of the entire maintenance decision-making method.

[0098] Through the above technical solution, this application provides a specific and quantitative method for generating basic predicted values ​​for wind turbine pitch bearings. First, by using a weighted summation method, multi-source heterogeneous state characteristic data is integrated into a single wind turbine pitch bearing state value, effectively solving the problem that multi-dimensional data is difficult to directly use for prediction, thus clearly quantifying the overall health status of the bearing. Second, by using an autoregressive model to model and predict historical state values, the inherent degradation patterns and trends of wind turbine pitch bearings can be accurately captured, eliminating the interference of external environmental factors, thereby generating more accurate and reliable basic predicted values. This significantly improves the accuracy and stability of wind turbine pitch bearing state prediction, providing high-quality input for subsequent maintenance urgency assessment and optimal maintenance time decision-making, thereby optimizing the efficiency and effectiveness of the entire maintenance decision-making process and avoiding over-maintenance or maintenance delays caused by inaccurate predictions.

[0099] Preferably, the present invention further proposes a method for generating the environmental impact factors, specifically including:

[0100] Through the formula: ;

[0101] Generate environmental impact factors ;

[0102] In the formula, environmental impact factors It refers to Environmental influencing factors at any given time It means At any given time, the ambient wind speed This represents the wind speed threshold. This represents the wind speed sensitivity coefficient, and the calculation formula is subject to the following conditions: Otherwise, environmental impact factors The value is 1;

[0103] The environmental impact factor refers to the degree of influence of ambient wind speed at a specific point in time on the predicted condition of wind turbine pitch bearings. Its function is to quantify the accelerating or decelerating effect of environmental factors on bearing health, thereby correcting the basic predictions derived solely from historical condition data. This factor can be a dimensionless multiplier used to adjust the bearing's degradation rate or health index. Besides calculation via an exponential function, the environmental impact factor can also be obtained through table lookup, training a machine learning model on historical environmental data and bearing degradation data, or evaluation using an expert experience system.

[0104] Ambient wind speed at any given time refers to the predicted wind speed value at a future forecast point in time. Accurate ambient wind speed forecasts are crucial for assessing the operating load of bearings over a future period. Ambient wind speed can be obtained in various ways, such as short-term forecasts using data from the wind farm's own anemometer towers, or medium- to long-term forecasts combining regional weather forecasts provided by meteorological departments.

[0105] The wind speed threshold refers to a preset critical wind speed point. When the ambient wind speed exceeds this threshold, the wind turbine pitch bearing is considered to bear greater loads and stresses, which may accelerate its degradation. This threshold is typically determined based on the design parameters, operating experience, and safety specifications of the wind turbine pitch bearing. For example, it can be the rated wind speed of the wind turbine or a critical wind speed point determined based on the stress-life curve provided by the bearing manufacturer.

[0106] The wind speed sensitivity coefficient is a parameter used to adjust the degree of influence of ambient wind speed on the calculation formula of environmental impact factors. This coefficient reflects the sensitivity of wind turbine pitch bearings to changes in wind speed, that is, how much an increase in wind speed will lead to an increase in the bearing degradation rate. The value of the wind speed sensitivity coefficient can be set through historical operating data analysis, bearing fatigue life model fitting, or through expert experience.

[0107] Limitations of the calculation formula Otherwise, the environmental impact factor is set to 1. This constraint indicates that only when the predicted wind speed exceeds a preset wind speed threshold will the ambient wind speed have an additional effect on the bearing's condition, accelerating its degradation. When the predicted wind speed is below or equal to this threshold, the wind speed's effect on the bearing's additional stress is insignificant; therefore, the environmental impact factor is set to 1, meaning no additional correction is applied, and the bearing's degradation is primarily determined by its inherent characteristics. This segmented processing mechanism simplifies the model and focuses on risk assessment under high wind speeds.

