A health assessment method for wind turbine transmission chain

Abstract

The present invention relates to a method for assessing the health of a wind turbine transmission chain, comprising the following steps: Step 1. Collecting vibration data; Step 2. Transforming the collected data into the frequency domain, establishing a data matrix W, and calculating a set Q based on the mutual information method; Step 3. Establishing an evaluation matrix E based on the set Q; Step 4. Performing principal component analysis on the matrix E to obtain quantitative evaluation values ​​for n wind turbines, and ranking the wind turbines based on the evaluation values; Step 5. Based on the ranking of the wind turbines, comprehensively considering the number of wind farm maintenance personnel, time, and weather factors, formulating a special inspection plan for the wind turbine transmission chain applicable to the wind farm. This method provides a scientific and reasonable basis for on-site special inspection plans for transmission chains, saving on-site manpower and reducing workload.

Description

Technical Field

[0001] The present invention relates to a method for evaluating the health of a wind turbine drive chain. This method quantitatively evaluates the drive chains of wind turbines in a wind farm, thereby determining a dedicated cycle plan for the drive chains within the wind farm. This quantitative analysis of the health of the wind turbines reduces the workload of on-site maintenance personnel and enables focused monitoring of wind turbines at varying levels. The present invention belongs to the technical field of equipment evaluation. Background Art

[0002] As an important component of wind turbines, the transmission chain can transfer wind energy to the generator in the form of mechanical energy, and then convert it into electrical energy for use in life and industry. Taking a doubly fed asynchronous wind turbine as an example, the main structure of the transmission chain includes bearings of different sizes, main shaft, gearbox, high-speed shaft, coupling and generator. The actual connection method is shown in the attached figure. Figure 1 The health of the transmission chain directly affects the power generation efficiency of the wind turbine and also determines whether the wind turbine can operate safely and stably. Therefore, timely understanding of the health of the wind turbine transmission chain is crucial for on-site operation and maintenance. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for evaluating the health of a wind turbine transmission chain that saves on-site manpower and reduces workload.

[0004] A method for health assessment of a wind turbine transmission chain comprises the following steps:

[0005] Step 1. Collect vibration data;

[0006] Step 2. Transform the collected data into the frequency domain, establish the data matrix W, and calculate the set Q based on the mutual information method;

[0007] Step 3. Establish the evaluation matrix E based on the set Q;

[0008] Step 4. Perform principal component analysis on the evaluation matrix E to obtain the quantitative evaluation values ​​of the n wind turbines, and rank the wind turbines according to the evaluation values;

[0009] Step 5. Based on the order of wind turbines, taking into account the number of wind farm maintenance personnel, time and weather factors, formulate a special inspection plan for the wind turbine drive chain that is suitable for the wind farm.

[0010] Preferably, the wind turbine transmission chain includes a main shaft, a gearbox, a high-speed shaft and a generator; a first bearing is provided on the main shaft, a second bearing is provided between the main shaft and the gearbox, a third bearing is provided between the gearbox and the high-speed shaft, and a fourth bearing is provided between the high-speed shaft and the generator.

[0011] Preferably, the vibration data in step 1 includes vibration data on the first bearing, the second bearing, the third bearing and the fourth bearing.

[0012] Preferably, the vibration data collection method is: using an acceleration sensor to collect vibration data in different directions of bearings of different sizes at the connection of the transmission chain; when the power output of the wind turbine is 20% of the rated power, vibration data is collected, and the acceleration sensor is used to collect vibration data in the x, y and z directions for the first bearing, and the collection method for the second bearing is the same as that for the first bearing. The vibration data for the third and fourth bearings are collected in the x and z directions respectively. For n wind turbines in a certain wind farm, the data collection method for each wind turbine is carried out as described above.

[0013] Preferably, the method for establishing the data matrix W is to transform the collected vibration data of the wind turbines from the time domain to the frequency domain, and calculate the mutual information of the vibration data at the same position; by transforming the vibration data of n wind turbines into the frequency domain, an n×10 data matrix W can be obtained, and the form of W is as follows:

[0014]

[0015] In W, x, y, and z represent the column vectors obtained by transforming the vibration data in the three directions into the frequency domain. The first subscript of each vector is the number of the wind turbine, ranging from [1, n], and the second subscript represents the bearing position of the wind turbine.

[0016] Preferably, the set Q is based on the calculation method of mutual information. Taking the k-th wind turbine as an example, the k-th wind turbine can obtain a set of 10 mutual information values, each set has n-1 mutual information values, and the calculation method is as follows:

[0017]

[0018] Preferably, the expression of the evaluation matrix E is:

[0019]

[0020] e n1 e n2 ……e n10 is the element of the evaluation matrix, where the first subscript represents the wind turbine number and the second subscript represents the evaluation index number;

[0021] The calculation method of each element in the evaluation matrix E is derived from the set of mutual information values ​​in step 2. The calculation formula is as follows:

[0022]

[0023] MaxQk1 means taking the maximum value in the set Qk1, and minQk1 means taking the minimum value in the set Qk1;

[0024] Where λ is a defined correction factor, which is related to the year and the number of wind turbines. The calculation formula of λ is as follows:

[0025]

[0026] In formula (5), year is the number of years the wind turbine has been in operation, and λ is the integer part of the calculated result within [].

