Abnormal monitoring and early warning method and device for temperature of generator bearing of wind turbine generator

Abstract

The invention relates to the field of wind turbine generator fault diagnosis, and discloses an abnormal monitoring and early warning method and device for the temperature of a generator bearing of a wind turbine generator. The method comprises the following steps: acquiring and preprocessing online SCADA data of a wind turbine generator; bearing temperature related variables are extracted and input into a preset model to obtain a temperature theoretical value; extracting an actual temperature value, calculating the temperature difference between the actual temperature value and the theoretical temperature value, and judging abnormity when the temperature difference exceeds a preset value; according to the method and the device, the efficiency and the accuracy of fault diagnosis of the wind turbine generator are improved, and accurate monitoring and early warning of the temperature abnormity of the generator bearing of the wind turbine generator are realized.

Description

Technical Field

[0001] The present application relates to the technical field of wind turbine fault diagnosis, and in particular to a method and device for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator. Background Art

[0002] With the rapid development of renewable energy, wind power generation, as a clean energy, has accounted for an increasing proportion of the global energy structure year by year. The core components of wind power generation equipment include wind rotors, blades, towers, bases, transmission systems and control systems. With the rapid development of the wind power industry, the capacity of single units continues to increase, and the stability and reliability of wind turbines have become particularly important. The failure of wind turbines not only increases maintenance costs, but also has a negative impact on the healthy development of the wind power industry. During the operation of wind turbines, abnormal generator bearing temperature is one of the common faults. If it is not discovered and handled in time, it may cause system failures, reduce wind energy utilization, and even cause significant economic losses. Therefore, accurately identifying abnormal generator bearing temperature and realizing early warning of faults is of great significance for reducing safety risks and power generation losses.

[0003] Similar prior art includes a Chinese patent application with publication number CN112577739A, which discloses a method for diagnosing and warning over-temperature faults of the engine drive end bearing of a wind turbine. Based on the multivariate regression analysis method, a nonlinear causal relationship between the bearing temperature and related variables is established to obtain a reference curve of the engine drive end bearing temperature in a healthy state of the wind turbine; the deviation degree of the real-time regression curve of the bearing temperature relative to the healthy reference curve is quantified to obtain a health index reflecting the potential over-temperature fault of the bearing; based on the obtained health index reflecting the potential over-temperature fault of the bearing, the over-temperature fault of the bearing is monitored and warned by setting a threshold. This method relies on the accuracy of the reference curve. If the reference curve is not established accurately, it will directly affect the reliability of the health index. Moreover, its quantification method does not take into account the abnormal values ​​or noise in the data. When the bearing temperature fluctuates abnormally, the accuracy of the health index is affected. There is also a Chinese patent application with publication number CN112598172A, which discloses a wind turbine bearing temperature early warning method, which obtains the wind turbine SCADA temperature data set; preprocesses the wind turbine SCADA temperature data set and screens the model input variables; divides the data set and establishes a wind turbine bearing temperature prediction model based on GS-LightGBM; predicts the wind turbine bearing operation data based on the prediction model, performs residual analysis on it, calculates the EWMA control chart parameters, obtains the alarm threshold, and analyzes the working status. This method only provides abnormal early warning based on temperature data, lacks comprehensive consideration of bearing temperature abnormality factors, and is not comprehensive enough.

[0004] Therefore, it is an urgent problem to provide a method and device for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator to improve the effectiveness and accuracy of abnormal monitoring of the wind turbine generator. Summary of the invention

[0005] The present application provides a method and device for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator, which are used to improve the efficiency and accuracy of abnormal monitoring and early warning of the bearing temperature of a wind turbine generator.

[0006] In a first aspect, the present application provides a method for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator. The method for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator comprises:

[0007] Step 1, obtaining online SCADA data of the wind turbine generator, and preprocessing the online SCADA data;

[0008] Step 2: extract bearing temperature related variables from the preprocessed online SCADA data, input the related variables into the preset model, and obtain the theoretical value of bearing temperature;

[0009] Step 3, extracting the actual value of the bearing temperature from the preprocessed online SCADA data, calculating the temperature difference between the actual value of the bearing temperature and the theoretical value of the bearing temperature, and determining that an abnormality exists when the absolute value of the temperature difference is greater than a first preset value;

[0010] Step 4: Acquire the operating data of the first generator within the first preset time, diagnose the generator bearing based on the operating data of the first generator, and issue a warning message.

[0011] In combination with the first aspect, in a first implementation of the first aspect of the present application, preprocessing the online SCADA data includes:

[0012] Step 11, determining whether there is missing data or abnormal data in the online SCADA data, and if so, defining the variable corresponding to the missing data or abnormal data as the first variable;

[0013] Step 12: extract any first variable, obtain the theoretical number of collections of any first variable within the second preset time, and the invalid number of times when the data corresponding to any first variable is missing data or abnormal data, calculate the ratio of the invalid number of times to the theoretical number of collections, and define it as inefficiency;

[0014] Step 13: After traversing all first variables, based on inefficiency, divide all first variables into a preset number of variable groups according to preset rules;

[0015] Step 14: Set a corresponding data generation strategy for each variable group, extract any variable group, use the data generation strategy corresponding to any variable group, generate replacement values ​​for each first variable in any variable group, and use the replacement values ​​to complete the online SCADA data.

