A gear box oil pool temperature early warning method, system, device and medium

Abstract

The present application belongs to the technical field of wind power generation, and specifically discloses a gear box oil pool temperature early warning method, system, device and medium. The present application obtains all historical operation data of wind turbines, analyzes the correlation of wind turbines, establishes a healthy unit group, establishes positive and negative sample data according to historical operation data of a wind farm, selects a feature with relatively large correlation with the gear box oil pool temperature as a main characteristic value for input, establishes a gear box oil pool temperature early warning model through the positive sample data, determines the average value and threshold of the standard deviation of the gear box oil pool temperature residual error, predicts the model through the negative sample data, calculates the deviation between the predicted value and the actual value of the current unit gear box oil pool temperature in real time, and compares it with the set threshold. When the temperature deviation is greater than the set threshold, an out-of-limit early warning of the gear box oil pool temperature is issued. Through this method, the gear box oil pool temperature can be early warned through real-time data, and the method has the advantages of real-time and strong operability.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wind power generation, and particularly relates to a gear box oil pool temperature early warning method, system, device and medium. BACKGROUND

[0002] The daily operation environment of a wind turbine is relatively harsh and complex, and long-term operation can cause certain impact on the inside of the wind turbine, so that the operation reliability of the wind turbine is reduced and the operation life is shortened. Once the wind turbine fails and stops, a lot of economic losses will be caused. The gear box is an important component of the wind turbine, wind energy drives the wind wheel to rotate to generate mechanical energy, but the generator needs to input a high enough rotating speed, and the gear box is a device for increasing the rotating speed of the mechanical energy transmitted from the wind wheel to the generator to a corresponding standard. In view of the function of the gear box, if the gear box fails and cannot be repaired in time, a large amount of power loss will be caused, so it is very important to monitor the real-time operation state of the gear box and give an early warning before failure occurs.

[0003] At present, the wind turbine is equipped with a supervisory control and data acquisition system (SCADA system), which can monitor the power, rotating speed of the wind turbine, and the temperature and other parameters of the gear box, generator and other main components, and simply determine the operation state of the main components, but generally only a fixed threshold is set for a single parameter, and the timeliness is poor. In addition, the monitoring of the gear box oil pool temperature is easily affected by the surrounding environment, so that the gear box oil pool temperature cannot be warned in real time. SUMMARY

[0004] In order to overcome the defects of the prior art, the purpose of the present application is to provide a gear box oil pool temperature early warning method, system, device and medium, so as to solve the problem that the existing monitoring system has poor timeliness and cannot warn the gear box oil pool temperature in real time when warning the gear box oil pool temperature.

[0005] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0006] In a first aspect, the present application provides a gear box oil pool temperature early warning method, comprising:

[0007] Obtaining all historical operation data of the wind turbine, performing data preprocessing, analyzing the correlation of the wind turbine, establishing a healthy unit group, and selecting the main characteristic value of the healthy unit group;

[0008] Selecting the historical operation data of the healthy unit group as positive sample data, establishing a gear box oil pool temperature early warning model through the main characteristic value of the healthy unit group, and determining the threshold value of the average value and the standard deviation of the gear box oil pool temperature residual error;

[0009] The operation data of the unit with the gearbox temperature early warning is selected as the negative sample data, and is input into the established gearbox oil pool temperature early warning model to obtain the gearbox oil pool temperature prediction value of the early warning unit, and the average value and the standard deviation of the gearbox oil pool temperature residual of the negative sample data are calculated;

[0010] The average value and the standard deviation of the gearbox oil pool temperature residual of the calculated negative sample data are compared with the threshold value of the determined average value and the standard deviation of the gearbox oil pool temperature residual, and the gearbox oil pool temperature is judged.

[0011] Further, the historical operation data of all wind turbines is obtained, and data preprocessing is performed, specifically including:

[0012] The historical operation data of all wind turbines in the wind farm is obtained, and the data is preprocessed;

[0013] The historical operation data is taken from the SCADA system data, the data acquisition period is 5 minutes, and the data source contains 14 characteristic parameters of the wind turbine running state, generator active power, instantaneous wind speed, environmental temperature, main shaft front bearing temperature, main shaft rear bearing temperature, generator speed, gearbox high-speed shaft front end temperature, gearbox high-speed shaft rear end temperature, gearbox inlet oil temperature, gearbox oil pool temperature, gearbox inlet pressure, gearbox oil pump outlet pressure and gearbox cooling water temperature in the same time sequence;

[0014] The obtained historical operation data is cleaned, and the cleaning principle is: excluding points with gearbox oil pool temperature less than-50℃ and greater than 120℃; selecting time points with wind turbine active power greater than zero for analysis, if there is missing data for a characteristic parameter at a certain time point, all data collected at the corresponding time are deleted.

