A wind turbine negative power fault detection and identification method

Abstract

The application discloses a wind turbine negative power fault detection and identification method. The wind turbine negative power fault detection and identification method comprises the following steps: obtaining and preprocessing wind turbine data; constructing a wind turbine negative power fault detection model; analyzing the possibility of the current wind turbine negative power fault; analyzing the wind turbine negative power fault degree of the current wind turbine, generating wind turbine negative power alarm information corresponding to the level; and analyzing the comprehensive detection and identification precision evaluation index of the current wind turbine. The application analyzes the comprehensive detection and identification precision evaluation index. When the comprehensive detection and identification precision evaluation index is less than or equal to the fourth threshold value, it indicates that the wind turbine fault detection and identification accuracy is insufficient, thereby issuing a warning for further optimization of the wind turbine negative power fault detection and identification method, and improving the wind turbine fault detection and identification precision, thereby solving the problem of low wind turbine fault detection and identification accuracy in the prior art.

Description

Technical Field

[0001] This invention relates to the field of wind turbine fault detection technology, and in particular to a method for detecting and identifying negative power faults in wind turbines. Background Technology

[0002] With the increasing demand for energy in daily production and life, and the continuous depletion of non-renewable energy reserves, renewable energy has become a key area of ​​research and development in the energy and environmental protection fields. Under the background of pursuing environmental protection and energy conservation, wind power generation has become an important renewable energy power generation method. Wind turbines are devices that use wind energy to generate electricity, also known as wind turbine generators or wind turbines. They are a type of renewable energy power generation equipment that converts wind energy into mechanical energy, and then the mechanical energy is converted into electrical energy by a generator. Negative power of a wind turbine refers to the fact that the power generated by the wind turbine is negative under certain operating conditions. Negative power of a wind turbine is not necessarily a fault, but it usually indicates the existence of some problems. If the wind turbine frequently experiences negative power, or if the negative power lasts for too long, it needs to be inspected and maintained. Therefore, when facing a negative power fault, a detailed fault diagnosis and troubleshooting are usually required to ensure the reliable operation of the wind turbine.

[0003] Currently, there is extensive research on wind turbine fault detection. For example, Chinese invention patent CN110805534B discloses a method, device, and equipment for wind turbine fault detection. This method not only removes non-primary characteristic variables from the operating data that have little linear correlation with the primary characteristic variables (used to characterize preset fault types), but also removes non-primary characteristic variables that have little nonlinear correlation with the primary characteristic variables. Because these operating data characteristic variables that do not significantly contribute to the fault detection results are removed, the number of characteristic variables in the wind turbine operating data is reduced, thus improving the detection speed of wind turbine fault detection and reducing the false alarm and missed alarm rates. Another example is Chinese invention patent CN116104710A. A method and system for detecting wind turbine faults are disclosed. This method relates to the field of wind turbine fault detection and includes: acquiring discrete and continuous quantities of the wind turbine from SCADA; determining a first correlation degree between pairs of discrete quantities; if the first correlation degree meets a first preset requirement, then determining that a migration correlation exists between the pairs of discrete quantities; determining a second correlation degree between pairs of continuous quantities; if the second correlation degree meets a second preset requirement, then determining that a migration correlation exists between the pairs of continuous quantities; distinguishing the source quantity and target quantity among the pairs of discrete and continuous quantities with migration correlation, and ensuring that the source quantity and target quantity correspond; establishing a classification model based on the source quantity; predicting the predicted quantities of the source quantity and target quantity over a future period based on the classification model; and determining the fault status of the wind turbine based on the predicted quantities.

[0004] However, the internal structure of wind turbines is currently quite complex, involving multiple components and subsystems. Each component may have different types of faults, and they are coupled with each other. A fault in one component may cause a fault in other components, which increases the complexity of fault detection and identification. It also makes the amount of fault detection and identification data large and diverse, making accurate fault detection and identification of wind turbines a challenge.

[0005] Therefore, in summary, the existing technology suffers from low accuracy in wind turbine fault detection and identification. Summary of the Invention

[0006] This application provides a method for detecting and identifying negative power faults in wind turbines, which solves the problem of low accuracy in the detection and identification of wind turbine faults in the prior art and improves the accuracy of wind turbine fault detection and identification.

[0007] This application provides a method for detecting and identifying negative power faults in wind turbines, including the following steps: acquiring real-time monitoring data of the wind turbine through wind turbine sensors, and simultaneously acquiring historical negative power fault data and historical monitoring data of the wind turbine; and preprocessing the acquired real-time monitoring data, historical negative power fault data, and historical non-fault monitoring data of the wind turbine.

[0008] Based on preprocessed historical negative power fault data, a wind turbine negative power fault detection model is constructed. When the output power of the current wind turbine is detected to be negative, the preprocessed real-time monitoring data of the current wind turbine is input into the wind turbine negative power fault detection model to analyze the wind turbine negative power fault detection index. The wind turbine negative power fault detection index is used to detect the probability of the current wind turbine experiencing a negative power fault. When a negative power fault is detected, the degree of the wind turbine negative power fault is analyzed based on the severity and urgency of the fault. The type of wind turbine negative power fault is identified through the degree of the fault, and a corresponding level of wind turbine negative power alarm information is generated based on the fault type. Based on the preprocessed real-time monitoring data of the wind turbine, a comprehensive detection and identification accuracy evaluation index is analyzed. This comprehensive detection and identification accuracy evaluation index is used to evaluate the accuracy of wind turbine fault detection and identification. When the comprehensive detection and identification accuracy evaluation index is less than or equal to the fourth threshold, it indicates that the accuracy of wind turbine fault detection and identification is insufficient, and an early warning is issued for further optimization of the wind turbine negative power fault detection and identification method.