[0108] This application's solution refines the generation process of wind turbine pitch bearing condition prediction values ​​by introducing an environmental impact factor. Specifically, the solution first obtains the ambient wind speed at a future time. Then, it compares this predicted wind speed with a preset wind speed threshold. When the predicted wind speed exceeds the threshold, it indicates that the wind turbine pitch bearing will bear a higher operating load, and its degradation process may be accelerated. At this point, an environmental impact factor is calculated using an exponential function combined with a wind speed sensitivity coefficient. This exponential function can non-linearly reflect the aggravating effect of wind speed exceeding the threshold on bearing degradation. If the predicted wind speed does not exceed the threshold, the ambient wind speed is considered to have no significant impact on the additional stress on the bearing, and the environmental impact factor is set to 1, meaning no additional correction is made. In this way, the environmental impact factor can accurately quantify the corrective effect of ambient wind speed on the wind turbine pitch bearing condition and, combined with the basic predicted value, jointly establish a condition prediction model, thereby generating more accurate wind turbine pitch bearing condition prediction values. This method, based on a dynamic adjustment prediction model for environmental wind speed, enables the predicted value of the wind turbine pitch bearing condition to more accurately reflect the health status of the bearing under actual operating conditions, providing a more reliable basis for subsequent maintenance decisions.

[0109] Through the above technical solution, this application can dynamically and quantitatively generate environmental impact factors based on the ambient wind speed in the area where the wind turbine is located. This method overcomes the limitation of traditional condition prediction models that may ignore the impact of environmental factors on bearing degradation, especially under high wind speed conditions, it can more accurately capture the trend of accelerated bearing degradation due to additional stress. By introducing wind speed thresholds and wind speed sensitivity coefficients, and using exponential functions for calculation, the environmental impact factors can more precisely reflect the degree of influence of different wind speeds on bearing condition, thereby correcting the basic prediction values ​​and generating more accurate wind turbine pitch bearing condition prediction values. This significantly improves the accuracy and reliability of the condition prediction model, provides a more forward-looking and targeted basis for wind turbine pitch bearing maintenance decisions, helps avoid sudden failures caused by environmental factors, and optimizes maintenance resource allocation.

[0110] Preferably, the present invention further proposes the following expression for the state prediction model: ;

[0111] In the expression, It means Predicted values ​​of wind turbine pitch bearing condition at any given time. It means The basic predicted value of the wind turbine pitch bearing at any given time. It means Environmental influencing factors at any given time;

[0112] This expression defines the calculation method for the predicted value of the wind turbine pitch bearing condition. Its function is to quantitatively combine the basic prediction based on historical data with the influence of environmental factors, thereby obtaining a comprehensive assessment of the future condition.

[0113] in, This represents the expected health condition of the wind turbine pitch bearing at a future point in time. It is a quantitative assessment of the bearing's future performance and reliability, and serves as the core basis for calculating subsequent maintenance urgency and making maintenance decisions.

[0114] This indicates the predicted health status of the wind turbine pitch bearing based solely on historical operating data and inherent degradation patterns, without considering the influence of external environmental factors. It reflects the natural wear trend of the bearing under ideal or average operating conditions.

[0115] It quantifies the impact of environmental conditions (wind speed) on the health status of wind turbine pitch bearings at a future point in time. It serves as a correction factor to adjust the baseline forecast value, reflecting the accelerating or decelerating effect of the external environment on the bearing degradation process.

[0116] This application's solution introduces a clear state prediction model expression, organically combining the basic predicted value of wind turbine pitch bearings with environmental impact factors. Specifically, the model generates the final state prediction value of the wind turbine pitch bearing by dividing the basic predicted value by the environmental impact factor. This division relationship cleverly reflects the corrective effect of environmental factors on the bearing's health status: when the environmental impact factor increases (e.g., wind speed exceeding a threshold leading to increased environmental pressure), the final state prediction value decreases accordingly, indicating that the bearing's health status deteriorates more rapidly due to environmental degradation; conversely, when the environmental impact factor decreases (e.g., environmental pressure decreases), the final state prediction value increases accordingly, indicating that the bearing's health status is less affected by the environment. In this way, the model can comprehensively consider the bearing's inherent degradation trend and the dynamic impact of the external environment, thereby generating a more accurate and comprehensive wind turbine pitch bearing state prediction value, effectively solving the problem of how to effectively integrate basic predictions with environmental impacts.

