A method, medium, and system for extracting a first order natural frequency of a wind turbine tower

Abstract

The application discloses a kind of wind turbine tower first natural frequency extraction method, medium and system, the method includes steps: S1, obtains unit operating state and cabin front and rear acceleration time series data;According to unit operating state data, identify the state switching time point T of unit from operating state to shutdown state i ;S2, according to each shutdown time point T i Of unit, respectively extract before shutdown time point continuous data point, and after shutdown time point continuous data point, obtain n time continuous data segment containing N i Data point;Wherein S3, based on time domain feature or / and frequency domain feature identification effective data segment in time continuous data segment;S4, according to effective data segment to cabin front and rear acceleration time series data carries out frequency spectrum analysis, obtains cabin front and rear acceleration frequency spectrum, and extracts tower first natural frequency from cabin front and rear acceleration frequency spectrum.The application has the advantages of high accuracy, strong anti-interference capability etc..

Description

Technical Field

[0001] This invention mainly relates to the field of wind turbine technology, specifically to a method, medium, and system for extracting the first-order natural frequency of a wind turbine tower. Background Technology

[0002] The tower's base connects to the foundation, and its top connects to the nacelle via a yaw bearing. The wind turbine system is installed at the front of the nacelle. The tower is a crucial supporting component in the wind turbine system, bearing the weight of the wind turbine system and the nacelle, and is also subject to highly random aerodynamic loads. Wind turbine tower structures are generally connected using high-strength bolts. When these bolts break, loosen, or cracks appear in the tower structure, the tower's structural stiffness will change, causing a shift in its first-order natural frequency. Similarly, cracks in the tower foundation will also cause a shift in the first-order natural frequency. Broken or loose tower bolts, and cracks in the tower structure and foundation, are significant defects and safety hazards in wind turbines, potentially leading to tower collapse. Therefore, it is necessary to monitor the first-order natural frequency of the wind turbine tower to promptly detect tower and foundation defects and prevent major losses and safety risks.

[0003] Adding vibration monitoring devices to existing vibration sensors on wind turbines is an effective method for monitoring the first-order natural frequency of wind turbine towers. However, with the arrival of the era of grid parity for wind power in my country, cost reduction and efficiency improvement have become the main theme of the wind power industry. Large-scale installation of dedicated vibration monitoring transducers will increase the overall cost of the turbine and wind farm construction. Therefore, it is necessary to propose an accurate and efficient method for extracting the first-order natural frequency of the tower based on existing vibration sensors on wind turbines, so as to identify abnormal deviations in the first-order natural frequency of the tower.

[0004] Wind turbines are equipped with acceleration sensors that monitor the fore-and-aft vibration of the nacelle. The first-order natural frequency of the wind turbine tower can be extracted from the fore-and-aft acceleration signals of the nacelle. However, due to aerodynamic loads and loads caused by turbine start-up, shutdown, pitch control, and yaw under the control system, the frequency components of the fore-and-aft acceleration signals of the nacelle are complex and diverse. Accurately and efficiently extracting the first-order natural frequency of the tower from the complex fore-and-aft acceleration signals of the nacelle is quite difficult. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a method, medium, and system for extracting the first-order natural frequency of wind turbine towers with high accuracy and strong anti-interference capability, in response to the technical problems existing in the prior art.

[0006] To solve the above-mentioned technical problems, the technical solution proposed by this invention is as follows:

[0007] A method for extracting the first-order natural frequency of a wind turbine tower includes the following steps:

[0008] S1. Acquire the unit's operating status and the timing data of acceleration before and after the nacelle; based on the unit's operating status data, identify the state transition point T from the operating state to the shutdown state. i ;

[0009] S2, based on the unit's shutdown time points T i Extract the time before the shutdown time respectively A series of consecutive data points, and after the shutdown time From n consecutive data points, n data points are obtained, each containing N. i A continuous time data segment of data points; among which

[0010] S3. Identify valid data segments in continuous time data segments based on time-domain features and / or frequency-domain features;

[0011] S4. Perform spectrum analysis on the front and rear acceleration time series data of the nacelle based on the effective data segments to obtain the front and rear acceleration spectrum of the nacelle, and then extract the first-order natural frequency of the tower from the front and rear acceleration spectrum of the nacelle.