[0009] Furthermore, the specific method for preprocessing the acquired real-time monitoring data, historical negative power fault data, and historical non-fault monitoring data of the wind turbine is as follows: based on a similarity algorithm, similar and duplicate data in the real-time monitoring data and historical negative power fault data of the wind turbine are identified, and the similar and duplicate data are merged; outliers in the data are identified using statistical methods and replaced with missing values; missing values ​​in the data are estimated using interpolation methods; noise in the data is removed using smoothing techniques; data with inconsistent units and numerical ranges are standardized and normalized to transform the data into a unified standard scale.

[0010] Furthermore, the specific method for constructing the wind turbine negative power fault detection model is as follows: Before constructing the wind turbine negative power fault detection model, the number of negative power faults of the wind turbine in the historical negative power fault data is numbered; the internal structure temperature of the wind turbine, the output power at each negative power fault, the measured humidity, and the measured wind speed data are extracted from the historical negative power fault data; at the same time, the average internal structure temperature and average internal structure humidity of the wind turbine are extracted from the historical non-fault monitoring data, and the wind turbine negative power fault detection model is constructed based on these data; the wind turbine negative power fault detection index is analyzed through the wind turbine negative power fault detection model.

[0011] Furthermore, the specific method for analyzing the negative power fault detection index of the current wind turbine is as follows: when the output power of the current wind turbine is detected to be negative, the internal structural temperature, measured humidity, current output power, and measured wind speed data of the wind turbine are extracted from the real-time monitoring data. These data are then input into the wind turbine negative power fault detection model to analyze and calculate the negative power fault detection index of the current wind turbine. The higher the negative power fault detection index, the more likely the current wind turbine is to experience a negative power fault. If the negative power fault detection index of the current wind turbine is greater than or equal to a first threshold, it indicates that the current wind turbine has experienced a negative power fault. If the negative power fault detection index of the current wind turbine is less than the first threshold, it indicates that the current wind turbine has not experienced a negative power fault.

[0012] Furthermore, the specific analysis method for determining the degree of negative power failure of the wind turbine is as follows: The severity and urgency of the negative power failure when the wind turbine experiences a negative power failure are extracted from the real-time monitoring data, and a calculation formula for the degree of negative power failure is constructed. This degree of negative power failure is used to identify the type of negative power failure and generate corresponding wind turbine negative power alarm information. The higher the degree of negative power failure, the higher the processing priority of the current wind turbine negative power failure. The degree of negative power failure is calculated according to the formula: GZ = -sech(YC*χ1 + JC*χ2) + 1, where GZ is the degree of negative power failure when the wind turbine experiences a negative power failure, YC is the severity of the negative power failure, JC is the urgency of the negative power failure, and χ1 and χ2 are the weight percentages of the severity and urgency of the negative power failure in the degree of negative power failure.

[0013] Furthermore, the specific method for analyzing the severity of the negative power fault is as follows: extract the negative power duration, negative power frequency, and output power data when the wind turbine experiences a negative power fault from the real-time monitoring data, and construct a formula for analyzing the severity of the negative power fault; analyze the severity of the negative power fault based on the formula for analyzing the severity of the negative power fault.

[0014] Furthermore, the specific method for analyzing the urgency of the negative power fault is as follows: extract the average vibration amplitude of the wind turbine blades and the average temperature of the internal structure of the wind turbine from the historical non-fault monitoring data, and simultaneously extract the blade vibration amplitude, the current internal structure temperature of the wind turbine, and the wind turbine blade speed data when the current wind turbine experiences a negative power fault from the real-time monitoring data, and construct a formula for analyzing the urgency of the negative power fault based on these data; analyze the urgency of the negative power fault according to the formula for analyzing the urgency of the negative power fault.

[0015] Furthermore, the specific method for generating corresponding levels of wind turbine negative power alarm information based on the wind turbine negative power fault type is as follows: The wind turbine negative power fault type is identified by analyzing the degree of the fault, and a corresponding level of wind turbine negative power alarm is generated based on the identified fault type. The wind turbine negative power alarm includes Level 1, Level 2, and Level 3 alarms. When the degree of the wind turbine negative power fault is less than a second threshold, it is classified as a Level 1 wind turbine negative power fault, and a Level 1 alarm is generated accordingly. When the degree of the wind turbine negative power fault is greater than or equal to the second threshold and less than the third threshold, it is classified as a Level 2 wind turbine negative power fault, and a Level 2 alarm is generated accordingly. When the degree of the wind turbine negative power fault is greater than or equal to the third threshold, it is classified as a Level 3 wind turbine negative power fault, and a Level 3 alarm is generated accordingly.

