Safety control method, system, device and medium applied to wind turbine generator system

Abstract

The application relates to the technical field of fault protection, and discloses a safety control method, system, equipment and medium applied to a wind turbine generator set, which comprises the following steps: setting a monitoring node and performing initialization configuration; the monitoring node collects voltage data of a battery in real time, and uses a sliding time window algorithm to obtain a voltage change trend; the voltage change trend is compared with a preset under-voltage threshold value, and an under-voltage early warning mechanism is triggered according to a comparison result; a main controller combines data of multiple monitoring nodes, judges whether a protection measure is started or not through a voltage prediction model; if a prediction result shows that the voltage will continuously decrease, a hierarchical protection strategy is executed; in a protection mode, a voltage recovery condition is monitored by increasing a sampling frequency, and when the voltage recovers to a safety range and a voltage recovery rate meets a condition, the protection mode is exited. The application significantly improves the stability and safety of a variable pitch system, prolongs the service life of a backup power supply, and provides protection for the safe operation of the wind turbine generator set.

Description

Technical Field

[0001] This invention relates to the field of pitch system safety control technology, and in particular to safety control methods, systems, equipment and media applied to wind turbine generator sets. Background Technology

[0002] With the rapid development of wind power technology, the pitch system, as the core control unit of a wind turbine generator, directly affects the overall performance and safety of the turbine. Pitch systems are typically equipped with backup battery packs to maintain short-term system operation in the event of external power failure. However, in actual operation, pitch batteries are prone to voltage drops due to factors such as sudden load changes, ambient temperature variations, and battery aging. If these drops are not detected and protected in time, they may lead to over-discharge, damage, or even pitch system failure, threatening the safety of the turbine.

[0003] Currently, the management of backup power for pitch control systems focuses primarily on charging efficiency and battery status maintenance, lacking a rapid diagnostic and tiered protection mechanism for undervoltage faults. Existing technologies often rely on a single voltage threshold for undervoltage detection, which is prone to false alarms or missed alarms due to instantaneous voltage fluctuations, failing to accurately identify the actual degradation trend of the battery. Furthermore, the protection measures after undervoltage occurs are relatively simple and lack dynamic response based on actual system conditions. This can lead to battery damage due to untimely protection or reduced generator efficiency due to premature load shedding. Therefore, it is necessary to design a method that can monitor battery voltage status in real time, intelligently identify undervoltage risks, and initiate tiered protection based on actual system conditions. This would improve the reliability of backup power for pitch control systems, extend battery life, and ensure the safe and stable operation of wind turbine generators. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a safety control method, system, equipment and medium for wind turbine generator sets, which solves the problem that the existing technology lacks an effective dynamic monitoring and graded response mechanism in the protection of undervoltage faults of pitch batteries, making it difficult to accurately identify the real undervoltage risk and achieve rapid and appropriate protection.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a safety control method for wind turbine generator sets, comprising:

[0008] A monitoring node is set up and initialized on each battery pack of the pitch system. The monitoring node collects the battery voltage data in real time and uses a sliding time window algorithm to obtain the voltage change trend.

[0009] The voltage change trend is compared with a preset undervoltage threshold. Based on the comparison result, an undervoltage warning mechanism is triggered, and the undervoltage warning information is sent to the wind turbine main controller.

[0010] After receiving the undervoltage warning information, the main controller of the wind turbine generator set combines data from multiple monitoring nodes and uses a voltage prediction model to determine whether to activate protection measures.

[0011] If the voltage prediction model predicts that the voltage will continue to drop, a graded protection strategy is implemented. In protection mode, the voltage recovery is monitored by increasing the sampling frequency. When the voltage recovers to a safe range and the voltage recovery rate meets the conditions, the protection mode is exited.

[0012] As a preferred embodiment of the safety control method for wind turbine generator sets described in this invention, the step of obtaining the voltage change trend using a sliding time window algorithm includes:

[0013] The real-time voltage data is stored in a fixed-length buffer in chronological order. The length of the buffer is determined by the sampling frequency and the window time.

