Wind driven generator collector ring carbon brush operation monitoring method

Abstract

The invention discloses a method for monitoring operation of a collector ring carbon brush of a wind driven generator. The method comprises the following steps: establishing an intelligent monitoring system, acquiring monitoring data of monitoring modes contained in the intelligent monitoring system, performing credibility analysis on various monitoring data, and determining characteristic monitoring data of the collector ring carbon brush; analyzing the feature monitoring data to obtain a state monitoring result of each feature monitoring data, and generating a state abnormal coefficient of the collector ring carbon brush according to the plurality of state monitoring results; and determining the running state of the collector ring carbon brush according to the abnormal state coefficient, thereby improving the accuracy of the abnormal state coefficient, accurately reflecting the running state of the collector ring carbon brush in the current preset time period and the future preset time period, and reducing the fault frequency of the wind driven generator.

Description

Technical Field

[0001] The present application relates to the technical field of operation monitoring, and in particular to a method for monitoring the operation of a collector ring carbon brush of a wind turbine. Background Art

[0002] The collector ring carbon brush is an important component of the wind turbine. For a long time, the collector ring carbon brush failure has been frequent, which is the main difficulty that plagues the stable operation of the power station. The traditional operation monitoring method used for the collector ring carbon brush is mainly through manual inspection or single sensor monitoring. The inspection workload is large or the monitoring results of the single sensor are inaccurate, resulting in a decrease in the accuracy of the collector ring carbon brush operation monitoring results. It cannot timely reflect the operating status of the collector ring carbon brush at a certain moment, affecting the normal operation of the wind turbine. Summary of the invention

[0003] To solve the above technical problems, the present application provides a method for monitoring the operation of the collector ring carbon brush of a wind turbine. By constructing an intelligent monitoring system, characteristic monitoring data of multiple monitoring methods are obtained, and the state monitoring results of each characteristic monitoring data are determined, and a predicted state monitoring result is generated based on the predicted characteristic monitoring data of a future preset time period, a state abnormality coefficient is generated based on the state monitoring result, a compensation coefficient is generated based on the predicted state monitoring result and the state abnormality coefficient is corrected to improve the accuracy of the state abnormality coefficient, timely reflect the operating status of the collector ring carbon brush in the current preset time period and the future preset time period, and reduce the failure frequency of the wind turbine.

[0004] In some embodiments of the present application, a method for monitoring the operation of a collector ring carbon brush of a wind turbine is provided, comprising:

[0005] Establish an intelligent monitoring system, obtain monitoring data of the monitoring methods included in the intelligent monitoring system, conduct credibility analysis on various monitoring data, and determine the characteristic monitoring data of the collector ring carbon brush;

[0006] Analyze the characteristic monitoring data to obtain the status monitoring results of each characteristic monitoring data, and generate the status abnormality coefficient of the collector ring carbon brush from multiple status monitoring results;

[0007] The operating state of the collector ring carbon brush is determined according to the state abnormality coefficient.

[0008] In some embodiments of the present application, an intelligent monitoring system is established, including:

[0009] Obtain a historical monitoring log of the collector ring carbon brush, wherein the historical monitoring log includes multiple monitoring methods and corresponding historical monitoring timeliness, historical monitoring accuracy, and historical monitoring cost;

[0010] Generate historical monitoring evaluation values ​​for corresponding monitoring methods based on historical monitoring timeliness, historical monitoring accuracy and historical monitoring costs;

[0011] Eliminate the monitoring methods whose historical monitoring evaluation values ​​are less than the preset monitoring evaluation value threshold, and obtain the remaining monitoring methods in the historical monitoring log and the types and quantities of sensors required for the corresponding monitoring methods;

[0012] Install the corresponding sensors at the preset positions, monitor the operation of the collector ring carbon brushes according to the corresponding monitoring methods, and transmit the monitoring data to the intelligent monitoring terminal to form an intelligent monitoring system.

[0013] In some embodiments of the present application, a credibility analysis is performed on a variety of monitoring data to determine characteristic monitoring data of the collector ring carbon brush, including:

[0014] Determine the sensor type of the sensor corresponding to each type of monitoring data, obtain the data preprocessing methods corresponding to sensors of different sensor types, and perform data preprocessing on the corresponding monitoring data based on the data preprocessing methods;

[0015] Preset a preset collection time interval for the monitoring data after data preprocessing, obtain the monitoring data of the corresponding monitoring method according to the corresponding preset collection time interval, and generate a monitoring data set within the preset time period of the corresponding monitoring method;

[0016] Generate monitoring data differences according to the monitoring data of adjacent preset collection time intervals corresponding to the monitoring data set, compare the multiple monitoring data differences with the corresponding preset monitoring data differences, and generate a fluctuation evaluation value of the corresponding monitoring data according to the comparison result;

[0017] Set the credibility of the corresponding monitoring data according to the fluctuation evaluation value;

[0018] Pre-set credibility threshold;

[0019] If the credibility is less than the credibility threshold, the corresponding monitoring data is eliminated;

[0020] If the credibility is greater than the credibility threshold, the corresponding monitoring data is set as the characteristic monitoring data of the collector ring carbon brush.

