Wind turbine generator wind wheel imbalance fault diagnosis method and related device

Abstract

The invention discloses a wind turbine generator wind wheel imbalance fault diagnosis method and a related device, and the method comprises the steps: obtaining wind wheel rotation frequency amplitudes in a preset period under the condition that a wind turbine generator is in a grid-connected power generation state, and sequentially putting the wind wheel rotation frequency amplitudes into a wind wheel imbalance change trend detection array according to a time sequence, when a preset period is larger than a wind wheel imbalance change trend long-period threshold value, under the condition that a wind wheel imbalance change trend abnormity suspicion condition is met, wind wheel imbalance change trend abnormity long-period fault diagnosis is carried out through a trend detection method; and when the preset period is smaller than a short-period threshold value of the wind wheel imbalance variation trend, drawing a wind wheel rotation frequency amplitude cumulative deviation curve according to the cumulative deviation of each wind wheel rotation frequency amplitude relative to a historical wind wheel rotation frequency target amplitude, so as to carry out abnormal short-time fault diagnosis on the wind wheel imbalance variation trend. The method not only can capture the trend enhancement characteristics with a long period, but also can keep sensitive to the abnormity of instantaneous sudden increase, and effectively avoids late report and missing report of early faults.

Description

Technical Field

[0001] The present invention relates to the field of wind power technology, and more particularly to a method for diagnosing an imbalance fault of a wind turbine rotor and a related device. Background Art

[0002] As a clean energy technology, wind power generation plays a vital role in the global energy transition. As core equipment, wind turbines ("wind turbines") have a significant impact on the power generation efficiency and economic benefits of wind farms due to their operational reliability and stability. During wind turbine operation, rotor imbalance is a common fault, posing a serious threat to the safety and lifespan of the equipment. When rotor imbalance develops to the mid- to late-stage, vibration intensifies on key components such as the blades, tower, yaw mechanism, and main shaft. This increases fatigue loads, shortening the operational lifespan and, in severe cases, can even lead to blade breakage or tower collapse.

[0003] Currently, the traditional method for diagnosing wind turbine rotor imbalance faults relies on static threshold diagnosis based on analysis of nacelle vibration characteristics. This method relies particularly on monitoring the rotor rotational frequency (1P) amplitude within the nacelle vibration signal. However, wind turbine blades are susceptible to structural damage such as bulging, debonding, and cracking under long-term, complex loads. Because the magnitude of such damage is minimal in the early stages of degradation, and the rotor rotational frequency amplitude fluctuates due to operating conditions such as wind speed variations, yaw, and pitch control strategies, effective identification is difficult using standard vibration alarms or fixed fault thresholds.

[0004] The traditional diagnostic method based on static threshold is only applicable to the stage when the wind rotor imbalance fault develops to the middle and late stages, causing the nacelle vibration signal to be significantly enhanced. Therefore, there is an inevitable risk of missed reporting or delayed reporting of early wind rotor imbalance faults. Summary of the Invention

[0005] In view of this, the present invention discloses a wind turbine rotor imbalance fault diagnosis method and related devices, so as to effectively avoid late reporting and missed reporting of early faults and significantly improve the early perception capability of wind turbine rotor imbalance faults.

[0006] A method for diagnosing an imbalance fault of a wind turbine rotor, comprising:

[0007] When the wind turbine is in a grid-connected power generation state, obtaining the wind rotor frequency amplitude within a preset period, and sequentially placing each of the wind rotor frequency amplitudes into a wind rotor imbalance change trend test array according to a time sequence;

[0008] Determining that the preset period is greater than a long-period threshold value of the wind rotor imbalance change trend or less than a short-period threshold value of the wind rotor imbalance change trend;

[0009] If the preset period is greater than the long-period threshold value of the wind rotor imbalance change trend, and the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend detection array meets the suspected abnormal wind rotor imbalance change trend condition, a trend detection method is used to perform an abnormally long-period fault diagnosis of the wind rotor imbalance change trend on the wind rotor imbalance change trend detection array;

[0010] If the preset period is less than the short-period threshold value of the wind rotor imbalance change trend, the cumulative deviation of each wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude is calculated in turn according to the time sequence of each wind rotor frequency amplitude in the wind rotor imbalance change trend inspection array, and a wind rotor frequency amplitude cumulative deviation curve is drawn. The abnormal short-term fault diagnosis of the wind rotor imbalance change trend is performed according to the change trend of the wind rotor frequency amplitude cumulative deviation curve, wherein each cumulative deviation is: the sum of the deviation of the current time sequence wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude and the cumulative deviation of the previous time sequence wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude.

[0011] Optionally, obtaining the wind rotor rotation frequency amplitude within a preset period includes:

[0012] When the wind turbine operating state characteristics in any sliding window within the preset period meet the following judgment conditions, the wind rotor rotation frequency amplitude corresponding to the sliding window is obtained, and the judgment conditions include:

[0013] The standard deviation of the wind rotor speed is less than the wind rotor speed stability judgment threshold;

[0014] The absolute value of the difference between the wind rotor rotation frequency and the tower first-order frequency is greater than a frequency difference threshold, where the frequency difference threshold is determined based on a frequency deviation adjustment coefficient;

[0015] The turbulence intensity does not exceed the turbulence intensity threshold;

[0016] The wind turbine operating state characteristics include: the standard deviation of the wind rotor speed, the wind rotor rotation frequency and the turbulence intensity.

[0017] Optionally, the process of determining the wind turbine operating status characteristics in any sliding window includes:

[0018] Determine the standard deviation of the wind rotor speed and the average wind rotor speed within a preset time period based on the wind rotor speed of the wind turbine generator set;

[0019] Obtaining a calculated value of the wind rotor rotation frequency based on the average wind rotor rotation speed;

[0020] Finding a rotor frequency identification frequency interval within a preset range of the rotor frequency calculation value, finding a maximum amplitude within a range corresponding to the rotor frequency identification frequency interval in a nacelle vibration signal spectrum as the final identified rotor frequency amplitude, and using a frequency value corresponding to the maximum amplitude in the nacelle vibration signal spectrum as the final identified rotor frequency, wherein the nacelle vibration signal spectrum is obtained by performing a fast Fourier transform on the nacelle vibration signal;

[0021] The turbulence intensity within the preset time period is determined based on the wind speed of the wind turbine.

[0022] Optionally, the process of determining whether the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend inspection array meets the abnormal suspicion condition of the wind rotor imbalance change trend includes:

[0023] Determine the number of points in the wind rotor imbalance change trend inspection array where the wind rotor rotation frequency amplitude is greater than a wind rotor rotation frequency amplitude change trend abnormality determination reference value;

[0024] Calculating the ratio of the number of points to the total number of points of the wind wheel rotation frequency amplitude;

[0025] When the ratio exceeds a preset point proportion threshold, it is determined that the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend inspection array meets the abnormal suspicion condition of the wind rotor imbalance change trend.

[0026] Optionally, the process of determining the abnormal reference value for determining the change trend of the wind rotor frequency amplitude includes:

[0027] After the wind turbine generator set is connected to the grid for the first time for power generation, historical wind rotor rotation frequency amplitude values ​​within a preset statistical period before the current moment are obtained, and a statistical sample set of historical wind rotor rotation frequency amplitude values ​​is formed;

[0028] Performing kernel density estimation on the historical wind rotor rotation frequency amplitude statistical sample set to obtain a wind rotor rotation frequency amplitude probability density function;

[0029] Constructing a cumulative distribution function of the wind rotor rotation frequency amplitude probability density function by integrating and accumulating;

[0030] The target wind rotor rotation frequency amplitude corresponding to the preset quantile position is found from the cumulative distribution function, and the target wind rotor rotation frequency amplitude is used as a reference value for determining abnormality in the wind rotor rotation frequency amplitude change trend.