[0016] In combination with the first aspect, in a second implementation of the first aspect of the present application, the preset number is 3, the first variable with an inefficiency less than the second preset value is divided into the first variable group, the first variable with an inefficiency greater than or equal to the second preset value and less than or equal to the third preset value is divided into the second variable group, and the first variable with an inefficiency greater than the third preset value is divided into the third variable group, and step 14 includes:

[0017] For the first variable group, define the variables in the first variable group as second variables, obtain the data sequence of any second variable within the third preset time, calculate the estimated value of any second variable at each time point based on formula 1, and use the estimated value of any second variable at time T-1 as the replacement value of any second variable at time T. Formula 1 is:

[0018]

[0019] Among them, E(t) is the estimated value at time t, α is the parameter weight coefficient, D(t) is the actual value at time t, and T is the current time;

[0020] For the third variable group, the variables in the third variable group are defined as third variables, and the data obtained at time T-1 for any third variable is used as the replacement value of any third variable at time T;

[0021] For the second variable group, the variables in the second variable group are defined as the fourth variables. First, based on the data generation strategy corresponding to the third variable group, a first replacement value at time T is generated for any fourth variable, and the first replacement value is used as the proposed numerical value of any fourth variable at time T. Subsequently, based on the data generation strategy corresponding to the first variable group, an estimated value of any fourth variable at time T is generated and defined as the second replacement value. The average of the first replacement value and the second replacement value is set as the replacement value of any first variable at time T.

[0022] In combination with the first aspect, in a third implementation of the first aspect of the present application, the method for establishing the preset model is:

[0023] Step 21, obtaining historical SCADA data of the wind turbine bearing, the historical SCADA data including N1 data sequences, each data sequence including the second generator operation data, wherein N1 is a positive integer;

[0024] Step 22, traversing the historical SCADA data, when the data corresponding to any characteristic variable in the second generator operation data is missing, defining any characteristic variable as the first characteristic variable, and at the same time, dividing the historical SCADA data into a training data set and a verification data set according to a preset ratio;

[0025] Step 23: arbitrarily combine all first characteristic variables to obtain N2 characteristic variable combinations, where N2 is a positive integer;

[0026] Step 24, extract any characteristic variable combination, define the variable in any characteristic variable combination as the third characteristic variable, define the variable outside any characteristic variable combination as the second characteristic variable, delete all data corresponding to any second characteristic variable in the training data set, and delete the data sequence lacking the data corresponding to the third characteristic variable, generate a first training data set, repeat step 24 until all characteristic variable combinations are traversed, and generate N3 first training data sets, where N3 is a positive integer less than or equal to N2;

[0027] Step 25, performing the same operation as step 24 on the verification data set to generate N4 first verification data sets, where N4 is a positive integer less than or equal to N2;

[0028] Step 26: Perform machine learning on any first training data set to obtain a first preset model corresponding to any first training data set, and after traversing all first training data sets, generate N3 first preset models;

[0029] Step 27: Use the first verification data set to verify each first preset model, obtain the accuracy of each first preset model respectively, and use the first preset model corresponding to the maximum accuracy as the preset model.

[0030] In combination with the first aspect, in a fourth implementation of the first aspect of the present application, step 27 includes:

[0031] Extract any first preset model, and define the feature variable used in training any first preset model as a fourth feature variable;

[0032] Extract any first validation data set, define the characteristic variable contained in any first validation data set as a fifth characteristic variable, and when the fifth characteristic variable contains the fourth characteristic variable, define any first validation data set as a second validation data set;

[0033] After traversing all the first validation data sets, each second validation data set is used to validate any first preset model, multiple model accuracies are obtained, a first statistical value of the accuracies of all models is calculated, and the first statistical value is used as the accuracy of any first preset model.

[0034] In combination with the first aspect, in a fifth implementation of the first aspect of the present application, in step 2, extracting bearing temperature related variables from the preprocessed online SCADA data includes:

[0035] The characteristic variable used in the preset model training is defined as the sixth characteristic variable, and the relevant variables are extracted from the first SCADA data based on the sixth characteristic variable.

[0036] In combination with the first aspect, in the sixth implementation method of the first aspect of the present application, the second generator operating data includes bearing temperature data, cabin temperature data, ambient temperature data, wind speed, wind direction, and the power, lubrication status data, speed, torque and power of each bearing component.

[0037] In combination with the first aspect, in a seventh implementation of the first aspect of the present application, step 4 includes:

[0038] dividing the first generator operation data into a plurality of variable data sequences based on the variable type, respectively calculating a second statistical value of each variable data sequence, and generating a first characteristic vector corresponding to the first generator operation data based on the second statistical value;

[0039] Extracting a second eigenvector corresponding to each fault event within a third preset time, and calculating the Euclidean distance between any second eigenvector and the first eigenvector, wherein the second eigenvector of any fault event is generated based on historical generator operation data within the first preset time before a specific moment, and the specific moment is a moment at a specific time away from the moment when any fault event occurs;

[0040] It is determined whether the minimum Euclidean distance is less than a fourth preset value. If so, it is determined that there is a fault risk.

