Wind turbine tower bolt failure state monitoring method, system, medium and device

Abstract

The application discloses a wind turbine tower drum bolt fault state monitoring method, system, medium and equipment, the method is according to the tower drum bolt pretightening force data, after K-means data preprocessing, combined with the pretightening force threshold design formula, the upper and lower limit threshold of pretightening force is output, and the Sen slope estimation method and Mann-Kendall test method are used, the change trend in the pretightening force cycle is obtained, and the significance of the trend is calculated, and then the obtained result is used for long-term and short-term fault monitoring of the wind turbine tower drum bolt;The application eliminates the instrument fault data caused by instrument failure or interference in the pretightening force data, calculates the trend by comparing the relative size relationship between data points, instead of directly using data points, so that the monitoring is more accurate;It can realize multidimensional monitoring of bolts, monitor the bolt state from two dimensions of short term and long term, and timely and accurately master the bolt working condition;High compatibility, can be applied to bolt monitoring of different models of wind turbines.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind turbine tower bolt monitoring, in particular to a wind turbine tower bolt fault state monitoring method, system, medium and equipment. BACKGROUND

[0002] High-strength bolts in wind turbines are one of the main components connecting the tower, hub, blades and other components, playing a role in maintaining structural stability and transmitting force and torque. Due to the fact that wind turbines are operated in harsh conditions all year round, bolts are prone to loosening, fatigue damage and fracture failure. Bolt failure needs to be discovered and addressed in a timely manner, otherwise it will damage related components of the wind turbine, and even cause the wind turbine to be damaged, resulting in huge economic losses.

[0003] Online monitoring of bolt operating state is the main method to determine whether the bolt is failed. Generally, monitoring instruments are used to identify the state of the bolt, mainly including reverse rotation method, displacement method and stress method. The reverse rotation method measures the relative rotation angle of the nut and the screw rod to determine whether the bolt is loose, but it cannot identify the pre-tightening force decay and bolt fracture in non-rotating conditions. The displacement method determines whether the bolt is loose according to the gap size between the flanges, but it cannot identify the failure trend of the bolt in advance. The stress method measures the bolt pre-tightening force according to the relationship between the ultrasonic axial propagation speed and the stress size, which is the main method to directly monitor the bolt pre-tightening force. Limited by the extreme operating environment of the wind turbine, the data collector is easily disturbed, and the collected data has large noise. In addition, the forms of bolt failure include: a) sudden failure under extreme conditions; b) gradual failure of the bolt due to long-term fatigue loss.

[0004] Chinese patent application CN112097706A discloses a short-term monitoring scheme for bolts, while there is no related technology for long-term trend early warning monitoring of bolt pre-tightening force in the prior art, and there is a lack of a scheme design that includes both long-term and short-term bolt failure monitoring in the current research. SUMMARY

[0005] The purpose of the present application is to overcome the shortcomings of the prior art, and a wind turbine tower bolt fault state monitoring method, system, medium and equipment are proposed. According to the tower bolt pre-tightening force data, after K-means data preprocessing, combined with the pre-tightening force threshold design formula, the upper and lower limit thresholds of the pre-tightening force are output, and the Sen slope estimation method and the Mann-Kendall test method are used to obtain the change trend in the pre-tightening force period and calculate the significance of the trend, and then the obtained results are used to monitor the long-term and short-term faults of the wind turbine tower bolts.

[0006] The purpose of the present application is achieved by the following technical solution: a wind turbine tower bolt fault state monitoring method, comprising the following steps:

[0007] S1, obtaining real-time pre-tightening force data of a wind turbine tower bolt;

[0008] S2, pre-processing the pre-tightening force data obtained in step S1 according to a K-means method;

[0009] S3, daily short-term monitoring of the bolt state according to the pre-processed pre-tightening force data, comparing the pre-tightening force data obtained on the day with a preset threshold range, triggering a pre-tightening force value out-of-limit alarm if the pre-tightening force data obtained on the day exceeds the threshold range, and vice versa; and long-term monitoring of the bolt state within a preset period according to the pre-processed pre-tightening force data, calculating a change trend of the pre-tightening force within the preset period and a significance of the change trend;

[0010] S4, judging the long-term state of the bolt pre-tightening force according to the change trend and the significance of the change trend calculated in step S3, triggering a pre-tightening force trend decline alarm if the bolt pre-tightening force trend is in a decline state, indicating that the bolt pre-tightening force gradually fails under long-term fatigue loss, and vice versa.