[0016] Furthermore, the specific method for analyzing the comprehensive detection and identification accuracy evaluation index is as follows: Extract the sensor comprehensive monitoring accuracy, monitoring time interval, and electromagnetic radiation intensity when the wind turbine experiences a negative power fault from the real-time monitoring data, and construct a calculation formula for the comprehensive detection and identification accuracy evaluation index; calculate the comprehensive detection and identification accuracy evaluation index according to the formula; the higher the comprehensive detection and identification accuracy evaluation index, the higher the accuracy of the current wind turbine's negative power fault monitoring and identification; when the comprehensive detection and identification accuracy evaluation index is lower than the fourth threshold, it indicates insufficient accuracy in wind turbine negative power fault detection and identification, and an early warning is issued for further optimization of the wind turbine negative power fault detection and identification method; the calculation formula for the comprehensive detection and identification accuracy evaluation index is: In the formula, JD is the comprehensive detection and identification accuracy evaluation index when the wind turbine has a negative power fault, CJ is the comprehensive sensor monitoring accuracy when the wind turbine has a negative power fault, SJ is the monitoring time interval, FS is the electromagnetic radiation intensity when the wind turbine has a negative power fault, and ε1, ε2 and ε3 are the weight ratios of the comprehensive sensor monitoring accuracy, monitoring time interval and electromagnetic radiation intensity in the comprehensive detection and identification accuracy evaluation index, respectively.

[0017] Furthermore, the specific analysis method for the comprehensive monitoring accuracy of the sensors is as follows: extract the average signal-to-noise ratio and average resolution of all sensors of the wind turbine when the wind turbine experiences a negative power fault from the real-time monitoring data, and construct a comprehensive monitoring accuracy analysis formula; analyze the comprehensive monitoring accuracy of the sensors based on the comprehensive monitoring accuracy analysis formula.

[0018] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0019] 1. By analyzing the real-time monitoring data after wind turbine preprocessing, a comprehensive detection and identification accuracy evaluation index is obtained. When the comprehensive detection and identification accuracy evaluation index is less than or equal to the fourth threshold, it indicates that the accuracy of wind turbine fault detection and identification is insufficient. This leads to an early warning for further optimization of the wind turbine negative power fault detection and identification method, thereby improving the accuracy of wind turbine fault detection and identification and solving the problem of low accuracy of wind turbine fault detection and identification in the existing technology.

[0020] 2. By analyzing the severity of the wind turbine's negative power fault, the type of wind turbine negative power fault can be identified. Based on the identified type of wind turbine negative power fault, a corresponding level of wind turbine negative power alarm prompt can be generated. This enables timely detection and handling of negative power faults, which can prevent serious accidents or unexpected events and help improve the safety of wind turbine operation.

[0021] 3. By acquiring the duration, amplitude, frequency, and output power data of the negative power fault when the wind turbine experiences a negative power fault, the severity of the wind turbine's negative power fault can be analyzed. This allows for the assessment of the impact of the current negative power fault on the wind turbine's performance and safe operation based on the severity of the fault. Attached Figure Description

[0022] Figure 1 A flowchart of the wind turbine negative power fault detection and identification method provided in the embodiments of this application. Detailed Implementation

[0023] This application provides a method for detecting and identifying negative power faults in wind turbines, which solves the problem of low accuracy in wind turbine fault detection and identification in the prior art. By analyzing the real-time monitoring data of the wind turbine after preprocessing, a comprehensive detection and identification accuracy evaluation index is obtained. When the comprehensive detection and identification accuracy evaluation index is less than or equal to the fourth threshold, it indicates that the accuracy of wind turbine fault detection and identification is insufficient, thereby issuing an early warning for further optimization and improvement of the wind turbine negative power fault detection and identification method, thus improving the accuracy of wind turbine fault detection and identification.

[0024] The technical solution in this application embodiment aims to address the problem of low accuracy in wind turbine fault detection and identification. The overall approach is as follows:

[0025] Real-time monitoring data and historical negative power fault data of wind turbines are acquired through wind turbine sensors. The acquired real-time monitoring data and historical negative power fault data are preprocessed to ensure data accuracy and consistency. Based on the preprocessed historical negative power fault data, a wind turbine negative power fault detection model is constructed. When the wind turbine's output power is detected as negative, the preprocessed real-time monitoring data is acquired. Based on the wind turbine negative power fault detection model, a negative power fault detection index is calculated. When the index is higher than a first threshold, it indicates that a negative power fault has occurred; when it is lower, it indicates that no negative power fault has occurred. The duration, amplitude, and frequency of the negative power fault are acquired to analyze the severity of the fault. The vibration amplitude and wind speed under the current negative power fault are also measured. The urgency of a negative power fault is analyzed using data on the average vibration amplitude of the wind turbine blades, the current internal structural temperature of the wind turbine, and the wind turbine blade rotation speed. When a negative power fault is detected, the severity and urgency of the fault are analyzed to determine its degree. Based on this degree, the type of negative power fault is identified, and a corresponding alarm message is generated. Wind turbine maintenance personnel are then notified to handle the fault according to its type. The comprehensive monitoring accuracy, monitoring time interval, and electromagnetic radiation intensity of the sensors are acquired to analyze the comprehensive detection and identification accuracy evaluation index. This index is used to evaluate the results of the wind turbine negative power fault detection. If the comprehensive detection and identification accuracy evaluation index is lower than the fourth threshold, it indicates insufficient accuracy in wind turbine fault detection and identification. Further optimization of the wind turbine negative power fault detection and identification method is then performed to improve the accuracy of wind turbine fault detection and identification.