[0041] In combination with the first aspect, in an eighth implementation of the first aspect of the present application, after determining that there is a failure risk, the following steps are further included:

[0042] The statistical value in the second eigenvector corresponding to the minimum Euclidean distance is defined as the third statistical value, and the sensitivity corresponding to any variable type is calculated based on Formula 2, which is:

[0043]

[0044] Among them, G i is the sensitivity corresponding to the i-th variable type, ST i,2 is the second statistical value of the i-th variable type, ST i,3 is the third statistical value of the i-th variable type, and d is the minimum Euclidean distance;

[0045] When the absolute value of the sensitivity corresponding to any variable type is greater than the preset threshold, any variable type is determined as an abnormal influencing factor;

[0046] After traversing all variable types, early warning information is generated based on all abnormal influencing factors.

[0047] In a second aspect, the present application provides a device for monitoring and warning the abnormality of the bearing temperature of a wind turbine generator, wherein the device comprises: a preprocessing module, a temperature analysis module, an abnormality judgment module and an abnormality diagnosis module;

[0048] A preprocessing module is used to obtain online SCADA data of wind turbine generators and preprocess the online SCADA data;

[0049] The temperature analysis module is used to extract bearing temperature related variables from the pre-processed online SCADA data, input the related variables into the preset model, and obtain the theoretical value of bearing temperature;

[0050] An abnormality judgment module is used to extract the actual value of the bearing temperature from the pre-processed online SCADA data, calculate the temperature difference between the actual value of the bearing temperature and the theoretical value of the bearing temperature, and judge that an abnormality exists when the absolute value of the temperature difference is greater than a first preset value;

[0051] The abnormality diagnosis module is used to obtain the operating data of the first generator within a first preset time, diagnose the generator bearing based on the first generator operating data, and issue a warning message.

[0052] The technical solution provided by this application has at least the following beneficial effects:

[0053] 1. After acquiring online SCADA data, the missing or abnormal data is identified and processed. Through variable grouping and data generation strategies, the data can be effectively completed or replaced to reduce the impact of missing data on the analysis results. Different data generation strategies are used for different inefficient data to ensure the rationality, accuracy and completeness of data completion, providing a reliable basis for subsequent analysis.

[0054] 2. The method for establishing the preset model takes into account the combination of multiple feature variables, and is trained and verified through machine learning methods, which can screen out the optimal model and improve the accuracy of the prediction.

[0055] 3. The fault diagnosis method based on generator operation data comprehensively considers the influence of multiple variable types and accurately determines the fault risk by calculating the characteristic vector and Euclidean distance. The sensitivity calculation can further identify abnormal influencing factors and provide strong support for finding the cause of the fault and generating early warning information. It can issue a warning at the early stage of the fault, so that maintenance personnel can take timely measures to reduce downtime and maintenance costs caused by the fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0057] Figure 1 A schematic diagram of an embodiment of a method for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator in an embodiment of the present application;

[0058] Figure 2 Schematic diagram of an embodiment of a device for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The embodiment of the present application provides a method and device for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the abnormal monitoring and early warning method of the bearing temperature of the wind turbine generator includes:

[0061] Step 1: Obtain online SCADA data of wind turbine generators and preprocess the online SCADA data.

[0062] Specifically, online SCADA data includes but is not limited to data collection time, wind turbine generator identification, bearing temperature data, cabin temperature data, ambient temperature data, wind speed, wind direction, and the power of each bearing component, lubrication status data, speed, torque and power, bearing vibration frequency, generator shaft torque, etc.

[0063] In a specific embodiment, in step 1, preprocessing the online SCADA data includes:

[0064] Step 11: determine whether there is missing data or abnormal data in the online SCADA data. If yes, define the variable corresponding to the missing data or abnormal data as the first variable.

[0065] Step 12: extract any first variable, obtain the theoretical collection times of any first variable within the second preset time, and the invalid times when the data corresponding to any first variable is missing data or abnormal data, calculate the ratio of the invalid times to the theoretical collection times, and define it as inefficiency.

[0066] Step 13: After traversing all first variables, based on inefficiency, all first variables are divided into a preset number of variable groups according to preset rules.

[0067] Step 14: Set a corresponding data generation strategy for each variable group, extract any variable group, use the data generation strategy corresponding to any variable group, generate replacement values ​​for each first variable in any variable group, and use the replacement values ​​to complete the online SCADA data.

[0068] The second preset time and preset number are set according to the experience of those skilled in the art or according to actual application scenarios, and are not limited in the embodiments of the present application.

[0069] Specifically, the above abnormal data is caused by data acquisition equipment failure or transmission error. For example, the temperature is 4000°C and the wind speed is -5m / s. These values ​​are obviously inconsistent with the actual situation and can be identified as noise or abnormal values. Furthermore, the above abnormal data can also be identified based on statistics such as standard deviation or interquartile range (IQR).