[0031] Optionally, the adopting a trend inspection method to perform fault diagnosis of an abnormally long-duration wind rotor imbalance change trend on the wind rotor imbalance change trend inspection array includes:

[0032] The wind rotor rotation frequency amplitude time series data in the inspection array is used to calculate the wind rotor rotation frequency amplitude statistics according to the wind rotor imbalance change trend;

[0033] Calculating the statistical variance of the wind rotor rotation frequency amplitude according to the size of the wind rotor imbalance change trend test array;

[0034] Calculating a standardized test statistic according to the wind rotor rotation frequency amplitude statistic and the variance of the wind rotor rotation frequency amplitude statistic;

[0035] A one-tailed hypothesis test is performed on the standardized test statistic. If the standardized test statistic is not less than the Z quantile corresponding to the preset percentile of the standard normal distribution, and the significance of the wind rotor imbalance change trend is less than the preset significance level, it is determined that the risk unit has an abnormally long-term wind rotor imbalance change trend fault; otherwise, it is determined that the risk unit has not an abnormally long-term wind rotor imbalance change trend fault, wherein the preset percentile is determined based on the preset significance level, and the Z quantile refers to the value corresponding to the cumulative probability equal to the preset percentile in the standard normal distribution, reflecting the degree of deviation from the sample mean.

[0036] Optionally, the process of determining the historical wind rotor rotation frequency target amplitude includes:

[0037] After the wind turbine generator set is connected to the grid for the first time for power generation, historical wind rotor rotation frequency amplitude values ​​within a preset statistical period before the current moment are obtained, and a statistical sample set of historical wind rotor rotation frequency amplitude values ​​is formed;

[0038] Calculating the average value of all historical wind rotor rotation frequency amplitudes in the historical wind rotor rotation frequency amplitude statistical sample set;

[0039] The average value is determined as the historical wind wheel rotation frequency target amplitude.

[0040] Optionally, the performing of short-term fault diagnosis of abnormal short-term wind rotor imbalance change trend according to the change trend of the wind rotor rotation frequency amplitude cumulative deviation curve includes:

[0041] Comparing each cumulative deviation data point in the cumulative deviation curve of the wind rotor rotation frequency amplitude with the wind rotor imbalance short-term trend abnormality judgment threshold in sequence according to time sequence;

[0042] If there are a continuous preset number of cumulative deviation data points that all exceed the wind rotor imbalance short-term trend abnormality judgment threshold, it is determined that the risk unit has an abnormal short-term fault of wind rotor imbalance change trend.

[0043] A wind turbine rotor imbalance fault diagnosis device, comprising:

[0044] An array acquisition unit is used to acquire the wind rotor frequency amplitude within a preset period when the wind turbine is in a grid-connected power generation state, and to sequentially place each of the wind rotor frequency amplitude values ​​into a wind rotor imbalance change trend detection array according to a time sequence;

[0045] A judging unit, configured to judge whether the preset period is greater than a long-period threshold value of the wind rotor imbalance change trend or less than a short-period threshold value of the wind rotor imbalance change trend;

[0046] a long-term fault diagnosis unit, configured to, if the preset period is greater than the long-term threshold value of the wind rotor imbalance change trend and the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend detection array meets a suspected condition of abnormal wind rotor imbalance change trend, perform a long-term fault diagnosis of abnormal wind rotor imbalance change trend on the wind rotor imbalance change trend detection array using a trend detection method;

[0047] A short-term fault diagnosis unit is used to calculate the cumulative deviation of each wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude according to the time sequence of each wind wheel frequency amplitude in the wind wheel imbalance change trend inspection array if the preset period is less than the short-term threshold of the wind wheel imbalance change trend, and draw a wind wheel frequency amplitude cumulative deviation curve, and perform abnormal short-term fault diagnosis of the wind wheel imbalance change trend according to the change trend of the wind wheel frequency amplitude cumulative deviation curve, wherein each of the cumulative deviations is: the sum of the deviation of the current time series wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude and the cumulative deviation of the previous time series wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude.

[0048] A computer storage medium stores at least one instruction, and when the at least one instruction is executed by a processor, any method for diagnosing an imbalance fault of a wind turbine rotor is implemented.

[0049] An electronic device, comprising: a memory and a processor;

[0050] The memory is used to store at least one instruction;

[0051] The processor is configured to execute the at least one instruction to implement any one of the wind turbine rotor imbalance fault diagnosis methods.

[0052] From the above technical solution, it can be seen that the present invention discloses a method and related device for diagnosing wind turbine rotor imbalance faults. When the wind turbine is in a grid-connected power generation state, the wind rotor frequency amplitude within a preset period is obtained, and each wind rotor frequency amplitude is sequentially placed in a wind rotor imbalance change trend inspection array according to a time sequence. When the preset period is greater than the long-period threshold value of the wind rotor imbalance change trend, and the wind rotor frequency amplitude in the wind rotor imbalance change trend inspection array meets the abnormal suspicion condition of the wind rotor imbalance change trend, the wind rotor imbalance change trend inspection array is subjected to an abnormal long-term fault diagnosis of the wind rotor imbalance change trend using a trend inspection method; when the preset period is less than the short-period threshold value of the wind rotor imbalance change trend, the cumulative deviation of each wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude is calculated in sequence according to the time sequence of each wind rotor frequency amplitude in the wind rotor imbalance change trend inspection array, and a wind rotor frequency amplitude cumulative deviation curve is drawn. According to the change trend of the wind rotor frequency amplitude cumulative deviation curve, a wind rotor imbalance change trend abnormal short-term fault diagnosis is performed. This invention constructs a rotor imbalance trend detection array, simultaneously considering two abnormal evolution patterns of rotor imbalance (long-term, slow, cumulative changes and short-term, rapid, sudden changes). It then employs a dual-path parallel fusion strategy for pattern analysis, enabling long-term and short-term fault diagnosis for rotor imbalance trend anomalies. This strategy captures long-term trend enhancement characteristics while remaining sensitive to transient anomalies, effectively avoiding late and missed early fault detection, significantly improving the ability to proactively detect rotor imbalance faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] 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 merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the disclosed drawings without any creative work.

[0054] Figure 1 This is a flow chart of a method for diagnosing wind turbine rotor imbalance faults disclosed in an embodiment of the present invention;

[0055] Figure 2 A schematic diagram of an example of determining an abnormally long-term trend of wind rotor imbalance change disclosed in an embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of an example of calculating a reference value for determining abnormality in a wind rotor frequency amplitude change trend and a historical wind rotor frequency target amplitude, as disclosed in an embodiment of the present invention;

[0057] Figure 4 This is a schematic diagram of an example of determining an abnormal short-term trend of wind rotor imbalance change disclosed in an embodiment of the present invention;

[0058] Figure 5 This is a schematic structural diagram of a wind turbine rotor imbalance fault diagnosis device disclosed in an embodiment of the present invention;

[0059] Figure 6 The figure is a schematic structural diagram of an electronic device disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] 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.

[0061] Existing methods for monitoring rotor imbalance generally employ a static threshold strategy, triggering an alarm only when the rotor frequency amplitude exceeds a set threshold. This approach only responds when the imbalance has reached its mid-to-late stages and vibration has significantly increased, often missing the optimal maintenance window. This leads to further damage to the blades, increasing maintenance costs and the risk of downtime. The fundamental reason for this is that wind turbine operating conditions are complex, and minor changes caused by early imbalance are easily masked by factors such as wind speed fluctuations and pitch control. This makes static threshold strategies ineffective in identifying trend changes.