[0026] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0027] like Figure 1The diagram shows a flowchart of a wind turbine negative power fault detection and identification method provided in this application embodiment. The method includes the following steps: acquiring and preprocessing wind turbine data: acquiring real-time monitoring data of the wind turbine through wind turbine sensors, and simultaneously acquiring historical negative power fault data of the wind turbine; preprocessing the acquired real-time monitoring data and historical negative power fault data of the wind turbine; constructing a wind turbine negative power fault detection model: constructing a wind turbine negative power fault detection model based on the preprocessed historical negative power fault data; analyzing the probability of a current wind turbine experiencing a negative power fault: when the output power of the current wind turbine is detected to be negative, inputting the preprocessed real-time monitoring data of the current wind turbine into the wind turbine negative power fault detection model, and analyzing the wind turbine negative power fault detection index of the current wind turbine. The wind turbine negative power fault detection index is used to detect the probability of a current wind turbine experiencing a negative power fault; The system analyzes the severity of negative power faults in current wind turbines and generates corresponding levels of negative power alarm information. When a negative power fault is detected in a wind turbine, the severity and urgency of the fault are analyzed to determine the fault level. The fault type is then identified based on the fault level, and a corresponding level of negative power alarm information is generated. The system also analyzes the comprehensive detection and identification accuracy evaluation index of the current wind turbine. Based on the pre-processed real-time monitoring data, a comprehensive detection and identification accuracy evaluation index is calculated. This index is used to assess the accuracy of wind turbine fault detection and identification. When the comprehensive detection and identification accuracy evaluation index is less than or equal to the fourth threshold, it indicates insufficient accuracy in wind turbine fault detection and identification, and a warning is issued for further optimization of the wind turbine negative power fault detection and identification method.

[0028] In this embodiment, when the comprehensive detection and identification accuracy evaluation index is less than or equal to the fourth threshold, it indicates that the accuracy of wind turbine fault detection and identification is insufficient, and the wind turbine negative power fault detection and identification method needs further optimization and improvement. For cases where the comprehensive detection and identification accuracy evaluation index is low, further analysis can be conducted. Through in-depth analysis, new possible negative power fault modes or characteristics can be discovered. This information can be used to optimize the model and improve the detection accuracy of negative power faults. More feedback mechanisms are added to the early warning system to record relevant information from the early warning for further analysis and optimization. The threshold of the early warning system is adjusted and optimized to ensure accurate judgment on the accuracy of wind turbine fault detection and identification.

[0029] Furthermore, the specific method for preprocessing the acquired real-time monitoring data, historical negative power fault data, and historical non-fault monitoring data of wind turbines is as follows: based on the similarity algorithm, similar and duplicate data in the real-time monitoring data and historical negative power fault data of wind turbines are identified, and the similar and duplicate data are merged; outliers in the data are identified using statistical methods and replaced with missing values; missing values ​​in the data are estimated using interpolation methods; noise in the data is removed using smoothing techniques; data with inconsistent units and numerical ranges are standardized and normalized to transform the data into a unified standard scale.

[0030] In this embodiment, the similarity algorithm is an algorithm for calculating the similarity between data. Appropriate similarity metrics, such as Euclidean distance, cosine similarity, and Manhattan distance, are selected to identify similar and duplicate data in the real-time monitoring data and historical negative power fault data of Fengdi Anji. These similar and duplicate data are then merged. Outliers in the data are identified using statistical methods and replaced with missing values. Statistical methods include outlier detection, Z-score methods, and box plots to identify outliers. Digital signal processing techniques can be applied to remove high-frequency or low-frequency noise from the data. Outliers can be replaced with missing values, data imputation can be performed using interpolation methods, or samples containing outliers can be directly deleted. If the number of missing values ​​is small and the sample size is large enough, samples containing missing values ​​can be directly deleted. For cases with many missing values, interpolation methods, such as linear interpolation, polynomial interpolation, and KNN interpolation, can be used to estimate the missing values. Additionally, other relevant data can be used to infer missing values ​​based on the similarity or time series relationship of wind turbine data.

[0031] Furthermore, the specific method for constructing the wind turbine negative power fault detection model is as follows: Before constructing the wind turbine negative power fault detection model, the number of negative power faults of the wind turbine in the historical negative power fault data is numbered; the internal structure temperature of the wind turbine, the output power at each negative power fault, the measured humidity, and the measured wind speed data are extracted from the historical negative power fault data; at the same time, the average internal structure temperature and average internal structure humidity of the wind turbine are extracted from the historical non-fault monitoring data, and the wind turbine negative power fault detection model is constructed based on these data; the wind turbine negative power fault detection index is analyzed through the wind turbine negative power fault detection model.