[0012] Preferably, in step S3, the specific process of identifying valid data segments in a continuous time data segment based on time-domain features is as follows:

[0013] S301. By using envelope detection, extract the positive peak envelope of the fore-and-aft acceleration signal of the cabin from the continuous time data segment, and perform sliding window averaging filtering on the envelope data to obtain the filtered fore-and-aft acceleration envelope of the cabin; wherein the start and end times of the filtered fore-and-aft acceleration envelope of the cabin are t and t, respectively. st t end ;

[0014] S302. Extract the maximum amplitude E of the acceleration envelope from the filtered front and rear acceleration envelopes of the engine compartment for the period from t1 seconds before the shutdown time to t2 seconds after the shutdown time. max , and t st Until t1 seconds before the shutdown time, and until t2 seconds after the shutdown time. end The maximum amplitude U of the envelope between two time periods max ;

[0015] S303. Will E max with U max A comparison is made, and the results are used to determine whether a time-continuous data segment is a valid data segment.

[0016] Preferably, in step S303, if the time-continuous data segment satisfies E max ≥δ×U maxIf the acceleration signal of the cabin in the continuous time data segment is dominated by the energy during the shutdown process, then the continuous time data segment is a valid data segment; δ is a coefficient.

[0017] Preferably, in step S3, the specific process of identifying valid data segments in a continuous time data segment based on frequency domain features is as follows:

[0018] S31. Based on the continuous time data segment, perform spectrum analysis on the front and rear acceleration signals of the cabin to obtain the front and rear acceleration spectrum of the cabin.

[0019] S32, Peak factor Cres and maximum amplitude x of the acceleration spectrum before and after the computer compartment peak The second largest value x sec The frequency f corresponding to the maximum amplitude top ;

[0020] S33, based on the peak factor Cres and maximum amplitude x of the fore-and-aft acceleration spectrum of the cabin. peak The second largest value x sec The frequency f corresponding to the maximum amplitude top This is used to determine whether a continuous time data segment is a valid data segment.

[0021] Preferably, in step S33, if Cres ≥ σ, x sec <η·x peak 0.9f ref ≤f top ≤1.1f ref If one or more of the following are true, then the continuous time data segment is determined to be a valid data segment; where f ref The first-order frequency value of the tower is obtained from the initial identification of the wind turbine unit at the beginning of grid connection; σ and η are constants.

[0022] Preferably, in step S3, preliminary valid data segments in a continuous time data segment are first identified based on time domain features, and then the final valid data segments are obtained by identifying the preliminary valid data segments based on frequency domain features.

[0023] Preferably, in step S2, the number N of the time-continuous data segment data is... i Based on the sampling frequency f of the acceleration signals at the front and rear of the cabin s The selection is based on a combination of the frequency resolution Δf and the overall spectrum.

[0024] Preferably, in step S4, the frequency f corresponding to the maximum amplitude is extracted from the acceleration spectrum of the fore and aft of the cabin. top This is the first-order natural frequency of the tower.

[0025] The present invention also discloses a computer-readable storage medium having a computer program stored thereon, the computer program performing the steps of the method described above when run by a processor.

[0026] The present invention further discloses a system for extracting the first-order natural frequency of a wind turbine tower, comprising a memory and a processor. The memory stores a computer program, which, when run by the processor, executes the steps of any of the methods described above.

