Wind power generation state monitoring and early warning method and system

Abstract

The invention discloses a wind power generation state monitoring and early warning method and system, and the method comprises the steps: obtaining and analyzing the historical operation data of a wind power generator, determining the operation parameters related to power generation, extracting the features of the operation parameters, building a power generation prediction model of the wind power generator based on the features and a preset neural network model, and carrying out the prediction. Obtaining power generation prediction data; analyzing the power generation prediction data to determine abnormal parameters, and analyzing the abnormal parameters to determine abnormal parameter characteristics; and evaluating the abnormal degree of each abnormal parameter based on the abnormal parameter characteristics to obtain an abnormal degree evaluation value, and comprehensively determining the state coefficient of the wind driven generator based on the abnormal degree evaluation value, so as to determine the early warning strategy of the wind driven generator. The method can accurately warn potential risks and assess the state of the equipment before a fault occurs, and provides scientific decision support for an operation and maintenance team, so that the non-planned downtime is reduced to the maximum extent, the power generation availability rate is improved, the service life of the equipment is prolonged, the operation and maintenance cost is reduced, and stable operation of a power grid is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wind power generation, in particular to a wind power generation state monitoring and early warning method and system. BACKGROUND

[0002] With the accelerated promotion of global energy transformation, wind power, as the main force of clean energy, its installed capacity and single machine scale continue to grow, and wind farms are increasingly expanding to complex environmental areas such as the sea, highlands, and remote areas. However, the large-scale wind turbines and complex sites have significantly increased the difficulty and cost of their operation and maintenance while improving power generation efficiency. Wind turbines, as complex mechanical and electrical systems exposed to harsh working conditions for a long time, their blades, gearboxes, generators, and main shafts are subjected to the continuous impact of alternating loads, fatigue stress, and natural erosion, resulting in high failure rates.

[0003] However, the traditional "periodic maintenance" and "after-maintenance" mode has been difficult to meet the needs of the development of modern wind power industry. Periodic maintenance has the risk of over-maintenance or insufficient maintenance, causing resource waste or failure to detect faults; and after-maintenance means unplanned downtime, heavy loss of power generation, and even may cause serious secondary damage, resulting in high repair costs. SUMMARY

[0004] To solve the above technical problems, the present application provides a wind power generation state monitoring and early warning method and system, comprising: obtaining historical operation data of a wind turbine and analyzing the historical operation data to determine operation parameters related to power generation; extracting features of the operation parameters, constructing a power generation prediction model of the wind turbine based on the features and a preset neural network model for prediction to obtain power generation prediction data; performing data analysis on the power generation prediction data to determine abnormal parameters, and performing data feature analysis on the abnormal parameters to determine abnormal parameter features; evaluating the abnormal degree of each abnormal parameter based on the abnormal parameter features to obtain an abnormal degree evaluation value of each abnormal parameter; determining a state coefficient of the wind turbine based on the abnormal degree evaluation values of the abnormal parameters, and determining an early warning strategy of the wind turbine based on the state coefficient.

[0005] Further, the obtaining of the historical operation data of the wind turbine and the analysis of the historical operation data to determine the operation parameters related to power generation comprises: obtaining historical operation data of a wind turbine and dividing the historical operation data into multiple operation parameter data groups according to parameter types; The historical power generation data is determined, the correlation between the operation parameter data set and the historical power generation data is calculated, and the parameter type corresponding to the operation parameter data set with a correlation exceeding a preset threshold is determined as the operation parameter related to power generation.

[0006] Further, the feature of the operation parameter is extracted, the power generation prediction model of the wind turbine is constructed based on the feature and the preset neural network model for prediction, and power generation prediction data is obtained. The feature of the operation parameter corresponding to the operation parameter data set is extracted, a data set is constructed based on the operation parameter data set of the operation parameter and the corresponding feature, and the data set is input into the preset neural network model to construct a power generation prediction initial model. The data set is divided into a training set and a test set according to a preset ratio, and the training set and the test set are input into the power generation prediction initial model; The power generation prediction initial model is trained and tested until the power generation prediction initial model meets a preset convergence condition, and a power generation prediction model is obtained. Real-time operation data of the wind turbine is obtained, the real-time operation data is input into the power generation prediction model for prediction, and power generation prediction data in a future period of time is obtained.

[0007] Further, the power generation prediction data is analyzed to determine abnormal parameters, including: Normal power generation data of the wind turbine is determined, and the normal power generation data is divided into a plurality of normal power generation parameter data sets according to the parameter type; Each normal power generation parameter data set is analyzed based on a K-Means clustering algorithm to obtain a plurality of clustering clusters and cluster centers corresponding to each normal power generation parameter data set; The maximum distance of all points in each clustering cluster to each cluster center is calculated, and the maximum distance is determined as the abnormal threshold of each clustering cluster; The power generation prediction data is divided into a plurality of power generation parameter prediction data sets according to the parameter type, and the power generation parameter prediction data sets are mapped into each clustering cluster of the corresponding normal power generation parameter data set in the same way; The distance of the newly added point in each clustering cluster to each cluster center is calculated, and the data corresponding to the newly added point with a distance greater than the maximum distance is determined as the abnormal data in each power generation parameter prediction data set; The number of abnormal data in each power generation parameter prediction data set is determined, and the parameter type corresponding to the power generation parameter prediction data set with a number of abnormal data greater than a preset threshold is determined as the abnormal parameter.