[0015] Further, the correlation of the wind turbine is analyzed, and a healthy unit group is established, and the main characteristic value of the healthy unit group is selected, specifically including:

[0016] The Pearson correlation coefficient is used to calculate the similarity between the same characteristic parameters in the wind turbine which has occurred gearbox oil pool temperature early warning and the adjacent unit in the same time period;

[0017] (x i ,y i ) is any one group of values in the total sample, the Pearson correlation coefficient p pearson between the variable X and the variable Y is:

[0018]

[0019] The Pearson coefficient is the ratio of the covariance of the two groups of variables and the standard deviation product, and the value range is [-1, 1];

[0020] A healthy unit group is established based on the wind turbines with high correlation; high correlation is defined as an absolute value of the correlation coefficient greater than or equal to 0.6.

[0021] Correlation analysis was performed on the gearbox oil sump temperature parameter and all other characteristic parameters of the fan in the healthy unit group. The feature with a large correlation was selected as the main characteristic value of the healthy unit group; a large correlation was defined as an absolute value of the correlation coefficient greater than or equal to 0.6.

[0022] Furthermore, the step of selecting historical operating data of healthy unit groups as positive sample data and establishing a gearbox oil sump temperature early warning model through the main feature values ​​of healthy unit groups specifically includes:

[0023] The main characteristic values ​​of the healthy unit group are normalized using the maximum and minimum values, and the specific formula is as follows:

[0024]

[0025] Where: X is a certain characteristic value; X max X min X represents the maximum and minimum values ​​of a certain characteristic. std These are the principal eigenvalues ​​after normalizing the maximum and minimum values;

[0026] The normalized principal eigenvalues ​​are used as the input to the model, and the gearbox oil sump temperature is used as the output to establish a gearbox oil sump temperature early warning model.

[0027] The accuracy of the model is evaluated using decision coefficients, as shown in the following formula:

[0028]

[0029] In the formula: R 2 y is the decision coefficient; true The actual temperature of the oil sump in the wind turbine gearbox is obtained through the SCADA system. true_mean y represents the average value of the actual oil sump temperature in the fan gearbox. pred is the predicted temperature of the wind turbine gearbox oil sump, and is the output value of the gearbox oil sump temperature early warning model.

[0030] Furthermore, the threshold for determining the average value and standard deviation of the gearbox oil sump temperature residuals specifically includes:

[0031] The sliding window analysis method is used to process the prediction residuals. The width of the sliding window is N, and the formulas for calculating the mean and standard deviation of the residuals containing the sampling points under this window are as follows:

[0032]

[0033]

[0034] In the formula, err i Let N be the i-th residual in the window, and N be the number of sampling points in the current sliding window. S is the average residual of the current sliding window, and S is the standard deviation of the residual of the current sliding window.

[0035] err i =y pred -y true

[0036] Use the sliding window method to calculate the maximum absolute value of the residual mean for the entire positive sample data. The maximum value S_max of the residual standard deviation is used as the threshold for the residual mean and residual standard deviation.

[0037] Furthermore, the operating data of the selected units with gearbox temperature early warning is used as negative sample data and input into the established gearbox oil sump temperature prediction model to obtain the predicted gearbox oil sump temperature of the units with early warning. The average and standard deviation of the gearbox oil sump temperature residuals of the negative sample data are calculated, specifically including:

[0038] The negative sample data is input into the established gearbox oil sump temperature early warning model for prediction, and the gearbox oil sump temperature of the early warning unit is obtained.

[0039] Based on the predicted and actual values ​​of the gearbox oil sump temperature of the early warning unit, the mean and standard deviation of the gearbox oil sump temperature residuals of the negative sample data are calculated using the sliding window analysis method, with a sliding window width of N.

[0040] Furthermore, the step of comparing the average and standard deviation of the calculated negative sample data gearbox oil sump temperature residuals with the determined threshold values ​​for the average and standard deviation of the gearbox oil sump temperature residuals to make an early warning judgment on the gearbox oil sump temperature specifically includes:

[0041] When the average value and standard deviation of the calculated negative sample data gearbox oil sump temperature residuals exceed the predetermined threshold values ​​for the average value and standard deviation of the gearbox oil sump temperature residuals for k consecutive time periods, a gearbox temperature over-limit warning is issued.