[0027] Compared with the prior art, the advantages of the present invention are as follows:

[0028] This invention employs a time-domain method based on the outer envelope of the acceleration signal before and after the nacelle and a frequency-domain method based on the spectral characteristics of the acceleration before and after the nacelle. It comprehensively identifies continuous time data segments dominated by the first-order natural frequency component of the tower, performs spectral analysis on the identified continuous time data segments to obtain the acceleration spectrum before and after the nacelle, calculates the frequency corresponding to the largest amplitude in the spectrum, which is the first-order natural frequency of the tower, and provides early warning for abnormal frequency deviations. This helps maintenance personnel to promptly detect defects in the wind turbine tower and foundation, avoiding tower collapse accidents. The above prediction process has high accuracy and strong anti-interference capability. Attached Figure Description

[0029] Figure 1 This is a flowchart of the extraction method of the present invention in an embodiment.

[0030] Figure 2 This is a block diagram of the extraction system of the present invention in an embodiment. Detailed Implementation

[0031] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0032] like Figure 1 As shown, the method for extracting the first-order natural frequency of a wind turbine tower according to an embodiment of the present invention includes the following steps:

[0033] S1. Acquire the unit's operating status and the timing data of acceleration before and after the nacelle; based on the unit's operating status data, identify the state transition point T from the operating state to the shutdown state. i ;

[0034] S2, based on the unit's shutdown time points T i Extract the time before the shutdown time respectively A series of consecutive data points, and after the shutdown time From n consecutive data points, n data points are obtained, each containing N. i A continuous time data segment of data points; among which

[0035] S3. Identify valid data segments in continuous time data segments based on time-domain features and / or frequency-domain features;

[0036] S4. Perform spectrum analysis on the front and rear acceleration time series data of the nacelle based on the effective data segments to obtain the front and rear acceleration spectrum of the nacelle, and then extract the first-order natural frequency of the tower from the front and rear acceleration spectrum of the nacelle.

[0037] In one specific embodiment, the specific process of identifying valid data segments in a time-continuous data segment based on time-domain features in step S3 is as follows:

[0038] S301. By using envelope detection, extract the positive peak envelope of the fore-and-aft acceleration signal of the cabin from the continuous time data segment, and perform sliding window averaging filtering on the envelope data to obtain the filtered fore-and-aft acceleration envelope of the cabin; wherein the start and end times of the filtered fore-and-aft acceleration envelope of the cabin are t and t, respectively. st t end ;

[0039] S302. Extract the maximum amplitude E of the acceleration envelope from the filtered front and rear acceleration envelopes of the engine compartment for the period from t1 seconds before the shutdown time to t2 seconds after the shutdown time. max , and t st Until t1 seconds before the shutdown time, and until t2 seconds after the shutdown time. end The maximum amplitude U of the envelope between two time periods max ;

[0040] S303. Will E max with U max A comparison is made, and the results are used to determine whether a time-continuous data segment is a valid data segment.

[0041] This invention extracts the envelope of the acceleration signal before and after the nacelle using envelope detection, and identifies the effective data segment dominated by energy during the shutdown process based on the amplitude of the envelope. It can filter out the acceleration signal before and after the nacelle dominated by the first-order natural frequency component of the tower. Based on this acceleration signal before and after the nacelle, the first-order natural frequency of the tower can be extracted more accurately, avoiding the extraction of the first-order natural frequency of the tower from deviating from the actual value due to interference from other frequency components, thus avoiding false alarms.

[0042] In one specific embodiment, the specific process of identifying valid data segments in a time-continuous data segment based on frequency domain features in step S3 is as follows:

[0043] S31. Based on the continuous time data segment, perform spectrum analysis on the front and rear acceleration signals of the cabin to obtain the front and rear acceleration spectrum of the cabin.

[0044] S32, Peak factor Cres and maximum amplitude x of the acceleration spectrum before and after the computer compartmentpeak The second largest value x sec The frequency f corresponding to the maximum amplitude top ;

[0045] S33, based on the peak factor Cres and maximum amplitude x of the fore-and-aft acceleration spectrum of the cabin. peak The second largest value x sec The frequency f corresponding to the maximum amplitude top This is used to determine whether a time-continuous data segment is a valid data segment; the above spectral characteristic indicators can effectively and quickly screen high-quality spectra.