[0032] In this embodiment, the internal structural temperature of the wind turbine is obtained through a temperature sensor installed inside the wind turbine. Common temperature sensors include thermocouples, thermistors, and infrared temperature sensors. Temperature has a certain impact on the operation of the wind turbine and the performance of its components. High temperatures can cause the temperature of the internal structural components of the wind turbine to rise, potentially leading to overheating or damage. Low temperatures may affect the effectiveness of the wind turbine's lubrication and cooling systems. Humidity has a certain impact on the durability and corrosiveness of the wind turbine's components and parts. High humidity environments may lead to corrosion and frost problems in the internal structure of the wind turbine, affecting its operation and performance. Low humidity environments may increase the risk of electrostatic discharge. The wind turbine negative power fault detection model is as follows: In the formula, GJ i WD represents the wind turbine negative power fault detection index for the i-th historical negative power fault, where i is the historical negative power fault number, i = 1, 2, 3, ..., S, and S is the total number of historical negative power faults. i Let LW be the internal structural temperature of the wind turbine during the i-th historical negative power fault, and GL be the average internal structural temperature of the wind turbine. i Let SD be the output power of the i-th historical negative power fault. i Let represent the measured humidity during the i-th historical negative power fault, LS represent the average internal temperature of the wind turbine, and FS represent... i Let α1 be the measured wind speed of the i-th historical negative power fault, and α2, α3 and α4 be the weight percentages of the temperature difference between the internal structure temperature of the wind turbine and the average internal structure temperature of the wind turbine, the output power, the humidity difference between the measured humidity and the average humidity of the internal structure of the wind turbine, and the measured wind speed in the wind turbine negative power fault detection index, respectively.

[0033] Furthermore, the specific method for analyzing the negative power fault detection index of the current wind turbine is as follows: When the output power of the current wind turbine is detected to be negative, the internal structural temperature, measured humidity, current output power, and measured wind speed data of the wind turbine are extracted from the real-time monitoring data. These data are then input into the wind turbine negative power fault detection model to analyze and calculate the wind turbine negative power fault detection index. The higher the wind turbine negative power fault detection index, the more likely the current wind turbine is to experience a negative power fault. If the wind turbine negative power fault detection index is greater than or equal to the first threshold, it indicates that the current wind turbine has experienced a negative power fault. If the wind turbine negative power fault detection index is less than the first threshold, it indicates that the current wind turbine has not experienced a negative power fault.

[0034] In this embodiment, output power is the core indicator of a wind turbine, directly reflecting its power generation capacity. Changes in output power can provide important signals to determine whether a wind turbine has a negative power fault. However, when the current output power of a wind turbine is negative, it may be during the wind turbine shutdown and startup process, and does not necessarily mean that a negative power fault has occurred. Therefore, if the current output power of a wind turbine is detected to be negative, it is necessary to further analyze the wind turbine negative power fault detection index to determine whether a negative power fault has actually occurred.

[0035] Furthermore, the specific analysis method for determining the degree of negative power failure of wind turbines is as follows: The severity and urgency of the negative power failure when it occurs are extracted from real-time monitoring data, and a calculation formula for the degree of negative power failure is constructed. This degree of negative power failure is used to identify the type of negative power failure and generate corresponding alarm information. The higher the degree of negative power failure, the higher the priority for handling the current negative power failure. The degree of negative power failure is calculated according to the formula: GZ = -sech(YC*χ1 + JC*χ2) + 1, where GZ is the degree of negative power failure when it occurs, YC is the severity of the negative power failure, JC is the urgency of the negative power failure, and χ1 and χ2 are the weighted proportions of the severity and urgency of the negative power failure in the degree of negative power failure.

[0036] In this embodiment, the severity of a wind turbine negative power fault refers to the degree of impact of the current negative power fault on the performance and safe operation of the wind turbine. The severity can be assessed by the power loss caused by the fault. If the negative power fault causes the wind turbine to be unable to generate electricity normally or the output power is significantly reduced, then its severity will be very high. The urgency of a negative power fault refers to how quickly measures need to be taken to repair the negative power fault. The urgency can be assessed by the impact of the negative power fault on the operational stability and reliability of the wind turbine. If the negative power fault causes the wind turbine to be in an unstable state or may further lead to the occurrence of other faults, then its urgency will be very high.

[0037] Furthermore, the specific analysis method for the severity of negative power faults is as follows: extract the duration, frequency, and output power data of the negative power fault when the wind turbine experiences a negative power fault from the real-time monitoring data, and construct a formula for analyzing the severity of negative power faults; analyze the severity of the negative power faults according to the formula, and the higher the severity of the negative power faults, the more serious the current wind turbine negative power fault problem.

[0038] In this embodiment, the formula for calculating the severity of a negative power fault when the current wind turbine experiences a negative power fault is: YC=-sech(CS*β1+FD*β2+PL*β3+|GL|*β4)+1, where YC is the severity of the current wind turbine's negative power fault, CS is the duration of the negative power fault when the current wind turbine experiences a negative power fault, FD is the amplitude of the negative power fault when the current wind turbine experiences a negative power fault, PL is the frequency of the negative power fault when the current wind turbine experiences a negative power fault, GL is the output power when the current wind turbine experiences a negative power fault, and β1, β2, β3 and β4 are the weight ratios of the duration of the negative power fault, the amplitude of the negative power fault, the frequency of the negative power fault, and the output power in the severity of the negative power fault, respectively.

[0039] Furthermore, the specific analysis method for the urgency of negative power faults is as follows: extract the average vibration amplitude of wind turbine blades and the average temperature of the internal structure of the wind turbine from historical non-fault monitoring data; simultaneously extract the blade vibration amplitude, the current internal structure temperature of the wind turbine, and the wind turbine blade speed data when a negative power fault occurs from real-time monitoring data; construct a formula for analyzing the urgency of negative power faults based on these data; analyze the urgency of negative power faults according to the formula; the higher the urgency of negative power faults, the more urgent the current negative power fault problem of the wind turbine.