[0011] Further, the step S1 comprises:

[0012] An automatic detection of the latest data mechanism is set to update the bolt pre-tightening force data in real time every day, an automatic detection of data update function module is set to ensure the bolt state is updated every day, and thus the real-time pre-tightening force data of the wind turbine tower bolt is obtained.

[0013] Further, the step S2 comprises:

[0014] The K-means clustering method is used to divide the pre-tightening force data into two categories, whether the two categories of data exist abnormal fluctuations is judged according to the clustering results, one category of data with abnormal fluctuations is eliminated, and finally normal data is selected;

[0015] The K-means clustering method is used to divide the pre-tightening force data into two categories, including the following steps:

[0016] S2.1, initialization: randomly selecting K data points in the pre-tightening force data obtained in step S1 as initial cluster centers;

[0017] S2.2, dividing data: traversing each data point, dividing it to the nearest cluster center, and finally making each data point marked as belonging to the corresponding cluster;

[0018] S2.3, updating cluster center: updating the center of each cluster, so that the mean value of all points in each cluster is the new cluster center;

[0019] S2.4, iteration convergence: repeat step S2.2 and S2.3 until the clustering result has no significant change, then reach the convergence state, at this time the cluster is the final clustering result.

[0020] Further, the step S3 comprises:

[0021] According to the pretreated preload data, the short-term monitoring of the bolt state is performed daily; according to the diameter and the cross-sectional area of the bolt to be monitored, the threshold range of the bolt preload is calculated, as shown in the following formula:

[0022] F=MxD;

[0023] Wherein, F is the threshold range of the bolt preload, D is the bolt diameter, M is the target stress of the bolt, and the range of M is 450-700 MPa;

[0024] The preload data obtained on the day is compared with the threshold range, if the preload data obtained on the day exceeds the threshold range, the preload value out-of-limit alarm is triggered, and the bolt is checked in time, otherwise no alarm is given.

[0025] Further, the step S3 comprises:

[0026] According to the pretreated preload data, the long-term monitoring of the bolt state is performed within a preset period, and the change trend of the preload within the preset period is calculated, including:

[0027] The change trend of the preload within the preset period is calculated based on the Sen's slope estimation method, the slope between any two preload data points is calculated, the obtained slope is sorted, and the median slope is selected as the final trend estimation value, as shown in the following formula:

[0028]

[0029] Wherein, n is a constant, Q i is a slope sequence with a sample size of n, x j is the preload data point at the jth moment in the preset time sequence X=(x1, x2,..., x n ), and x k is the preload data point at the kth moment in the preset time sequence X=(x1, x2,..., x n ), j and k are both constants and j>k;

[0030] According to the obtained slope sequence Q i , the median of the slope sequence is calculated as the final trend estimation value Q med of the time sequence, as shown in the following formula:

[0031]

[0032] Further, the step S3 comprises:

[0033] The significance of the change trend of the pretightening force in the preset period is calculated, comprising:

[0034] The significance of the final trend estimate value obtained by the Sen's slope estimation is tested based on the Mann-Kendall test method;

[0035] According to the preset time sequence X=(x1, x2,..., x n ), the test statistic S is calculated, as shown in the following formula:

[0036]

[0037] Wherein, the function sgn is a sign function:

[0038]

[0039] The variance calculation formula is as shown in the following formula:

[0040]

[0041] x j is the pretightening force data point of the jth moment in the preset time sequence X=(x1, x2,..., x n ), x i is the pretightening force data point of the ith moment in the preset time sequence X=(x1, x2,..., x n ), j and i are constants and j>i; m is the number of repeated data groups, t i is the number of repeated data in the ith repeated data group;

[0042] The test statistic S is standardized to obtain the statistic Z s , as shown in the following formula:

[0043]

[0044] When Z S >0, it represents that the trend of the bolt pretightening force is upward; when Z s <0, it represents that the trend of the bolt pretightening force is downward;

[0045] When 1.645≤|Z S |<1.960, the variable trend passes the 90% reliability test, and the P value range is (0.05, 0.1]; when 1.960≤|Z s |<2.576, the variable trend passes the 95% reliability test, and the P value range is (0.01, 0.05]; when 2.576≤|ZS | time, variable trends were tested at the 99% confidence level, with P values ranging from P < 0.01.