[0070] Inefficiency is an important indicator of data quality, which can help us understand the data integrity and reliability of each variable. By calculating the inefficiency, we can have a quantitative understanding of the quality of the data. The inefficiency of different variables may vary greatly. Dividing all first variables into a preset number of variable groups based on inefficiency can be processed more targeted. For example, variables with low inefficiency may only require simple processing, while variables with higher inefficiency may require more complex data generation strategies.

[0071] Use replacement values ​​to replace abnormal data or use replacement values ​​to supplement missing data to complete SCADA data.

[0072] In a specific embodiment, the preset number is 3, the first variable with an inefficiency less than the second preset value is divided into the first variable group, the first variable with an inefficiency greater than or equal to the second preset value and less than or equal to the third preset value is divided into the second variable group, and the first variable with an inefficiency greater than the third preset value is divided into the third variable group, and step 14 includes:

[0073] (1) For the first variable group, define the variables in the first variable group as second variables, obtain the data sequence of any second variable within the third preset time, calculate the estimated value of any second variable at each time point based on formula 1, and use the estimated value of any second variable at time T-1 as the replacement value of any second variable at time T. Formula 1 is:

[0074]

[0075] Among them, E(t) is the estimated value at time t, α is the parameter weight coefficient, D(t) is the actual value at time t, and T is the current time.

[0076] (2) For the third variable group, the variables in the third variable group are defined as third variables, and the data obtained at time T-1 for any third variable is used as the replacement value of any third variable at time T.

[0077] (3) For the second variable group, the variables in the second variable group are defined as fourth variables. First, based on the data generation strategy corresponding to the third variable group, a first replacement value is generated for any fourth variable at time T, and the first replacement value is used as the proposed numerical value of any fourth variable at time T. Subsequently, based on the data generation strategy corresponding to the first variable group, an estimated value of any fourth variable at time T is generated and defined as a second replacement value. The average of the first replacement value and the second replacement value is set as the replacement value of any first variable at time T.

[0078] The second preset value, the third preset value and the third preset time are set according to the experience of those skilled in the art or according to actual application scenarios, and the embodiments of the present application are not limited thereto.

[0079] Exemplarily, the data sequence A of temperature within the third preset time is [20, 21, 22], where a1 is 20, a2 is 21, a3 is 22, a4 is an abnormal value, and a replacement value is generated for a4. When the data sequence A belongs to the first variable group, if α is 0.2, a1 is the initial value, its corresponding estimated value is 20, the corresponding estimated value of a2 is 20.2, the corresponding estimated value of a3 is 20.56, and the actual value of a4 is missing, then 20.56 is used as its replacement value. When the data sequence A belongs to the third variable group, the replacement value of a4 is 22. When the data sequence A belongs to the second variable group, the first replacement value generated for a4 is 22 based on the data generation strategy corresponding to the third variable group, and then on the basis of taking 22 as the proposed value of a4, the data generation strategy corresponding to the first variable group is used to generate a second replacement value of 20.848 for a4, and the average value of the first replacement value and the second replacement value 21.424 is used as the replacement value of a4.

[0080] If the inefficiency is less than or equal to the second preset value, it means that the variable has a high availability rate. If the inefficiency is greater than the third preset value, it means that the variable has a low availability rate. The variables in the first variable group have high availability, less missing data, and contain sufficient information to estimate trends and patterns. The replacement values ​​generated according to Formula 1 can consider data from multiple time points, maintain the continuity and consistency of the data, reduce data noise caused by random fluctuations, and improve the accuracy of generating replacement values. The variables in the third variable group have low availability and more missing data. The parameter does not change frequently or the change trend is not obvious over time. If a complex interpolation method is used, it may lead to overfitting and introduce too many estimation errors. By simply continuing the last observation, the model complexity and the risk of overfitting are reduced, and data can be processed quickly and easily, reducing the consumption of computing resources. The variables in the second variable group have a certain number of missing values, and their availability is not very high. The integrity of the data set is affected to a certain extent, but there is still enough data for analysis. Such variables may not change as frequently as the variables in the first variable group, nor as slowly or stably as the variables in the third variable group. They may be somewhere in between and have a certain degree of dynamics. By combining the results of two different data generation strategies, the bias that may be introduced by a single method can be reduced and the accuracy of generating replacement values ​​can be improved.

[0081] By customizing the data generation strategy for each variable group, it is possible to process data with different characteristics more accurately and effectively, improve the quality of data processing, enhance the accuracy and robustness of data, reduce performance fluctuations due to data quality issues, thereby improving the accuracy of subsequent analysis and providing reliable early warning information.

[0082] Step 2: Extract bearing temperature related variables from the preprocessed online SCADA data, input the related variables into the preset model, and obtain the theoretical value of bearing temperature.

[0083] Specifically, the above-mentioned preset model is a model for obtaining the theoretical temperature value of the bearing by training temperature-related variables.

[0084] In a specific embodiment, the method for establishing the preset model is:

[0085] Step 21: Acquire historical SCADA data of the wind turbine bearing, the historical SCADA data includes N1 data sequences, each data sequence includes the second generator operation data, wherein N1 is a positive integer.