[0117] Through the above technical solution, this application provides a clear and quantitative method to integrate the basic predicted values ​​and environmental impact factors of wind turbine pitch bearings, thereby generating more accurate predicted values ​​for the bearing condition. This precise condition prediction can more realistically reflect the health status of the bearing under actual operating conditions, avoiding the biases caused by predictions based solely on historical data or single environmental factors. Therefore, it provides a solid foundation for subsequent assessments of the urgency of wind turbine pitch bearing maintenance, significantly improving the scientific nature and effectiveness of maintenance decisions, helping to achieve more accurate predictive maintenance, and reducing the risk of unplanned downtime.

[0118] Preferably, the present invention further proposes a method for generating the urgency of the maintenance personnel scheduling task, specifically including:

[0119] Through the formula: ;

[0120] Generate maintenance personnel scheduling task urgency ;

[0121] In the formula, It means The number of existing tasks for maintenance personnel. It means Total existing task man-hours for maintenance personnel This indicates that maintenance personnel are... The maximum number of tasks that can be tolerated at any given time. This indicates that maintenance personnel are... The maximum working hours that can be tolerated at any time. The weighting coefficients for the number of existing tasks. The weighting coefficients for existing task hours are as follows: ;

[0122] The urgency of maintenance personnel scheduling tasks is a quantitative indicator used to measure the workload and availability of maintenance personnel at a specific future point in time. A higher urgency value indicates a heavier workload and lower scheduleability for maintenance personnel. This urgency can be a continuous value between 0 and 1, or it can be divided into different levels based on actual needs.

[0123] The number of existing tasks refers to the number of tasks at a given time point. This refers to the total number of tasks assigned to maintenance personnel. This can be obtained by querying the maintenance personnel's task management system or scheduling database. For example, it can be used to count the number of work orders assigned to a specific team or individual within a specific time period.

[0124] The total work hours of existing tasks refer to the total work hours at a given time point. This is the estimated total working time for the tasks assigned to maintenance personnel. This is typically calculated by adding up the estimated working hours for each assigned task. For example, if one task is estimated to take 8 hours and another 4 hours, the total working time is 12 hours.

[0125] The maximum number of tasks threshold is the maximum number of tasks that maintenance personnel can reasonably handle within a specific time period. This threshold can be set based on factors such as historical data analysis, industry best practices, team size, and skill levels. For example, a team can handle a maximum of 5 tasks per day.

[0126] The maximum working hours threshold is the maximum total working time that a maintenance worker can reasonably undertake within a specific time period. This threshold can be set based on factors such as standard working hours, overtime policies, and worker fatigue levels. For example, the maximum working hours for a maintenance worker in a day may be 8 or 10 hours (including overtime).

[0127] Weighting coefficient and These factors are used to adjust the relative importance of the number of existing tasks and the total man-hours of existing tasks when calculating the urgency of maintenance personnel scheduling tasks. These coefficients are typically determined based on practical operational experience, management strategies, or through optimization algorithms. For example, if the number of tasks is considered to have a more significant impact on scheduling urgency, then... It can be set to greater than The value of . The condition ensures that the sum of the weights of the two factors is 1, making the calculation of the urgency value consistent.