[0021] In some embodiments of the present application, the characteristic monitoring data is analyzed to obtain a status monitoring result of each characteristic monitoring data, including:

[0022] Compare each characteristic monitoring data with a preset standard monitoring data interval respectively, and if the characteristic monitoring data is not in the corresponding standard monitoring data interval, calculate the first data difference value between the current characteristic monitoring data and the nearest critical value of the corresponding standard monitoring data interval;

[0023] If the first data difference value is less than the first preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a first-level abnormal state;

[0024] If the first data difference value is greater than the first preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a second-level abnormal state;

[0025] If the characteristic monitoring data is in the corresponding standard monitoring data interval, calculating a second data difference value between the current characteristic monitoring data and the nearest critical value of the corresponding standard monitoring data interval;

[0026] If the second data difference value is less than the second preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a first-level normal state;

[0027] If the second data difference value is greater than the second preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a secondary normal state;

[0028] The state abnormality coefficient of the collector ring carbon brush is generated according to the state monitoring results of various characteristic monitoring data and the credibility of the corresponding characteristic monitoring data.

[0029] In some embodiments of the present application, the state abnormality coefficient of the collector ring carbon brush is generated according to the state monitoring results of various characteristic monitoring data, including:

[0030] The calculation formula of the abnormal state coefficient of the collector ring carbon brush is:

[0031]

[0032] Among them, y is the state abnormality coefficient, a is the state abnormality conversion coefficient, n3 is the number of feature monitoring data of the secondary normal state, n4 is the number of feature monitoring data of the primary normal state, m is the total number of multiple feature monitoring data, k1v is the second data difference value of the vth feature monitoring data of the secondary normal state, uv is the credibility of the vth feature monitoring data of the secondary normal state, p0 is the second preset difference value threshold, k2s is the second data difference value of the sth feature monitoring data of the primary normal state, us is the credibility of the sth feature monitoring data of the primary normal state, n1 is the number of feature monitoring data of the primary abnormal state, n2 is the number of feature monitoring data of the secondary normal state, h0 is the first preset difference value threshold, L1i is the first data difference value of the i-th feature monitoring data of the primary abnormal state, ui is the credibility of the i-th feature monitoring data of the primary abnormal state, L2c is the first data difference value of the c-th feature monitoring data of the secondary abnormal state, and uc is the credibility of the c-th feature monitoring data of the secondary abnormal state.

[0033] In some embodiments of the present application, before determining the operating state of the collector ring carbon brush according to the abnormal state coefficient, the method includes:

[0034] Extract the corresponding sensor type and the historical status monitoring results of the historical feature monitoring data of the corresponding sensor type according to the monitoring data feature library;

[0035] Compare the status monitoring results of the feature monitoring data within a preset period with the historical status monitoring results of the same historical feature monitoring data extracted from the monitoring data feature library, and select multiple historical feature monitoring data with the same status monitoring results as each feature monitoring data;

[0036] Construct an actual data change curve according to each characteristic monitoring data within a preset time period, construct multiple historical data change curves according to the historical monitoring time periods of multiple historical characteristic monitoring data of the same state monitoring result of each characteristic monitoring data, and divide the historical data change curve into multiple historical data change curve segments according to the preset time period;

[0037] Perform similarity analysis on each actual data change curve and the corresponding multiple historical data change curve segments, and set the historical feature monitoring data in an adjacent preset time period after the historical data change curve segment with a similarity greater than a preset similarity threshold as the predicted feature monitoring data of the feature monitoring data corresponding to the current actual data change curve in the future preset time period;

[0038] Analyze the predicted characteristic monitoring data of each characteristic monitoring data in a preset future period to determine the predicted state monitoring results of the predicted characteristic monitoring data;

[0039] A compensation coefficient is set according to the differences between the predicted state monitoring results of multiple characteristic monitoring data and the current state monitoring results, the state abnormality coefficient is corrected according to the compensation coefficient, and the operating state of the collector ring carbon brush is determined according to the corrected state abnormality coefficient.