[0086] Step 22, traverse the historical SCADA data, when the data corresponding to any characteristic variable in the second generator operation data is missing, define any characteristic variable as the first characteristic variable, and at the same time, divide the historical SCADA data into a training data set and a verification data set according to a preset ratio.

[0087] Step 23: arbitrarily combine all first characteristic variables to obtain N2 characteristic variable combinations, where N2 is a positive integer.

[0088] Step 24, extract any combination of feature variables, define the variables in any combination of feature variables as the third feature variable, define the variables outside any combination of feature variables as the second feature variable, delete all data corresponding to any second feature variable in the training data set, and delete the data sequence that lacks the data corresponding to the third feature variable to generate a first training data set, repeat step 24 until all feature variable combinations are traversed, and generate N3 first training data sets, where N3 is a positive integer less than or equal to N2.

[0089] Step 25, performing the same operation as step 24 on the verification data set to generate N4 first verification data sets, where N4 is a positive integer less than or equal to N2.

[0090] Step 26: Perform machine learning on any first training data set to obtain a first preset model corresponding to any first training data set, and after traversing all first training data sets, generate N3 first preset models.

[0091] Step 27: Use the first verification data set to verify each first preset model, obtain the accuracy of each first preset model respectively, and use the first preset model corresponding to the maximum accuracy as the preset model.

[0092] Specifically, the second generator operation data includes but is not limited to bearing temperature data, cabin temperature data, ambient temperature data, wind speed, wind direction, and power of various bearing components, lubrication status data, speed, torque and power. The second generator operation data includes various equipment operation data related to bearing temperature.

[0093] Exemplarily, historical SCADA data includes 11 data sequences, x1, x2, ..., x11, each of which has 9 variables, b1, b2, ..., b9, wherein the data corresponding to the variable b1 in the x1 sequence is missing, and the data corresponding to the variable b5 in the x9 sequence is missing, then b1 and b5 are defined as the first characteristic variables. There are two first characteristic variables, b1 and b5, and any combination of the two first characteristic variables can be performed, and there are 3 combinations, b1, b5, (b1, b5). When b1 is used as the third characteristic variable and b5 is the second characteristic variable, if the training data set is (x1, x2, ..., x7) and the verification data set is (x8, x9, x10, x11), for the training data set, the variable b5 is deleted from 7 data sequences (remaining b1, b2, b3, b4, b6, ..., b9, 8 variables), and the data sequence x1 with the third characteristic variable b1 missing is deleted (remaining x2, ..., x7, 6 data sequences); for the verification data set, the variable b5 is deleted from 4 data sequences (remaining b1, b2, b3, b4, b6, ..., b9, 8 variables). Since there is no data sequence missing the third characteristic variable b1 in the verification data set, no further operation is required.

[0094] The technical solution of the present application provides flexibility and adjustability of model training by training multiple models according to different combinations of characteristic variables in the presence of missing data, and maximizes the use of available data for effective learning and prediction. By creating multiple training data sets, training multiple first preset models, and further evaluating and comparing multiple first preset models, and selecting the best preset model, the influence of different combinations of characteristic variables on the model prediction ability can be explored, which helps to reduce the influence of variable noise that does not contribute to the prediction of theoretical bearing temperature on the preset model, thereby improving the robustness and accuracy of the model.

[0095] In a specific embodiment, step 27 includes:

[0096] (1) Extract any first preset model, and define the feature variable used when training any first preset model as the fourth feature variable.

[0097] (2) Extract any first validation data set, define the feature variable contained in any first validation data set as the fifth feature variable, and when the fifth feature variable contains the fourth feature variable, define any first validation data set as the second validation data set.

[0098] (3) After traversing all the first verification data sets, use each second verification data set to verify any first preset model, obtain multiple model accuracies, calculate the first statistical value of the accuracies of all models, and use the first statistical value as the accuracy of any first preset model.

[0099] Specifically, the first statistical value is the average of the accuracy rates of all models.

[0100] Using a validation data set that includes the variables used in training the first preset model to validate the first preset model can ensure the pertinence, accuracy, and effectiveness of the model validation, and help improve the accuracy and reliability of model selection.

[0101] In a specific embodiment, in step 2, extracting bearing temperature related variables from the pre-processed online SCADA data includes:

[0102] The characteristic variable used in the preset model training is defined as the sixth characteristic variable, and the relevant variables are extracted from the first SCADA data based on the sixth characteristic variable.

[0103] Specifically, the variables most relevant to bearing temperature prediction are extracted more specifically according to the sixth characteristic variable, reducing the model performance fluctuation caused by inconsistent characteristic variables, reducing the impact of non-contributory variable noise on model prediction, improving the robustness of the model and the accuracy of data processing, and at the same time accelerating data processing speed and improving data processing efficiency.

[0104] Step 3: extract the actual value of the bearing temperature from the preprocessed online SCADA data, calculate the temperature difference between the actual value of the bearing temperature and the theoretical value of the bearing temperature, and when the absolute value of the temperature difference is greater than the first preset value, determine that an abnormality exists.

[0105] The first preset value is set according to the experience of those skilled in the art or according to an actual application scenario, and the embodiments of the present application are not limited to this.