[0062] Based on the concept of trend analysis, this paper discloses a method for diagnosing wind turbine rotor imbalance faults. By constructing a rotor imbalance trend detection array and simultaneously considering two abnormal evolution modes of rotor imbalance (long-term slow cumulative changes and short-term rapid mutations), a dual-path parallel fusion strategy is employed for pattern analysis, enabling both long-term and short-term fault diagnosis of rotor imbalance trend anomalies. This strategy captures the long-term trend enhancement characteristics while remaining sensitive to transient anomalies, effectively avoiding late and missed detection of early faults and significantly improving the ability to detect rotor imbalance faults in advance.

[0063] See also Figure 1 , a flow chart of a method for diagnosing wind turbine rotor imbalance faults disclosed in an embodiment of the present invention, the method comprising:

[0064] Step S101: When the wind turbine is in a grid-connected power generation state, obtain the wind rotor frequency amplitude within a preset period, and put each of the wind rotor frequency amplitudes into a wind rotor imbalance change trend detection array in a time sequence.

[0065] When the wind turbine operating status characteristics under any sliding window within a preset period meet the following judgment conditions, the wind rotor frequency amplitude corresponding to the sliding window is obtained. The wind turbine operating status characteristics include: wind rotor speed standard deviation, wind rotor frequency and turbulence intensity.

[0066] The judgment conditions include:

[0067] (1) The standard deviation of the wind rotor speed is less than the wind rotor speed stability judgment threshold.

[0068] The value of the wind wheel speed stability judgment threshold is determined according to actual needs, for example, it is set to 0.4 revolutions per minute.

[0069] (2) The absolute value of the difference between the wind rotor frequency and the first-order frequency of the tower is greater than the frequency difference threshold.

[0070] The frequency difference threshold is determined based on the frequency deviation adjustment coefficient. For example, the frequency difference threshold = tower first-order frequency × frequency deviation adjustment coefficient. The value of the frequency deviation adjustment coefficient is determined based on actual needs, for example, the frequency deviation adjustment coefficient is 10%.

[0071] (3) The turbulence intensity does not exceed the turbulence intensity threshold.

[0072] The turbulence intensity threshold may be a Class A turbulence intensity at an average wind speed over a preset time period in the IEC (International Electrotechnical Commission) standard. The Class A turbulence intensity is also the turbulence intensity in the IEC turbulence intensity curve.

[0073] In practical applications, the sliding window algorithm can be used to determine the operating status characteristics of the wind turbine based on the actual operating status parameters of the wind turbine. The actual operating status parameters of the wind turbine are provided in real time by the wind turbine main control system, including at least: rotor speed, wind speed and nacelle vibration signal

[0074] The wind rotor speed includes all wind rotor speeds collected within a preset time period before the current moment.

[0075] Likewise, the wind speed includes all wind speeds collected within a preset time period before the current moment.

[0076] The cabin vibration signal includes the cabin front-to-back vibration signal and the cabin left-to-right vibration signal.

[0077] In this embodiment, the value of the preset time period is determined according to actual needs, for example, 10 minutes.

[0078] It should be noted that, in this embodiment, the preset period includes multiple sliding windows, and each sliding window is also the preset time period in this embodiment.

[0079] Step S102: Determine whether the preset period is greater than the long-period threshold of the wind rotor imbalance change trend or less than the short-period threshold of the wind rotor imbalance change trend. If the preset period is greater than the long-period threshold of the wind rotor imbalance change trend, execute step S103; if the preset period is less than the short-period threshold of the wind rotor imbalance change trend, execute step S104.

[0080] The inventors' research has revealed that damage to wind turbine blade tips, internal bulging, and trailing edge cracking typically takes days, weeks, or even longer to accumulate from the initial onset to significant rotor imbalance. However, in some exceptional cases, once cracks reach a critical size or the adhesive layer suddenly fails, the vibration level can surge and rapidly trigger severe rotor imbalance. Consequently, two abnormal patterns of rotor imbalance change trends exist: short, rapid changes and longer, slower changes.

[0081] Based on this, the present invention sets a long-term and short-term threshold for the rotor imbalance trend. When the preset period is greater than the long-term threshold, fault diagnosis for an abnormally long rotor imbalance trend is performed; when the preset period is less than the short-term threshold, fault diagnosis for an abnormally short rotor imbalance trend is performed. This present invention simultaneously considers both abnormal modes through a dual-path parallel fusion approach, ensuring early warning protection in scenarios where rotor imbalance is rapidly deteriorating.

[0082] Among them, the values ​​of the long-term threshold value of the wind rotor imbalance change trend and the short-term threshold value of the wind rotor imbalance change trend are determined according to actual needs. For example, the long-term threshold value of the wind rotor imbalance change trend is 30 days, and the short-term threshold value of the wind rotor imbalance change trend is 1 day.

[0083] Step S103: When the rotor rotation frequency amplitude in the rotor imbalance change trend detection array meets the suspicion condition of abnormal rotor imbalance change trend, a trend detection method is used to perform abnormal long-term rotor imbalance change trend fault diagnosis on the rotor imbalance change trend detection array.

[0084] This application mainly includes two steps to diagnose the abnormal long-term imbalance trend of the wind rotor:

[0085] Step 1: Detect the unbalance change trend of the wind rotor and check whether the wind rotor rotation frequency amplitude in the array meets the abnormal suspicion condition of the unbalance change trend of the wind rotor.

[0086] The second step will only be executed if the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend test array meets the abnormal suspicion condition of the wind rotor imbalance change trend. Otherwise, the observation state will be maintained within the judgment period when the current wind rotor imbalance change trend is abnormally long.

[0087] The second step: fault diagnosis of abnormally long-term unbalanced change trend of the wind rotor, specifically: adopting the trend test method to perform fault diagnosis of abnormally long-term unbalanced change trend of the wind rotor on the wind rotor unbalanced change trend test array.

[0088] In practical applications, trend test methods may include the Mann-Kendall trend test, the Spearman rank correlation test, and the linear fitting trend test.

[0089] The Mann-Kendall trend test method is a climate diagnosis and prediction technology. The Mann-Kendall trend test method can be used to determine whether there is a mutation in the sequence. If so, the time when the mutation occurred can be determined.

[0090] When the present application determines that the wind wheel rotation frequency amplitude in the wind wheel imbalance change trend inspection array meets the abnormal suspicion condition of the wind wheel imbalance change trend, and determines that the wind wheel imbalance change trend is abnormally long-term fault, it is finally determined that the wind wheel imbalance change trend is abnormally long-term fault.

[0091] Step S104: check the timing of each wind rotor frequency amplitude in the array according to the wind rotor imbalance change trend, calculate the cumulative deviation of each wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude in turn, and draw a wind rotor frequency amplitude cumulative deviation curve; and perform abnormal short-term fault diagnosis of wind rotor imbalance change trend according to the change trend of the wind rotor frequency amplitude cumulative deviation curve.

[0092] Each of the cumulative deviations is the sum of the deviation of the current time series wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude and the cumulative deviation of the previous time series wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude.

[0093] The process of drawing the cumulative deviation curve of the wind rotor frequency amplitude in this embodiment is as follows:

[0094] The time series data of the wind wheel rotation frequency amplitude in the wind wheel imbalance change trend test array are set to be: , ,…, , m is the size of the wind wheel imbalance change trend test array, and then perform the following steps:

[0095] (1);

[0096] (2);

[0097] Where, It indicates the initial value of the cumulative deviation of the wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude, which is generally 0.

[0098] It represents the rotational frequency amplitude of the i-th wind rotor in the short-term trend test array of wind rotor imbalance, and the value range of i is 1~m.

[0099] Indicates the historical wind rotor rotation frequency target amplitude.