[0128] This application's solution quantifies the urgency of maintenance personnel's scheduling tasks by comprehensively considering the existing number of tasks and total working hours of existing tasks, and comparing them with the maximum task quantity threshold and maximum working hour threshold that maintenance personnel can bear. Specifically, firstly, the ratio of the existing number of tasks to the maximum task quantity threshold is calculated to obtain the task quantity load ratio; simultaneously, the ratio of the existing total task working hours to the maximum working hour threshold is calculated to obtain the working hour load ratio. Subsequently, by introducing weighting coefficients, these two load ratios are weighted and summed to generate a comprehensive maintenance personnel scheduling task urgency. This calculation method allows for the accurate quantification of the actual workload of maintenance personnel and reflects their availability at future points in time. By using this quantified maintenance personnel scheduling task urgency as input and combining it with the maintenance urgency of wind turbine pitch bearings, the feasibility of maintenance tasks can be assessed more comprehensively and objectively. Thus, when generating the optimal maintenance time for wind turbine pitch bearings, not only the equipment status is considered, but also the limitations of human resources are fully taken into account, avoiding maintenance delays or inefficiencies caused by insufficient or overloaded personnel.

[0129] Through the above technical solution, this application provides a quantitative and objective method for assessing the urgency of maintenance personnel scheduling tasks. This method comprehensively considers the number of existing tasks and the total working hours of existing tasks for maintenance personnel, combined with their maximum capacity, and generates a numerical urgency index through weighted summation. This allows the maintenance decision-making process to more accurately reflect the actual workload and availability of maintenance personnel, avoiding biases caused by subjective judgment. Therefore, when generating the optimal maintenance time for wind turbine pitch bearings, it can more effectively balance equipment maintenance needs with human resource constraints, thereby optimizing maintenance scheduling, improving maintenance efficiency, and ultimately ensuring the stable operation of wind turbine units.

[0130] Preferably, the present invention further proposes a method for generating the maintenance urgency of the wind turbine pitch bearing, specifically including:

[0131] Through the formula: ;

[0132] The urgency of generating wind turbine pitch bearing maintenance ;

[0133] In the formula, the urgency of wind turbine pitch bearing maintenance It refers to The urgency of wind turbine pitch bearing maintenance at that time It means Predicted values ​​of wind turbine pitch bearing condition at any given time. This represents the state threshold of the wind turbine pitch bearing;

[0134] The urgency of wind turbine pitch bearing maintenance refers to the degree or priority at which the wind turbine pitch bearing needs maintenance at a future point in time. Its purpose is to transform abstract bearing condition predictions into a quantifiable, easily understood, and comparable indicator to facilitate subsequent maintenance decisions. This urgency can be a dimensionless numerical value indicating the degree to which the current bearing condition deviates from its health threshold.

[0135] The bearing condition threshold for wind turbine pitch control is a preset reference value used to determine whether the bearing requires maintenance or has reached a critical state. When the predicted bearing condition value reaches or exceeds this threshold, it usually means that the bearing's health condition is not optimistic and maintenance measures are required. This threshold can be set based on the equipment manufacturer's recommendations, industry standards, historical failure data analysis, or expert experience. For example, it can be set as a warning line for the bearing health index or the upper limit of the probability of a certain failure mode.

[0136] This application's solution, based on the obtained predicted values ​​of wind turbine pitch bearing conditions, introduces a wind turbine pitch bearing condition threshold as a reference benchmark. By comparing the predicted value with this threshold, the maintenance urgency of the wind turbine pitch bearing is generated. This ratio calculation method makes maintenance urgency a standardized, dimensionless indicator that can intuitively reflect the gap between the current predicted bearing condition and the critical condition requiring maintenance. When the predicted value approaches or exceeds the threshold, the maintenance urgency increases accordingly, thus providing a quantitative and comparable input for generating the optimal maintenance time, making maintenance decisions more scientific and accurate. In this way, the abstract bearing health prediction is transformed into a specific maintenance priority, effectively solving the problem of how to quantitatively assess maintenance urgency.

[0137] Through the aforementioned technical solution, this application can transform the abstract state prediction value of wind turbine pitch bearings into a standardized and quantifiable maintenance urgency index. This quantification process makes maintenance decision-making more objective and accurate, avoiding the limitations of subjective judgment based solely on raw prediction values. Maintenance personnel can clearly understand the gap between the bearing's health condition and maintenance needs based on this urgency index, thereby more effectively prioritizing maintenance tasks and allocating resources. This not only improves the scientific nature of maintenance decisions but also provides a solid data foundation for generating optimal maintenance times, ultimately enhancing the operational reliability and maintenance efficiency of wind turbine pitch bearings.