[0042] In a second aspect, the present invention provides a gearbox oil sump temperature early warning system, comprising:

[0043] The data acquisition and main feature value selection module is used to acquire all historical operating data of wind turbines, perform data preprocessing, analyze the correlation of wind turbines, establish a healthy unit group, and select the main feature value of the healthy unit group.

[0044] The module for model establishment and threshold determination is used to select historical operating data of healthy unit groups as positive sample data, establish a gearbox oil sump temperature early warning model through the main feature values ​​of healthy unit groups, and determine the threshold values ​​of the average value and standard deviation of the gearbox oil sump temperature residuals.

[0045] The module for obtaining the predicted oil pool temperature and the module for calculating the average and standard deviation are used to select the operating data of the unit with gearbox temperature warning as negative sample data, input it into the established gearbox oil pool temperature warning model, obtain the predicted gearbox oil pool temperature of the warning unit, and calculate the average and standard deviation of the gearbox oil pool temperature residual of the negative sample data.

[0046] The gearbox oil sump temperature early warning module is used to compare the average and standard deviation of the calculated gearbox oil sump temperature residuals with the determined threshold values ​​of the average and standard deviation of the gearbox oil sump temperature residuals, and to make an early warning judgment on the gearbox oil sump temperature.

[0047] Thirdly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a gearbox oil sump temperature early warning method as described above.

[0048] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a gearbox oil sump temperature early warning method as described above.

[0049] The present invention has at least the following beneficial effects:

[0050] 1. This invention acquires all historical operating data of wind turbines, analyzes the correlation of wind turbines, establishes a healthy unit group, and establishes positive and negative sample data for the historical operating data of wind farms. Features with a high correlation with gearbox oil sump temperature are selected as the main feature values ​​for input. A gearbox oil sump temperature early warning model is established using positive sample data, and the threshold values ​​of the average value and standard deviation of the gearbox oil sump temperature residual are determined. The model is predicted using negative sample data, and the deviation between the predicted value and the actual value of the current unit's gearbox oil sump temperature is calculated in real time and compared with the set threshold. When the temperature deviation exceeds the set threshold, a gearbox oil sump temperature over-limit warning is issued. This method can provide early warning of gearbox oil sump temperature through real-time data, and has the advantages of real-time performance and strong operability.

[0051] 2. This invention does not require the installation of additional sensors; it only requires modeling and analyzing historical operating data of the wind farm, making it more feasible and scalable. Attached Figure Description

[0052] The accompanying drawings, which form part of this specification, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0053] Figure 1 A schematic diagram of the gearbox oil sump temperature early warning method;

[0054] Figure 2 Schematic diagram of a gearbox oil sump temperature early warning system module;

[0055] Figure 3 A schematic diagram for predicting the temperature of the gearbox oil sump;

[0056] Figure 4 This is a schematic diagram of the residual temperature in the gearbox oil sump. Detailed Implementation

[0057] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0058] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this invention is for describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0059] Example 1

[0060] like Figure 1 As shown, a method for early warning of gearbox oil sump temperature includes:

[0061] S1: Obtain all historical operating data of wind turbines, perform data preprocessing, analyze the correlation of wind turbines, establish a healthy unit group, and select the main characteristic value of the healthy unit group;

[0062] S11: Obtain the wind farm layout plan, understand the wind turbine layout, and select the target wind turbine and surrounding wind turbines for correlation analysis;

[0063] S12: Obtain historical operating data of all wind turbines in the wind farm and preprocess the data;

[0064] Historical operating data is taken from SCADA system data, with a data acquisition cycle of 5 minutes. The data source includes 14 characteristic parameters in the same time series, including wind turbine operating status, generator active power, instantaneous wind speed, ambient temperature, main shaft front bearing temperature, main shaft rear bearing temperature, generator speed, gearbox high-speed shaft front end temperature, gearbox high-speed shaft rear end temperature, gearbox inlet oil temperature, gearbox oil sump temperature, gearbox inlet pressure, gearbox oil pump outlet pressure, and gearbox cooling water temperature.

[0065] The acquired historical operating data is cleaned for anomalies. The cleaning principles are as follows: exclude points where the gearbox oil sump temperature is less than -50℃ or greater than 120℃; select points where the active power of the wind turbine is greater than zero for analysis, and remove anomaly points where the active power is less than or equal to zero due to no power generation at night or other times; in addition, if there is missing data for a certain characteristic parameter at a certain point in time, all data collected at that time are deleted.

[0066] S13: Use Pearson correlation coefficient to analyze the correlation of wind turbines, identify wind turbines with high correlation, and establish a healthy unit group;

[0067] Pearson correlation coefficient was used to calculate the similarity of the same characteristic parameters between wind turbines that had experienced gearbox oil sump temperature warnings and neighboring units in the same time period.