[0008] Further, the data feature of the abnormal parameter is analyzed to determine the abnormal parameter feature, including: Identify the power generation parameter prediction data group corresponding to each abnormal parameter, and identify the abnormal data in the power generation parameter prediction data group; Determine the proportion of abnormal data in the power generation parameter prediction data set, and define this proportion as the first abnormal parameter feature corresponding to each abnormal parameter; Identify the maximum and minimum values ​​in the outlier data, calculate the difference between the maximum and minimum values ​​in the outlier data, and determine the difference as the second outlier parameter feature corresponding to each outlier parameter; The first and second abnormal parameter features are determined as the abnormal parameter features of each abnormal parameter.

[0009] Furthermore, the evaluation of the degree of anomalousness of each anomalous parameter based on its characteristics to obtain an evaluation value for the degree of anomalousness of each parameter includes: The first and second abnormal parameter features of each abnormal parameter are evaluated and their values ​​are obtained respectively to obtain the first abnormal evaluation value corresponding to the first abnormal parameter feature and the second abnormal evaluation value corresponding to the second abnormal parameter feature. Determine the first standard anomaly evaluation value corresponding to the first anomaly parameter feature and the second standard anomaly evaluation value corresponding to the second anomaly parameter feature, and calculate the difference between the first anomaly evaluation value and the first standard anomaly evaluation value, as well as the difference between the second anomaly evaluation value and the second standard anomaly evaluation value, to obtain the first difference and the second difference respectively; The first difference and the second difference are scored based on the preset scoring model, and the scores are normalized to obtain the weights corresponding to the first abnormal parameter feature and the second abnormal parameter feature. The abnormality assessment value of each abnormal parameter is obtained by calculating the first abnormality assessment value corresponding to the first abnormal parameter feature and the second abnormality assessment value corresponding to the second abnormal parameter feature, as well as the corresponding weight.

[0010] Furthermore, the formula for calculating the abnormality assessment value of the abnormal parameter is as follows: , Where K is the abnormality assessment value of the abnormal parameter, α is the weight corresponding to the first abnormal parameter feature, P1 is the first abnormality assessment value corresponding to the first abnormal parameter feature, β is the weight corresponding to the second abnormal parameter feature, and P2 is the second abnormality assessment value corresponding to the second abnormal parameter feature.

[0011] Furthermore, the determination of the state coefficient of the wind turbine based on the comprehensive assessment values ​​of the anomaly degree of each anomaly parameter includes: Determine the preset weights of each abnormal parameter, and then add the preset weights of each abnormal parameter to the abnormality assessment value to obtain the comprehensive abnormality assessment value of the wind turbine. A preset state coefficient-comprehensive anomaly assessment value range correspondence is set in advance. For each comprehensive anomaly assessment value range, a corresponding preset state coefficient is associated with it. Obtain the comprehensive anomaly assessment value of the wind turbine, and based on the mapping relationship between the comprehensive anomaly assessment value interval to which the comprehensive anomaly assessment value belongs and the preset state coefficient-comprehensive anomaly assessment value interval correspondence, select the preset state coefficient corresponding to the comprehensive anomaly assessment value interval as the state coefficient of the wind turbine.

[0012] Furthermore, the early warning strategy for determining wind turbines based on state coefficients includes: If the state coefficient of the wind turbine is less than the first preset value, the early warning strategy for the wind turbine is to mark the turbine on the monitoring interface, list the specific abnormal parameters and the abnormality assessment value, and include the turbine in the key inspection objects of the next regular inspection, reminding the inspection personnel to pay attention to specific components. If the state coefficient of the wind turbine is greater than the first preset value and less than the second preset value, the early warning strategy for the wind turbine is to send a formal warning to the operation and maintenance team, automatically generate a preliminary maintenance work order, and start a deep diagnostic model to further locate the root cause of the fault. If the state coefficient of the wind turbine is greater than the second preset value and less than the third preset value, the early warning strategy for the wind turbine is to dispatch the operation and maintenance team to conduct on-site inspection within a specified time, automatically adjust the wind turbine operation strategy, and formulate specific maintenance plans and schedules. If the state coefficient of the wind turbine exceeds the third preset value, the early warning strategy for the wind turbine is to automatically execute a protective shutdown, initiate an emergency maintenance process, and immediately dispatch a team to the site for handling.

