A method for monitoring the operating status of a wind turbine generator set

Abstract

The present invention discloses a method for monitoring the operating status of a wind turbine generator set, belonging to the technical field of wind turbine generator set monitoring. The method comprises the following steps: S1, preprocessing vibration data of a wind turbine tower and real-time vibration data of blades to obtain the latest vibration data of the wind turbine tower and blades; S2, extracting a tower feature matrix and a blade feature matrix; S3, determining tower feature values ​​and blade feature values; and S4, constructing a state detection model, inputting the tower feature values ​​and blade feature values ​​into the state detection model, and determining the operating status of the wind turbine generator set. The method for monitoring the operating status of a wind turbine generator set determines the operating status of the wind turbine generator set by constructing a state detection model, performing feature fusion on tower features and blade features, and performing value comparison. The process is simple and easy to implement.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind turbine generator set monitoring, and in particular relates to a method for monitoring the operating status of a wind turbine generator set. Background Art

[0002] Wind turbines are devices that convert wind energy into electricity. Their proper operation directly affects wind power production. Failures in wind turbines can damage their components and even lead to more serious consequences. To ensure the safe operation of wind turbines, it is necessary to monitor their operating status and promptly diagnose any potential failures. Summary of the Invention

[0003] In order to solve the above problems, the present invention proposes a method for monitoring the operating status of a wind turbine generator set.

[0004] The technical solution of the present invention is: a method for monitoring the operating status of a wind turbine generator system comprises the following steps:

[0005] S1. Collecting real-time vibration data of the wind turbine tower and blades at different times, and preprocessing the vibration data of the wind turbine tower and the real-time vibration data of the blades to obtain the latest vibration data of the wind turbine tower and blades;

[0006] S2. Extracting a tower feature matrix based on the latest vibration data of the wind turbine tower; extracting a blade feature matrix based on the latest vibration data of the blade;

[0007] S3. Determine the tower characteristic value according to the tower characteristic matrix; determine the blade characteristic value according to the blade characteristic matrix;

[0008] S4. Construct a state detection model, input the tower characteristic value and the blade characteristic value into the state detection model, and determine the operating state of the wind turbine.

[0009] The beneficial effects of the present invention are:

[0010] (1) The wind turbine operating status monitoring method pre-processes the real-time vibration data of the wind turbine tower and blades through a sliding window, eliminates abnormal vibration data, and ensures the accuracy of the data processed in subsequent steps;

[0011] (2) The wind turbine operating status monitoring method can intuitively reflect the characteristics of the tower and blades by constructing the tower characteristic matrix and the blade characteristic matrix. The variance and standard deviation can reflect the discrete degree of vibration data, which is convenient for objectively describing the vibration situation.

[0012] (3) The wind turbine operating status monitoring method constructs a status detection model, fuses the tower characteristics and blade characteristics, and compares the size values ​​to determine the wind turbine operating status. The process is simple and easy to implement.

[0013] Furthermore, in S1, the methods for preprocessing the real-time vibration data of the wind turbine tower and blades are the same, which are: using a sliding window to evenly divide the real-time vibration data at different times into several vibration data groups, extracting the mean of all vibration data in each vibration data group as the vibration group eigenvalue of each vibration data group, and calculating the mean of all vibration group eigenvalues ​​as the vibration eigenvalue, eliminating the real-time vibration data that is less than the vibration eigenvalue, and completing the preprocessing.

[0014] Furthermore, the calculation formula for the sliding window size is:

[0015]

[0016]

[0017] In the formula, a represents the length of the sliding window, b represents the width of the sliding window, and K max Indicates the maximum real-time vibration data, K min Indicates the minimum real-time vibration data, K ave represents the mean of all real-time vibration data, n represents the number of basic windows that the sliding window can accommodate, and T represents the duration of vibration data collection.

[0018] Furthermore, S2 includes the following sub-steps:

[0019] S21. Sort the latest vibration data of the wind turbine tower from small to large to generate a wind turbine tower vibration sequence;

[0020] S22. Generate a tower characteristic matrix according to the wind turbine tower vibration sequence;

[0021] S23, sorting the latest vibration data of the blade tower from small to large to generate a wind turbine blade vibration sequence;

[0022] S24. Generate a blade feature matrix based on the wind turbine blade vibration sequence.