[0040] In some embodiments of the present application, each actual data change curve is subjected to similarity analysis with a plurality of corresponding historical data change curve segments, including:

[0041] Compare the characteristic monitoring data of multiple preset collection time intervals in each actual data change curve with the historical characteristic monitoring data of multiple preset collection time intervals in the corresponding multiple historical data change curve segments to obtain the comparison data difference of each preset collection time interval;

[0042] Generate characteristic monitoring data differences according to characteristic monitoring data of adjacent preset collection time intervals in each actual data change curve, and generate historical characteristic monitoring data differences according to historical characteristic monitoring data of adjacent preset collection time intervals in multiple historical data change curve segments corresponding to each actual data change curve;

[0043] Determine the similarity between each actual data change curve and the corresponding historical data change curve segment according to the comparison data difference of each preset collection time interval, the feature monitoring data difference corresponding to adjacent preset collection time intervals, and the historical feature monitoring data difference;

[0044] The calculation formula of the similarity is:

[0045]

[0046] Among them, D is the similarity, j1 is the first weight coefficient, j2 is the second weight coefficient, w is the number of preset collection time intervals in the preset period, △Gr is the comparison data difference of the rth preset collection time interval, △Q1r is the feature monitoring data difference of the rth adjacent preset collection time interval, △Q2r is the historical feature monitoring data difference of the rth adjacent preset collection time interval, and w-1 is the number of adjacent preset collection time intervals.

[0047] In some embodiments of the present application, the compensation coefficient is set according to the difference between the predicted state monitoring results of the multiple characteristic monitoring data and the current state monitoring results, including:

[0048] Determine the predicted change level of the current state monitoring result according to the result difference between the predicted state monitoring result and the current state monitoring result of each characteristic monitoring data;

[0049] The predicted change level includes a predicted increase level and a predicted decrease level, wherein the predicted increase level includes a first preset predicted increase level, a second preset predicted increase level and a third preset predicted increase level, the predicted decrease level includes a first preset predicted decrease level, a second preset predicted decrease level and a third preset predicted decrease level, and each preset predicted increase level and preset predicted decrease level is configured with a corresponding preset compensation coefficient;

[0050] A plurality of preset compensation coefficients are determined according to the predicted change level of the current state monitoring result of each characteristic monitoring data, and the plurality of preset compensation coefficients are averaged to obtain the compensation coefficient.

[0051] In some embodiments of the present application, determining the operating state of the collector ring carbon brush according to the corrected state abnormality coefficient includes:

[0052] Presetting a first preset state abnormal coefficient threshold and a second preset state abnormal coefficient threshold;

[0053] When the corrected state abnormality coefficient is less than the first preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is a normal state;

[0054] When the corrected state abnormality coefficient is between the first preset state abnormality coefficient threshold and the second preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is an abnormal state, and an early warning instruction is sent;

[0055] When the corrected state abnormality coefficient is greater than the second preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is a fault state, and an alarm instruction is sent.

[0056] Compared with the prior art, the wind turbine collector ring carbon brush operation monitoring method of the present application has the following beneficial effects:

[0057] By constructing an intelligent monitoring system, characteristic monitoring data of various monitoring methods are obtained, and the state monitoring results of each characteristic monitoring data are determined, and the predicted state monitoring results are generated according to the predicted characteristic monitoring data of the future preset time period, and the state abnormality coefficient is generated according to the state monitoring results. The compensation coefficient is generated according to the predicted state monitoring results and the state abnormality coefficient is corrected to improve the accuracy of the state abnormality coefficient, timely reflect the operating status of the collector ring carbon brush in the current preset time period and the future preset time period, and reduce the failure frequency of wind turbines. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flow chart of a method for monitoring the operation of carbon brushes in a wind turbine collector ring in a preferred embodiment of the present application. DETAILED DESCRIPTION

[0059] The specific implementation methods of the present application are further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application but are not intended to limit the scope of the present application.

[0060] In the description of the present application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0061] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0062] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0063] like Figure 1 As shown, a method for monitoring the operation of a carbon brush of a wind turbine collector ring according to a preferred embodiment of the present application includes:

[0064] Step S101: Establish an intelligent monitoring system, obtain monitoring data of the monitoring methods included in the intelligent monitoring system, perform credibility analysis on a variety of monitoring data, and determine characteristic monitoring data of the collector ring carbon brush;

[0065] Step S102: Analyze the characteristic monitoring data to obtain the state monitoring result of each characteristic monitoring data, and generate the state abnormality coefficient of the collector ring carbon brush from the multiple state monitoring results;

[0066] Step S103: determining the operating state of the collector ring carbon brush according to the abnormal state coefficient.

[0067] In this embodiment, the monitoring methods include current monitoring, temperature monitoring, vibration monitoring, visual monitoring, etc. Current monitoring refers to using a current sensor to monitor the output current of the wind turbine. When the carbon brush is severely worn or in poor contact, it will cause abnormal current. The state of the carbon brush can be judged by monitoring the change in current. Vibration monitoring refers to installing a vibration sensor on the rotating part of the wind turbine to monitor the vibration of the carbon brush. Carbon brush wear or poor contact may cause abnormal vibration. The state of the carbon brush can be judged by monitoring vibration. Temperature monitoring refers to installing a temperature sensor near the collector ring of the wind turbine to monitor the working temperature of the carbon brush. When the carbon brush is worn or in poor contact, too much heat will be generated, causing the temperature to rise. The state of the carbon brush can be judged by monitoring the temperature change. Visual monitoring refers to installing a camera or an optical fiber sensor at the collector ring position of the wind turbine to monitor the wear and contact state of the carbon brush in real time. The working state of the carbon brush can be judged by image processing or data analysis of optical sensors. Analyze a variety of monitoring methods, obtain monitoring evaluation values, select appropriate monitoring methods, and build an intelligent monitoring system to achieve accurate monitoring and early warning of the state of the collector ring carbon brush.