[0013] The present invention also provides a wind power generation status monitoring and early warning system, comprising: The acquisition module is used to acquire historical operating data of wind turbines and analyze the historical operating data to determine operating parameters related to power generation. The prediction module is used to extract features of operating parameters, construct a power generation prediction model for wind turbines based on the features and a preset neural network model, and obtain power generation prediction data. The analysis module is used to perform data analysis on power generation forecast data, identify abnormal parameters, and perform data feature analysis on the abnormal parameters to determine the characteristics of the abnormal parameters. The evaluation module is used to evaluate the degree of abnormality of each abnormal parameter based on its characteristics, and obtain the evaluation value of the degree of abnormality of each abnormal parameter. The early warning module is used to comprehensively determine the state coefficient of the wind turbine based on the anomaly severity assessment values ​​of each abnormal parameter, and to determine the early warning strategy for the wind turbine based on the state coefficient.

[0014] Compared with the prior art, the wind power generation status monitoring and early warning method and system of this invention have the following advantages: This invention, by constructing a high-precision power generation prediction model, not only enables accurate prediction of power generation, providing reliable data support for grid dispatch and power trading, but more importantly, it establishes a dynamic and personalized "health baseline" that can immediately capture subtle abnormal signs when the predicted value deviates significantly from the normal value. This invention transforms the general concept of "abnormality" into a quantifiable "abnormality assessment value" by performing deep feature extraction and quantitative evaluation of abnormal parameters. It can issue early warnings at the incipient stage of minor faults, long before they cause downtime, and clearly indicate the nature, location, and severity of the problem to maintenance personnel, greatly shortening the fault diagnosis time. This invention integrates the abnormal assessment values ​​of multiple parameters to calculate a comprehensive "state coefficient," which can intuitively reflect the overall health of the wind turbine. This allows for the intelligent formulation of differentiated early warning strategies, effectively avoiding false alarms and missed alarms. This makes operation and maintenance decisions more scientific and accurate, ultimately achieving the core objectives of significantly reducing unplanned downtime, reducing the risk of damage to expensive components, and optimizing the allocation of operation and maintenance resources, thereby comprehensively improving the economic benefits and safety level of wind farm operation. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the process structure of the wind power generation status monitoring and early warning method in an embodiment of the present invention; Figure 2 This is a schematic diagram of the composition of the wind power generation status monitoring and early warning system in an embodiment of the present invention. Detailed Implementation

[0016] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0017] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the platform or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0018] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0019] like Figure 1 As shown in the embodiments of this application, a wind power generation status monitoring and early warning method is provided, including: S100: acquiring historical operating data of a wind turbine generator, analyzing the historical operating data, and determining operating parameters related to power generation; S200: extracting features of the operating parameters, constructing a power generation prediction model for the wind turbine generator based on the features and a preset neural network model, and obtaining power generation prediction data; S300: performing data analysis on the power generation prediction data, determining abnormal parameters, and performing data feature analysis on the abnormal parameters to determine abnormal parameter features; S400: evaluating the degree of abnormality of each abnormal parameter based on the abnormal parameter features, and obtaining an abnormality degree evaluation value for each abnormal parameter; S500: comprehensively determining the state coefficient of the wind turbine generator based on the abnormality degree evaluation values ​​of each abnormal parameter, and determining an early warning strategy for the wind turbine generator based on the state coefficient.

[0020] Furthermore, this invention, by constructing a high-precision power generation prediction model, not only achieves accurate prediction of power generation, providing reliable data support for grid dispatch and power trading, but more importantly, it establishes a dynamic and personalized "health baseline." When the predicted value deviates significantly from the normal value, it can immediately capture these subtle abnormal signs. This invention, through deep feature extraction and quantitative evaluation of abnormal parameters, transforms the general "abnormality" into a quantifiable "abnormality assessment value," enabling early warnings at the nascent stage of minor faults, long before they cause shutdowns. It also clearly indicates the nature, location, and severity of the problem to maintenance personnel, greatly shortening fault diagnosis time. By integrating the abnormal assessment values ​​of multiple parameters, this invention systematically calculates a comprehensive "state coefficient," which can intuitively reflect the overall health of the wind turbine. This allows for the intelligent formulation of differentiated early warning strategies, effectively avoiding false alarms and missed alarms, making maintenance decisions more scientific and accurate. Ultimately, this achieves the core objectives of significantly reducing unplanned downtime, reducing the risk of damage to expensive components, and optimizing the allocation of maintenance resources, comprehensively improving the economic efficiency and safety level of wind farm operations.

[0021] In the embodiments of this application, a method for monitoring and early warning of wind power generation status is provided. The method involves acquiring historical operating data of a wind turbine and analyzing the historical operating data to determine operating parameters related to power generation. The method includes: acquiring historical operating data of the wind turbine and dividing the historical operating data into multiple operating parameter data groups according to parameter type; determining historical power generation data; calculating the correlation between the operating parameter data groups and the historical power generation data; and determining the parameter type corresponding to the operating parameter data group whose correlation exceeds a preset threshold as the operating parameter related to power generation.