[0068] Pearson correlation analysis can quantitatively analyze the similarity between two sets of data, assuming (x i ,y i Let X be any set of values ​​in the total sample. Then, the Pearson correlation coefficient ρ between variable X and variable Y is... pearson for:

[0069]

[0070] The Pearson coefficient is the ratio of the product of the covariance and the standard deviation of two sets of variables, with a range of [-1, 1]. When two variables have a strong linear relationship, the Pearson coefficient is 1 or -1; the larger the absolute value of the correlation coefficient, the higher the correlation between X and Y.

[0071] Table 1. Relationship between correlation coefficient and correlation degree

[0072] Correlation coefficient Correlation 0.8-1.0 Very strong correlation 0.6-0.8 Strong correlation 0.4-0.6 Moderate correlation 0.2-0.4 Weak correlation 0.0-0.2 Very weak correlation or no correlation

[0073] In this invention, a correlation coefficient with an absolute value greater than or equal to 0.6 is considered to have a high correlation. Based on the wind turbines with high correlation obtained, a healthy unit group is established.

[0074] S14: In the healthy unit group, the correlation between the gearbox oil sump temperature parameter and all other characteristic parameters of the fan is analyzed, and the feature with a large correlation is selected as the main characteristic value of the healthy unit group; a large correlation is defined as the absolute value of the correlation coefficient being greater than or equal to 0.6.

[0075] S2: Select historical operating data of healthy unit groups as positive sample data, establish a gearbox oil sump temperature early warning model through the main characteristic values ​​of healthy unit groups, and determine the threshold of the average value and standard deviation of the gearbox oil sump temperature residuals;

[0076] S21: Normalize the main characteristic values ​​of the healthy unit group;

[0077] To avoid deviations in the final results due to inconsistencies in the ranges of different characteristic parameters, it is necessary to normalize the characteristic parameters that are highly correlated with the gearbox oil sump temperature. Here, we choose to normalize by the maximum and minimum values, and the specific formula is as follows:

[0078]

[0079] Where: X is a certain characteristic value; X max X min X represents the maximum and minimum values ​​of a certain characteristic. std These are the principal eigenvalues ​​after normalizing the maximum and minimum values.

[0080] S22: Using the normalized principal eigenvalues ​​as the input of the model and the gearbox oil sump temperature as the output of the model, a gearbox oil sump temperature early warning model is established to predict the gearbox oil sump temperature value of a normal fan.

[0081] S23: Introduce decision coefficients to evaluate the accuracy of the model, as shown in the following formula:

[0082]

[0083] In the formula: R 2 y is the decision coefficient; true The actual temperature of the oil sump in the wind turbine gearbox is obtained through the SCADA system. true_mean y represents the average value of the actual oil sump temperature in the fan gearbox. pred , where is the predicted temperature of the wind turbine gearbox oil sump, and is the output value of the gearbox oil sump temperature early warning model; the closer the decision coefficient is to 1, the higher the accuracy of the model.

[0084] S24: The sliding window analysis method is used to process the prediction residuals in order to eliminate the influence of random factors on the variation of residuals;

[0085] The sliding window has a width of N. The formulas for calculating the mean and standard deviation of the residuals of the sampled points within this window are as follows:

[0086]

[0087]

[0088] In the formula, err i Let N be the i-th residual in the window, and N be the number of sampling points in the current sliding window. S is the average residual of the current sliding window, and S is the standard deviation of the residual of the current sliding window.

[0089] err i =y pred -y true

[0090] Use the sliding window method to calculate the maximum absolute value of the residual mean for the entire positive sample data. The maximum value S_max of the residual standard deviation is used as the threshold for the residual mean and residual standard deviation.

[0091] S3: Select the operating data of the unit with gearbox temperature warning as negative sample data, input it into the established gearbox oil sump temperature warning model, obtain the predicted value of gearbox oil sump temperature of the warning unit, and calculate the average value and standard deviation of the gearbox oil sump temperature residual of the negative sample data.

[0092] S31: Input the negative sample data into the established gearbox oil sump temperature early warning model for prediction, and obtain the gearbox oil sump temperature of the early warning unit.

[0093] S32: Based on the predicted and actual values ​​of the gearbox oil sump temperature of the early warning unit, use the sliding window analysis method to calculate the average and standard deviation of the gearbox oil sump temperature residuals of the negative sample data. The width of the sliding window needs to be consistent with step S24.