[0027] Preferably, step 4 calculates the established evaluation matrix E to obtain quantitative evaluation values ​​of n wind turbines, where the evaluation value range is [0,1]. The wind turbines are sorted according to the evaluation value. The health status of the transmission chain of the wind turbine with a higher ranking is worse than that of other wind turbines, and special attention needs to be paid during the daily inspection of the transmission chain of the wind farm.

[0028] Preferably, the special inspection plan described in step 5 is formulated based on the ranking of the fans, and the fans with the highest ranking are selected. The wind turbines are the focus of inspection. The number of people responsible for maintenance is η. The number of wind turbines inspected weekly according to the order of wind turbines is 3η. The time required to complete the special inspection of the wind farm is T, in days. The calculation formula of T is:

[0029]

[0030] The wind farm shall set the interval for each special inspection of the transmission chain according to the results of the special inspection and weather conditions.

[0031] Beneficial effects of the present invention:

[0032] This paper proposes a method for quantitatively evaluating wind turbine drive trains. By collecting vibration data from bearings at different directions at the connection, this method performs mutual information analysis along the same dimensions for similar wind turbines to generate an evaluation matrix. Principal component analysis (PCA) is then used to reduce the matrix's dimensionality and assign different weights to data columns with different degrees of discreteness, enabling a quantitative evaluation of the wind turbines. The evaluation results can be used to rank the health of wind turbine drive trains across a wind farm, providing a scientific and rational basis for developing specific inspection plans for wind turbine drive trains, saving on-site manpower and reducing workload. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 This is a schematic diagram of the transmission chain structure of a doubly-fed asynchronous wind turbine generator disclosed in the present invention;

[0035] Figure 2 This is a flow chart of a wind turbine transmission chain health assessment method disclosed in the present invention.

[0036] In the accompanying drawings, 1-first bearing, 2-second bearing, 3-third bearing, 4-fourth bearing, 5-main shaft, 6-gearbox, 7-high-speed shaft, 8-generator. DETAILED DESCRIPTION

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0038] like Figure 2 FIG. 1 is a flow chart of a method for health assessment of a wind turbine transmission chain according to the present invention, wherein the steps include:

[0039] Step 1. Use the acceleration sensor to collect vibration data of bearings of different sizes and directions at the connection of the transmission chain. When the power output of the fan is 20% of the rated power, vibration data collection is carried out, such as Figure 2 As shown, for the first bearing, the acceleration sensor is used to collect vibration data in the x, y and z directions respectively. Similarly, the second bearing is collected in the same way as the first bearing. The third and fourth bearings collect vibration data in the x and z directions respectively. For n wind turbines in a certain wind farm, the data collection method is carried out as described above;

[0040] Step 2. Transform the collected wind turbine vibration data from the time domain to the frequency domain and calculate the mutual information of the vibration data at the same location. Transforming the vibration data of n wind turbines into the frequency domain can produce an n×10 data matrix W, which is in the following form:

[0041]

[0042] In W, x, y, and z represent the column vectors obtained by transforming the vibration data in the three directions in the frequency domain. The first subscript of each vector is the number of the wind turbine, ranging from [1, n]. The second subscript represents the bearing position of the wind turbine. The specific situation is shown in the attached figure. Figure 1 As shown;

[0043] Mutual information can reflect the correlation between two event sets, where p(x,y) is the joint probability distribution function of X and Y, and p(x) and p(y) are the marginal probability distribution functions of X and Y, respectively. The simplified formula for mutual information can be expressed as:

[0044]

[0045] The x and y in this formula only represent discrete random variables, not the x and y in W. Based on formula (2), the mutual information value between each wind turbine and the indicators of other wind turbines can be calculated. Taking the k-th wind turbine as an example, the k-th wind turbine can obtain a set of 10 mutual information values, and each set has n-1 mutual information values. The calculation method is as follows:

[0046]

[0047] Step 3. Based on the set of mutual information values ​​in step 2, establish the evaluation matrix E. E is an n×10 matrix, expressed as:

[0048]

[0049] The calculation method of each element in the evaluation matrix E is derived from the set of mutual information values ​​in step 2. The calculation formula is as follows:

[0050]

[0051] Where λ is a defined correction factor, which is related to the year and the number of wind turbines. The calculation formula of λ is as follows:

[0052]

[0053] In formula (6), year is the number of years the wind turbine has been in operation, and λ is the integer part of the calculated result within [].

[0054] Step 4. Evaluate the E matrix based on principal component analysis (PCA), calculate the score of each wind turbine, and sort the wind turbines in descending order according to the numerical value;

[0055] The PCA algorithm is a general algorithm

[0056] PCA algorithm steps:

[0057] (1) Calculate the correlation coefficient matrix R

[0058] Correlation coefficient matrix R=(r ij ) m×m

[0059]

[0060] In the formula, r ij=1, r ij =r ji , r ij is the correlation coefficient between the i-th goal and the j-th indicator;

[0061] (2) Calculate eigenvalues ​​and eigenvectors

[0062] Calculate the eigenvalues ​​of the correlation coefficient matrix R λ1≥λ2≥…≥λ m ≥0, and the corresponding eigenvectors, μ1, μ2, …, μ m .