[0046] A wind turbine tower bolt failure state monitoring system for implementing the wind turbine tower bolt failure state monitoring method described above, comprising:

[0047] A pre-tightening force data acquisition module for acquiring real-time wind turbine tower bolt pre-tightening force data;

[0048] A pre-tightening force data preprocessing module for preprocessing the pre-tightening force data according to the K-means method;

[0049] A short-term monitoring module for short-term monitoring of the bolt state daily based on the preprocessed pre-tightening force data;

[0050] A long-term monitoring module for long-term monitoring of the bolt state within a preset period based on the preprocessed pre-tightening force data;

[0051] An alarm module for alarming the tower bolt failure state based on the monitoring results of the short-term monitoring module and the long-term monitoring module.

[0052] Further, the short-term monitoring module performs the following steps:

[0053] The pre-tightening force data acquired on the day is compared with the preset threshold range, and if the pre-tightening force data acquired on the day exceeds the threshold range, it is considered that the bolt is in a pre-tightening force value overrun state, and the monitoring result is transmitted to the alarm module, otherwise it is considered that the bolt state is normal;

[0054] The long-term monitoring module performs the following steps:

[0055] The long-term monitoring module performs the following steps:

[0056] A non-transitory computer readable medium storing instructions, when executed by a processor, performs the steps of the wind turbine tower bolt failure state monitoring method described above.

[0057] A kind of computing device, comprising processor and the memory for storing the executable program of processor, when the processor executes the program stored in memory, the wind turbine tower bolt failure state monitoring method described above is realized.

[0058] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0059] 1、The present application can realize multidimensional monitoring of bolts, monitor bolt status from two dimensions of short-term and long-term, and accurately grasp bolt working condition in time;

[0060] 2、The present application can eliminate instrument failure data caused by instrument failure or interference in pre-tightening force data, so that monitoring is more accurate;

[0061] 3、The present application uses Sen slope estimation to calculate the change trend of pre-tightening force within a preset time in order to capture long-term loosening condition of bolts, and calculates the trend by comparing the relative size relationship between data points, instead of directly using the value of data points themselves, so that the influence of abnormal values is effectively excluded. At the same time, the Mann-Kendall test method is used to test whether the time series has a significant upward or downward trend;

[0062] 4、The present application has high compatibility and is not limited by the model of wind turbine, and can be applied to bolt monitoring of different models of wind turbine. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 It is a flow chart of the wind turbine tower bolt failure state monitoring method.

[0064] Figure 2 It is a pre-tightening diagram of pre-tightening force exceeding limit in short-term monitoring.

[0065] Figure 3 It is a trend chart of pre-tightening force in long-term monitoring.

[0066] Figure 4 It is a trend chart of pre-tightening force in long-term monitoring. DETAILED DESCRIPTION

[0067] The present application will be further described below in combination with specific embodiments.

[0068] Embodiment 1

[0069] Referring to Figure 1 The wind turbine tower bolt failure state monitoring method provided by the present embodiment includes the following steps:

[0070] S1, obtaining real-time wind turbine tower bolt pre-tightening force data; returning the bolt pre-tightening force data at a regular time, the data time resolution is 1s, and the data specifically includes the pre-tightening force sampling data of 5 tower flange surfaces and 6 bolts on each flange surface.

[0071] S2, pre-processing the pre-tightening force data obtained in step S1 according to the K-means method; checking whether there are invalid values, and the pre-tightening force values that have no long-term fluctuation or have abnormally large data due to interference of the data collector are invalid values that need to be removed;

[0072] The K-means clustering method is used to divide the pre-tightening force data into two categories, whether the two categories of data exist abnormal fluctuation is judged according to the clustering result, one category of data with abnormal fluctuation is removed, and finally the normal data is selected;

[0073] The K-means clustering method is used to divide the pre-tightening force data into two categories, including the following steps:

[0074] S2.1, initialization: randomly selecting K data points in the pre-tightening force data obtained in step S1 as initial cluster centers;

[0075] S2.2, dividing data: traversing each data point, and dividing it to the nearest cluster center, so that each data point is finally marked as belonging to the corresponding cluster;

[0076] S2.3, updating cluster center: updating the center of each cluster, so that the mean value of all points in each cluster is the new cluster center;

[0077] S2.4, iterative convergence: repeating steps S2.2 and S2.3 until the clustering result no longer changes significantly, then reaching the convergence state, and the cluster at this time is the final clustering result.