[0094] S4: Compare the average value and standard deviation of the gearbox oil sump temperature residuals calculated in step S3 with the threshold values ​​of the average value and standard deviation of the gearbox oil sump temperature residuals determined in step S2, and make an early warning judgment on the gearbox oil sump temperature.

[0095] When the average value and standard deviation of the calculated negative sample data gearbox oil sump temperature residuals are greater than the threshold determined in step S24 for k consecutive time periods (ranging from 3 to 6), a gearbox temperature over-limit warning is issued.

[0096] Example 2

[0097] like Figure 2 As shown, a gearbox oil sump temperature early warning system includes:

[0098] The data acquisition and main feature value selection module is used to acquire all historical operating data of wind turbines, perform data preprocessing, analyze the correlation of wind turbines, establish a healthy unit group, and select the main feature value of the healthy unit group.

[0099] The module for model establishment and threshold determination is used to select historical operating data of healthy unit groups as positive sample data, establish a gearbox oil sump temperature early warning model through the main feature values ​​of healthy unit groups, and determine the threshold values ​​of the average value and standard deviation of the gearbox oil sump temperature residuals.

[0100] The module for obtaining the predicted oil pool temperature and the module for calculating the average and standard deviation are used to select the operating data of the unit with gearbox temperature warning as negative sample data, input it into the established gearbox oil pool temperature warning model, obtain the predicted gearbox oil pool temperature of the warning unit, and calculate the average and standard deviation of the gearbox oil pool temperature residual of the negative sample data.

[0101] The gearbox oil sump temperature early warning module is used to compare the average and standard deviation of the calculated gearbox oil sump temperature residuals with the determined threshold values ​​of the average and standard deviation of the gearbox oil sump temperature residuals, and to make an early warning judgment on the gearbox oil sump temperature.

[0102] Example 3

[0103] The present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a gearbox oil sump temperature early warning method as described in Embodiment 1.

[0104] Example 4

[0105] The present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a gearbox oil sump temperature early warning method as described in Embodiment 1.

[0106] Example 5

[0107] S1: Obtain all historical operating data of wind turbines, perform data preprocessing, analyze the correlation of wind turbines, establish a healthy unit group, and select the main characteristic value of the healthy unit group;

[0108] S11: Obtain the wind farm layout plan, understand the wind turbine layout, and select the target wind turbine and surrounding wind turbines for correlation analysis;

[0109] S12: Obtain historical operating data of all wind turbines in the wind farm and preprocess the data;

[0110] The operational data was obtained from the SCADA system of a wind farm in Xinjiang, totaling 99,065 historical data points from November 14, 2021 to October 29, 2022. The data collection period was 5 minutes. The data source included 14 characteristic parameters for the same time series, including wind turbine operating status, generator active power, instantaneous wind speed, ambient temperature, main shaft front bearing temperature, main shaft rear bearing temperature, generator speed, gearbox high-speed shaft front end temperature, gearbox high-speed shaft rear end temperature, gearbox inlet oil temperature, gearbox oil sump temperature, gearbox inlet pressure, gearbox oil pump outlet pressure, and gearbox cooling water temperature.

[0111] The acquired historical operating data is cleaned for abnormal data. The cleaning principles are as follows: exclude points where the gearbox oil sump temperature is less than -50℃ or greater than 120℃; select points where the active power of the wind turbine is greater than zero for analysis, and remove abnormal points where the active power is less than or equal to zero due to no power generation at night or other times; in addition, if there is missing data for a certain indicator at a certain point in time, all data collected at that time are deleted.

[0112] S13: Use Pearson correlation coefficient to analyze the correlation of wind turbines, identify wind turbines with high correlation, and establish a healthy unit group;

[0113] Pearson correlation coefficient was used to calculate the similarity of the same characteristic parameters between wind turbines that had experienced gearbox oil sump temperature warnings and neighboring units in the same time period.

[0114] Pearson correlation analysis can quantitatively analyze the similarity between two sets of data, assuming (x i ,y i Let X be any set of values ​​in the total sample. Then, the Pearson correlation coefficient ρ between variable X and variable Y is... pearson for:

[0115]

[0116] The Pearson coefficient is the ratio of the product of the covariance and the standard deviation of two variables, with a range of [-1, 1]. When the two variables have a strong linear relationship, the Pearson coefficient is 1 or -1. The larger the absolute value of the correlation coefficient, the stronger the correlation between X and Y.