[0023] Furthermore, in S22, the specific method for generating the tower characteristic matrix is: calculating the standard deviation between every two adjacent latest vibration data in the wind turbine tower vibration sequence, generating a tower standard deviation sequence, and generating the tower characteristic matrix according to the tower standard deviation sequence.

[0024] Furthermore, the element X in the tower characteristic matrix i The calculation formula is:

[0025]

[0026] Where s irepresents the i-th standard deviation in the tower standard deviation sequence, x1 and x2 represent the two adjacent latest vibration data corresponding to the i-th standard deviation in the wind turbine tower vibration sequence, and S0 represents the mean of all standard deviations in the tower standard deviation sequence.

[0027] Furthermore, in S24, the specific method for generating the blade characteristic matrix is: calculating the standard deviation and variance between every two adjacent latest vibration data in the wind turbine blade vibration sequence, generating a blade standard deviation sequence and a blade variance sequence, and generating a blade characteristic matrix according to the blade standard deviation sequence and the blade variance sequence.

[0028] Furthermore, the element Y in the leaf feature matrix j The calculation formula is:

[0029]

[0030] Where, t j represents the jth standard deviation in the blade standard deviation sequence, y1 and y2 represent the two adjacent latest vibration data corresponding to the jth standard deviation in the blade vibration sequence, T1 represents the mean of all standard deviations in the blade standard deviation sequence, and T2 represents the mean of all variances in the blade standard deviation sequence.

[0031] Furthermore, in S3, a specific method for determining the tower eigenvalue is: calculating an error coefficient of the latest vibration data of the wind turbine tower, and taking the product of the error coefficient of the latest vibration data of the wind turbine tower and the eigenvalue of the tower characteristic matrix as the tower eigenvalue;

[0032] The calculation formula for the error coefficient U of the latest vibration data of the wind turbine tower is:

[0033]

[0034] Where, P represents the number of all the latest vibration data of the wind turbine tower, u p represents the latest vibration data of the p-th wind turbine tower, X represents the tower characteristic matrix, and E represents the unit matrix;

[0035] In S3, the specific method for determining the blade eigenvalue U is: calculating the error coefficient of the latest vibration data of the blade, and taking the product of the error coefficient of the latest vibration data of the blade and the eigenvalue of the blade characteristic matrix as the blade eigenvalue;

[0036] The calculation formula of the error coefficient V of the latest vibration data of the blade is:

[0037]

[0038] Where Q represents the number of all the latest vibration data of the blade, v qrepresents the latest vibration data of the qth blade, and Y represents the tower characteristic matrix.

[0039] Furthermore, in S4, a specific method for determining the operating state of the wind turbine generator set is as follows: inputting the tower characteristic value and the blade characteristic value into a state detection model to generate an operating characteristic value of the wind turbine generator set; if the operating characteristic value of the wind turbine generator set is greater than a wind turbine generator set operating characteristic threshold, the wind turbine generator set is operating abnormally; otherwise, the wind turbine generator set is operating normally;

[0040] Among them, the expression of the state detection model Z is:

[0041]

[0042] Where X represents the tower characteristic matrix, Y represents the blade characteristic matrix, U represents the error coefficient of the latest vibration data of the wind turbine tower, V represents the error coefficient of the latest vibration data of the blade, u represents the tower characteristic value, and v represents the blade characteristic value. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Flowchart of the wind turbine operating status monitoring method. DETAILED DESCRIPTION

[0044] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0045] like Figure 1 As shown, the present invention provides a method for monitoring the operating status of a wind turbine generator set, comprising the following steps:

[0046] S1. Collecting real-time vibration data of the wind turbine tower and blades at different times, and preprocessing the vibration data of the wind turbine tower and the real-time vibration data of the blades to obtain the latest vibration data of the wind turbine tower and blades;

[0047] S2. Extracting a tower feature matrix based on the latest vibration data of the wind turbine tower; extracting a blade feature matrix based on the latest vibration data of the blade;

[0048] S3. Determine the tower characteristic value according to the tower characteristic matrix; determine the blade characteristic value according to the blade characteristic matrix;

[0049] S4. Construct a state detection model, input the tower characteristic value and the blade characteristic value into the state detection model, and determine the operating state of the wind turbine.