[0046] The aforementioned effective data segment identification method based on frequency domain features can filter out the nacelle front and rear acceleration spectrum dominated by the first-order natural frequency component of the tower. Based on this nacelle front and rear acceleration spectrum, the first-order natural frequency of the tower can be extracted more accurately, avoiding the extraction of the first-order natural frequency of the tower from deviating from the actual value due to interference from other frequency components, thus preventing false alarms from the early warning module.

[0047] Specifically, in step S3, valid data segments in a continuous time data segment can be identified based on either the time-domain feature or frequency-domain feature identification method described above; alternatively, time-domain features can be used first to identify preliminary valid data segments in the continuous time data segment, and then frequency-domain features can be used to identify the preliminary valid data segments to obtain the final valid data segments. By simultaneously using time-domain and frequency-domain identification methods, higher-quality valid data segments can be selected, enabling accurate extraction of the tower's first-order natural frequency and avoiding false alarms.

[0048] This invention employs a time-domain method based on the outer envelope of the acceleration signal before and after the nacelle and a frequency-domain method based on the spectral characteristics of the acceleration before and after the nacelle. It comprehensively identifies continuous time data segments dominated by the first-order natural frequency component of the tower, performs spectral analysis on the identified continuous time data segments to obtain the acceleration spectrum before and after the nacelle, calculates the frequency corresponding to the largest amplitude in the spectrum, which is the first-order natural frequency of the tower, and provides early warning for abnormal frequency deviations. This helps maintenance personnel to promptly detect defects in the wind turbine tower and foundation, avoiding tower collapse accidents. The above prediction process has high accuracy and strong anti-interference capability.

[0049] This invention also discloses a computer-readable storage medium storing a computer program thereon, which, when run by a processor, executes the steps of the method described above. Further, this invention discloses a first-order natural frequency extraction system for wind turbine towers, including a memory and a processor. The memory stores a computer program, which, when run by a processor, executes the steps of the method described above. This invention also discloses a device for extracting the first-order natural frequency of a wind turbine tower, comprising: a continuous data segment extraction module for obtaining the unit's operating status and the acceleration time series data before and after the nacelle; identifying the state transition point from the operating state to the shutdown state based on the unit's operating status data; extracting continuous data points before and after each shutdown time point to obtain continuous time data segments; a time-domain effective data segment identification module and a frequency-domain effective data segment identification module for identifying effective data segments in the continuous time data segments based on time-domain features and / or frequency-domain features; and an automatic extraction and early warning module for the first-order natural frequency of the tower, which performs spectral analysis on the acceleration time series data before and after the nacelle based on the effective data segments to obtain the acceleration spectrum before and after the nacelle, and then extracts the first-order natural frequency of the tower from the acceleration spectrum before and after the nacelle, thereby realizing an early warning of the first-order frequency deviation of the tower. The medium, system, and device of this invention correspond to the above method and also have the advantages described above.

[0050] The present invention will be further described below with reference to a complete specific embodiment:

[0051] (1) Data reading and wind turbine shutdown time point identification

[0052] Read the turbine's operating status and nacelle acceleration timing data from the wind turbine's operating data. Based on the turbine's operating status data, identify the state transition point T from the operating state to the shutdown state. i (T i =1,2,3,...,n);

[0053] (2) Extraction of continuous data slices before and after the wind turbine shutdown time

[0054] For the obtained shutdown time points T of each unit i (T i =1,2,3,...,n), extract the values ​​before the shutdown time point respectively. After consecutive data points and shutdown time From n consecutive data points, we obtain n data points that all contain A continuous time segment of data points.