[0078] S3, according to the pre-processed pre-tightening force data, daily short-term monitoring of the bolt state is performed, the pre-tightening force data obtained on the day is compared with the preset threshold range, if the pre-tightening force data obtained on the day exceeds the threshold range, the pre-tightening force value out-of-limit alarm is triggered, otherwise no alarm; at the same time, according to the pre-processed pre-tightening force data, long-term monitoring of the bolt state is performed in a preset period, the change trend of the pre-tightening force in the preset period and the significance of the change trend are calculated.

[0079] According to the pre-processed pre-tightening force data, daily short-term monitoring of the bolt state is performed; according to the diameter and cross-sectional area of the bolt to be monitored, the threshold range of the bolt pre-tightening force is calculated, as shown in the following formula:

[0080] F=MxD;

[0081] Wherein, F is the threshold range of bolt pretightening force, D is the bolt diameter, M is the target stress of the bolt, and the range of M is 450-700 MPa; the pretightening force data obtained on the same day is compared with the threshold range, if the pretightening force data obtained on the same day exceeds the threshold range, the over-limit warning of the pretightening force value is triggered, and the bolt is checked in time, otherwise no warning is given; see Figure 2 As shown, according to the pretreated pretightening force data, the bolt state is monitored daily for a short period, and the over-limit warning is triggered on March 9, 2023 and March 10, 2023, and the data accuracy checking rate is 100%;

[0082] According to the pretreated pretightening force data, the bolt state is monitored for a long period within 3 months, and the change trend of the pretightening force within the preset period is calculated, including:

[0083] Based on Sen's slope estimation method, the change trend of the pretightening force within the preset period is calculated, the slope between any two pretightening force data points is calculated, the obtained slope is sorted, and the median slope is selected as the final trend estimate value, as shown in the following formula:

[0084]

[0085] Wherein, n is a constant, Q i is the slope sequence with sample size n, x j is the pretightening force data point at j time in the preset time sequence X=(x1, x2,..., x n ), and x k is the pretightening force data point at k time in the preset time sequence X=(x1, x2,..., x n ), j and k are both constants and j>k;

[0086] According to the obtained slope sequence Q i , the median of the slope sequence is calculated as the final trend estimate value Q med of the time sequence, as shown in the following formula:

[0087]

[0088] Further, the step S3 comprises:

[0089] The significance of the change trend of the pretightening force within the preset period is calculated, including:

[0090] Based on the Mann-Kendall test method, the final trend estimate value obtained by Sen's slope estimation is subjected to significance test;

[0091] According to the preset time sequence X=(x1, x2,..., x n ), the test statistic S is calculated, as shown in the following formula:

[0092]

[0093] Among them, the function sgn is a symbolic function:

[0094]

[0095] The variance calculation formula is as follows:

[0096]

[0097] x j For the preset time series X=(x1, x2, ..., x n ) in the j-th moment of the preload data point, x i For the preset time series X=(x1, x2, ..., x n ) is the preload force data point at the i-th moment, j and i are both constants and j>i; m is the number of repeated data groups, t i is the number of repeated data in the i-th repeated data group;

[0098] The test statistic S is standardized to obtain the statistic Z s , as shown below:

[0099]

[0100] When Z S When Z > 0, it means that the bolt preload trend is increasing; when Z s When <0, it means that the bolt preload trend is decreasing;

[0101] When 1.645≤|Z S When |<1.960, the variable trend passed the 90% confidence test, and the P value range was (0.05, 0.1]; when 1.960≤|Z s When |<2.576, the variable trend passed the 95% confidence test, and the P value range was (0.01, 0.05]; when 2.576≤|Z S |When the variable trend passed the 99% confidence test, the P value range was P≤0.01.

[0102] S4. Determine the long-term state of the bolt preload based on the change trend calculated in step S3 and the significance of the change trend. If the bolt preload trend is in a downward state, it is determined that the bolt preload is gradually failing due to long-term fatigue loss, and a preload trend decrease alarm is triggered. Otherwise, no alarm is triggered.

[0103] See also Figures 3-4As shown, the trend of two bolts in April to July 2023 was calculated, and both of the two bolts showed a significant downward trend, with a significance of 99% (P<0.01).

[0104] Embodiment 2

[0105] The wind turbine tower bolt failure state monitoring system provided in this embodiment is used to implement the wind turbine tower bolt failure state monitoring method described in Embodiment 1, and comprises:

[0106] The pre-tightening force data acquisition module is configured to acquire real-time pre-tightening force data of the wind turbine tower bolt.