[0050] Wind turbine towers, as the masts of wind power generation, primarily support wind turbines and absorb vibrations. Tower vibrations can easily cause resonance between the tower and blades, affecting the stability of the entire wind turbine. Even minor changes or damage to the blades during operation can lead to more serious consequences, such as turbine failure or blade breakage. Therefore, tower condition monitoring is crucial.

[0051] In the embodiment of the present invention, in S1, the method for preprocessing the real-time vibration data of the wind turbine tower and the blade is the same, which is: using a sliding window to evenly divide the real-time vibration data at different times into several vibration data groups, extracting the mean of all vibration data in each vibration data group as the vibration group eigenvalue of each vibration data group, and calculating the mean of all vibration group eigenvalues ​​as the vibration eigenvalue, and eliminating the real-time vibration data smaller than the vibration eigenvalue to complete the preprocessing.

[0052] In the present invention, a sliding window is used to divide the vibration data at different times, and vibration data groups of the same size are determined. The mean of the mean values ​​of all vibration data groups is used as the vibration characteristic value, and a size comparison is performed. The vibration data with smaller reference value is eliminated, and the remaining vibration data is used as the latest vibration data, which can objectively and truly reflect the operation status of the wind turbine.

[0053] In the embodiment of the present invention, the calculation formula for the sliding window size is:

[0054]

[0055]

[0056] In the formula, a represents the length of the sliding window, b represents the width of the sliding window, and K max Indicates the maximum real-time vibration data, K min Indicates the minimum real-time vibration data, K ave represents the mean of all real-time vibration data, n represents the number of basic windows that the sliding window can accommodate, and T represents the duration of vibration data collection.

[0057] In this embodiment of the present invention, S2 includes the following sub-steps:

[0058] S21. Sort the latest vibration data of the wind turbine tower from small to large to generate a wind turbine tower vibration sequence;

[0059] S22. Generate a tower characteristic matrix according to the wind turbine tower vibration sequence;

[0060] S23, sorting the latest vibration data of the blade tower from small to large to generate a wind turbine blade vibration sequence;

[0061] S24. Generate a blade feature matrix based on the wind turbine blade vibration sequence.

[0062] The standard deviation and variance of vibration data can reflect the degree of data dispersion, so the feature matrix is ​​constructed based on the standard deviation. Since wind turbines generally have multiple blades, the variance is calculated to reflect the data characteristics.

[0063] In an embodiment of the present invention, in S22, the specific method for generating the tower characteristic matrix is: calculating the standard deviation between every two adjacent latest vibration data in the wind turbine tower vibration sequence, generating a tower standard deviation sequence, and generating the tower characteristic matrix according to the tower standard deviation sequence.

[0064] In the embodiment of the present invention, the element X in the tower characteristic matrix i The calculation formula is:

[0065]

[0066] Where s i represents the i-th standard deviation in the tower standard deviation sequence, x1 and x2 represent the two adjacent latest vibration data corresponding to the i-th standard deviation in the wind turbine tower vibration sequence, and S0 represents the mean of all standard deviations in the tower standard deviation sequence.

[0067] Tower feature matrix

[0068] In an embodiment of the present invention, in S24, the specific method for generating the blade characteristic matrix is: calculating the standard deviation and variance between every two adjacent latest vibration data in the wind turbine blade vibration sequence, generating a blade standard deviation sequence and a blade variance sequence, and generating a blade characteristic matrix according to the blade standard deviation sequence and the blade variance sequence.