[0022] Specifically, the historical operating data of wind turbines is systematically organized and divided into different data groups according to parameter type. Historical power generation data is introduced, and the statistical correlation (such as Pearson correlation coefficient) between each operating parameter data group and the power generation data sequence is calculated. Compared with the dilemma of "rich data but scarce information" faced by traditional monitoring systems, this step automatically filters out the core parameter groups that are strongly correlated with power generation performance by setting a correlation threshold, while filtering out irrelevant or weakly correlated redundant parameters. This greatly reduces the complexity and computational burden of subsequent modeling, allowing analysis resources to be focused on truly critical indicators. This step elevates wind turbine status monitoring from a crude "all-encompassing" mode to a new intelligent stage of "focused monitoring." By letting the data speak for itself, it scientifically points out the direction of monitoring work, providing crucial input for building a high-precision, high-efficiency predictive maintenance system, ultimately helping to increase power generation, reduce operation and maintenance costs, and ensure asset safety.

[0023] In embodiments of this application, a method for monitoring and early warning of wind power generation status is provided. The method involves extracting features of operating parameters, constructing a wind turbine power generation prediction model based on the features and a preset neural network model, and obtaining power generation prediction data. The method includes: extracting features of operating parameter data sets corresponding to the operating parameters; constructing a dataset based on the operating parameter data sets and corresponding features; inputting the dataset into a preset neural network model to construct an initial power generation prediction model; dividing the dataset into training and testing sets according to a preset ratio, and inputting the training and testing sets into the initial power generation prediction model; training and testing the initial power generation prediction model until it meets a preset convergence condition to obtain a power generation prediction model; acquiring real-time operating data of the wind turbine, and inputting the real-time operating data into the power generation prediction model for prediction to obtain power generation prediction data for a future period.

[0024] Specifically, key features are extracted from data sets corresponding to operating parameters strongly correlated with power generation. These features may include time-series statistics, frequency domain components, and trends, forming a dataset for model learning. A pre-defined neural network structure is used, dividing the dataset into training and testing sets for iterative training and validation. The training process continues until the model reaches a pre-defined convergence condition, ensuring it learns deep patterns in the data and possesses good generalization ability, ultimately resulting in a high-precision power generation prediction model. This step achieves high-precision, forward-looking power generation prediction, providing reliable power generation estimates for grid dispatch and electricity trading over a future period. This significantly enhances the reliability and dispatchability of wind farms as power sources and reduces the impact of power fluctuations on the grid. The data-driven modeling method is highly adaptive, continuously updating and optimizing with the age of wind turbines and the accumulation of new data, maintaining consistent prediction accuracy.

[0025] In embodiments of this application, a method for monitoring and early warning of wind power generation status is provided. The step of analyzing power generation prediction data to determine abnormal parameters includes: determining normal power generation data of wind turbines and dividing the normal power generation data into multiple normal power generation parameter data groups according to parameter type; performing cluster analysis on each normal power generation parameter data group based on the K-Means clustering algorithm to obtain multiple clusters and cluster centers corresponding to each normal power generation parameter data group; calculating the maximum distance from all points in each cluster to each cluster center and determining the maximum distance as the abnormal threshold of each cluster; dividing the power generation prediction data into multiple power generation parameter prediction data groups according to parameter type and mapping the power generation parameter prediction data groups to each cluster of the corresponding normal power generation parameter data group in the same way; calculating the distance from newly added points in each cluster to each cluster center, and determining the data corresponding to newly added points with distances greater than the maximum distance as abnormal data in each power generation parameter prediction data group; determining the number of abnormal data in each power generation parameter prediction data group, and determining the parameter type corresponding to the power generation parameter prediction data group with the number of abnormal data greater than a preset threshold as abnormal parameters.

[0026] Specifically, historical data of wind turbines in a healthy state are collected and grouped by parameter type to form multiple normal power generation parameter data groups. For each data group, the K-Means clustering algorithm is independently applied to automatically divide the normal operating conditions in the multi-dimensional parameter space into several clusters, each representing a specific healthy operating mode, and the geometric center of each cluster is calculated. By calculating the maximum Euclidean distance from all data points within each cluster to its cluster center, a dynamic, data-driven anomaly detection boundary is defined for each operating mode, achieving accurate anomaly identification through multi-condition adaptive analysis. This approach can simultaneously consider the coordinated changes of multiple parameters, establishing different normal ranges for different operating modes and significantly reducing the false alarm rate. When new power generation forecast data is input, it is mapped to the corresponding parameter group cluster, and the calculation... The system calculates the distance to the center of its cluster. If the distance exceeds the cluster's individual anomaly threshold, it is marked as abnormal data, enabling early detection of systemic deviations. By statistically analyzing the number of abnormal data points in the predicted data set for each parameter and comparing it with a preset threshold, the system can not only detect instantaneous anomalies of individual data points but also capture the overall trend of parameters deviating from the normal pattern. This judgment based on "anomaly density" makes the system more sensitive to potential performance degradation or progressive failures, and can issue early warnings at an earlier stage of failure development. It elevates anomaly detection from a simple "single parameter-single threshold" judgment to an intelligent diagnostic level of "multi-parameter-multi-mode-dynamic threshold," providing a more accurate and reliable decision-making basis for predictive maintenance of wind turbines, effectively improving operation and maintenance efficiency and wind turbine reliability.