[0106] The theoretical bearing temperature is calculated based on a preset model, representing the state of the bearing temperature under normal conditions, and provides a benchmark for temperature assessment. The actual bearing temperature is affected by many factors, including environmental changes, equipment aging, changes in operating conditions, etc. These factors may cause deviations between the actual temperature and the theoretical temperature. By comparing the two, these unknown abnormalities can be detected.

[0107] Abnormal bearing temperature may be a precursor to equipment failure. By comparing the actual temperature with the theoretical temperature, problems can be discovered before failure occurs, so that preventive maintenance measures can be taken to ensure the stable operation and safety of the wind turbine.

[0108] Step 4: Acquire the operating data of the first generator within the first preset time, diagnose the generator bearing based on the operating data of the first generator, and issue a warning message.

[0109] The first preset time is set according to the experience of those skilled in the art or according to an actual application scenario, and the embodiment of the present application is not limited to this.

[0110] The first generator operation data includes but is not limited to bearing temperature data, cabin temperature data, ambient temperature data, wind speed, wind direction, and power of various bearing components, lubrication status data, speed, torque and power. The first generator operation data includes various equipment status data related to bearing abnormality diagnosis.

[0111] In a specific embodiment, step 4 includes:

[0112] (1) Dividing the first generator operation data into a plurality of variable data sequences based on variable types, respectively calculating a second statistical value of each variable data sequence, and generating a first eigenvector corresponding to the first generator operation data based on the second statistical value.

[0113] (2) extracting a second eigenvector corresponding to each fault event within a third preset time, and calculating the Euclidean distance between any second eigenvector and the first eigenvector, wherein the second eigenvector of any fault event is generated based on historical generator operation data within the first preset time before a specific moment, and the specific moment is a moment at a specific time away from the moment when any fault event occurs.

[0114] (3) Determine whether the minimum Euclidean distance is less than a fourth preset value. If so, determine that there is a fault risk.

[0115] The third preset time and the fourth preset value are set according to the experience of those skilled in the art or according to actual application scenarios, and are not limited in the embodiments of the present application.

[0116] Specifically, the second statistical value is the average value. The variable type is the type of each variable in the first generator operation data, including but not limited to pressure, temperature, humidity, and rotation speed. Exemplarily, the first generator operation data includes three variables c1, c2 and c3, then the first generator operation data is divided into three variable data sequences, the variable digital sequence y1 corresponding to the variable c1: [c11, ..., c1m], the variable digital sequence y2 corresponding to the variable c2: [c21, ..., c2m], the variable digital sequence y3 corresponding to the variable c3: [c31, ..., c3m], the second statistical value of the variable digital sequence y1 is E1, the second statistical value of the variable digital sequence y2 is E2, and the second statistical value of the variable digital sequence y3 is E3, then the first eigenvector corresponding to the first generator operation data is [E1, E2, E3].

[0117] The second eigenvector is a eigenvector corresponding to the generator operation data within a period of time before the fault event occurs, and the second eigenvector is obtained using the same method as the first eigenvector. Exemplarily, the specific moment is the time when the generator operation data was last obtained before the fault event occurs. If the first preset time is 30 minutes, the occurrence time of the fault event P is 10:00, and the last time the generator operation data was obtained before the fault event P occurs is 9:55, then the second eigenvector corresponding to the fault event P is the eigenvector corresponding to the generator operation data within 9:25-9:55.

[0118] Preferably, the specific time may also be the time when the abnormal generator operation data is first acquired before the fault event occurs.

[0119] By calculating the Euclidean distance between the second eigenvector and the first eigenvector and comparing the similarity between the first eigenvector and the second eigenvector corresponding to the fault event, it is possible to quickly identify whether the current operating condition of the generator is similar to past fault conditions, thereby helping to determine whether a fault is about to occur and identifying fault events in advance, so that timely measures can be taken to intervene.

[0120] If the minimum Euclidean distance is less than the fourth preset value, it means that the current generator operation state is highly similar to the feature vector in the historical fault data, and may be in a high risk state of fault occurrence. By identifying the second feature vector with the minimum Euclidean distance, it can help analyze the cause of the fault, understand the potential factors leading to the fault, provide a basis for troubleshooting, simplify the decision-making process, and provide clear fault warnings and recommended action plans.

[0121] Preferably, the warning information also includes a fault prediction time, which is generated based on the time from a specific moment of the fault event corresponding to the minimum Euclidean distance to the moment of fault occurrence.

[0122] By predicting when a generator may fail in the future, it helps to effectively plan maintenance activities and improve operation and maintenance efficiency.

[0123] In a specific embodiment, after determining that there is a failure risk, the following steps are also included:

[0124] (1) The statistical value in the second eigenvector corresponding to the minimum Euclidean distance is defined as the third statistical value, and the sensitivity corresponding to any variable type is calculated based on Formula 2, which is:

[0125]

[0126] Among them, G i is the sensitivity corresponding to the i-th variable type, ST i,2 is the second statistical value of the i-th variable type, ST i,3is the third statistical value of the i-th variable type, and d is the minimum Euclidean distance.