[0138] Preferably, the present invention further proposes a method for generating the optimal maintenance time of the wind turbine pitch bearing, specifically including:

[0139] The maintenance time of wind turbine pitch bearings was taken as the independent variable, and the urgency of wind turbine pitch bearing maintenance and the urgency of maintenance personnel scheduling tasks were taken as the dependent variables.

[0140] Through the formula: ;

[0141] Optimal maintenance time for wind turbine pitch bearings ;

[0142] In the formula, This indicates the maintenance time for the wind turbine pitch bearing. The urgency of wind turbine pitch bearing maintenance at that time This indicates the maintenance time for the wind turbine pitch bearing. The urgency of the maintenance personnel scheduling task at that time This represents the weighting coefficient for the urgency of wind turbine pitch bearing maintenance. This represents the set of time windows that maintenance personnel can schedule;

[0143] The wind turbine pitch bearing maintenance time, as an independent variable, represents all potential time points or time periods within a future period where maintenance of the wind turbine pitch bearing can be considered. Its function is to serve as input to the optimization process, enabling the system to evaluate the overall benefits under different maintenance times. For example, this independent variable can be a series of discrete dates, such as "today," "tomorrow," and "the day after tomorrow," or a more specific time window, such as "Monday morning" or "Tuesday afternoon."

[0144] The urgency of wind turbine pitch bearing maintenance, as a dependent variable, indicates the level of urgency required for maintenance of the wind turbine pitch bearing at a specific maintenance time. Its value is typically calculated based on the predicted condition of the wind turbine pitch bearing; the worse the condition, the higher the urgency. This dependent variable reflects the equipment's health status and potential failure risks.

[0145] The urgency of maintenance personnel scheduling tasks, as a dependent variable, represents the level of workload or resource availability for maintenance personnel during a specific wind turbine pitch bearing maintenance period. Its value is typically calculated based on the number of existing tasks and total working hours of maintenance personnel; the more tasks and the longer the working hours, the higher the urgency. This dependent variable reflects the scheduling pressure on human resources.

[0146] Optimal maintenance time for wind turbine pitch bearings refers to the point in time, among all feasible maintenance times, where a predetermined comprehensive cost or risk function is minimized by comprehensively considering both the urgency of wind turbine pitch bearing maintenance and the urgency of maintenance personnel scheduling. The goal is to find a balance that allows for timely handling of equipment issues while making rational use of human resources.

[0147] argmin is a mathematical operator that represents the expression in a given set. In this context, the system seeks the independent variable that minimizes the subsequent function expression. In this context, it means the system will iterate through all feasible maintenance time points, calculate the overall urgency for each time point, and ultimately select the time point with the lowest overall urgency as the optimal maintenance time.

[0148] This represents the set of time windows during which maintenance personnel can be scheduled. This set includes all maintenance personnel available for maintenance of wind turbine pitch bearings within a future period. This set can be determined by a maintenance scheduling system, a personnel availability calendar, or preset maintenance windows. For example, it could be the mornings of all weekdays within the next two weeks, or specific dates when a particular shift is available.

[0149] It is a weighted summation function expression used to comprehensively evaluate the performance at a certain maintenance time. The overall urgency level. Among them, This indicates the urgency of equipment maintenance. The urgency of personnel scheduling tasks can be considered. By weighting these two dependent variables, their relative importance in the decision-making process can be adjusted according to actual needs. This represents the weighting coefficient for the urgency of wind turbine pitch bearing maintenance. It is a value between 0 and 1, used to adjust the influence of wind turbine pitch bearing maintenance urgency in the overall urgency calculation. A larger value indicates a greater focus on the health status and maintenance needs of the equipment; when... A smaller value indicates a greater emphasis on the scheduling pressure on maintenance personnel and resource utilization. For example, when equipment is about to experience a serious failure, It can be set to close to 1 to prioritize equipment safety; however, when the equipment is in acceptable condition but personnel resources are limited, It can be set to a smaller value to optimize personnel scheduling.