[0063] Among them, μ j =(μ 1j , μ 2j ,…,μ mj ) T , are m new indicator variables composed of eigenvectors.

[0064]

[0065] In the formula, y1 is the first principal component, y2 is the second principal component, ..., y m is the mth principal component.

[0066] (3) Eigenvalue λ j Variance contribution rate and cumulative contribution rate of (j=1, 2, ..., m)

[0067]

[0068] is the principal component y j The variance contribution rate of .

[0069]

[0070] are the principal components y1, y2, …, y p Cumulative contribution rate;

[0071] When the eigenvalue is less than 1 or the cumulative contribution rate is greater than 85%, the first P indicator variables y1, y2, ..., y p As p principal components, replacing the original m indicator variables, the p principal components can be comprehensively analyzed. The comprehensive formula is as follows:

[0072]

[0073] Among them, bj is the variance contribution rate of the jth principal component, and it can be evaluated based on the comprehensive score.

[0074] Step 5. Develop a special inspection plan for the wind farm based on the ranking of the wind turbines. The wind turbines at the front of the list are the ones that need to be paid special attention to during wind farm operation and maintenance. The wind turbines are the focus of inspection. The number of people responsible for maintenance is η. With a week as the cycle, the number of wind turbines inspected weekly according to the order of wind turbines is 3η. Then the wind farm completes the special inspection.

[0075] The inspection time is T (in days), and the calculation formula of T is:

[0076]

[0077] The wind farm shall set the interval for each special inspection of the transmission chain according to the results of the special inspection and weather conditions.

[0078] The above content explains the principles and implementation methods of the present invention. The above description is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A method for health assessment of a wind turbine transmission chain, characterized in that: The following steps are involved: Step 1. Collect vibration data; Step 2. Transform the collected data into the frequency domain and create a data matrix W , based on the mutual information method to calculate the set Q ; Step 3. Based on the collection Q Establish an evaluation matrix E ; Step 4. Target the evaluation matrix E Perform principal component analysis and obtain n Quantitative evaluation value of typhoon turbines, and ranking of wind turbines according to the evaluation value; Step 5. Based on the order of wind turbines, taking into account the number of wind farm maintenance personnel, time, and weather factors, develop a dedicated inspection plan for the wind turbine drive chains that is appropriate for the wind farm. The wind turbine transmission chain includes a main shaft, a gearbox, a high-speed shaft and a generator; a first bearing is provided on the main shaft, a second bearing is provided between the main shaft and the gearbox, a third bearing is provided between the gearbox and the high-speed shaft, and a fourth bearing is provided between the high-speed shaft and the generator; The vibration data in step 1 includes vibration data on the first bearing, the second bearing, the third bearing, and the fourth bearing; The vibration data collection method is as follows: using an acceleration sensor to collect vibration data of bearings of different sizes and directions at the connection of the transmission chain; when the power output of the fan is 20% of the rated power, vibration data collection is performed, and the acceleration sensor is used to collect the vibration data of the first bearing. x 、 y and z The vibration data of the second bearing and the first bearing are collected in the same way, and the third and fourth bearings are collected separately. x and z Vibration data in the direction of a wind field n The data collection method for each wind turbine is carried out as described above.

2. A wind turbine transmission chain health assessment method according to claim 1, characterized in that: The data matrix W The establishment method is to transform the collected vibration data of the fan from the time domain to the frequency domain, and calculate the mutual information of the vibration data at the same position; n The data matrix obtained by converting the vibration data of the typhoon to the frequency domain is W , W The form is as follows: （1） W middle, x 、 y and z They represent the column vectors obtained by transforming the vibration data in the three directions in the frequency domain. The first subscript of each vector is the number of the fan, and the range is [1, n ], the second subscript represents the bearing position of the wind turbine.

3. A wind turbine transmission chain health assessment method according to claim 2, characterized in that: Step 4: Establish an evaluation matrix E Calculate and get n The quantitative evaluation value of the typhoon turbine is in the range of [0,1]. The wind turbines are ranked according to the evaluation value. The health status of the transmission chain of the wind turbine with a higher ranking is worse than that of other wind turbines, and it is necessary to pay special attention to it during the daily inspection of the transmission chain of the wind farm.

4. A wind turbine transmission chain health assessment method according to claim 3, characterized in that: The special inspection plan described in step 5 is formulated based on the ranking of the fans, and the ones with the highest ranking are selected. The wind turbine is the focus of the inspection. The number of people responsible for maintenance is η. The number of wind turbines inspected weekly according to the order of wind turbines is 3η. The time to complete the special inspection of the wind farm is T , the unit is day, T The calculation formula is: （7） The wind farm shall set the interval for each special inspection of the transmission chain according to the results of the special inspection and weather conditions.

Images