[0107] The pre-tightening force data preprocessing module is configured to perform preprocessing on the pre-tightening force data according to the K-means method.

[0108] The short-term monitoring module is configured to perform short-term monitoring on the bolt state every day according to the pre-processed pre-tightening force data, and perform the following steps:

[0109] The pre-tightening force data acquired on the day is compared with the preset threshold range, and if the pre-tightening force data acquired on the day exceeds the threshold range, it is considered that the bolt is in a pre-tightening force value overrun state, and the monitoring result is transmitted to the alarm module; otherwise, it is considered that the bolt state is normal.

[0110] The long-term monitoring module is configured to perform long-term monitoring on the bolt state in a preset period according to the pre-processed pre-tightening force data, and perform the following steps:

[0111] The long-term monitoring module is configured to perform long-term monitoring on the bolt state in a preset period according to the pre-processed pre-tightening force data, and perform the following steps:

[0112] The alarm module is configured to alarm the tower bolt failure state according to the monitoring results of the short-term monitoring module and the long-term monitoring module.

[0113] Embodiment 3

[0114] The embodiment discloses a non-transitory computer readable medium storing instructions, when the instructions are executed by a processor, the steps of the wind turbine tower bolt failure state monitoring method according to Embodiment 1 are performed.

[0115] The non-transitory computer readable medium in the embodiment can be a disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), a U disk, a mobile hard disk, and the like.

[0116] Embodiment 4

[0117] The embodiment discloses a computing device, comprising a processor and a memory for storing a processor-executable program, when the processor executes the program stored in the memory, the wind turbine tower bolt fault state monitoring method of embodiment 1 is realized.

[0118] The computing device in the embodiment can be a desktop computer, a notebook computer, a smart phone, a PDA handheld terminal, a tablet computer, a programmable logic controller (PLC), or other terminal devices with processor functions.

[0119] The above-mentioned embodiments are only the preferred embodiments of the present application, and are not intended to limit the scope of the present application, so any changes made according to the shape and principle of the present application should be covered within the protection scope of the present application.

Claims

1. A method for monitoring the fault status of a wind turbine tower bolt, characterized in that: The following steps are involved: S1. Obtaining real-time preload data of wind turbine tower bolts; S2. Preprocessing the preload force data obtained in step S1 by screening according to the K-means method; S3. Based on the pre-processed preload data, the bolt status is monitored daily in the short term. The preload data obtained on that day is compared with the preset threshold range. If the preload data obtained on that day exceeds the threshold range, a preload value over-limit alarm is triggered; otherwise, no alarm is triggered. At the same time, based on the pre-processed preload data, the bolt status is monitored long-term within a preset period. The trend of the preload within the preset period is calculated based on the Sen's slope estimation method, and the final trend estimate obtained by the Sen's slope estimation is tested for significance based on the Mann-Kendall test method. S4. Determine the long-term state of the bolt preload based on the change trend calculated in step S3 and the significance of the change trend. If the bolt preload trend is in a downward state, it is determined that the bolt preload is gradually failing due to long-term fatigue loss, and a preload trend decrease alarm is triggered. Otherwise, no alarm is triggered.

2. A wind turbine tower bolt fault status monitoring method according to claim 1, characterized in that: The step S1 comprises: Set up an automatic detection mechanism for the latest data to update the bolt preload data in real time every day. Set up an automatic detection module for whether the data is updated to ensure that the bolt status is updated daily, thereby obtaining real-time wind turbine tower bolt preload data.

3. A wind turbine tower bolt fault status monitoring method according to claim 1, characterized in that: The step S2 comprises: The K-means clustering method is used to divide the preload data into two categories. Based on the clustering results, it is determined whether the two categories of data have abnormal fluctuations. The data with abnormal fluctuations are eliminated, and finally the normal data are selected. The K-means clustering method is used to classify the preload force data into two categories, including the following steps: S2.1, initialization: randomly select K data points from the preload data obtained in step S1 as the initial cluster centers; S2.

2. Divide the data: traverse each data point and assign it to the nearest cluster center, so that each data point is marked as belonging to the corresponding cluster; S2.

3. Update cluster centers: Update the center of each cluster so that the mean of all points in each cluster is used as the new cluster center. S2.4, iterative convergence: Repeat steps S2.2 and S2.3 until the clustering results no longer change significantly, then the convergence state is reached, and the cluster at this time is the final clustering result.