[0055] Number of data points in a continuous time data segment N i Based on the sampling frequency f of the acceleration signals at the front and rear of the cabins The selection is based on a combination of factors, including the spectral resolution Δf. Generally, for a cabin acceleration signal with a sampling frequency of 1Hz, N... i A value of 1200 is typically used, corresponding to a frequency resolution of 0.0008Hz, which meets the requirements. Of course, it can also be applied to other sampling frequencies of cabin acceleration signals, such as 5Hz, 10Hz, etc., and the number of data points in the data segment can also be 1300, 1400, 1500,... Typically, N... i The smaller the value, the lower the complexity of the wind turbine operating conditions contained in the continuous time data slice, which is more conducive to accurately extracting the first-order natural frequency of the tower.

[0056] (3) Identifying valid data segments based on time-domain features

[0057] By using envelope detection, the positive peak envelope of the fore-and-aft acceleration signal of the cabin is extracted from the continuous time data segment, and the envelope data is subjected to sliding window averaging filtering to obtain the filtered fore-and-aft acceleration envelope of the cabin.

[0058] The start and end times of the filtered front and rear acceleration envelopes of the cabin are t and t, respectively. st t end The maximum amplitude E of the envelope from the filtered front and rear acceleration envelopes of the engine room is extracted during the period from t1 seconds before the shutdown time to t2 seconds after the shutdown time. max , and t st Until t1 seconds before the shutdown time, and until t2 seconds after the shutdown time. end The maximum amplitude U of the envelope between two time periods max .

[0059] If a continuous time data segment satisfies E max ≥δ×U max In this case, the fore-and-aft acceleration signals of the cabin in the continuous time data segment are dominated by the energy during the shutdown process, and the continuous time data segment is considered a valid data segment. Generally, the coefficient δ is usually taken as 120%. Of course, in other embodiments, the coefficient δ can also be taken as 125%, 130%, 135%, ... etc.

[0060] (4) Identifying valid data segments based on frequency domain features

[0061] Based on the effective data segments obtained from time-domain feature recognition, spectral analysis was performed on the front and rear acceleration signals of the cabin to obtain the front and rear acceleration spectrum of the cabin.

[0062] Peak factor Cres and maximum amplitude x of the acceleration spectrum before and after the computer cabin peak The second largest value x sec The frequency f corresponding to the maximum amplitude top ;

[0063]

[0064] In the above formula, x j (j=1,2,3,...,m) represents the amplitude data of the fore-and-aft acceleration spectrum of the cabin.

[0065] If the acceleration spectrum characteristics of the wind turbine nacelle before and after the nacelle satisfy Cres ≥ σ, x sec <η·x peak 0.9f ref ≤f top ≤1.1f ref If one or more of the requirements are met, then a time-continuous data segment is a valid data segment; where f ref The first-order frequency value of the tower is obtained at the beginning of grid connection of the wind turbine; generally, σ is usually taken as 4.5 and η is usually taken as 0.9.

[0066] (5) Identification and automatic extraction of the first-order natural frequency of wind turbine tower

[0067] The effective data segment is obtained by combining time-domain and frequency-domain feature recognition methods, and spectral analysis is performed on the fore-and-aft acceleration signals of the cabin to obtain the fore-and-aft acceleration spectrum. The frequency f corresponding to the largest amplitude in the fore-and-aft acceleration spectrum of the cabin is... top This is the first-order natural frequency of the tower.

[0068] As shown in this disclosure and the claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. The terms "first," "second," and similar terms used in this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, the terms "comprising" or "including" mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, without excluding other elements or objects. The terms "connected" or "linked" are not limited to physical or mechanical connections but may include electrical connections, whether direct or indirect.