[0040] In this embodiment, the blade vibration amplitude when the wind turbine experiences a negative power fault is obtained through vibration sensors on the wind turbine blades; the internal structure temperature of the wind turbine is obtained through temperature sensors installed inside the wind turbine, such as thermocouples, thermistors, and infrared temperature sensors; the formula for analyzing the urgency of the negative power fault when the wind turbine experiences a negative power fault is as follows: In the formula, JC represents the urgency of the negative power fault when the wind turbine experiences a negative power fault, ZF represents the blade vibration amplitude when the wind turbine experiences a negative power fault, LF represents the average vibration amplitude of the wind turbine blades, WD represents the internal structure temperature of the wind turbine when the wind turbine experiences a negative power fault, LW represents the average internal structure temperature of the wind turbine, ZS represents the blade rotation speed of the wind turbine when the wind turbine experiences a negative power fault, and δ1, δ2, and δ3 represent the weighting percentages of the difference between the current vibration amplitude and the theoretical vibration amplitude of the wind turbine, the internal structure temperature and the average internal structure temperature of the wind turbine, and the blade rotation speed in the urgency of the negative power fault, respectively.

[0041] Furthermore, the specific method for generating corresponding levels of wind turbine negative power alarm information based on the type of wind turbine negative power fault is as follows: The type of wind turbine negative power fault is identified by analyzing the severity of the fault, and a corresponding level of wind turbine negative power alarm is generated based on the identified fault type. The wind turbine negative power alarm includes Level 1, Level 2, and Level 3 alarms, with the priority of negative power fault handling increasing sequentially from Level 1 to Level 3. When the severity of the wind turbine negative power fault is less than the second threshold, it is classified as a Level 1 wind turbine negative power fault, and a Level 1 alarm is generated accordingly. When the severity of the wind turbine negative power fault is greater than or equal to the second threshold and less than the third threshold, it is classified as a Level 2 wind turbine negative power fault, and a Level 2 alarm is generated accordingly. When the severity of the wind turbine negative power fault is greater than or equal to the third threshold, it is classified as a Level 3 wind turbine negative power fault, and a Level 3 alarm is generated accordingly.

[0042] In this embodiment, the type of negative power fault of the wind turbine is identified by analyzing the degree of negative power fault. Based on the identified negative power fault type, a corresponding level of negative power alarm prompt is generated. This helps to quickly take measures to repair or resolve the negative power fault, avoid further deterioration of the fault or greater impact on the operation of the wind turbine, and help improve the safety of wind turbine operation and protect the safety of personnel and equipment.

[0043] Furthermore, the specific method for analyzing the comprehensive detection and identification accuracy evaluation index is as follows: Extract the sensor comprehensive monitoring accuracy, monitoring time interval, and electromagnetic radiation intensity when the wind turbine experiences a negative power fault from real-time monitoring data, and construct a formula for calculating the comprehensive detection and identification accuracy evaluation index; calculate the comprehensive detection and identification accuracy evaluation index according to this formula; the higher the comprehensive detection and identification accuracy evaluation index, the higher the accuracy of the current wind turbine's negative power fault monitoring and identification; when the comprehensive detection and identification accuracy evaluation index is below the fourth threshold, it indicates insufficient accuracy in detecting and identifying the wind turbine's negative power fault, issuing a warning for further optimization of the wind turbine's negative power fault detection and identification method; the formula for calculating the comprehensive detection and identification accuracy evaluation index is: In the formula, JD is the comprehensive detection and identification accuracy evaluation index when the wind turbine has a negative power fault, CJ is the comprehensive sensor monitoring accuracy when the wind turbine has a negative power fault, SJ is the monitoring time interval, FS is the electromagnetic radiation intensity when the wind turbine has a negative power fault, and ε1, ε2 and ε3 are the weight ratios of the comprehensive sensor monitoring accuracy, monitoring time interval and electromagnetic radiation intensity in the comprehensive detection and identification accuracy evaluation index, respectively.

[0044] In this embodiment, the wind turbine uses multiple sensors, such as temperature sensors, humidity sensors, and vibration sensors, to monitor its operation in real time. The overall sensor monitoring accuracy when the wind turbine experiences a negative power fault refers to the accuracy of the real-time monitoring data acquired by multiple sensors when the wind turbine experiences a negative power fault. When the wind turbine experiences a negative power fault, it may generate electromagnetic radiation, which may mix with electromagnetic radiation from other environments and cause interference. Therefore, electromagnetic radiation can affect the accurate detection and identification of negative power faults. The electromagnetic radiation when the wind turbine experiences a negative power fault is acquired using electromagnetic radiation testers, magnetic field detectors, etc. Based on the feedback results of the comprehensive detection and identification accuracy evaluation index, the existing wind turbine negative power fault detection model is optimized by adjusting the model parameters, modifying the model structure, or using a more advanced algorithm model to improve the accuracy of fault detection.

[0045] Furthermore, the specific analysis method for the comprehensive monitoring accuracy of sensors is as follows: extract the average signal-to-noise ratio and average resolution of all sensors of the wind turbine when a negative power fault occurs from the real-time monitoring data, and construct a formula for analyzing the comprehensive monitoring accuracy of sensors; analyze the comprehensive monitoring accuracy of sensors according to the formula; the higher the comprehensive monitoring accuracy of sensors, the more accurate the data monitored by the wind turbine sensors.