[0068] In some embodiments of the present application, an intelligent monitoring system is established, including:

[0069] Obtain a historical monitoring log of the collector ring carbon brush, wherein the historical monitoring log includes multiple monitoring methods and corresponding historical monitoring timeliness, historical monitoring accuracy, and historical monitoring cost;

[0070] Generate historical monitoring evaluation values ​​for corresponding monitoring methods based on historical monitoring timeliness, historical monitoring accuracy and historical monitoring costs;

[0071] Eliminate the monitoring methods whose historical monitoring evaluation values ​​are less than the preset monitoring evaluation value threshold, and obtain the remaining monitoring methods in the historical monitoring log and the types and quantities of sensors required for the corresponding monitoring methods;

[0072] Install the corresponding sensors at the preset positions, monitor the operation of the collector ring carbon brushes according to the corresponding monitoring methods, and transmit the monitoring data to the intelligent monitoring terminal to form an intelligent monitoring system.

[0073] In this embodiment, by analyzing the monitoring logs of various monitoring methods of the collector ring carbon brushes, a monitoring evaluation value of each monitoring method is obtained, and a reasonable monitoring method is selected according to the monitoring evaluation value, that is, a monitoring method with high monitoring accuracy, low monitoring degree and timely monitoring, so as to improve the accuracy of subsequent characteristic monitoring data.

[0074] In some embodiments of the present application, a credibility analysis is performed on a variety of monitoring data to determine characteristic monitoring data of the collector ring carbon brush, including:

[0075] Determine the sensor type of the sensor corresponding to each type of monitoring data, obtain the data preprocessing methods corresponding to sensors of different sensor types, and perform data preprocessing on the corresponding monitoring data based on the data preprocessing methods;

[0076] Preset a preset collection time interval for the monitoring data after data preprocessing, obtain the monitoring data of the corresponding monitoring method according to the corresponding preset collection time interval, and generate a monitoring data set within the preset time period of the corresponding monitoring method;

[0077] Generate monitoring data differences according to the monitoring data of adjacent preset collection time intervals corresponding to the monitoring data set, compare the multiple monitoring data differences with the corresponding preset monitoring data differences, and generate a fluctuation evaluation value of the corresponding monitoring data according to the comparison result;

[0078] Set the credibility of the corresponding monitoring data according to the fluctuation evaluation value;

[0079] Pre-set credibility threshold;

[0080] If the credibility is less than the credibility threshold, the corresponding monitoring data is eliminated;

[0081] If the credibility is greater than the credibility threshold, the corresponding monitoring data is set as the characteristic monitoring data of the collector ring carbon brush.

[0082] In this embodiment, the data preprocessing method includes data cleaning, data standardization and normalization, data feature selection, etc. The data preprocessing method is used to improve the accuracy and reliability of the monitoring data collected by the corresponding sensor, laying the foundation for the subsequent determination of the operating status of the collector ring carbon brush and improving the status accuracy.

[0083] In this embodiment, the fluctuation evaluation value refers to the degree of fluctuation of the monitoring data within a preset time period. A smaller degree of fluctuation indicates that the stability of the monitoring data is higher, that is, the smaller the fluctuation evaluation value, the greater the credibility. A larger degree of fluctuation indicates that the stability of the monitoring data is lower, that is, the larger the fluctuation evaluation value, the smaller the credibility.

[0084] In this embodiment, characteristic monitoring data is determined by calculating the accuracy of the monitoring method and the degree of fluctuation of the monitoring data, thereby reducing the amount of data processing and analysis. Based on the analysis results of the characteristic monitoring data, the operating status of the collector ring carbon brush is accurately and quickly judged.

[0085] In some embodiments of the present application, the characteristic monitoring data is analyzed to obtain a status monitoring result of each characteristic monitoring data, including:

[0086] Compare each characteristic monitoring data with a preset standard monitoring data interval respectively, and if the characteristic monitoring data is not in the corresponding standard monitoring data interval, calculate the first data difference value between the current characteristic monitoring data and the nearest critical value of the corresponding standard monitoring data interval;

[0087] If the first data difference value is less than the first preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a first-level abnormal state;

[0088] If the first data difference value is greater than the first preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a second-level abnormal state;

[0089] If the characteristic monitoring data is in the corresponding standard monitoring data interval, calculating a second data difference value between the current characteristic monitoring data and the nearest critical value of the corresponding standard monitoring data interval;

[0090] If the second data difference value is less than the second preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a first-level normal state;

[0091] If the second data difference value is greater than the second preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a secondary normal state;

[0092] The state abnormality coefficient of the collector ring carbon brush is generated according to the state monitoring results of various characteristic monitoring data and the credibility of the corresponding characteristic monitoring data.