[0100] It represents the cumulative deviation of the i-th wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude, and the value range of i is 1~m.

[0101] It represents the cumulative deviation of the i-1th wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude, and the value range of i is 1~m.

[0102] Calculated Finally, these accumulated deviations are connected in time series to obtain the cumulative deviation curve of the wind rotor frequency amplitude within the determination period of abnormally short wind rotor imbalance change trend. Based on the changing trend of the cumulative deviation curve of the wind rotor frequency amplitude, the abnormally short wind rotor imbalance change trend fault diagnosis can be performed.

[0103] When the preset period is less than the short-period threshold value of the wind rotor imbalance change trend, the wind rotor imbalance change trend detection array may be defined as: a wind rotor imbalance short-time trend detection array.

[0104] In summary, the present invention discloses a method for diagnosing an imbalance fault of a wind turbine rotor. When the wind turbine rotor is in a grid-connected power generation state, the rotor frequency amplitude within a preset period is obtained, and each rotor frequency amplitude is sequentially placed into a rotor imbalance change trend detection array according to a time sequence. When the preset period is greater than a long-period threshold value of the rotor imbalance change trend, and the rotor frequency amplitude in the rotor imbalance change trend detection array meets a suspected condition of an abnormal rotor imbalance change trend, a trend detection method is used to perform an abnormal long-term rotor imbalance change trend fault diagnosis on the rotor imbalance change trend detection array. When the preset period is less than a short-period threshold value of the rotor imbalance change trend, the cumulative deviation of each rotor frequency amplitude relative to a historical rotor frequency target amplitude is calculated in sequence according to the time sequence of each rotor frequency amplitude in the rotor imbalance change trend detection array, and a rotor frequency amplitude cumulative deviation curve is drawn. According to the change trend of the rotor frequency amplitude cumulative deviation curve, an abnormal short-term rotor imbalance change trend fault diagnosis is performed. This invention constructs a rotor imbalance trend detection array, simultaneously considering two abnormal evolution patterns of rotor imbalance (long-term, slow, cumulative changes and short-term, rapid, sudden changes). It then employs a dual-path parallel fusion strategy for pattern analysis, enabling long-term and short-term fault diagnosis for rotor imbalance trend anomalies. This strategy captures long-term trend enhancement characteristics while remaining sensitive to transient anomalies, effectively avoiding late and missed early fault detection, significantly improving the ability to proactively detect rotor imbalance faults.

[0105] It should be noted that the present invention requires two types of input signals when diagnosing wind turbine rotor imbalance faults, namely:

[0106] (1) Main control signal: including wind speed, rotor speed, cabin vibration signal (front and back and left and right directions) and other actual operating status parameters of the wind turbine. These parameters are provided in real time by the wind turbine main control system.

[0107] (2) Algorithm signal: the first-order frequency of the tower, provided by the wind turbine control strategy module running in the PLC (Programmable Logic Controller).

[0108] The tower's first-order frequency is the natural frequency of the tower. The first-order frequency of different tower models can be determined during the design phase through BLADE (Blade Load and Dynamic Environment) simulation. The frequency range of the tower's first-order frequency is typically between 0.1 and 0.5 Hz.

[0109] Among them, if the algorithm is deployed in the wind turbine edge computing terminal or field-level control system, all input data are directly transmitted from the master control to the edge computing terminal or field-level control system.

[0110] In one embodiment, based on the actual operating state parameters of the wind turbine generator set, a sliding window algorithm is used to determine the operating state characteristics of the wind turbine generator set. The process of determining the operating state characteristics of the wind turbine generator set under any sliding window includes:

[0111] (1) Based on the rotor speed of the wind turbine, the standard deviation and the average rotor speed within a preset time period are determined.

[0112] The process of determining the average wind rotor speed is as follows: adding all wind rotor speed values ​​in the wind rotor speed to obtain a total, and then dividing the total by the total number of wind rotor speeds to obtain the average wind rotor speed.

[0113] The standard deviation of the wind rotor speed reflects the degree to which the wind rotor speed deviates from the average wind rotor speed. The larger the value, the more severe the speed fluctuation.

[0114] The calculation formula of the standard deviation of the wind wheel speed is as follows:

[0115] (3);

[0116] Where, is the standard deviation of the wind wheel speed, is the total number of wind wheel speeds, is the i-th wind wheel speed value, is the average wind wheel speed.

[0117] (2) The calculated value of the wind wheel rotation frequency is obtained based on the average value of the wind wheel speed.

[0118] The wind wheel rotation frequency refers to the frequency of 1 times the wind wheel speed (unit: Hz).

[0119] In this embodiment, the calculated value of the wind wheel rotation frequency = the average value of the wind wheel rotation speed / 60.

[0120] (3) Finding a wind rotor frequency identification frequency interval within a preset range of the wind rotor frequency calculation value, and finding a maximum amplitude value within a range corresponding to the wind rotor frequency identification frequency interval in the cabin vibration signal spectrum as the final identified wind rotor frequency amplitude value, and using the frequency value corresponding to the maximum amplitude value in the cabin vibration signal spectrum as the final identified wind rotor frequency, wherein the cabin vibration signal spectrum is obtained by performing FFT (Fast Fourier Transform) on the cabin vibration signal.

[0121] The core principle of FFT is to decompose the discrete Fourier transform into smaller sub-problems through a divide-and-conquer strategy, and use the symmetry and periodicity of the rotation factors to reduce the amount of repeated calculations, thereby significantly reducing the time complexity.

[0122] (4) Determining the turbulence intensity within the preset time period based on the wind speed of the wind turbine generator system.

[0123] With reference to the calculation process of the average wind speed value and the standard deviation of the wind speed value, the average wind speed value and the standard deviation of the wind speed value within a preset time period are determined based on the wind speed.

[0124] Turbulence intensity = wind speed standard deviation ÷ wind speed average.

[0125] According to step S103, the fault diagnosis of abnormal long-term imbalance trend of the wind rotor in this application mainly includes two steps:

[0126] Step 1: Detect the wind rotor imbalance change trend to check whether the wind rotor rotation frequency amplitude in the array meets the abnormal suspicion condition of the wind rotor imbalance change trend. The specific judgment process is as follows:

[0127] Determine the number of points in the wind rotor imbalance change trend test array where the wind rotor rotation frequency amplitude is greater than the wind rotor rotation frequency amplitude change trend abnormality judgment reference value;

[0128] Calculating the ratio of the number of points to the total number of points of the wind wheel rotation frequency amplitude;

[0129] When the ratio exceeds a preset point proportion threshold, it is determined that the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend inspection array meets the abnormal suspicion condition of the wind rotor imbalance change trend.

[0130] The value of the preset point ratio threshold is determined according to actual needs, for example, 30%.

[0131] It should be noted that the present application will only proceed to the second step if the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend inspection array meets the abnormal suspicion condition of the wind rotor imbalance change trend. Otherwise, the observation state will be maintained within the judgment period when the current wind rotor imbalance change trend is abnormally long.

[0132] Step 2: Use the trend test method to perform fault diagnosis on the abnormally long-term unbalance change trend of the wind wheel on the wind wheel unbalance change trend test array.