[0046] In this embodiment, the formula for analyzing the overall monitoring accuracy of the wind turbine sensors when a negative power fault occurs is as follows: In the formula, CJ is the overall monitoring accuracy of the sensors when the current wind turbine experiences a negative power fault, XZ is the average signal-to-noise ratio of all sensors when the current wind turbine experiences a negative power fault, FB is the average resolution of all sensors when the current wind turbine experiences a negative power fault, φ1 and φ2 are the weight ratios of the average signal-to-noise ratio of the sensors and the average resolution of the sensors in the overall monitoring accuracy of the sensors, respectively, and e is a natural constant.

[0047] In summary, by utilizing the above-mentioned technical solution of the present invention, a comprehensive detection and identification accuracy evaluation index is analyzed based on the real-time monitoring data after wind turbine preprocessing. When the comprehensive detection and identification accuracy evaluation index is less than or equal to the fourth threshold, it indicates that the accuracy of wind turbine fault detection and identification is insufficient, thereby issuing an early warning for further optimization of the wind turbine negative power fault detection and identification method, and thus improving the accuracy of wind turbine fault detection and identification.

[0048] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the specific device can be divided into different functional modules to complete all or part of the functions described above.

[0049] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0050] In the embodiments covered by this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the division of modules is merely a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connections shown or discussed may be indirect coupling or communication connections through some interfaces, devices, or modules, or they may be electrical, mechanical, or other forms of connection.

[0051] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0052] Furthermore, the functional modules in the various embodiments of this application can be implemented either in hardware or as software functional modules. If these functional modules are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a computer program product, which includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer program product is stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid state disks (SSDs)).

[0053] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

Claims

1. A method for detecting and identifying negative power faults in wind turbines, characterized in that, Includes the following steps: Real-time monitoring data of wind turbines is acquired through wind turbine sensors, along with historical negative power fault data and historical monitoring data. The acquired real-time monitoring data, historical negative power fault data, and historical non-fault monitoring data of wind turbines are then preprocessed. Based on the preprocessed historical negative power fault data, a wind turbine negative power fault detection model is constructed. When the output power of the current wind turbine is detected to be negative, the pre-processed real-time monitoring data of the current wind turbine is input into the wind turbine negative power fault detection model to analyze the wind turbine negative power fault detection index of the current wind turbine. The wind turbine negative power fault detection index is used to detect the possibility of the current wind turbine experiencing a negative power fault. When a negative power fault is detected in the current wind turbine, the severity and urgency of the negative power fault are analyzed to determine the degree of the wind turbine's negative power fault. The type of wind turbine's negative power fault is identified by the degree of the wind turbine's negative power fault. The higher the degree of the wind turbine's negative power fault, the higher the processing priority of the current wind turbine's negative power fault. The degree of negative power failure of the wind turbine is calculated according to the formula for calculating the degree of negative power failure of the wind turbine. The formula for calculating the degree of negative power failure of the wind turbine is as follows: ， In the formula, This represents the severity of a negative power fault in a wind turbine when it experiences a negative power fault. This represents the severity of a negative power fault when the wind turbine experiences a negative power fault. This indicates the urgency level of a negative power fault when it occurs in a wind turbine. and These represent the weighting percentages of the severity and urgency of negative power faults in the overall severity of wind turbine negative power faults; corresponding levels of wind turbine negative power alarm information are generated based on the type of wind turbine negative power fault. Based on the pre-processed real-time monitoring data of the wind turbine, a comprehensive detection and identification accuracy evaluation index is analyzed. This index is used to evaluate the accuracy of wind turbine fault detection and identification. When the comprehensive detection and identification accuracy evaluation index is less than or equal to the fourth threshold, it indicates that the accuracy of wind turbine fault detection and identification is insufficient, and an early warning is issued to further optimize the wind turbine negative power fault detection and identification method. The sensor comprehensive monitoring accuracy, monitoring time interval, and electromagnetic radiation intensity when the wind turbine experiences a negative power fault are extracted from the real-time monitoring data, and a formula for calculating the comprehensive detection and identification accuracy evaluation index is constructed. The formula for calculating the comprehensive detection and recognition accuracy evaluation index is as follows: ， In the formula, This is an evaluation index for the comprehensive detection and identification accuracy when a wind turbine experiences a negative power fault. To improve the overall accuracy of sensor monitoring when a wind turbine experiences a negative power fault, For monitoring time intervals, This represents the electromagnetic radiation intensity when a wind turbine experiences a negative power fault. , and These represent the weighting percentages of sensor comprehensive monitoring accuracy, monitoring time interval, and electromagnetic radiation intensity in the comprehensive detection and identification accuracy evaluation index.

2. The wind turbine negative power fault detection and identification method as described in claim 1, characterized in that: The specific method for preprocessing the acquired real-time monitoring data, historical negative power fault data, and historical non-fault monitoring data of the wind turbine is as follows: Based on the similarity algorithm, similar and duplicate data in the real-time monitoring data and historical negative power fault data of wind turbines are identified, and the similar and duplicate data are merged. Outliers in the data are identified using statistical methods and replaced with missing values. Missing values ​​are estimated using interpolation methods, and noise in the data is removed using smoothing techniques. Data with inconsistent units and numerical ranges is standardized and normalized to transform it into a unified standard scale.