[0093] In this embodiment, the abnormality degree of the secondary abnormal state is greater than that of the primary abnormal state, and the normality degree of the primary normal state is less than that of the secondary normal state, that is, the monitoring data of the secondary abnormal state indicates a greater probability of abnormality of the collector ring carbon brush than the monitoring data of the primary abnormal state, and the monitoring data of the secondary normal state indicates a better operating state of the collector ring carbon brush than the monitoring data of the primary normal state.

[0094] In this embodiment, the first data difference value refers to the difference value of the critical data closest to the endpoint of the standard monitoring data interval when the characteristic monitoring data is not in the standard monitoring data interval, and the second data difference value refers to the difference value of the critical data closest to the endpoint of the standard monitoring data interval when the characteristic monitoring data is in the standard monitoring data interval. The first preset difference value threshold is set based on the minimum difference value between the historical monitoring data belonging to the secondary abnormal state and the critical data closest to the standard monitoring data interval, and the second preset difference value threshold is set based on the minimum difference value between the historical monitoring data belonging to the secondary normal state and the critical data closest to the standard monitoring data interval.

[0095] In some embodiments of the present application, the state abnormality coefficient of the collector ring carbon brush is generated according to the state monitoring results of various characteristic monitoring data, including:

[0096] The calculation formula of the abnormal state coefficient of the collector ring carbon brush is:

[0097]

[0098] Among them, y is the state abnormality coefficient, a is the state abnormality conversion coefficient, n3 is the number of feature monitoring data of the secondary normal state, n4 is the number of feature monitoring data of the primary normal state, m is the total number of multiple feature monitoring data, k1v is the second data difference value of the vth feature monitoring data of the secondary normal state, uv is the credibility of the vth feature monitoring data of the secondary normal state, p0 is the second preset difference value threshold, k2s is the second data difference value of the sth feature monitoring data of the primary normal state, us is the credibility of the sth feature monitoring data of the primary normal state, n1 is the number of feature monitoring data of the primary abnormal state, n2 is the number of feature monitoring data of the secondary normal state, h0 is the first preset difference value threshold, L1i is the first data difference value of the i-th feature monitoring data of the primary abnormal state, ui is the credibility of the i-th feature monitoring data of the primary abnormal state, L2c is the first data difference value of the c-th feature monitoring data of the secondary abnormal state, and uc is the credibility of the c-th feature monitoring data of the secondary abnormal state.

[0099] In some embodiments of the present application, before determining the operating state of the collector ring carbon brush according to the abnormal state coefficient, the method includes:

[0100] Extract the corresponding sensor type and the historical status monitoring results of the historical feature monitoring data of the corresponding sensor type according to the monitoring data feature library;

[0101] Compare the status monitoring results of the feature monitoring data within a preset period with the historical status monitoring results of the same historical feature monitoring data extracted from the monitoring data feature library, and select multiple historical feature monitoring data with the same status monitoring results as each feature monitoring data;

[0102] Construct an actual data change curve according to each characteristic monitoring data within a preset time period, construct multiple historical data change curves according to the historical monitoring time periods of multiple historical characteristic monitoring data of the same state monitoring result of each characteristic monitoring data, and divide the historical data change curve into multiple historical data change curve segments according to the preset time period;

[0103] Perform similarity analysis on each actual data change curve and the corresponding multiple historical data change curve segments, and set the historical characteristic monitoring data in an adjacent preset time period after the historical data change curve segment with the greatest similarity as the predicted characteristic monitoring data of the characteristic monitoring data corresponding to the current actual data change curve in the future preset time period;

[0104] Analyze the predicted characteristic monitoring data of each characteristic monitoring data in a preset future period to determine the predicted state monitoring results of the predicted characteristic monitoring data;

[0105] A compensation coefficient is set according to the differences between the predicted state monitoring results of multiple characteristic monitoring data and the current state monitoring results, the state abnormality coefficient is corrected according to the compensation coefficient, and the operating state of the collector ring carbon brush is determined according to the corrected state abnormality coefficient.

[0106] In this embodiment, the monitoring data feature library refers to the historical monitoring data collected by the sensor types corresponding to all monitoring methods and the corresponding historical status monitoring results. The historical monitoring data with the same feature monitoring data and the same status monitoring results are screened out according to the monitoring data feature library, and multiple historical data change curve segments are constructed. According to the similarity between the actual data change curve and the historical data change curve segments, the feature monitoring data is predicted for a preset time period in the future to obtain the predicted feature monitoring data and the predicted status monitoring results.

[0107] In this embodiment, by predicting the characteristic monitoring data of a preset time period in the future, the predicted characteristic monitoring data and the predicted state monitoring results are obtained, the result difference is obtained and a compensation coefficient is set, and the state abnormality coefficient is corrected according to the compensation coefficient to improve the accuracy of the state abnormality coefficient, that is, to improve the accuracy of the diagnosis of the operating status of the collector ring carbon brush.