[0133] Specifically, the wind rotor rotation frequency amplitude time series data in the inspection array is tested according to the wind rotor imbalance change trend, and the wind rotor rotation frequency amplitude statistics are calculated;

[0134] Calculating the statistical variance of the wind rotor rotation frequency amplitude according to the size of the wind rotor imbalance change trend test array;

[0135] Calculating a standardized test statistic based on the wind rotor rotation frequency amplitude statistic and the variance of the wind rotor rotation frequency amplitude statistic;

[0136] A one-tailed hypothesis test is performed on the standardized test statistic. If the standardized test statistic is not less than the Z quantile corresponding to the preset percentile of the standard normal distribution, and the significance of the wind rotor imbalance change trend is less than the preset significance level, it is determined that the risk unit has an abnormally long-term wind rotor imbalance change trend fault; otherwise, it is determined that the risk unit has not an abnormally long-term wind rotor imbalance change trend fault, wherein the preset percentile is determined based on the preset significance level, and the Z quantile refers to the value corresponding to the cumulative probability equal to the preset percentile in the standard normal distribution, reflecting the degree of deviation from the sample mean.

[0137] Taking the Mann-Kendall trend test method as an example to diagnose the abnormal long-term fault of the wind rotor imbalance change trend, the diagnosis process is as follows:

[0138] The time series data of the wind wheel rotation frequency amplitude in the wind wheel imbalance change trend test array are set as: Amp_1P1, Amp_1P2, ..., Amp_1P n , n is the size of the wind wheel imbalance change trend test array, and then perform the following steps:

[0139] A1. Calculate the wind rotor frequency amplitude statistic S;

[0140] (4);

[0141] The symbolic function is defined as:

[0142] (5);

[0143] A2. Calculate the variance Var(S) of the wind rotor frequency amplitude statistic S;

[0144] (6);

[0145] A3. Calculate the standardized test statistic Z MK ;

[0146] (7);

[0147] It should be noted that this application converts the wind rotor frequency amplitude statistic S into a standardized test statistic Z MK , so that for large sample data Z MK It can approximately satisfy the standard normal distribution.

[0148] A4. Determine trend significance (one-tail hypothesis test, calculate p-value);

[0149] Null hypothesis H a : When the current wind rotor imbalance change trend is abnormally long, it is determined that the wind rotor imbalance degree has a significant increasing trend within the period.

[0150] Assuming the significance level α (usually α=0.05), calculate the significance p value as follows:

[0151] (8);

[0152] Where function Φ is the cumulative distribution function of the standard normal distribution, represents the standardized test statistic Z MK The absolute value of .

[0153] If Z MK ≥Z 1-α If p < 0.05, then accept the null hypothesis H a Otherwise, there is no trend of increasing wind wheel imbalance. 1-α Indicates the Z quantile corresponding to the 100(1-α)th percentile of the standard normal distribution.

[0154] In summary, the present invention adopts a serial fusion diagnosis strategy that combines the abnormal point proportion results with the Mann-Kendall test results. It not only focuses on "instantaneously exceeding the abnormal reference value for judging the trend of the wind wheel frequency amplitude change", but also considers "whether the trend offset is significant". It can form a comprehensive algorithm for quickly discovering abnormal fluctuations and rigorously judging trend growth, which greatly improves the accuracy and robustness of the diagnosis of abnormal wind wheel imbalance change trends.

[0155] See also Figure 2 A schematic diagram of an example of determining an abnormally long-term rotor imbalance trend, disclosed in an embodiment of the present invention, shows the horizontal axis representing the data point number and the vertical axis representing the rotor frequency amplitude. Taking a faulty wind turbine as an example, a rotor long-term imbalance trend detection array containing 491 rotor frequency amplitude values ​​is first obtained using a continuous sliding window within the abnormally long-term rotor imbalance trend determination period. An abnormally long-term rotor imbalance trend diagnosis is then performed. This involves two sub-steps. The first step is "suspecting an abnormal rotor imbalance trend." The calculation shows that the percentage of points in the rotor frequency amplitude detection array with rotor frequency amplitudes greater than the abnormal rotor frequency amplitude trend determination threshold is 35.64%, exceeding the preset point percentage threshold (30%). The second step is to "determine the abnormal trend of the wind rotor imbalance change". The Mann-Kendall trend test method is used to calculate the wind rotor frequency amplitude statistic S = 56400, the variance of the wind rotor frequency amplitude statistic S Var(S) = 13192351.67 and the standardized statistic Z = 15.528. At the significance level α = 0.05, Z is much larger than Z 1-α And p<0.05, therefore, within the current determination period for when the unbalance change trend of the wind rotor is abnormally long, it is determined that a fault of when the unbalance change trend of the wind rotor is abnormally long occurs.

[0156] Through research, the inventors found that there would be certain differences in the distribution of wind rotor frequency amplitude of different wind turbines. In order to improve the accuracy and robustness of the diagnosis of abnormal wind rotor imbalance change trend, when calculating the abnormal determination benchmark value of wind rotor frequency amplitude change trend and the historical wind rotor frequency target amplitude, a "one machine, one policy" strategy was adopted for each wind turbine separately.

[0157] (1) The process of determining the reference value for abnormal wind rotor frequency amplitude change trend includes:

[0158] B1. After the wind turbine is connected to the grid for the first time, obtain the historical wind rotor frequency and amplitude values ​​within the preset statistical period before the current moment, and form a statistical sample set of historical wind rotor frequency and amplitude values. The rules are as follows:

[0159] The continuous sliding window makes the following judgments:

[0160] (1) The standard deviation of the wind rotor speed is less than the wind rotor speed stability judgment threshold.

[0161] The value of the wind wheel speed stability judgment threshold is determined according to actual needs, for example, it is set to 0.4 revolutions per minute.

[0162] (2) The absolute value of the difference between the wind rotor frequency and the first-order frequency of the tower is greater than the frequency difference threshold.

[0163] The frequency difference threshold is determined based on the frequency deviation adjustment coefficient. For example, the frequency difference threshold = tower first-order frequency × frequency deviation adjustment coefficient. The value of the frequency deviation adjustment coefficient is determined based on actual needs, for example, the frequency deviation adjustment coefficient is 10%.

[0164] (3) The turbulence intensity does not exceed the turbulence intensity threshold.

[0165] The turbulence intensity threshold may be a Class A turbulence intensity at an average wind speed over a preset time period in the IEC (International Electrotechnical Commission) standard. The Class A turbulence intensity is also the turbulence intensity in the IEC turbulence intensity curve.

[0166] If all the above conditions are met in the current sliding window, the wind rotor frequency amplitude value in the current sliding window is added to the historical wind rotor frequency amplitude value statistical sample set, and then the next sliding window is entered to continue the above judgment.

[0167] B2. Perform kernel density estimation on the statistical sample set of historical wind rotor rotation frequency and amplitude values ​​to obtain a probability density function of the wind rotor rotation frequency and amplitude values.

[0168] The core idea of ​​Kernel Density Estimation (KDE) is to place a "kernel function" at each data point and superimpose all kernel functions to obtain the overall probability density estimate.

[0169] The core idea of ​​the Probability Density Function (PDF) is to describe the "probability size" of a continuous random variable taking values ​​near a certain value through "density" rather than direct probability.

[0170] B3. Constructing a cumulative distribution function of the wind rotor rotation frequency amplitude probability density function by integrating and accumulating.

[0171] The Cumulative Distribution Function (CDF) is a key tool in probability theory used to describe the probability distribution of a random variable. It represents the cumulative probability that a random variable takes a value less than or equal to a specific value.

[0172] The core idea of ​​the cumulative distribution function is to describe the sum of the probabilities of a random variable within a certain value and below by means of cumulative probability.

[0173] B4. Find the target wind rotor frequency amplitude corresponding to the preset quantile position from the cumulative distribution function, and use the target wind rotor frequency amplitude as a reference value for determining abnormality in the wind rotor frequency amplitude change trend.

[0174] The value of the preset quantile position is determined according to actual needs, for example, the 95% quantile position.

[0175] B5. Calculate the average value of all historical wind rotor rotation frequency amplitudes in the historical wind rotor rotation frequency amplitude statistical sample set, and determine the average value as the historical wind rotor rotation frequency target amplitude.