[0117] Table 1. Relationship between correlation coefficient and correlation degree

[0118] Correlation coefficient Correlation 0.8-1.0 Very strong correlation 0.6-0.8 Strong correlation 0.4-0.6 Moderate correlation 0.2-0.4 Weak correlation 0.0-0.2 Very weak correlation or no correlation

[0119] In this invention, a correlation coefficient with an absolute value greater than or equal to 0.6 is considered to have a high correlation. Based on the wind turbines with high correlation obtained, a healthy unit group is established.

[0120] S14: In the healthy unit group, the correlation between the gearbox oil sump temperature parameter and all other characteristic parameters of the fan is analyzed, and the feature with a large correlation is selected as the main characteristic value of the healthy unit group; a large correlation is defined as the absolute value of the correlation coefficient being greater than or equal to 0.6.

[0121] For example, if a correlation analysis is performed on the annual operating data of a certain wind turbine, the following characteristics are found to have relatively high correlation coefficients;

[0122] XX fan influence factor Correlation coefficient Gearbox inlet oil temperature 0.96 Generator active power 0.7 Gearbox oil pump outlet pressure 0.83 Gearbox inlet pressure -0.9

[0123] As can be seen from the table, the gearbox inlet pressure and the gearbox oil sump temperature are negatively correlated, while all other characteristics are positively correlated.

[0124] S2: Select historical operating data of healthy unit groups as positive sample data, establish a gearbox oil sump temperature early warning model through the main characteristic values ​​of healthy unit groups, and determine the threshold of the average value and standard deviation of the gearbox oil sump temperature residuals;

[0125] S21: Normalize the main characteristic values ​​of the healthy unit group;

[0126] To avoid discrepancies in the final results caused by inconsistent ranges of different characteristic parameters, it is necessary to normalize the characteristic parameters that are highly correlated with the gearbox oil sump temperature. Here, we choose to normalize by the maximum and minimum values, and the specific formula is as follows:

[0127]

[0128] Where: X is a certain characteristic value; X max X min X represents the maximum and minimum values ​​of a certain characteristic. std These are the principal eigenvalues ​​after normalizing the maximum and minimum values.

[0129] S22: Using the normalized principal eigenvalues ​​as the input of the model and the gearbox oil sump temperature as the output of the model, a gearbox oil sump temperature early warning model is established to predict the gearbox oil sump temperature value of a normal fan.

[0130] A gradient boosting decision tree model, i.e., a gearbox oil sump temperature early warning model, is established to predict the gearbox oil sump temperature value of a normal fan; the corresponding prediction model is:

[0131] y = f(T) jk ,P jk ,P,P ck )

[0132] Where: y is the predicted gearbox oil sump temperature, T jk ,P jk ,P,P ckThese are the gearbox inlet oil temperature, gearbox inlet pressure, generator active power, and gearbox oil pump outlet pressure, respectively.

[0133] The gradient boosting decision tree utilizes a serial model and a forward stepwise algorithm to optimize the learning process. Based on the boosting method, base classifiers are stacked layer by layer. During each training iteration, samples that were misclassified in the previous iteration are given higher weights. Finally, the results of each layer of classifiers are weighted to obtain the final prediction result. The next weak classifier is then used to fit the residual between the predicted value and the true value.

[0134] S23: Introduce decision coefficients to evaluate the accuracy of the model, as shown in the following formula:

[0135]

[0136] In the formula: R 2 y is the decision coefficient; true The actual temperature of the oil sump in the wind turbine gearbox is obtained through the SCADA system. true_mean y represents the average value of the actual oil sump temperature in the fan gearbox. pred , where is the predicted temperature of the wind turbine gearbox oil sump, and is the output value of the gearbox oil sump temperature early warning model; the closer the decision coefficient is to 1, the higher the accuracy of the model.

[0137] S24: The sliding window analysis method is used to process the prediction residuals in order to eliminate the influence of random factors on the variation of residuals;

[0138] The sliding window has a width of N. The formulas for calculating the mean and standard deviation of the residuals of the sampled points within this window are as follows:

[0139]

[0140]

[0141] In the formula, err i Let N be the i-th residual in the window, and N be the number of sampling points in the current sliding window. S is the average residual of the current sliding window, and S is the standard deviation of the residual of the current sliding window.

[0142] err i =y pred -y true

[0143] Use the sliding window method to calculate the maximum absolute value of the residual mean for the entire positive sample data. The maximum value S_max of the residual standard deviation is used as the threshold for the residual mean and residual standard deviation.

[0144] S3: Select the operating data of the unit with gearbox temperature warning as negative sample data, input it into the established gearbox oil sump temperature warning model, obtain the predicted value of gearbox oil sump temperature of the warning unit, and calculate the average value and standard deviation of the gearbox oil sump temperature residual of the negative sample data.