[0027] In embodiments of this application, a method for monitoring and early warning of wind power generation status is provided. The step of performing data feature analysis on abnormal parameters to determine abnormal parameter features includes: determining the power generation parameter prediction data group corresponding to each abnormal parameter, and determining the abnormal data in the power generation parameter prediction data group; determining the proportion of the abnormal data in the power generation parameter prediction data group, and determining the proportion as the first abnormal parameter feature corresponding to each abnormal parameter; determining the maximum and minimum values ​​in the abnormal data, calculating the difference between the maximum and minimum values ​​in the abnormal data, and determining the difference as the second abnormal parameter feature corresponding to each abnormal parameter; and determining the first abnormal parameter feature and the second abnormal parameter feature as the abnormal parameter feature of each abnormal parameter.

[0028] Specifically, features are constructed from two key dimensions. First, the proportion of anomalous data is calculated. This indicator reflects the frequency and prevalence of anomalies. A high proportion means that the parameter has experienced a systematic and widespread deviation throughout the prediction period, rather than isolated instantaneous fluctuations. Second, the range of anomalous data intervals is calculated, which is the difference between the maximum and minimum values ​​in the anomalous data set. This indicator reveals the severity and range boundaries of anomalous fluctuations. A large range indicates that the parameter not only deviates from normal but also exhibits significant fluctuations or remains at an extremely high level under anomalous conditions. This step enables in-depth insight and quantitative assessment of anomalous states, transforming the vague concept of "anomaly" into precise "breadth" and "intensity" indicators, providing operations and maintenance personnel with richer and more accurate information about failure modes. Multi-feature analysis lays a solid foundation for subsequent failure severity classification and root cause analysis. By integrating these two features, the priority of different anomalies can be intelligently assessed, thereby guiding the operations and maintenance team to formulate more targeted maintenance strategies and effectively improving the allocation efficiency of operations and maintenance resources and the accuracy of fault handling.

[0029] In embodiments of this application, a method for monitoring and early warning of wind power generation status is provided. The method involves evaluating the degree of abnormality of each abnormal parameter based on its abnormal parameter characteristics to obtain an evaluation value for the degree of abnormality of each parameter. This includes: evaluating and taking values ​​for a first abnormal parameter characteristic and a second abnormal parameter characteristic of each abnormal parameter to obtain a first abnormal evaluation value corresponding to the first abnormal parameter characteristic and a second abnormal evaluation value corresponding to the second abnormal parameter characteristic; determining a first standard abnormal evaluation value corresponding to the first abnormal parameter characteristic and a second standard abnormal evaluation value corresponding to the second abnormal parameter characteristic, and calculating the differences between the first abnormal evaluation value and the first standard abnormal evaluation value, as well as between the second abnormal evaluation value and the second standard abnormal evaluation value, to obtain a first difference and a second difference, respectively; scoring the first difference and the second difference based on a preset scoring model, and normalizing the scores to obtain weights corresponding to the first abnormal parameter characteristic and the second abnormal parameter characteristic, respectively; and calculating the degree of abnormality evaluation value of each abnormal parameter based on the first abnormal evaluation value corresponding to the first abnormal parameter characteristic, the second abnormal evaluation value corresponding to the second abnormal parameter characteristic, and the corresponding weights.

[0030] Specifically, two key features of each anomaly parameter—the anomaly percentage (first feature) and the anomaly range (second feature)—are independently evaluated and converted into specific numerical values ​​(first and second anomaly evaluation values). By introducing a benchmark "standard anomaly evaluation value," the severity of the deviation of the current anomaly state from the benchmark is accurately quantified by calculating the "difference" between each feature evaluation value and its corresponding standard value. The difference is mapped to a score through a preset scoring model and normalized to unify features (percentage and range) with different dimensions and physical meanings onto a comparable scale. Appropriate weights are dynamically allocated based on the contribution of the difference to the overall anomaly. By weighting the evaluation values ​​of each feature with their calculated weights, a single, quantitative anomaly severity evaluation value for each anomaly parameter is obtained, enabling precise prioritization of anomalies. This allows for clear identification of which of multiple simultaneously occurring anomaly parameters needs to be prioritized due to its wider anomaly range and more drastic fluctuations. Thus, multidimensional and complex anomaly information is extracted into intelligent signals that directly support decision-making, greatly improving the accuracy of early warning and the efficiency of operational actions.

[0031] In an embodiment of this application, a method for monitoring and early warning of wind power generation status is provided, wherein the formula for calculating the anomaly degree assessment value of the abnormal parameter is: , Where K is the abnormality assessment value of the abnormal parameter, α is the weight corresponding to the first abnormal parameter feature, P1 is the first abnormality assessment value corresponding to the first abnormal parameter feature, β is the weight corresponding to the second abnormal parameter feature, and P2 is the second abnormality assessment value corresponding to the second abnormal parameter feature.