[0069] In the embodiment of the present invention, the element Y in the blade feature matrix j The calculation formula is:

[0070]

[0071] Where, t j represents the jth standard deviation in the blade standard deviation sequence, y1 and y2 represent the two adjacent latest vibration data corresponding to the jth standard deviation in the blade vibration sequence, T1 represents the mean of all standard deviations in the blade standard deviation sequence, and T2 represents the mean of all variances in the blade standard deviation sequence.

[0072] Leaf feature matrix

[0073] In the embodiment of the present invention, in S3, the specific method for determining the tower eigenvalue is: calculating the error coefficient of the latest vibration data of the wind turbine tower, and taking the product of the error coefficient of the latest vibration data of the wind turbine tower and the eigenvalue of the tower characteristic matrix as the tower eigenvalue;

[0074] The calculation formula for the error coefficient U of the latest vibration data of the wind turbine tower is:

[0075]

[0076] Where, P represents the number of all the latest vibration data of the wind turbine tower, u p represents the latest vibration data of the p-th wind turbine tower, X represents the tower characteristic matrix, and E represents the unit matrix;

[0077] In S3, the specific method for determining the blade eigenvalue U is: calculating the error coefficient of the latest vibration data of the blade, and taking the product of the error coefficient of the latest vibration data of the blade and the eigenvalue of the blade characteristic matrix as the blade eigenvalue;

[0078] The calculation formula of the error coefficient V of the latest vibration data of the blade is:

[0079]

[0080] Where Q represents the number of all the latest vibration data of the blade, v q represents the latest vibration data of the qth blade, and Y represents the tower characteristic matrix.

[0081] In the embodiment of the present invention, in S4, the specific method for determining the operating state of the wind turbine generator set is as follows: inputting the tower characteristic value and the blade characteristic value into the state detection model to generate the wind turbine generator set operating characteristic value; if the wind turbine generator set operating characteristic value is greater than the wind turbine generator set operating characteristic threshold, the wind turbine generator set is operating abnormally; otherwise, the wind turbine generator set is operating normally;

[0082] Among them, the expression of the state detection model Z is:

[0083]

[0084] Where X represents the tower characteristic matrix, Y represents the blade characteristic matrix, U represents the error coefficient of the latest vibration data of the wind turbine tower, V represents the error coefficient of the latest vibration data of the blade, u represents the tower characteristic value, and v represents the blade characteristic value.

[0085] The state detection model calculates the tower characteristic matrix, the blade characteristic matrix, the error coefficient of the latest vibration data of the wind turbine tower, the error coefficient of the latest vibration data of the blade, the tower characteristic value and the blade characteristic value to enrich the model parameters, which can more accurately compare the size with the set wind turbine operation characteristic threshold.

[0086] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific descriptions and embodiments. Those skilled in the art can make various other specific variations and combinations based on the technical teachings disclosed in the present invention without departing from the essence of the present invention, and such variations and combinations are still within the scope of protection of the present invention.