[0176] See also Figure 3 , an embodiment of the present invention discloses a reference value for determining abnormality in a wind rotor frequency amplitude change trend and a schematic diagram of an example of calculating a historical wind rotor frequency target amplitude. Taking a certain actual faulty wind turbine as an example, a historical wind rotor frequency amplitude statistical sample set containing 6015 wind rotor frequency amplitudes is obtained through a continuous sliding window within a preset statistical period and a probability distribution histogram is plotted. The probability density function of the wind rotor frequency amplitude is obtained through kernel density estimation; then, a cumulative distribution function of the wind rotor frequency amplitude is constructed through an accumulation method; finally, the position of the 95% quantile is found in the cumulative distribution function as the reference value for determining abnormality in a wind rotor frequency amplitude change trend of the faulty turbine. In this embodiment, the reference value for determining abnormality in a wind rotor frequency amplitude change trend is 0.0073g (i.e. Figure 3The vertical dotted line 11 corresponds to 0.0073g), and the average value of the historical wind rotor frequency amplitude statistical sample set is used as the historical wind rotor frequency target amplitude of the faulty unit. In this embodiment, the historical wind rotor frequency target amplitude is 0.0027g (i.e. Figure 3 The middle vertical dashed line 12 corresponds to 0.0027g).

[0177] In one embodiment, the process of diagnosing an abnormal short-term fault of a wind rotor imbalance change trend based on a change trend of a wind rotor rotation frequency amplitude cumulative deviation curve includes:

[0178] Compare the cumulative deviation data points in the cumulative deviation curve of the wind rotor frequency amplitude with the wind rotor imbalance short-term trend abnormality judgment threshold in sequence according to the time sequence;

[0179] If there are a continuous preset number of cumulative deviation data points that all exceed the wind rotor imbalance short-term trend abnormality judgment threshold, it is determined that the risk unit has an abnormal short-term fault of wind rotor imbalance change trend.

[0180] The value of the threshold value for determining abnormal short-term trend of wind rotor imbalance is determined according to actual needs, for example, 0.05g, where g is the unit of acceleration.

[0181] The value of the continuous preset number is determined according to actual needs. For example, the value of the continuous preset number is 5, which is not limited in the present invention.

[0182] See also Figure 4 The present invention discloses a schematic diagram of an example of determining an abnormally short-term wind rotor imbalance change trend. Taking a certain actual faulty wind turbine as an example, a wind rotor imbalance change trend test array containing 43 wind rotor rotation frequency amplitudes is first obtained through a continuous sliding window within the wind rotor imbalance change trend abnormally short-term determination period; then, an abnormally short-term wind rotor imbalance change trend diagnosis is performed, and the calculation results are obtained. Then the cumulative deviation curve of the wind rotor frequency amplitude is drawn (the blue curve in the figure); finally, each cumulative deviation data point in the cumulative deviation curve of the wind rotor frequency amplitude is compared with the threshold value for judging the short-term abnormal trend of the wind rotor imbalance. It can be seen from the figure that starting from sequence number 19 (the position corresponding to the pink vertical dotted line in the figure, the corresponding cumulative deviation value is 0.05225g), the cumulative deviation curve of the wind rotor frequency amplitude exceeds the threshold value for judging the short-term abnormal trend of the wind rotor imbalance (0.05g) for multiple consecutive points. At sequence number 23 (the position corresponding to the orange vertical dotted line in the figure), it meets the logic for triggering the short-term trend abnormal warning, so the abnormal short-term judgment of the wind rotor imbalance change trend is established.

[0183] Corresponding to the above method embodiment, the present invention further discloses a device for diagnosing wind turbine rotor imbalance faults.

[0184] See also Figure 5, a schematic structural diagram of a wind turbine rotor imbalance fault diagnosis device disclosed in an embodiment of the present invention, the device may include:

[0185] The array acquisition unit 201 is used to acquire the wind rotor frequency amplitude within a preset period when the wind turbine is in a grid-connected power generation state, and to sequentially place each of the wind rotor frequency amplitude values ​​into a wind rotor imbalance change trend detection array according to a time sequence;

[0186] A judgment unit 202 is configured to judge whether the preset period is greater than a long-period threshold value of a wind rotor imbalance change trend or less than a short-period threshold value of a wind rotor imbalance change trend;

[0187] The long-term fault diagnosis unit 203 is configured to, if the preset period is greater than the long-term threshold value of the rotor imbalance change trend, and if the rotor rotation frequency amplitude in the rotor imbalance change trend detection array meets the condition for suspecting abnormality of the rotor imbalance change trend, perform abnormal long-term fault diagnosis of the rotor imbalance change trend on the rotor imbalance change trend detection array using a trend detection method;

[0188] The short-term fault diagnosis unit 204 is used to calculate the cumulative deviation of each wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude according to the time sequence of each wind wheel frequency amplitude in the wind wheel imbalance change trend inspection array if the preset period is less than the short-term threshold of the wind wheel imbalance change trend, and draw a wind wheel frequency amplitude cumulative deviation curve, and perform abnormal short-term fault diagnosis of the wind wheel imbalance change trend according to the change trend of the wind wheel frequency amplitude cumulative deviation curve, wherein each of the cumulative deviations is: the sum of the deviation of the current time series wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude and the cumulative deviation of the previous time series wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude.

[0189] In summary, the present invention discloses a device for diagnosing an imbalance fault of a wind turbine rotor. When the wind turbine rotor is in a grid-connected power generation state, the device obtains the rotor frequency amplitude within a preset period, and sequentially places each rotor frequency amplitude into a rotor imbalance change trend detection array according to a time sequence. When the preset period is greater than a long-period threshold value of the rotor imbalance change trend, and the rotor frequency amplitude in the rotor imbalance change trend detection array meets a suspected condition of an abnormal rotor imbalance change trend, the device adopts a trend detection method to perform an abnormal long-term rotor imbalance change trend fault diagnosis on the rotor imbalance change trend detection array. When the preset period is less than a short-period threshold value of the rotor imbalance change trend, the device calculates the cumulative deviation of each rotor frequency amplitude relative to a historical rotor frequency target amplitude according to the time sequence of each rotor frequency amplitude in the rotor imbalance change trend detection array, draws a rotor frequency amplitude cumulative deviation curve, and performs an abnormal short-term rotor imbalance change trend fault diagnosis according to the change trend of the rotor frequency amplitude cumulative deviation curve. This invention constructs a rotor imbalance trend detection array, simultaneously considering two abnormal evolution patterns of rotor imbalance (long-term, slow, cumulative changes and short-term, rapid, sudden changes). It then employs a dual-path parallel fusion strategy for pattern analysis, enabling long-term and short-term fault diagnosis for rotor imbalance trend anomalies. This strategy captures long-term trend enhancement characteristics while remaining sensitive to transient anomalies, effectively avoiding late and missed early fault detection, significantly improving the ability to proactively detect rotor imbalance faults.

[0190] In one embodiment, the array acquisition unit 201 may be specifically configured to:

[0191] When the wind turbine operating state characteristics in any sliding window within the preset period meet the following judgment conditions, the wind rotor rotation frequency amplitude corresponding to the sliding window is obtained, and the judgment conditions include:

[0192] The standard deviation of the wind rotor speed is less than the wind rotor speed stability judgment threshold;

[0193] The absolute value of the difference between the wind rotor rotation frequency and the tower first-order frequency is greater than a frequency difference threshold, where the frequency difference threshold is determined based on a frequency deviation adjustment coefficient;

[0194] The turbulence intensity does not exceed the turbulence intensity threshold;

[0195] The wind turbine operating state characteristics include: the standard deviation of the wind rotor speed, the wind rotor rotation frequency and the turbulence intensity.