[0032] In embodiments of this application, a method for monitoring and early warning of wind power generation status is provided. The method for comprehensively determining the state coefficient of a wind turbine based on the anomaly severity assessment values ​​of various abnormal parameters includes: determining a preset weight for each abnormal parameter, and weighting and adding the preset weights of each abnormal parameter with the anomaly severity assessment values ​​to obtain a comprehensive anomaly severity assessment value for the wind turbine; pre-setting a preset state coefficient-comprehensive anomaly severity assessment value interval correspondence relationship, wherein each comprehensive anomaly severity assessment value interval is associated with a corresponding preset state coefficient; obtaining the comprehensive anomaly severity assessment value of the wind turbine, and based on the mapping relationship between the comprehensive anomaly severity assessment value interval to which the comprehensive anomaly severity assessment value belongs and the preset state coefficient corresponding to the comprehensive anomaly severity assessment value interval correspondence relationship, selecting the preset state coefficient corresponding to the comprehensive anomaly severity assessment value interval as the state coefficient of the wind turbine.

[0033] Specifically, each identified abnormal parameter is assigned a preset weight. A comprehensive abnormality assessment value is calculated by weighting and summing the abnormality assessment values ​​of each parameter with their respective weights. Through a preset state coefficient mapping relationship, different intervals of comprehensive abnormality assessment values ​​are mapped to a standardized and easily understood state coefficient. Based on the calculated comprehensive assessment value, the corresponding interval is found, automatically determining the final state coefficient. This step condenses complex, multi-parameter abnormal information into a single, clear state coefficient, enabling maintenance personnel to instantly grasp the overall health of the wind turbine, greatly improving the intuitiveness of status monitoring. The preset, fixed mapping relationship eliminates the subjectivity and arbitrariness of human judgment, ensuring stable and consistent assessment results for the same abnormal situation, thus improving the standardization and normalization of maintenance management.

[0034] In the embodiments of this application, a method for monitoring and early warning of wind power generation status is provided. The method for determining the early warning strategy of the wind turbine based on the status coefficient includes: if the status coefficient of the wind turbine is less than a first preset value, the early warning strategy is to mark the wind turbine on the monitoring interface, list the specific abnormal parameters and abnormality assessment values, and include the wind turbine in the key inspection objects of the next regular inspection, reminding the inspection personnel to pay attention to specific components; if the status coefficient of the wind turbine is greater than the first preset value and less than the second preset value, the early warning strategy is to send a formal warning to the operation and maintenance team, automatically generate a preliminary maintenance work order, and start a deep diagnostic model to further locate the root cause of the fault; if the status coefficient of the wind turbine is greater than the second preset value and less than the third preset value, the early warning strategy is to dispatch the operation and maintenance team to conduct on-site inspection within a specified time, automatically adjust the wind turbine operation strategy, and formulate a specific maintenance plan and schedule; if the status coefficient of the wind turbine is greater than the third preset value, the early warning strategy is to automatically execute a protective shutdown, initiate an emergency maintenance process, and immediately dispatch a team to the site for handling.

[0035] like Figure 2As shown in the embodiments of this application, a wind power generation status monitoring and early warning system is provided, comprising: an acquisition module for acquiring historical operating data of a wind turbine, analyzing the historical operating data, and determining operating parameters related to power generation; a prediction module for extracting features of the operating parameters, constructing a power generation prediction model for the wind turbine based on the features and a preset neural network model, and obtaining power generation prediction data; an analysis module for performing data analysis on the power generation prediction data, determining abnormal parameters, and performing data feature analysis on the abnormal parameters to determine the abnormal parameter features; an evaluation module for evaluating the degree of abnormality of each abnormal parameter based on the abnormal parameter features, and obtaining an abnormality degree evaluation value for each abnormal parameter; and an early warning module for comprehensively determining the state coefficient of the wind turbine based on the abnormality degree evaluation values ​​of each abnormal parameter, and determining an early warning strategy for the wind turbine based on the state coefficient.

[0036] In summary, this invention provides a method and system for wind power generation status monitoring and early warning, comprising: acquiring and analyzing historical operating data of a wind turbine, determining operating parameters related to power generation, extracting their features, and constructing a power generation prediction model for the wind turbine based on the features and a preset neural network model to make predictions, thereby obtaining power generation prediction data; analyzing the power generation prediction data to determine abnormal parameters, and analyzing and determining the characteristics of these abnormal parameters; evaluating the degree of abnormality of each abnormal parameter based on the abnormal parameter characteristics to obtain an abnormality degree evaluation value, and comprehensively determining the state coefficient of the wind turbine based on this evaluation value, thereby determining the early warning strategy for the wind turbine. This invention can accurately predict potential risks before a fault occurs, assess the status of equipment, and provide scientific decision support for the operation and maintenance team, thereby minimizing unplanned downtime, improving power generation availability, extending equipment life, reducing operation and maintenance costs, and ensuring stable grid operation.