Claims

1. A method for monitoring the operating status of a wind turbine generator system, characterized in that: The following steps are involved: S1. Collecting real-time vibration data of the wind turbine tower and blades at different times, and preprocessing the vibration data of the wind turbine tower and the real-time vibration data of the blades to obtain the latest vibration data of the wind turbine tower and blades; S2. Extract the tower feature matrix based on the latest vibration data of the wind turbine tower; Extract blade feature matrix based on the latest vibration data of the blade; S3. Determine the tower characteristic value according to the tower characteristic matrix; Determine the leaf eigenvalue according to the leaf characteristic matrix; S4. Build a state detection model, input the tower characteristic value and the blade characteristic value into the state detection model, and determine the operating state of the wind turbine; In S1, the methods for preprocessing the real-time vibration data of the wind turbine tower and the blades are the same, which are: using a sliding window to evenly divide the real-time vibration data at different times into a number of vibration data groups, extracting the mean of all vibration data in each vibration data group as the vibration group eigenvalue of each vibration data group, and calculating the mean of all vibration group eigenvalues ​​as the vibration eigenvalue, and eliminating the real-time vibration data with a value smaller than the vibration eigenvalue to complete the preprocessing; The S2 includes the following sub-steps: S21. Sort the latest vibration data of the wind turbine tower from small to large to generate a wind turbine tower vibration sequence; S22. Generate a tower characteristic matrix according to the wind turbine tower vibration sequence; S23, sorting the latest vibration data of the blades from small to large to generate a wind turbine blade vibration sequence; S24. Generate a blade feature matrix based on the wind turbine blade vibration sequence; In said S3, the specific method for determining the tower eigenvalue is: calculating the error coefficient of the latest vibration data of the wind turbine tower, and taking the product of the error coefficient of the latest vibration data of the wind turbine tower and the eigenvalue of the tower characteristic matrix as the tower eigenvalue; Among them, the error coefficient of the latest vibration data of the wind turbine tower is U The calculation formula is: Where, P Indicates the number of all the latest vibration data of the wind turbine tower. u p Indicates the p The latest vibration data of wind turbine towers, X represents the tower characteristic matrix, E represents the identity matrix; In said S3, the specific method of determining the blade eigenvalue is: calculating the error coefficient of the latest vibration data of the blade, and taking the product of the error coefficient of the latest vibration data of the blade and the eigenvalue of the blade characteristic matrix as the blade eigenvalue; Among them, the error coefficient of the latest vibration data of the blade is V The calculation formula is: Where, Q Indicates the number of all the latest vibration data of the blade, v q Indicates the q The latest vibration data of each blade, Y represents the tower characteristic matrix; In said S4, the specific method for determining the operating state of the wind turbine generator set is: inputting the tower characteristic value and the blade characteristic value into the state detection model to generate the wind turbine generator set operating characteristic value; if the wind turbine generator set operating characteristic value is greater than the wind turbine generator set operating characteristic threshold, the wind turbine generator set is operating abnormally; otherwise, the wind turbine generator set is operating normally; Among them, the state detection model Z The expression is: Where, X represents the tower characteristic matrix, Y represents the leaf feature matrix, U Indicates the error coefficient of the latest vibration data of the wind turbine tower, V represents the error coefficient of the latest vibration data of the blade, u represents the tower characteristic value, v Represents the leaf eigenvalue.

2. The method for monitoring the operating status of a wind turbine generator system according to claim 1, wherein: The calculation formula of the sliding window size is: Where, a Indicates the length of the sliding window, b Indicates the width of the sliding window, K max Indicates the maximum real-time vibration data, K min Indicates the minimum real-time vibration data, K ave Represents the mean value of all real-time vibration data, n Indicates the number of windows that the sliding window can accommodate, T Indicates the duration of collecting vibration data.

3. The method for monitoring the operating status of a wind turbine generator system according to claim 1, wherein: In said S22, the specific method of generating the tower characteristic matrix is: calculating the standard deviation between every two adjacent latest vibration data in the wind turbine tower vibration sequence, generating a tower standard deviation sequence, and generating the tower characteristic matrix according to the tower standard deviation sequence.

4. The method for monitoring the operating status of a wind turbine generator set according to claim 3, wherein: The elements in the tower characteristic matrix X i The calculation formula is: Where, s i Indicates the first i standard deviations, x 1 and x 2 respectively represent the first i The two adjacent latest vibration data corresponding to the standard deviation, S 0 represents the mean of all standard deviations in the tower standard deviation series.

5. The method for monitoring the operating status of a wind turbine generator system according to claim 1, wherein: In said S24, the specific method of generating the blade characteristic matrix is: calculating the standard deviation and variance between every two adjacent latest vibration data in the wind turbine blade vibration sequence, generating a blade standard deviation sequence and a blade variance sequence, and generating the blade characteristic matrix according to the blade standard deviation sequence and the blade variance sequence.

6. The method for monitoring the operating status of a wind turbine generator system according to claim 5, wherein: The elements in the leaf feature matrix Y j The calculation formula is: Where, t j represents the first j standard deviations, y 1 and y 2 respectively represent the first j The two adjacent latest vibration data corresponding to the standard deviation, T 1 represents the mean of all standard deviations in the leaf standard deviation sequence, T 2 represents the mean of all variances in the leaf standard deviation series.

Images