[0108] In some embodiments of the present application, each actual data change curve is subjected to similarity analysis with a plurality of corresponding historical data change curve segments, including:

[0109] Compare the characteristic monitoring data of multiple preset collection time intervals in each actual data change curve with the historical characteristic monitoring data of multiple preset collection time intervals in the corresponding multiple historical data change curve segments to obtain the comparison data difference of each preset collection time interval;

[0110] Generate characteristic monitoring data differences according to characteristic monitoring data of adjacent preset collection time intervals in each actual data change curve, and generate historical characteristic monitoring data differences according to historical characteristic monitoring data of adjacent preset collection time intervals in multiple historical data change curve segments corresponding to each actual data change curve;

[0111] Determine the similarity between each actual data change curve and the corresponding historical data change curve segment according to the comparison data difference of each preset collection time interval, the feature monitoring data difference corresponding to adjacent preset collection time intervals, and the historical feature monitoring data difference;

[0112] The calculation formula of the similarity is:

[0113]

[0114] Among them, D is the similarity, j1 is the first weight coefficient, j2 is the second weight coefficient, w is the number of preset collection time intervals in the preset period, △Gr is the comparison data difference of the rth preset collection time interval, △Q1r is the feature monitoring data difference of the rth adjacent preset collection time interval, △Q2r is the historical feature monitoring data difference of the rth adjacent preset collection time interval, and w-1 is the number of adjacent preset collection time intervals.

[0115] In this embodiment, the comparison data difference refers to the data difference between the feature monitoring data and the historical feature monitoring data at the same preset collection time interval. The difference similarity is determined based on multiple comparison data differences, and the fluctuation degree similarity of adjacent preset collection time intervals is obtained based on multiple feature monitoring data differences and historical feature monitoring data differences, thereby screening out the historical data change curve segment with the greatest similarity to the actual data change curve, and setting the historical feature monitoring data of the preset time period adjacent to the historical data change curve segment with the greatest similarity as the predicted feature detection data of the current feature monitoring data.

[0116] In some embodiments of the present application, the compensation coefficient is set according to the difference between the predicted state monitoring results of the multiple characteristic monitoring data and the current state monitoring results, including:

[0117] Determine the predicted change level of the current state monitoring result according to the result difference between the predicted state monitoring result and the current state monitoring result of each characteristic monitoring data;

[0118] The predicted change level includes a predicted increase level and a predicted decrease level, wherein the predicted increase level includes a first preset predicted increase level, a second preset predicted increase level and a third preset predicted increase level, the predicted decrease level includes a first preset predicted decrease level, a second preset predicted decrease level and a third preset predicted decrease level, and each preset predicted increase level and preset predicted decrease level is configured with a corresponding preset compensation coefficient;

[0119] A plurality of preset compensation coefficients are determined according to the predicted change level of the current state monitoring result of each characteristic monitoring data, and the plurality of preset compensation coefficients are averaged to obtain the compensation coefficient.

[0120] In this embodiment, the predicted rising level means that the predicted state monitoring result is better than the state in the current state monitoring result. For example, if the current state monitoring result is a level one abnormal state and the predicted state monitoring result is a level two normal state, then the predicted rising level is the second predicted rising level; or if the current state monitoring result is a level one normal state and the predicted state monitoring result is a level two abnormal state, then the predicted falling level is the second predicted falling level.

[0121] In this embodiment, the values ​​of the preset compensation coefficient and the compensation coefficient are both (0.65, 1.35). When the predicted decrease level is greater, the preset compensation coefficient is smaller, and when the predicted increase level is greater, the preset compensation coefficient is larger.

[0122] In this embodiment, by predicting the characteristic monitoring data of the future preset time period, the predicted state monitoring results of each characteristic monitoring data in the future preset time period are obtained, and compared with the current state monitoring results, so as to determine the compensation coefficient to correct the state abnormality coefficient, improve the monitoring accuracy of the operating status of the collector ring carbon brush in the current preset time period and the future preset time period, and formulate adjustment strategies in time to reduce the probability of failure.

[0123] In some embodiments of the present application, determining the operating state of the collector ring carbon brush according to the corrected state abnormality coefficient includes:

[0124] Presetting a first preset state abnormal coefficient threshold and a second preset state abnormal coefficient threshold;

[0125] When the corrected state abnormality coefficient is less than the first preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is a normal state;

[0126] When the corrected state abnormality coefficient is between the first preset state abnormality coefficient threshold and the second preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is an abnormal state, and an early warning instruction is sent;

[0127] When the corrected state abnormality coefficient is greater than the second preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is a fault state, and an alarm instruction is sent.

[0128] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present application. These improvements and substitutions should also be regarded as the scope of protection of the present application.