[0037] Finally, it should be noted that those skilled in the art can obviously make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims and their equivalents, this invention also intends to include these modifications and variations.

[0038] The above description is merely one embodiment of the present invention, and should not be construed as limiting the scope of the invention. Any structural changes made based on the present invention, as long as they do not depart from the essence of the invention, should be considered as falling within the protection scope of the present invention and subject to its restrictions. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process and related descriptions of the platform described above can be referred to the corresponding processes in the foregoing platform embodiments, and will not be repeated here.

[0039] The term "comprising" or any other similar term is intended to cover non-exclusive inclusion, such that a process, platform, article, or device / platform that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to those processes, platforms, articles, or devices / platforms.

[0040] The technical solutions of the present invention have been described in conjunction with the accompanying drawings and further embodiments. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to closely related technical features, and the technical solutions resulting from such changes or substitutions will all fall within the scope of protection of the present invention.

[0041] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention.

Claims

1. A method for monitoring and early warning of wind power generation status, characterized in that, include: Acquire historical operating data of wind turbines and analyze the historical operating data to determine operating parameters related to power generation; Features of operating parameters are extracted, and a power generation prediction model for wind turbines is constructed based on the features and a preset neural network model to make predictions and obtain power generation prediction data. Data analysis is performed on power generation forecast data to identify abnormal parameters, and data feature analysis is conducted on the abnormal parameters to determine their characteristics. The degree of abnormality of each abnormal parameter is evaluated based on its characteristics, and an evaluation value of the degree of abnormality of each abnormal parameter is obtained. The state coefficient of the wind turbine is determined by comprehensively evaluating the degree of abnormality of each abnormal parameter, and the early warning strategy of the wind turbine is determined based on the state coefficient.

2. A method for monitoring and early warning of wind power generation status according to claim 1, characterized in that, The process of acquiring historical operating data of wind turbines and analyzing this data to determine operating parameters related to power generation includes: Acquire historical operating data of wind turbines and divide the historical operating data into multiple operating parameter data groups according to parameter type; Historical power generation data is determined, the correlation between the operating parameter data set and the historical power generation data is calculated, and the parameter types corresponding to the operating parameter data set whose correlation exceeds a preset threshold are determined as operating parameters related to power generation.

3. A method for monitoring and early warning of wind power generation status according to claim 2, characterized in that, The extracted operating parameters are used to construct a power generation prediction model for the wind turbine based on the features and a preset neural network model, resulting in power generation prediction data, including: Features of the operating parameter data set corresponding to the operating parameters are extracted. A dataset is constructed based on the operating parameter data set and the corresponding features. The dataset is then input into a preset neural network model to construct an initial model for power generation prediction. The dataset is divided into training and test sets according to a preset ratio, and the training and test sets are input into the initial model for power generation prediction. The initial power generation prediction model is trained and tested until it meets the preset convergence conditions, thus obtaining the power generation prediction model. The real-time operating data of the wind turbine is obtained and input into the power generation prediction model for prediction, so as to obtain the power generation prediction data for a period of time in the future.

4. A method for monitoring and early warning of wind power generation status according to claim 3, characterized in that, The process of analyzing power generation forecast data to determine abnormal parameters includes: Determine the normal power generation data of the wind turbine and divide the normal power generation data into multiple normal power generation parameter data groups according to parameter type; Cluster analysis was performed on each normal power generation parameter data group based on the K-Means clustering algorithm to obtain multiple clusters and cluster centers corresponding to each normal power generation parameter data group; Calculate the maximum distance from all points within each cluster to the center of each cluster, and determine this maximum distance as the anomaly threshold for each cluster; According to the parameter type, the power generation forecast data is divided into multiple power generation parameter forecast data groups, and the power generation parameter forecast data groups are mapped to the corresponding clusters of normal power generation parameter data groups in the same way. Calculate the distance from newly added points within each cluster to the center of each cluster, and identify the data corresponding to newly added points whose distance is greater than the maximum distance as outliers in each power generation parameter prediction data group. Determine the number of abnormal data in each power generation parameter prediction data group, and identify the parameter type corresponding to the power generation parameter prediction data group whose number of abnormal data exceeds a preset threshold as abnormal parameters.

5. A method for monitoring and early warning of wind power generation status according to claim 4, characterized in that, The step of performing data feature analysis on abnormal parameters to determine the characteristics of abnormal parameters includes: Identify the power generation parameter prediction data group corresponding to each abnormal parameter, and identify the abnormal data in the power generation parameter prediction data group; Determine the proportion of abnormal data in the power generation parameter prediction data set, and define this proportion as the first abnormal parameter feature corresponding to each abnormal parameter; Identify the maximum and minimum values ​​in the outlier data, calculate the difference between the maximum and minimum values ​​in the outlier data, and determine the difference as the second outlier parameter feature corresponding to each outlier parameter; The first and second abnormal parameter features are determined as the abnormal parameter features of each abnormal parameter.