[0127] (2) When the absolute value of the sensitivity corresponding to any variable type is greater than the preset threshold, any variable type is determined to be an abnormal influencing factor.

[0128] (3) After traversing all variable types, early warning information is generated based on all abnormal influencing factors.

[0129] The preset threshold is set according to the experience of those skilled in the art or according to the actual application scenario, and the embodiments of the present application are not limited to this.

[0130] By calculating the sensitivity corresponding to any variable type, we can quantitatively describe the rate of change of the Euclidean distance relative to each variable data, identify which variables contribute more to the risk of failure, and these variables may be the key factors leading to failure, thereby reducing false alarms caused by non-critical factors, conducting more refined fault analysis, and improving the accuracy of the early warning system. At the same time, after identifying specific abnormal influencing factors, more targeted early warning information is generated to help maintenance personnel quickly locate problems.

[0131] The above describes the abnormal monitoring and early warning method for the bearing temperature of the wind turbine generator in the embodiment of the present application. The following describes the abnormal monitoring and early warning device for the bearing temperature of the wind turbine generator in the embodiment of the present application. Figure 2 In an embodiment of the present application, an abnormal monitoring and early warning device for the bearing temperature of a wind turbine generator includes: a preprocessing module 10, a temperature analysis module 20, an abnormality judgment module 30 and an abnormality diagnosis module 40.

[0132] The preprocessing module 10 is used to obtain the online SCADA data of the wind turbine generator and preprocess the online SCADA data.

[0133] The temperature analysis module 20 is used to extract bearing temperature related variables from the pre-processed online SCADA data, input the related variables into a preset model, and obtain the theoretical value of the bearing temperature.

[0134] The abnormality judgment module 30 is used to extract the actual value of the bearing temperature from the pre-processed online SCADA data, calculate the temperature difference between the actual value of the bearing temperature and the theoretical value of the bearing temperature, and judge that an abnormality exists when the absolute value of the temperature difference is greater than a first preset value.

[0135] The abnormality diagnosis module 40 is used to obtain the first generator operation data within a first preset time, diagnose the generator bearing based on the first generator operation data, and issue a warning message.

[0136] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0137] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0138] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for abnormal monitoring and early warning of the bearing temperature of a wind turbine generator, characterized in that: The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator comprises: Step 1, obtaining online SCADA data of the wind turbine generator, and preprocessing the online SCADA data; Step 2, extracting bearing temperature related variables from the pre-processed online SCADA data, inputting the related variables into a preset model, and obtaining a theoretical value of bearing temperature; Step 3, extracting the actual value of the bearing temperature from the pre-processed online SCADA data, calculating the temperature difference between the actual value of the bearing temperature and the theoretical value of the bearing temperature, and determining that an abnormality exists when the absolute value of the temperature difference is greater than a first preset value; Step 4: Acquire the first generator operation data within a first preset time, diagnose the generator bearing based on the first generator operation data, and issue a warning message.

2. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 1 is characterized in that: In the step 1, the pre-processing of the online SCADA data comprises: Step 11, determining whether there is missing data or abnormal data in the online SCADA data, and if so, defining a variable corresponding to the missing data or the abnormal data as a first variable; Step 12: extract any first variable, obtain the theoretical number of collections of any first variable within a second preset time, and the invalid number of times when the data corresponding to any first variable is the missing data or the abnormal data, calculate the ratio of the invalid number of times to the theoretical number of collections, and define it as inefficiency; Step 13: After traversing all the first variables, based on the inefficiency, divide all the first variables into a preset number of variable groups according to a preset rule; Step 14: Set a corresponding data generation strategy for each variable group, extract any variable group, use the data generation strategy corresponding to any of the variable groups, generate replacement values ​​for each first variable in any of the variable groups, and use the replacement values ​​to complete the online SCADA data.

3. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 2 is characterized in that: The preset number is 3, the first variable whose inefficiency is less than the second preset value is divided into the first variable group, the first variable whose inefficiency is greater than or equal to the second preset value and less than or equal to the third preset value is divided into the second variable group, and the first variable whose inefficiency is greater than the third preset value is divided into the third variable group, and the step 14 includes: For the first variable group, the variables in the first variable group are defined as second variables, a data sequence of any second variable within a third preset time is obtained, an estimated value of any second variable at each time point is calculated based on formula 1, and the estimated value of any second variable at time T-1 is used as the replacement value of any second variable at time T, and formula 1 is: Wherein, E(t) is the estimated value at time t, α is the parameter weight coefficient, D(t) is the actual value at time t, and T is the current time; For the third variable group, the variables in the third variable group are defined as third variables, and the data obtained at time T-1 for any third variable is used as the replacement value of any third variable at time T; For the second variable group, the variables in the second variable group are defined as fourth variables, and a first replacement value at time T is generated for any fourth variable based on the data generation strategy corresponding to the third variable group, and the first replacement value is used as the proposed numerical value of any fourth variable at time T. Subsequently, the estimated value of any fourth variable at time T is generated based on the data generation strategy corresponding to the first variable group and defined as a second replacement value, and the average of the first replacement value and the second replacement value is set as the replacement value of any first variable at time T.

4. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 1 is characterized in that: The method for establishing the preset model is: Step 21, obtaining historical SCADA data of the wind turbine bearing, wherein the historical SCADA data includes N1 data sequences, each data sequence includes the second generator operation data, wherein N1 is a positive integer; Step 22, traversing the historical SCADA data, when data corresponding to any characteristic variable in the second generator operation data is missing, defining any of the characteristic variables as the first characteristic variable, and at the same time, dividing the historical SCADA data into a training data set and a verification data set according to a preset ratio; Step 23: arbitrarily combine all first characteristic variables to obtain N2 characteristic variable combinations, where N2 is a positive integer; Step 24, extract any characteristic variable combination, define the variable in any of the characteristic variable combinations as the third characteristic variable, define the variable outside any of the characteristic variable combinations as the second characteristic variable, delete all data corresponding to any of the second characteristic variables in the training data set, and delete the data sequence lacking the data corresponding to the third characteristic variable, generate a first training data set, repeat step 24 until all characteristic variable combinations are traversed, and generate N3 first training data sets, where N3 is a positive integer less than or equal to N2; Step 25, performing the same operation as step 24 on the verification data set to generate N4 first verification data sets, where N4 is a positive integer less than or equal to N2; Step 26: Perform machine learning on any first training data set to obtain a first preset model corresponding to any first training data set, and after traversing all first training data sets, generate N3 first preset models; Step 27: Use the first verification data set to verify each first preset model, obtain the accuracy of each first preset model respectively, and use the first preset model corresponding to the maximum accuracy as the preset model.

5. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 4 is characterized in that: The step 27 comprises: Extract any first preset model, and define a feature variable used when training any first preset model as a fourth feature variable; Extract any first validation data set, define a feature variable included in any first validation data set as a fifth feature variable, and when the fifth feature variable includes the fourth feature variable, define any first validation data set as a second validation data set; After traversing all the first verification data sets, each second verification data set is used to verify any of the first preset models, multiple model accuracies are obtained, a first statistical value of the accuracies of all models is calculated, and the first statistical value is used as the accuracy of any of the first preset models.

6. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 4 is characterized in that: In step 2, extracting bearing temperature related variables from the pre-processed online SCADA data includes: The characteristic variable used in the training of the preset model is defined as a sixth characteristic variable, and the relevant variable is extracted from the first SCADA data based on the sixth characteristic variable.

7. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 4 is characterized in that: The second generator operation data includes bearing temperature data, cabin temperature data, ambient temperature data, wind speed, wind direction, and power of various bearing components, lubrication status data, speed, torque and power.

8. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 1 is characterized in that: The step 4 comprises: dividing the first generator operation data into a plurality of variable data sequences based on variable types, calculating a second statistical value of each variable data sequence respectively, and generating a first feature vector corresponding to the first generator operation data based on the second statistical value; Extracting a second feature vector corresponding to each fault event within a third preset time, and calculating the Euclidean distance between any second feature vector and the first feature vector, wherein the second feature vector of any fault event is generated based on historical generator operation data within the first preset time before a specific moment, and the specific moment is a moment at a specific time away from the moment when any of the fault events occurs; It is determined whether the minimum Euclidean distance is less than a fourth preset value. If so, it is determined that there is a fault risk.

9. The abnormal monitoring and early warning method for the bearing temperature of a wind turbine generator according to claim 8 is characterized in that: After determining that there is a risk of failure, the following steps are also included: The statistical value in the second eigenvector corresponding to the minimum Euclidean distance is defined as a third statistical value, and the sensitivity corresponding to any variable type is calculated based on Formula 2, where Formula 2 is: Among them, G i is the sensitivity corresponding to the i-th variable type, ST i,2 is the second statistical value of the i-th variable type, ST i,3 is the third statistical value of the i-th variable type, and d is the minimum value of the Euclidean distance; When the absolute value of the sensitivity corresponding to any of the variable types is greater than a preset threshold, any of the variable types is determined as an abnormal influencing factor; After traversing all variable types, the warning information is generated based on all abnormal influencing factors.

10. The abnormal monitoring and early warning device for the bearing temperature of a wind turbine generator is characterized in that: The abnormal monitoring and early warning device for the bearing temperature of the wind turbine generator comprises: a preprocessing module, a temperature analysis module, an abnormal judgment module and an abnormal diagnosis module; The preprocessing module is used to obtain online SCADA data of the wind turbine generator and preprocess the online SCADA data; The temperature analysis module is used to extract bearing temperature related variables from the pre-processed online SCADA data, input the related variables into a preset model, and obtain a theoretical value of the bearing temperature; The abnormality judgment module is used to extract the actual value of the bearing temperature from the pre-processed online SCADA data, calculate the temperature difference between the actual value of the bearing temperature and the theoretical value of the bearing temperature, and judge that an abnormality exists when the absolute value of the temperature difference is greater than a first preset value; The abnormality diagnosis module is used to obtain the first generator operation data within a first preset time, diagnose the generator bearing based on the first generator operation data, and issue a warning message.

Images