Claims

1. A method for monitoring the operation of a collector ring carbon brush of a wind turbine, characterized in that: include: Establish an intelligent monitoring system, obtain monitoring data of the monitoring methods included in the intelligent monitoring system, conduct credibility analysis on various monitoring data, and determine the characteristic monitoring data of the collector ring carbon brush; Analyze the characteristic monitoring data to obtain the status monitoring results of each characteristic monitoring data, and generate the status abnormality coefficient of the collector ring carbon brush from multiple status monitoring results; The operating state of the collector ring carbon brush is determined according to the state abnormality coefficient.

2. The wind turbine collector ring carbon brush operation monitoring method according to claim 1, characterized in that: Establish an intelligent monitoring system, including: Obtain a historical monitoring log of the collector ring carbon brush, wherein the historical monitoring log includes multiple monitoring methods and corresponding historical monitoring timeliness, historical monitoring accuracy, and historical monitoring cost; Generate historical monitoring evaluation values ​​for corresponding monitoring methods based on historical monitoring timeliness, historical monitoring accuracy and historical monitoring costs; Eliminate the monitoring methods whose historical monitoring evaluation values ​​are less than the preset monitoring evaluation value threshold, and obtain the remaining monitoring methods in the historical monitoring log and the types and quantities of sensors required for the corresponding monitoring methods; Install the corresponding sensors at the preset positions, monitor the operation of the collector ring carbon brushes according to the corresponding monitoring methods, and transmit the monitoring data to the intelligent monitoring terminal to form an intelligent monitoring system.

3. The wind turbine collector ring carbon brush operation monitoring method according to claim 2, characterized in that: Conduct credibility analysis on various monitoring data to determine the characteristic monitoring data of the collector ring carbon brush, including: Determine the sensor type of the sensor corresponding to each type of monitoring data, obtain the data preprocessing methods corresponding to sensors of different sensor types, and perform data preprocessing on the corresponding monitoring data based on the data preprocessing methods; Preset a preset collection time interval for the monitoring data after data preprocessing, obtain the monitoring data of the corresponding monitoring method according to the corresponding preset collection time interval, and generate a monitoring data set within the preset time period of the corresponding monitoring method; Generate monitoring data differences according to the monitoring data of adjacent preset collection time intervals corresponding to the monitoring data set, compare the multiple monitoring data differences with the corresponding preset monitoring data differences, and generate a fluctuation evaluation value of the corresponding monitoring data according to the comparison result; Set the credibility of the corresponding monitoring data according to the fluctuation evaluation value; Pre-set credibility threshold; If the credibility is less than the credibility threshold, the corresponding monitoring data is eliminated; If the credibility is greater than the credibility threshold, the corresponding monitoring data is set as the characteristic monitoring data of the collector ring carbon brush.

4. The wind turbine collector ring carbon brush operation monitoring method according to claim 3, characterized in that: Analyze the characteristic monitoring data to obtain the status monitoring results of each characteristic monitoring data, including: Compare each characteristic monitoring data with a preset standard monitoring data interval respectively, and if the characteristic monitoring data is not in the corresponding standard monitoring data interval, calculate the first data difference value between the current characteristic monitoring data and the nearest critical value of the corresponding standard monitoring data interval; If the first data difference value is less than the first preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a first-level abnormal state; If the first data difference value is greater than the first preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a second-level abnormal state; If the characteristic monitoring data is in the corresponding standard monitoring data interval, calculating a second data difference value between the current characteristic monitoring data and the nearest critical value of the corresponding standard monitoring data interval; If the second data difference value is less than the second preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a first-level normal state; If the second data difference value is greater than the second preset difference value threshold, the state monitoring result of the corresponding characteristic monitoring data is set to a secondary normal state; The state abnormality coefficient of the collector ring carbon brush is generated according to the state monitoring results of various characteristic monitoring data and the credibility of the corresponding characteristic monitoring data.

5. The wind turbine collector ring carbon brush operation monitoring method according to claim 4, characterized in that: The state abnormality coefficient of the collector ring carbon brush is generated based on the state monitoring results of various characteristic monitoring data, including: The calculation formula of the abnormal state coefficient of the collector ring carbon brush is: Among them, y is the state abnormality coefficient, a is the state abnormality conversion coefficient, n3 is the number of feature monitoring data of the secondary normal state, n4 is the number of feature monitoring data of the primary normal state, m is the total number of multiple feature monitoring data, k1v is the second data difference value of the vth feature monitoring data of the secondary normal state, uv is the credibility of the vth feature monitoring data of the secondary normal state, p0 is the second preset difference value threshold, k2s is the second data difference value of the sth feature monitoring data of the primary normal state, us is the credibility of the sth feature monitoring data of the primary normal state, n1 is the number of feature monitoring data of the primary abnormal state, n2 is the number of feature monitoring data of the secondary normal state, h0 is the first preset difference value threshold, L1i is the first data difference value of the i-th feature monitoring data of the primary abnormal state, ui is the credibility of the i-th feature monitoring data of the primary abnormal state, L2c is the first data difference value of the c-th feature monitoring data of the secondary abnormal state, and uc is the credibility of the c-th feature monitoring data of the secondary abnormal state.