[0196] In one embodiment, the array acquisition unit 201 may be specifically configured to:

[0197] Determine the standard deviation of the wind rotor speed and the average wind rotor speed within a preset time period based on the wind rotor speed of the wind turbine generator set;

[0198] Obtaining a calculated value of the wind rotor rotation frequency based on the average wind rotor rotation speed;

[0199] Finding a rotor frequency identification frequency interval within a preset range of the rotor frequency calculation value, finding a maximum amplitude within a range corresponding to the rotor frequency identification frequency interval in a nacelle vibration signal spectrum as the final identified rotor frequency amplitude, and using a frequency value corresponding to the maximum amplitude in the nacelle vibration signal spectrum as the final identified rotor frequency, wherein the nacelle vibration signal spectrum is obtained by performing a fast Fourier transform on the nacelle vibration signal;

[0200] The turbulence intensity within the preset time period is determined based on the wind speed of the wind turbine.

[0201] In one embodiment, the long-term fault diagnosis unit 203 may be specifically configured to:

[0202] Determine the number of points in the wind rotor imbalance change trend inspection array where the wind rotor rotation frequency amplitude is greater than a wind rotor rotation frequency amplitude change trend abnormality determination reference value;

[0203] Calculating the ratio of the number of points to the total number of points of the wind wheel rotation frequency amplitude;

[0204] When the ratio exceeds a preset point proportion threshold, it is determined that the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend inspection array meets the abnormal suspicion condition of the wind rotor imbalance change trend.

[0205] In one embodiment, the wind turbine rotor imbalance fault diagnosis device may further include:

[0206] The unit for determining the abnormal reference value for determining the change trend of the wind rotor frequency amplitude is used to:

[0207] After the wind turbine generator set is connected to the grid for the first time for power generation, historical wind rotor rotation frequency amplitude values ​​within a preset statistical period before the current moment are obtained, and a statistical sample set of historical wind rotor rotation frequency amplitude values ​​is formed;

[0208] Performing kernel density estimation on the historical wind rotor rotation frequency amplitude statistical sample set to obtain a wind rotor rotation frequency amplitude probability density function;

[0209] Constructing a cumulative distribution function of the wind rotor rotation frequency amplitude probability density function by integrating and accumulating;

[0210] The target wind rotor rotation frequency amplitude corresponding to the preset quantile position is found from the cumulative distribution function, and the target wind rotor rotation frequency amplitude is used as a reference value for determining abnormality in the wind rotor rotation frequency amplitude change trend.

[0211] In one embodiment, the long-term fault diagnosis unit 203 may be further configured to:

[0212] The wind rotor rotation frequency amplitude time series data in the inspection array is used to calculate the wind rotor rotation frequency amplitude statistics according to the wind rotor imbalance change trend;

[0213] Calculating the statistical variance of the wind rotor rotation frequency amplitude according to the size of the wind rotor imbalance change trend test array;

[0214] Calculating a standardized test statistic based on the wind rotor rotation frequency amplitude statistic and the variance of the wind rotor rotation frequency amplitude statistic;

[0215] A one-tailed hypothesis test is performed on the standardized test statistic. If the standardized test statistic is not less than the Z quantile corresponding to the preset percentile of the standard normal distribution, and the significance of the wind rotor imbalance change trend is less than the preset significance level, it is determined that the risk unit has an abnormally long-term wind rotor imbalance change trend fault; otherwise, it is determined that the risk unit has not an abnormally long-term wind rotor imbalance change trend fault, wherein the preset percentile is determined based on the preset significance level, and the Z quantile refers to the value corresponding to the cumulative probability equal to the preset percentile in the standard normal distribution, reflecting the degree of deviation from the sample mean.

[0216] In one embodiment, the wind turbine rotor imbalance fault diagnosis device may further include:

[0217] The unit for determining the target amplitude of the historical wind rotor rotation frequency is used to:

[0218] After the wind turbine generator set is connected to the grid for the first time for power generation, historical wind rotor rotation frequency amplitude values ​​within a preset statistical period before the current moment are obtained, and a statistical sample set of historical wind rotor rotation frequency amplitude values ​​is formed;

[0219] Calculating the average value of all historical wind rotor rotation frequency amplitudes in the historical wind rotor rotation frequency amplitude statistical sample set;

[0220] The average value is determined as the historical wind wheel rotation frequency target amplitude.

[0221] In one embodiment, the short-term fault diagnosis unit 204 may be specifically configured to:

[0222] Comparing each cumulative deviation data point in the cumulative deviation curve of the wind rotor rotation frequency amplitude with the wind rotor imbalance short-term trend abnormality judgment threshold in sequence according to time sequence;

[0223] If there are a continuous preset number of cumulative deviation data points that all exceed the wind rotor imbalance short-term trend abnormality judgment threshold, it is determined that the risk unit has an abnormal short-term fault of wind rotor imbalance change trend.

[0224] It should be noted that, for the specific working principles of the various components in the device embodiment, please refer to the corresponding part of the method embodiment, which will not be repeated here.

[0225] Corresponding to the above embodiment, the present invention further discloses a computer storage medium, which stores at least one instruction. When the at least one instruction is executed by a processor, the steps shown in the embodiment of the method for diagnosing wind turbine rotor imbalance fault are implemented.

[0226] A computer storage medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer storage medium may be a machine-readable signal medium or a machine-readable storage medium. A computer storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0227] Corresponding to the above embodiment, Figure 6 As shown, the present invention also provides a structural diagram of an electronic device, which may include: a processor 1 and a memory 2;

[0228] The processor 1 and the memory 2 communicate with each other via a communication bus 3.

[0229] Processor 1, configured to execute at least one instruction;

[0230] Memory 2, used to store at least one instruction;

[0231] The processor 1 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0232] The memory 2 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0233] The processor executes at least one instruction to implement the steps shown in the embodiment of the method for diagnosing wind turbine rotor imbalance fault.

[0234] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0235] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0236] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for diagnosing wind turbine rotor imbalance fault, characterized in that: include: When the wind turbine is in a grid-connected power generation state, obtaining the wind rotor frequency amplitude within a preset period, and sequentially placing each of the wind rotor frequency amplitudes into a wind rotor imbalance change trend test array according to a time sequence; Determining that the preset period is greater than a long-period threshold value of the wind rotor imbalance change trend or less than a short-period threshold value of the wind rotor imbalance change trend; If the preset period is greater than the long-period threshold value of the wind rotor imbalance change trend, and the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend detection array meets the suspected abnormal wind rotor imbalance change trend condition, a trend detection method is used to perform an abnormally long-period fault diagnosis of the wind rotor imbalance change trend on the wind rotor imbalance change trend detection array; If the preset period is less than the short-period threshold value of the wind rotor imbalance change trend, the cumulative deviation of each wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude is calculated in turn according to the time sequence of each wind rotor frequency amplitude in the wind rotor imbalance change trend inspection array, and a wind rotor frequency amplitude cumulative deviation curve is drawn. The abnormal short-term fault diagnosis of the wind rotor imbalance change trend is performed according to the change trend of the wind rotor frequency amplitude cumulative deviation curve, wherein each cumulative deviation is: the sum of the deviation of the current time sequence wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude and the cumulative deviation of the previous time sequence wind rotor frequency amplitude relative to the historical wind rotor frequency target amplitude.