[0150] This application's solution effectively addresses the problem of determining the optimal maintenance timing from known urgency levels of wind turbine pitch bearing maintenance and maintenance personnel scheduling tasks by introducing an optimization-based decision-making mechanism. The core of this mechanism lies in constructing a comprehensive evaluation function by using wind turbine pitch bearing maintenance time as the independent variable and both the urgency levels of wind turbine pitch bearing maintenance and maintenance personnel scheduling tasks as dependent variables. Specifically, the system first obtains a set of available time windows for maintenance personnel, containing all potential maintenance time points. For each maintenance time point in the set, the system calculates the urgency of wind turbine pitch bearing maintenance at that time point based on previously generated wind turbine pitch bearing state predictions and a wind turbine pitch bearing state threshold. Simultaneously, the system also calculates the urgency of maintenance personnel scheduling tasks based on the number of existing tasks and total existing task hours for maintenance personnel at that time point. Subsequently, a weighted summation formula is used to comprehensively consider these two urgency levels, where the weighting coefficients... This is used to balance equipment maintenance needs with personnel scheduling pressure. Ultimately, the system will... Within the set, the time point that minimizes the weighted overall urgency is identified and determined as the optimal maintenance time for the wind turbine pitch bearing. This method quantifies and comprehensively considers both the equipment's condition and human resources, ensuring that maintenance decisions are no longer based on a single dimension but rather on the optimization result of multiple factors working together. This allows for more scientific and efficient scheduling of maintenance work.

[0151] Through the above technical solution, this application effectively addresses the challenge of comprehensively balancing equipment health status and maintenance personnel availability in wind turbine pitch bearing maintenance decisions. By using wind turbine pitch bearing maintenance time as the independent variable and the urgency of wind turbine pitch bearing maintenance and the urgency of maintenance personnel scheduling tasks as dependent variables to construct an optimization model, this application can systematically evaluate the comprehensive benefits of different maintenance time points. This makes maintenance decisions no longer based on a single dimension or experience-based judgment, but on data-driven quantitative analysis, thereby generating a truly optimal maintenance time for wind turbine pitch bearings. This optimized decision-making mechanism avoids delays in critical maintenance due to personnel shortages and also avoids over-scheduling personnel when equipment condition is still acceptable, thus significantly improving maintenance efficiency, reducing operating costs, and extending the service life of wind turbine pitch bearings. This solution, combined with wind turbine pitch bearing condition prediction and maintenance personnel scheduling task urgency analysis, forms a closed-loop intelligent maintenance decision-making system, ensuring the timeliness, economy, and effectiveness of maintenance work.

[0152] Please see Figure 3In another embodiment, the present invention also proposes a data-driven wind turbine pitch bearing maintenance decision-making system, which is used to execute the above-described data-driven wind turbine pitch bearing maintenance decision-making method, specifically including:

[0153] Data acquisition unit 10 is used to acquire the status characteristic data of wind turbine pitch bearing, the ambient wind speed in the area where the wind turbine is located, and the task characteristic data of maintenance personnel; among which, the task characteristic data of maintenance personnel includes the number of existing tasks and the total working hours of existing tasks;

[0154] State prediction unit 20 is used to establish a state prediction model based on the state characteristic data of wind turbine pitch bearing and the ambient wind speed in the area where the wind turbine is located, and generate the state prediction value of wind turbine pitch bearing.

[0155] The personnel scheduling urgency analysis unit 30 is used to generate the urgency of maintenance personnel scheduling tasks based on the task characteristic data of maintenance personnel.

[0156] The maintenance urgency analysis unit 40 is used to generate the maintenance urgency of the wind turbine pitch bearing based on the predicted value of the wind turbine pitch bearing condition.