6. A method for monitoring and early warning of wind power generation status according to claim 5, characterized in that, The assessment of the abnormality degree of each abnormal parameter based on its characteristics, to obtain an assessment value for the abnormality degree of each abnormal parameter, includes: The first and second abnormal parameter features of each abnormal parameter are evaluated and their values ​​are obtained respectively to obtain the first abnormal evaluation value corresponding to the first abnormal parameter feature and the second abnormal evaluation value corresponding to the second abnormal parameter feature. Determine the first standard anomaly evaluation value corresponding to the first anomaly parameter feature and the second standard anomaly evaluation value corresponding to the second anomaly parameter feature, and calculate the difference between the first anomaly evaluation value and the first standard anomaly evaluation value, as well as the difference between the second anomaly evaluation value and the second standard anomaly evaluation value, to obtain the first difference and the second difference respectively; The first difference and the second difference are scored based on the preset scoring model, and the scores are normalized to obtain the weights corresponding to the first abnormal parameter feature and the second abnormal parameter feature. The abnormality assessment value of each abnormal parameter is obtained by calculating the first abnormality assessment value corresponding to the first abnormal parameter feature and the second abnormality assessment value corresponding to the second abnormal parameter feature, as well as the corresponding weight.

7. A method for monitoring and early warning of wind power generation status according to claim 6, characterized in that, The formula for calculating the abnormality assessment value of the abnormal parameter is as follows: ， Where K is the abnormality assessment value of the abnormal parameter, α is the weight corresponding to the first abnormal parameter feature, P1 is the first abnormality assessment value corresponding to the first abnormal parameter feature, β is the weight corresponding to the second abnormal parameter feature, and P2 is the second abnormality assessment value corresponding to the second abnormal parameter feature.

8. A method for monitoring and early warning of wind power generation status according to claim 6, characterized in that, The determination of the state coefficient of the wind turbine based on the anomaly severity assessment values ​​of each anomaly parameter includes: Determine the preset weights of each abnormal parameter, and then add the preset weights of each abnormal parameter to the abnormality assessment value to obtain the comprehensive abnormality assessment value of the wind turbine. A preset state coefficient-comprehensive anomaly assessment value range correspondence is set in advance. For each comprehensive anomaly assessment value range, a corresponding preset state coefficient is associated with it. Obtain the comprehensive anomaly assessment value of the wind turbine, and based on the mapping relationship between the comprehensive anomaly assessment value interval to which the comprehensive anomaly assessment value belongs and the preset state coefficient-comprehensive anomaly assessment value interval correspondence, select the preset state coefficient corresponding to the comprehensive anomaly assessment value interval as the state coefficient of the wind turbine.

9. A method for monitoring and early warning of wind power generation status according to claim 8, characterized in that, The early warning strategy for determining wind turbine generators based on state coefficients includes: If the state coefficient of the wind turbine is less than the first preset value, the early warning strategy for the wind turbine is to mark the turbine on the monitoring interface, list the specific abnormal parameters and the abnormality assessment value, and include the turbine in the key inspection objects of the next regular inspection, reminding the inspection personnel to pay attention to specific components. If the state coefficient of the wind turbine is greater than the first preset value and less than the second preset value, the early warning strategy for the wind turbine is to send a formal warning to the operation and maintenance team, automatically generate a preliminary maintenance work order, and start a deep diagnostic model to further locate the root cause of the fault. If the state coefficient of the wind turbine is greater than the second preset value and less than the third preset value, the early warning strategy for the wind turbine is to dispatch the operation and maintenance team to conduct on-site inspection within a specified time, automatically adjust the wind turbine operation strategy, and formulate specific maintenance plans and schedules. If the state coefficient of the wind turbine exceeds the third preset value, the early warning strategy for the wind turbine is to automatically execute a protective shutdown, initiate an emergency maintenance process, and immediately dispatch a team to the site for handling.

10. A wind power generation condition monitoring and early warning system, characterized in that, include: The acquisition module is used to acquire historical operating data of wind turbines and analyze the historical operating data to determine operating parameters related to power generation. The prediction module is used to extract features of operating parameters, construct a power generation prediction model for wind turbines based on the features and a preset neural network model, and obtain power generation prediction data. The analysis module is used to perform data analysis on power generation forecast data, identify abnormal parameters, and perform data feature analysis on the abnormal parameters to determine the characteristics of the abnormal parameters. The evaluation module is used to evaluate the degree of abnormality of each abnormal parameter based on its characteristics, and obtain the evaluation value of the degree of abnormality of each abnormal parameter. The early warning module is used to comprehensively determine the state coefficient of the wind turbine based on the anomaly severity assessment values ​​of each abnormal parameter, and to determine the early warning strategy for the wind turbine based on the state coefficient.

Images