6. The wind turbine collector ring carbon brush operation monitoring method according to claim 5, characterized in that: Before determining the operating state of the collector ring carbon brush according to the abnormal state coefficient, the method includes: Extract the corresponding sensor type and the historical status monitoring results of the historical feature monitoring data of the corresponding sensor type according to the monitoring data feature library; Compare the status monitoring results of the feature monitoring data within a preset period with the historical status monitoring results of the same historical feature monitoring data extracted from the monitoring data feature library, and select multiple historical feature monitoring data with the same status monitoring results as each feature monitoring data; Construct an actual data change curve according to each characteristic monitoring data within a preset time period, construct multiple historical data change curves according to the historical monitoring time periods of multiple historical characteristic monitoring data of the same state monitoring result of each characteristic monitoring data, and divide the historical data change curve into multiple historical data change curve segments according to the preset time period; Perform similarity analysis on each actual data change curve and the corresponding multiple historical data change curve segments, and set the historical feature monitoring data in an adjacent preset time period after the historical data change curve segment with a similarity greater than a preset similarity threshold as the predicted feature monitoring data of the feature monitoring data corresponding to the current actual data change curve in the future preset time period; Analyze the predicted characteristic monitoring data of each characteristic monitoring data in a preset future period to determine the predicted state monitoring results of the predicted characteristic monitoring data; A compensation coefficient is set according to the differences between the predicted state monitoring results of multiple characteristic monitoring data and the current state monitoring results, the state abnormality coefficient is corrected according to the compensation coefficient, and the operating state of the collector ring carbon brush is determined according to the corrected state abnormality coefficient.

7. The wind turbine collector ring carbon brush operation monitoring method according to claim 6, characterized in that: Perform similarity analysis on each actual data change curve and the corresponding multiple historical data change curve segments, including: Compare the characteristic monitoring data of multiple preset collection time intervals in each actual data change curve with the historical characteristic monitoring data of multiple preset collection time intervals in the corresponding multiple historical data change curve segments to obtain the comparison data difference of each preset collection time interval; Generate characteristic monitoring data differences according to characteristic monitoring data of adjacent preset collection time intervals in each actual data change curve, and generate historical characteristic monitoring data differences according to historical characteristic monitoring data of adjacent preset collection time intervals in multiple historical data change curve segments corresponding to each actual data change curve; Determine the similarity between each actual data change curve and the corresponding historical data change curve segment according to the comparison data difference of each preset collection time interval, the feature monitoring data difference corresponding to adjacent preset collection time intervals, and the historical feature monitoring data difference; The calculation formula of the similarity is: Among them, D is the similarity, j1 is the first weight coefficient, j2 is the second weight coefficient, w is the number of preset collection time intervals in the preset period, △Gr is the comparison data difference of the rth preset collection time interval, △Q1r is the feature monitoring data difference of the rth adjacent preset collection time interval, △Q2r is the historical feature monitoring data difference of the rth adjacent preset collection time interval, and w-1 is the number of adjacent preset collection time intervals.

8. The wind turbine collector ring carbon brush operation monitoring method according to claim 7, characterized in that: The compensation coefficient is set according to the differences between the predicted state monitoring results of the various characteristic monitoring data and the current state monitoring results, including: Determine the predicted change level of the current state monitoring result according to the result difference between the predicted state monitoring result and the current state monitoring result of each characteristic monitoring data; The predicted change level includes a predicted increase level and a predicted decrease level, wherein the predicted increase level includes a first preset predicted increase level, a second preset predicted increase level and a third preset predicted increase level, the predicted decrease level includes a first preset predicted decrease level, a second preset predicted decrease level and a third preset predicted decrease level, and each preset predicted increase level and preset predicted decrease level is configured with a corresponding preset compensation coefficient; A plurality of preset compensation coefficients are determined according to the predicted change level of the current state monitoring result of each characteristic monitoring data, and the plurality of preset compensation coefficients are averaged to obtain the compensation coefficient.

9. The wind turbine collector ring carbon brush operation monitoring method according to claim 8, characterized in that: Determine the operating status of the collector ring carbon brush according to the corrected abnormal status coefficient, including: Presetting a first preset state abnormal coefficient threshold and a second preset state abnormal coefficient threshold; When the corrected state abnormality coefficient is less than the first preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is a normal state; When the corrected state abnormality coefficient is between the first preset state abnormality coefficient threshold and the second preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is an abnormal state, and an early warning instruction is sent; When the corrected state abnormality coefficient is greater than the second preset state abnormality coefficient threshold, it is determined that the operating state of the collector ring carbon brush is a fault state, and an alarm instruction is sent.

Images