[0157] The optimal maintenance time analysis unit 50 is used to generate the optimal maintenance time for the wind turbine pitch bearing based on the urgency of the maintenance personnel scheduling task and the urgency of the wind turbine pitch bearing maintenance.

[0158] Decision unit 60 is used to make decisions on the maintenance of wind turbine pitch bearings based on the optimal maintenance time of the wind turbine pitch bearings.

[0159] For preferred options, please refer to [link / reference]. Figure 4 The present invention further proposes that the state prediction unit 20 specifically includes:

[0160] The basic prediction module 21 is used to generate basic prediction values ​​based on the state characteristic data of the wind turbine pitch bearing.

[0161] Environmental analysis module 22 is used to generate environmental impact factors based on the ambient wind speed in the area where the wind turbine is located;

[0162] The comprehensive analysis module 23 is used to establish a state prediction model based on the basic predicted values ​​and environmental impact factors, and generate the state prediction values ​​of the wind turbine pitch bearing.

[0163] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A data-driven decision-making method for wind turbine pitch bearing maintenance, characterized in that, The method specifically includes: Acquire the status characteristic data of the wind turbine pitch bearing, the ambient wind speed in the area where the wind turbine is located, and the task characteristic data of the maintenance personnel; among which, the task characteristic data of the maintenance personnel includes the number of existing tasks and the total man-hours of existing tasks; A state prediction model is established based on the state characteristic data of the wind turbine pitch bearing and the ambient wind speed in the area where the wind turbine is located, and the state prediction value of the wind turbine pitch bearing is generated. Based on the task characteristic data of maintenance personnel, the urgency of maintenance personnel scheduling tasks is generated; Based on the predicted condition of the wind turbine pitch bearing, the maintenance urgency of the wind turbine pitch bearing is generated. The optimal maintenance time for the wind turbine pitch bearing is generated based on the urgency of the maintenance personnel scheduling tasks and the urgency of the wind turbine pitch bearing maintenance. The maintenance of wind turbine pitch bearings is decided based on the optimal maintenance time.

2. The data-driven wind turbine pitch bearing maintenance decision-making method according to claim 1, characterized in that, The specific methods for generating the predicted state values ​​of the wind turbine pitch bearings include: Based on the condition characteristic data of wind turbine pitch bearings, basic predicted values ​​are generated; Based on the ambient wind speed in the area where the wind turbine is located, environmental impact factors are generated. A state prediction model is established based on the basic predicted values ​​and environmental impact factors to generate the state prediction values ​​of the wind turbine pitch bearing.

3. The data-driven wind turbine pitch bearing maintenance decision-making method according to claim 2, characterized in that, The specific methods for generating the basic predicted values ​​include: Through the formula: ; Generate wind turbine pitch bearing state values ; In the formula, This represents the normalized value of the i-th state characteristic data of the wind turbine pitch bearing. This represents the weight coefficient of the i-th state feature of the wind turbine pitch bearing, and n represents the number of state feature data of the wind turbine pitch bearing. Through the formula: ; Generate basic forecast values ; In the formula, the base forecast value It refers to The basic predicted value of the wind turbine pitch bearing at any given time. This represents the wind turbine pitch bearing state value at time t. This represents the wind turbine pitch bearing state value at time t-1. This represents the wind turbine pitch bearing state value at time t-p+1, and AR represents the autoregressive model.

4. The data-driven wind turbine pitch bearing maintenance decision-making method according to claim 2, characterized in that, The specific methods for generating the environmental impact factors include: Through the formula: ; Generate environmental impact factors ; In the formula, environmental impact factors It refers to Environmental influencing factors at any given time It means At any given time, the ambient wind speed This represents the wind speed threshold. This represents the wind speed sensitivity coefficient, and the calculation formula is subject to the following conditions: Otherwise, environmental impact factors The value is 1.

5. The data-driven wind turbine pitch bearing maintenance decision-making method according to claim 2, characterized in that, The specific expression of the state prediction model is as follows: ; In the expression, It means Predicted values ​​of wind turbine pitch bearing condition at any given time. It means The basic predicted value of the wind turbine pitch bearing at any given time. It means Environmental influencing factors at any given time.