[0145] S31: Input the negative sample data into the established gearbox oil sump temperature early warning model for prediction, and obtain the gearbox oil sump temperature of the early warning unit.

[0146] S32: Based on the predicted and actual values ​​of the gearbox oil sump temperature of the early warning unit, use the sliding window analysis method to calculate the average and standard deviation of the gearbox oil sump temperature residuals of the negative sample data. The size of the sliding window needs to be consistent with step S24.

[0147] S4: Compare the average value and standard deviation of the gearbox oil sump temperature residuals calculated in step S3 with the threshold values ​​of the average value and standard deviation of the gearbox oil sump temperature residuals determined in step S2, and make an early warning judgment on the gearbox oil sump temperature.

[0148] When the average value and standard deviation of the calculated negative sample data gearbox oil sump temperature residuals are greater than the threshold determined in step S24 for k consecutive time periods (ranging from 3 to 6), a gearbox temperature over-limit warning is issued.

[0149] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0150] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0151] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0152] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for early warning of gearbox oil sump temperature, characterized in that, include: Acquire all historical operating data of wind turbines, perform data preprocessing, analyze the correlation of wind turbines, establish a healthy unit group, and select the main characteristic value of the healthy unit group; Historical operating data of healthy unit groups were selected as positive sample data. A gearbox oil sump temperature early warning model was established using the main characteristic values ​​of the healthy unit groups to determine the threshold values ​​of the average value and standard deviation of the gearbox oil sump temperature residuals. The sliding window method is used to calculate the entire positive sample data, determining the maximum absolute value of the residual mean and the maximum residual standard deviation, which serve as thresholds for the residual mean and standard deviation. The gearbox oil sump temperature early warning model adopts a gradient boosting decision tree model, as shown in the following equation: Where: y is the predicted gearbox oil sump temperature, These are, respectively, the gearbox inlet oil temperature, the gearbox inlet pressure, the generator active power, and the gearbox oil pump outlet pressure; The gradient boosting decision tree model uses a serial model and a forward step-by-step algorithm for learning and optimization. The operating data of the units with gearbox temperature warnings are selected as negative sample data and input into the established gearbox oil sump temperature warning model to obtain the predicted gearbox oil sump temperature of the warning units. The mean and standard deviation of the gearbox oil sump temperature residuals of the negative sample data are calculated using the sliding window analysis method. The average and standard deviation of the calculated gearbox oil sump temperature residuals of the negative sample data are compared with the determined threshold values ​​for the average and standard deviation of the gearbox oil sump temperature residuals to make an early warning judgment on the gearbox oil sump temperature. When the average and standard deviation of the calculated gearbox oil sump temperature residuals of the negative sample data are greater than the determined threshold values ​​for the average and standard deviation of the gearbox oil sump temperature residuals for k consecutive time periods, an over-limit warning for gearbox temperature is issued. The process of acquiring all historical operating data of wind turbines and performing data preprocessing includes: acquiring all historical operating data of wind turbines and performing data preprocessing, including excluding points where the gearbox oil sump temperature is less than -50℃ and greater than 120℃, selecting points where the active power of the wind turbine is greater than zero for analysis, and deleting all data collected at the corresponding point if a certain characteristic parameter has missing data at a certain point in time. The correlation analysis of the wind turbines, the establishment of a healthy unit group, and the selection of the main characteristic values ​​of the healthy unit group specifically include: using the Pearson correlation coefficient to calculate the similarity between the same characteristic parameters of wind turbines that have experienced gearbox oil sump temperature warnings in the same time period and neighboring units; establishing a healthy unit group based on the wind turbines with high correlation; and normalizing the main characteristic values ​​of the healthy unit group by selecting the maximum and minimum values ​​for normalization.

2. The gearbox oil sump temperature early warning method according to claim 1, characterized in that, The process of acquiring all historical operating data of wind turbines and performing data preprocessing specifically includes: Obtain historical operating data of all wind turbines in the wind farm and preprocess the data; Historical operating data is taken from the SCADA system, with a data acquisition cycle of 5 minutes. The data source includes 14 characteristic parameters for the same time series, such as wind turbine operating status, generator active power, instantaneous wind speed, ambient temperature, main shaft front bearing temperature, main shaft rear bearing temperature, generator speed, gearbox high-speed shaft front end temperature, gearbox high-speed shaft rear end temperature, gearbox inlet oil temperature, gearbox oil sump temperature, gearbox inlet pressure, gearbox oil pump outlet pressure, and gearbox cooling water temperature.