4. A wind turbine tower bolt fault status monitoring method according to claim 1, characterized in that: The step S3 comprises: Based on the pre-processed preload data, the bolt status is monitored daily for a short period of time. Based on the diameter and cross-sectional area of ​​the bolt to be monitored, the threshold range of the bolt preload is calculated as shown in the following formula: F = M × D; Where F is the threshold range of bolt preload, D is the bolt diameter, and M is the target stress of the bolt, with the range of M being 450-700 MPa. The preload force data obtained on the day is compared with the threshold range. If the preload force data obtained on the day exceeds the threshold range, an over-limit preload force alarm is triggered and the bolts are checked in time. Otherwise, no alarm is triggered.

5. The method for monitoring the fault status of a wind turbine tower bolt according to claim 1, wherein: The step S3 comprises: Based on the pre-processed preload data, the bolt status is monitored over a long period of time within a preset period, and the change trend of the preload within the preset period is calculated, including: Based on Sen's slope estimation method, the trend of preload force changes within a preset period is calculated. The slope between any two preload force data points is calculated, the obtained slopes are sorted, and the median slope is selected as the final trend estimate, as shown in the following formula: Where n is a constant, Q i is the slope sequence with sample size n, x j For the preset time series X=(x1,x2,...,x n ) in the j-th moment of the preload data point, x k For the preset time series X=(x1,x2,...,x n ) is the preload force data point at the kth moment, where j and k are both constants and j>k; According to the slope sequence Q i Calculate the median of the slope series as the final trend estimate Q of the time series med , as shown in the following formula:

6. A wind turbine tower bolt fault status monitoring method according to claim 5, characterized in that: The step S3 comprises: Calculate the significance of the trend of preload force changes within a preset period, including: The significance test of the final trend estimate obtained by Sen's slope estimation was performed based on the Mann-Kendall test method; According to the preset time series X=(x1,x2,...,x n ), calculate the test statistic S as shown below: Among them, the function sgn is a symbolic function: The variance calculation formula is as follows: x j For the preset time series X=(x1,x2,…,x n ) in the j-th moment of the preload data point, x i For the preset time series X=(x1,x2,…,x n ) is the preload data point at the i-th moment, j and i are both constants and j>i; m is the number of repeated data groups, t i is the number of repeated data in the i-th repeated data group; The test statistic S is standardized to obtain the statistic Z s , as shown below: When Z s When Z > 0, it means that the bolt preload is on the rise. s When <0, it means that the bolt preload trend is decreasing.

7. A wind turbine tower bolt fault status monitoring system, characterized in that: A method for monitoring a wind turbine tower bolt fault state according to any one of claims 1 to 6, comprising: Preload data acquisition module, used to obtain real-time preload data of wind turbine tower bolts; Preload data preprocessing module, which pre-processes the preload data based on the K-means method; The short-term monitoring module performs short-term monitoring of the bolt status daily based on the pre-processed preload data. The short-term monitoring module performs the following steps: The preload force data obtained on the day is compared with the preset threshold range. If the preload force data obtained on the day exceeds the threshold range, the bolt is considered to be in a preload force over-limit state and the monitoring result is transmitted to the alarm module. Otherwise, the bolt is considered to be in a normal state. The long-term monitoring module performs long-term monitoring of the bolt status within a preset period based on the pre-processed preload data. The long-term monitoring module performs the following steps: Based on the pre-processed preload data, the bolt status is monitored over a long period of time within a preset period. The trend of preload variation within the preset period is calculated using the Sen's slope estimation method, and a significance test is performed on the final trend estimate obtained from the Sen's slope estimation using the Mann-Kendall test method. If the bolt preload trend is decreasing, it is judged that the bolt preload is gradually failing due to long-term fatigue loss. The bolt is then considered to be in a decreasing preload trend state, and the monitoring results are transmitted to the alarm module. Otherwise, the bolt is considered to be in normal condition. The alarm module issues an alarm for the tower bolt failure status based on the monitoring results of the short-term monitoring module and the long-term monitoring module.

8. A non-transitory computer-readable medium storing instructions, characterized in that: When the instruction is executed by the processor, the steps of the method for monitoring the fault status of a wind turbine tower bolt according to any one of claims 1 to 6 are executed.

9. A computing device comprising a processor and a memory for storing a program executable by the processor, characterized in that: When the processor executes the program stored in the memory, the method for monitoring the fault status of a wind turbine tower bolt according to any one of claims 1 to 6 is implemented.

Images