6. The data-driven wind turbine pitch bearing maintenance decision-making method according to claim 1, characterized in that, The specific methods for generating the urgency of maintenance personnel scheduling tasks include: Through the formula: ; Generate maintenance personnel scheduling task urgency ; In the formula, It means The number of existing tasks for maintenance personnel. It means Total existing task man-hours for maintenance personnel This indicates that maintenance personnel are... The maximum number of tasks that can be tolerated at any given time. This indicates that maintenance personnel are... The maximum working hours that can be tolerated at any time. The weighting coefficients for the number of existing tasks. The weighting coefficients for existing task hours are as follows: .

7. The data-driven wind turbine pitch bearing maintenance decision-making method according to claim 1, characterized in that, The specific methods for generating the urgency of wind turbine pitch bearing maintenance include: Through the formula: ; The urgency of generating wind turbine pitch bearing maintenance ; In the formula, the urgency of wind turbine pitch bearing maintenance It refers to The urgency of wind turbine pitch bearing maintenance at that time It means Predicted values ​​of wind turbine pitch bearing condition at any given time. This represents the state threshold of the wind turbine pitch bearing.

8. The data-driven wind turbine pitch bearing maintenance decision-making method according to claim 1, characterized in that, The specific method for generating the optimal maintenance time for the wind turbine pitch bearing includes: The maintenance time of wind turbine pitch bearings was taken as the independent variable, and the urgency of wind turbine pitch bearing maintenance and the urgency of maintenance personnel scheduling tasks were taken as the dependent variables. Through the formula: ; Optimal maintenance time for wind turbine pitch bearings ; In the formula, This indicates the maintenance time for the wind turbine pitch bearing. The urgency of wind turbine pitch bearing maintenance at that time This indicates the maintenance time for the wind turbine pitch bearing. The urgency of the maintenance personnel scheduling task at that time This represents the weighting coefficient for the urgency of wind turbine pitch bearing maintenance. This represents the set of time windows that maintenance personnel can schedule.

9. A data-driven wind turbine pitch bearing maintenance decision-making system, characterized in that, The system is used to execute the data-driven wind turbine pitch bearing maintenance decision-making method according to any one of claims 1-8, specifically including: The data acquisition unit is used to acquire the status characteristic data of the wind turbine pitch bearing, the ambient wind speed in the area where the wind turbine is located, and the task characteristic data of the maintenance personnel; among which, the task characteristic data of the maintenance personnel includes the number of existing tasks and the total working hours of existing tasks; The state prediction unit is used to establish a state prediction model based on the state characteristic data of the wind turbine pitch bearing and the ambient wind speed in the area where the wind turbine is located, and generate the state prediction value of the wind turbine pitch bearing. The personnel scheduling urgency analysis unit is used to generate the urgency of maintenance personnel scheduling tasks based on the task characteristic data of maintenance personnel; The maintenance urgency analysis unit is used to generate the maintenance urgency of the wind turbine pitch bearing based on the predicted value of the wind turbine pitch bearing condition. The optimal maintenance time analysis unit is used to generate the optimal maintenance time for the wind turbine pitch bearing based on the urgency of the maintenance personnel scheduling tasks and the urgency of the wind turbine pitch bearing maintenance. The decision-making unit is used to make decisions on the maintenance of wind turbine pitch bearings based on the optimal maintenance time.

10. The data-driven wind turbine pitch bearing maintenance decision-making system according to claim 9, characterized in that, The state prediction unit specifically includes: The basic prediction module is used to generate basic prediction values ​​based on the state characteristic data of wind turbine pitch bearings. The environmental analysis module is used to generate environmental impact factors based on the ambient wind speed in the area where the wind turbine is located. The comprehensive analysis module is used to establish a state prediction model based on the basic predicted values ​​and environmental impact factors, and generate the state prediction values ​​of the wind turbine pitch bearing.

Images