3. The gearbox oil sump temperature early warning method according to claim 1, characterized in that, The correlation analysis of the wind turbines, the establishment of a healthy turbine group, and the selection of the main characteristic values ​​of the healthy turbine group specifically include: ( Let X be any set of values ​​in the total sample, and let Y be the Pearson correlation coefficient between variables X and Y. for: The Pearson coefficient is the ratio of the product of the covariance and the standard deviation of two sets of variables, and its range is [-1, 1]. Correlation analysis was performed on the gearbox oil sump temperature parameter and all other characteristic parameters of the fan in the healthy unit group. The feature with a large correlation was selected as the main characteristic value of the healthy unit group; a large correlation was defined as an absolute value of the correlation coefficient greater than or equal to 0.

6.

4. The gearbox oil sump temperature early warning method according to claim 1, characterized in that, The step of selecting historical operating data from healthy generator sets as positive sample data and establishing a gearbox oil sump temperature early warning model based on the main feature values ​​of the healthy generator sets specifically includes: The main characteristic values ​​of the healthy unit group are normalized by using the maximum and minimum values ​​for normalization. The specific formula is as follows: Where: X is a certain characteristic value; , For the maximum and minimum values ​​of a certain characteristic, These are the principal eigenvalues ​​after normalizing the maximum and minimum values; The normalized principal eigenvalues ​​are used as the input to the model, and the gearbox oil sump temperature is used as the output to establish a gearbox oil sump temperature early warning model. The accuracy of the model is evaluated using decision coefficients, as shown in the following formula: In the formula: For decision coefficients; The actual temperature of the oil sump in the wind turbine gearbox is obtained through the SCADA system. This represents the average value of the actual oil sump temperature in the fan gearbox. is the predicted temperature of the wind turbine gearbox oil sump, and is the output value of the gearbox oil sump temperature early warning model.

5. A gearbox oil sump temperature early warning method according to claim 4, characterized in that, The threshold values ​​for determining the average value and standard deviation of the gearbox oil sump temperature residuals specifically include: The sliding window analysis method is used to process the prediction residuals. The width of the sliding window is N, and the formulas for calculating the mean and standard deviation of the residuals containing the sampling points under this window are as follows: In the formula, Let N be the i-th residual in the window, and N be the number of sampling points in the current sliding window. This represents the average residual of the current sliding window. This represents the standard deviation of the residuals in the current sliding window. 。 6. The gearbox oil sump temperature early warning method according to claim 5, characterized in that, The operating data of the selected units with gearbox temperature early warning is used as negative sample data and input into the established gearbox oil sump temperature prediction model to obtain the predicted gearbox oil sump temperature of the units with early warning. The mean and standard deviation of the gearbox oil sump temperature residuals of the negative sample data are calculated using the sliding window analysis method, specifically including: The negative sample data is input into the established gearbox oil sump temperature early warning model for prediction, and the gearbox oil sump temperature of the early warning unit is obtained. Based on the predicted and actual values ​​of the gearbox oil sump temperature of the early warning unit, the mean and standard deviation of the gearbox oil sump temperature residuals of the negative sample data are calculated using the sliding window analysis method, with a sliding window width of N.

7. A gearbox oil sump temperature early warning system, using the gearbox oil sump temperature early warning method according to any one of claims 1-6, characterized in that, include: The data acquisition and main feature value selection module is used to acquire all historical operating data of wind turbines, perform data preprocessing, analyze the correlation of wind turbines, establish a healthy unit group, and select the main feature value of the healthy unit group. The module for model establishment and threshold determination is used to select historical operating data of healthy unit groups as positive sample data, establish a gearbox oil sump temperature early warning model through the main feature values ​​of healthy unit groups, and determine the threshold values ​​of the average value and standard deviation of the gearbox oil sump temperature residuals. The module for obtaining the predicted oil pool temperature and the module for calculating the average and standard deviation are used to select the operating data of the unit with gearbox temperature warning as negative sample data, input it into the established gearbox oil pool temperature warning model, obtain the predicted gearbox oil pool temperature of the warning unit, and calculate the average and standard deviation of the gearbox oil pool temperature residual of the negative sample data. The gearbox oil sump temperature early warning module is used to compare the average and standard deviation of the calculated gearbox oil sump temperature residuals with the determined threshold values ​​of the average and standard deviation of the gearbox oil sump temperature residuals, and to make an early warning judgment on the gearbox oil sump temperature.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a gearbox oil sump temperature early warning method according to any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a gearbox oil sump temperature early warning method according to any one of claims 1-6.

Images