[0069] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for extracting the first-order natural frequency of a wind turbine tower, characterized in that, Including the following steps: S1. Acquire unit operating status and nacelle acceleration timing data; based on the unit operating status data, identify the state transition point from operating state to shutdown state. T i ; S2, based on the unit's shutdown times. T i Extract the time before the shutdown time respectively A series of consecutive data points, and after the shutdown time 1 consecutive data point, to obtain n Each contains A continuous time data segment of data points; among which ; S3. Identify valid data segments in continuous time data segments based on time-domain features and / or frequency-domain features; S4. Perform spectrum analysis on the front and rear acceleration time series data of the nacelle based on the effective data segments to obtain the front and rear acceleration spectrum of the nacelle, and then extract the first-order natural frequency of the tower from the front and rear acceleration spectrum of the nacelle. In step S3, the specific process of identifying valid data segments in a continuous time data segment based on frequency domain features is as follows: S31. Based on the continuous time data segment, perform spectrum analysis on the front and rear acceleration signals of the cabin to obtain the front and rear acceleration spectrum of the cabin. S32, Peak factor Cres of the acceleration spectrum before and after the computer compartment, Maximum amplitude Second largest value The frequency corresponding to the maximum amplitude ; S33, based on the peak factor Cres and maximum amplitude of the fore-and-aft acceleration spectrum of the cabin. Second largest value The frequency corresponding to the maximum amplitude To determine whether a time-continuous data segment is a valid data segment; In step S33, if the following conditions are met... , , If one or more of the following are true, then the time-continuous data segment is determined to be a valid data segment; where The first-order frequency value of the tower is obtained from the initial identification of the wind turbine unit at the beginning of grid connection; and It is a constant.

2. The method for extracting the first-order natural frequency of a wind turbine tower according to claim 1, characterized in that, In step S3, the specific process of identifying valid data segments in a continuous time data segment based on time-domain features is as follows: S301. By using envelope detection, extract the positive peak envelope of the fore-and-aft acceleration signal of the cabin from the continuous time data segment, and perform sliding window averaging filtering on the envelope data to obtain the filtered fore-and-aft acceleration envelope of the cabin. The start and end times of the filtered cabin acceleration envelope are respectively t st , t end ; S302. Extract the acceleration envelopes before the shutdown time from the filtered front and rear acceleration envelopes of the engine compartment. t 1 second until the shutdown time t Maximum amplitude of the envelope within a 2-second time period E max ,as well as t st Before the shutdown time t 1 second after the shutdown time t 2 seconds to t end Maximum amplitude of envelope between two time periods U max ; S303, will E max and U max A comparison is made, and the results are used to determine whether a time-continuous data segment is a valid data segment.

3. The method for extracting the first-order natural frequency of a wind turbine tower according to claim 2, characterized in that, In step S303, if the time-continuous data segment satisfies E max U max In the case of a continuous time data segment, the acceleration signal fore and aft of the cabin is dominated by the energy during the shutdown process, and the continuous time data segment is a valid data segment. is a coefficient.

4. The method for extracting the first-order natural frequency of a wind turbine tower according to claim 1, 2, or 3, characterized in that, In step S3, preliminary valid data segments are first identified based on time-domain features, and then the final valid data segments are obtained by identifying the preliminary valid data segments based on frequency-domain features.

5. The method for extracting the first-order natural frequency of a wind turbine tower according to claim 1, 2, or 3, characterized in that, In step S2, the number of points in the continuous time data segment. Based on the sampling frequency of the acceleration signals fore and aft of the cabin and spectral frequency resolution Choose from a comprehensive approach.

6. The method for extracting the first-order natural frequency of a wind turbine tower according to claim 1, 2, or 3, characterized in that, In step S4, the frequency corresponding to the maximum amplitude is extracted from the acceleration spectrum of the fore and aft of the cabin. This is the first-order natural frequency of the tower.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program, when run by a processor, performs the steps of the method as described in any one of claims 1 to 6.

8. A system for extracting the first-order natural frequency of a wind turbine tower, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, The computer program, when run by a processor, performs the steps of the method as described in any one of claims 1 to 6.

Images