3. The wind turbine negative power fault detection and identification method as described in claim 1, characterized in that: The specific method for constructing the wind turbine negative power fault detection model is as follows: Before constructing the wind turbine negative power fault detection model, the number of negative power faults of the wind turbine in the historical negative power fault data is numbered. The internal structure temperature of the wind turbine, the output power at each negative power fault, the measured humidity and the measured wind speed data are extracted from the historical negative power fault data. At the same time, the average internal structure temperature and average internal structure humidity of the wind turbine are extracted from the historical non-fault monitoring data. Based on these data, a wind turbine negative power fault detection model is constructed. The wind turbine negative power fault detection index was analyzed using a wind turbine negative power fault detection model.

4. The wind turbine negative power fault detection and identification method as described in claim 1, characterized in that: The specific method for analyzing the wind turbine negative power fault detection index is as follows: When the current output power of the wind turbine is detected to be negative, the internal structure temperature, measured humidity, current output power and measured wind speed of the wind turbine are extracted from the real-time monitoring data. These data are then input into the wind turbine negative power fault detection model to analyze the wind turbine negative power fault detection index. The higher the wind turbine negative power fault detection index, the more likely the wind turbine is to experience a negative power fault. If the wind turbine negative power fault detection index of the current wind turbine is greater than or equal to the first threshold, it indicates that the current wind turbine has a negative power fault. If the wind turbine negative power fault detection index is less than the first threshold, it means that the wind turbine has not experienced a negative power fault.

5. The wind turbine negative power fault detection and identification method as described in claim 1, characterized in that: The specific analytical method for analyzing the degree of negative power failure of the exhaust motor is as follows: The severity and urgency of the negative power fault when the wind turbine experiences a negative power fault are extracted from the real-time monitoring data. A formula for calculating the degree of the negative power fault of the wind turbine is constructed. The degree of the negative power fault of the wind turbine is used to identify the type of negative power fault of the wind turbine and generate corresponding type of wind turbine negative power alarm information.

6. The wind turbine negative power fault detection and identification method as described in claim 5, characterized in that: The specific method for analyzing the severity of the negative power fault is as follows: Extract the duration, frequency, and output power of the negative power fault when the wind turbine experiences a negative power fault from the real-time monitoring data, and construct a formula for analyzing the severity of the negative power fault. The severity of negative power faults is analyzed based on the formula for analyzing the severity of negative power faults.

7. The wind turbine negative power fault detection and identification method as described in claim 5, characterized in that: The specific method for analyzing the urgency of the negative power fault is as follows: The average vibration amplitude of the wind turbine blades and the average temperature of the internal structure of the wind turbine are extracted from the historical non-fault monitoring data. At the same time, the blade vibration amplitude, the current internal structure temperature of the wind turbine, and the wind turbine blade speed data when the wind turbine experiences a negative power fault are extracted from the real-time monitoring data. Based on these data, a formula for analyzing the urgency of the negative power fault is constructed. The urgency of a negative power fault is analyzed based on the formula for analyzing the urgency of a negative power fault.

8. The wind turbine negative power fault detection and identification method as described in claim 1, characterized in that: The specific method for generating corresponding levels of wind turbine negative power alarm information based on the wind turbine negative power fault type is as follows: The type of wind turbine negative power fault is identified by analyzing the degree of wind turbine negative power fault, and a corresponding level of wind turbine negative power alarm prompt is generated based on the identified wind turbine negative power fault type. The wind turbine negative power alarm message includes a first-level wind turbine negative power alarm message, a second-level wind turbine negative power alarm message, and a third-level wind turbine negative power alarm message; When the degree of wind turbine negative power failure is less than the second threshold, it is classified as a first-class wind turbine negative power failure, and a first-level wind turbine negative power alarm message is generated accordingly. When the degree of wind turbine negative power failure is greater than or equal to the second threshold and less than the third threshold, it is classified as a second-class wind turbine negative power failure, and a corresponding second-level wind turbine negative power alarm message is generated. When the degree of negative power failure of a wind turbine is greater than or equal to the third threshold, it is classified as a third-class wind turbine negative power failure, and a corresponding level three wind turbine negative power alarm message is generated.

9. The wind turbine negative power fault detection and identification method as described in claim 1, characterized in that: The specific method for analyzing the comprehensive detection and recognition accuracy evaluation index is as follows: The comprehensive detection and recognition accuracy evaluation index is calculated based on the formula for calculating the comprehensive detection and recognition accuracy evaluation index. The higher the comprehensive detection and identification accuracy evaluation index, the more accurate the current negative power fault monitoring and identification of wind turbines is. When the comprehensive detection and identification accuracy evaluation index is lower than the fourth threshold, it indicates that the accuracy of wind turbine negative power fault detection and identification is insufficient, and an early warning is issued to further optimize the wind turbine negative power fault detection and identification method.

10. The wind turbine negative power fault detection and identification method as described in claim 9, characterized in that: The specific analysis method for the overall monitoring accuracy of the aforementioned sensors is as follows: The average signal-to-noise ratio and average resolution of all sensors of the wind turbine when the wind turbine experiences a negative power fault are extracted from the real-time monitoring data, and a comprehensive sensor monitoring accuracy analysis formula is constructed. The accuracy of sensor comprehensive monitoring is analyzed based on the formula for analyzing the accuracy of sensor comprehensive monitoring.

Images