2. The wind turbine rotor imbalance fault diagnosis method according to claim 1, characterized in that: The step of obtaining the wind rotor rotation frequency amplitude within a preset period includes: When the wind turbine operating state characteristics in any sliding window within the preset period meet the following judgment conditions, the wind rotor rotation frequency amplitude corresponding to the sliding window is obtained, and the judgment conditions include: The standard deviation of the wind rotor speed is less than the wind rotor speed stability judgment threshold; The absolute value of the difference between the wind rotor rotation frequency and the tower first-order frequency is greater than a frequency difference threshold, where the frequency difference threshold is determined based on a frequency deviation adjustment coefficient; The turbulence intensity does not exceed the turbulence intensity threshold; The wind turbine operating state characteristics include: the standard deviation of the wind rotor speed, the wind rotor rotation frequency and the turbulence intensity.

3. The wind turbine rotor imbalance fault diagnosis method according to claim 2, characterized in that: The process of determining the operating status characteristics of the wind turbine generator system in any sliding window includes: Determine the standard deviation of the wind rotor speed and the average wind rotor speed within a preset time period based on the wind rotor speed of the wind turbine generator set; Obtaining a calculated value of the wind rotor rotation frequency based on the average wind rotor rotation speed; Finding a rotor frequency identification frequency interval within a preset range of the rotor frequency calculation value, finding a maximum amplitude within a range corresponding to the rotor frequency identification frequency interval in a nacelle vibration signal spectrum as the final identified rotor frequency amplitude, and using a frequency value corresponding to the maximum amplitude in the nacelle vibration signal spectrum as the final identified rotor frequency, wherein the nacelle vibration signal spectrum is obtained by performing a fast Fourier transform on the nacelle vibration signal; The turbulence intensity within the preset time period is determined based on the wind speed of the wind turbine.

4. The wind turbine rotor imbalance fault diagnosis method according to any one of claims 1 to 3, characterized in that: The process of determining whether the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend test array meets the abnormal suspicion condition of the wind rotor imbalance change trend includes: Determine the number of points in the wind rotor imbalance change trend inspection array where the wind rotor rotation frequency amplitude is greater than a wind rotor rotation frequency amplitude change trend abnormality determination reference value; Calculating the ratio of the number of points to the total number of points of the wind wheel rotation frequency amplitude; When the ratio exceeds a preset point proportion threshold, it is determined that the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend inspection array meets the abnormal suspicion condition of the wind rotor imbalance change trend.

5. The wind turbine rotor imbalance fault diagnosis method according to claim 4, characterized in that: The process of determining the abnormal reference value for determining the wind rotor frequency amplitude change trend includes: After the wind turbine generator set is connected to the grid for the first time for power generation, historical wind rotor rotation frequency amplitude values ​​within a preset statistical period before the current moment are obtained, and a statistical sample set of historical wind rotor rotation frequency amplitude values ​​is formed; Performing kernel density estimation on the historical wind rotor rotation frequency amplitude statistical sample set to obtain a wind rotor rotation frequency amplitude probability density function; Constructing a cumulative distribution function of the wind rotor rotation frequency amplitude probability density function by integrating and accumulating; The target wind rotor rotation frequency amplitude corresponding to the preset quantile position is found from the cumulative distribution function, and the target wind rotor rotation frequency amplitude is used as a reference value for determining abnormality in the wind rotor rotation frequency amplitude change trend.

6. The wind turbine rotor imbalance fault diagnosis method according to claim 1, characterized in that: The method of using a trend inspection method to perform fault diagnosis of an abnormally long-term wind rotor imbalance change trend on the wind rotor imbalance change trend inspection array includes: The wind rotor rotation frequency amplitude time series data in the inspection array is used to calculate the wind rotor rotation frequency amplitude statistics according to the wind rotor imbalance change trend; Calculating the statistical variance of the wind rotor rotation frequency amplitude according to the size of the wind rotor imbalance change trend test array; Calculating a standardized test statistic according to the wind rotor rotation frequency amplitude statistic and the variance of the wind rotor rotation frequency amplitude statistic; A one-tailed hypothesis test is performed on the standardized test statistic. If the standardized test statistic is not less than the Z quantile corresponding to the preset percentile of the standard normal distribution, and the significance of the wind rotor imbalance change trend is less than the preset significance level, it is determined that the risk unit has an abnormally long-term wind rotor imbalance change trend fault; otherwise, it is determined that the risk unit has not an abnormally long-term wind rotor imbalance change trend fault, wherein the preset percentile is determined based on the preset significance level, and the Z quantile refers to the value corresponding to the cumulative probability equal to the preset percentile in the standard normal distribution, reflecting the degree of deviation from the sample mean.

7. The wind turbine rotor imbalance fault diagnosis method according to claim 1, characterized in that: The process of determining the target amplitude of the historical wind rotor rotation frequency includes: After the wind turbine generator set is connected to the grid for the first time for power generation, historical wind rotor rotation frequency amplitude values ​​within a preset statistical period before the current moment are obtained, and a statistical sample set of historical wind rotor rotation frequency amplitude values ​​is formed; Calculating the average value of all historical wind rotor rotation frequency amplitudes in the historical wind rotor rotation frequency amplitude statistical sample set; The average value is determined as the historical wind wheel rotation frequency target amplitude.

8. The wind turbine rotor imbalance fault diagnosis method according to claim 1, characterized in that: The abnormal short-term fault diagnosis of the wind rotor imbalance change trend according to the change trend of the wind rotor rotation frequency amplitude cumulative deviation curve includes: Comparing each cumulative deviation data point in the cumulative deviation curve of the wind rotor rotation frequency amplitude with the wind rotor imbalance short-term trend abnormality judgment threshold in sequence according to time sequence; If there are a continuous preset number of cumulative deviation data points that all exceed the wind rotor imbalance short-term trend abnormality judgment threshold, it is determined that the risk unit has an abnormal short-term fault of wind rotor imbalance change trend.

9. A wind turbine rotor imbalance fault diagnosis device, characterized in that: include: An array acquisition unit is used to acquire the wind rotor frequency amplitude within a preset period when the wind turbine is in a grid-connected power generation state, and to sequentially place each of the wind rotor frequency amplitude values ​​into a wind rotor imbalance change trend detection array according to a time sequence; A judging unit, configured to judge whether the preset period is greater than a long-period threshold value of the wind rotor imbalance change trend or less than a short-period threshold value of the wind rotor imbalance change trend; a long-term fault diagnosis unit, configured to, if the preset period is greater than the long-term threshold value of the wind rotor imbalance change trend and the wind rotor rotation frequency amplitude in the wind rotor imbalance change trend detection array meets a suspected condition of abnormal wind rotor imbalance change trend, perform a long-term fault diagnosis of abnormal wind rotor imbalance change trend on the wind rotor imbalance change trend detection array using a trend detection method; A short-term fault diagnosis unit is used to calculate the cumulative deviation of each wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude according to the time sequence of each wind wheel frequency amplitude in the wind wheel imbalance change trend inspection array if the preset period is less than the short-term threshold of the wind wheel imbalance change trend, and draw a wind wheel frequency amplitude cumulative deviation curve, and perform abnormal short-term fault diagnosis of the wind wheel imbalance change trend according to the change trend of the wind wheel frequency amplitude cumulative deviation curve, wherein each of the cumulative deviations is: the sum of the deviation of the current time series wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude and the cumulative deviation of the previous time series wind wheel frequency amplitude relative to the historical wind wheel frequency target amplitude.

10. A computer storage medium, characterized in that The computer storage medium stores at least one instruction, and when the at least one instruction is executed by the processor, the wind turbine rotor imbalance fault diagnosis method according to any one of claims 1 to 8 is implemented.

11. An electronic device, characterized in that: The electronic device includes: a memory and a processor; The memory is used to store at least one instruction; The processor is used to execute the at least one instruction to implement the wind turbine rotor imbalance fault diagnosis method according to any one of claims 1 to 8.

Images