[0014] Whenever a new voltage sampling data is acquired and stored in the buffer, the historical voltage data with the longest storage time in the buffer is removed to maintain a constant total amount of data in the buffer.

[0015] After each data update, the average voltage and standard deviation within the time window are calculated based on all voltage data in the current buffer. The average voltage and standard deviation are then used as the output of the voltage change trend for subsequent undervoltage judgment. The average voltage is used to characterize the overall level of the current voltage, and the standard deviation is used to characterize the degree of fluctuation of the current voltage.

[0016] As a preferred embodiment of the safety control method for wind turbine generator sets described in this invention, the step of triggering the undervoltage warning mechanism based on the comparison result includes:

[0017] If the average voltage is less than a preset undervoltage threshold, the voltage drop rate is calculated and compared with a preset rate threshold.

[0018] When the average voltage is lower than the preset undervoltage threshold and the voltage drop rate exceeds the preset rate threshold, the undervoltage warning mechanism is triggered.

[0019] After determining that the undervoltage warning mechanism has been triggered, an undervoltage warning message containing the battery pack identifier, current voltage value, voltage drop rate and timestamp is generated, and the undervoltage warning message is sent to the main controller of the wind turbine generator set based on the communication network.

[0020] As a preferred embodiment of the safety control method for wind turbine generator sets described in this invention, the step of determining whether to activate protection measures through a voltage prediction model includes:

[0021] The wind turbine generator main controller receives undervoltage warning information from the monitoring nodes and simultaneously collects voltage data from all other monitoring nodes within the same time window. It then calculates the overall voltage level of the entire pitch battery system using a weighted average method.

[0022] Based on the comprehensive voltage level and historical voltage data time series, a voltage prediction model is constructed using a linear regression algorithm. The voltage prediction model is then used to predict the voltage change trend in the future and output the slope of the voltage change trend.

[0023] If the slope of the voltage change trend is less than the slope threshold, it is determined that the voltage has a continuous downward trend and protection measures need to be activated. Otherwise, the voltage trend is determined to be stable, protection measures are not activated for the time being, and the system status is monitored continuously.

[0024] As a preferred embodiment of the safety control method for wind turbine generator sets described in this invention, the implementation of the hierarchical protection strategy includes:

[0025] The trigger condition for a Level 1 response is a single battery pack that is continuously undervoltage and the predicted trend is downward; the trigger condition for a Level 2 response is multiple battery packs that are simultaneously undervoltage or the voltage does not recover after the Level 1 response; the trigger condition for a Level 3 response is a continuous deterioration of the system voltage and a prediction that it will endanger system safety.

[0026] The main controller of the wind turbine generator set determines the protection level that should be activated at the current time;

[0027] If the Level 1 response is met, the load on the undervoltage battery pack is reduced by adjusting the power distribution of non-critical equipment in the pitch system. If the voltage does not recover to a safe range within a set time after the Level 1 response is executed, the response is upgraded to Level 2, and the backup power supply is activated to power the critical control circuits of the pitch system. If the system voltage still does not improve after the Level 2 response or the prediction results show that the voltage will continue to drop to a dangerous threshold, the Level 3 response is activated, non-essential loads are disconnected, and a system shutdown command is executed.

[0028] During any level of protection response, the wind turbine main controller continuously monitors battery voltage changes at an increased sampling frequency. If the voltage shows an upward trend after the protection action, the current protection level is maintained and the system continuously assesses whether the protection can be downgraded or deactivated. If the voltage continues to deteriorate, the protection level is gradually upgraded according to the conditions.

[0029] As a preferred embodiment of the safety control method for wind turbine generator sets described in this invention, when the voltage recovers to a safe range and the voltage recovery rate meets the conditions, exiting the protection mode includes:

[0030] In protection mode, the main controller of the wind turbine generator monitors the battery voltage at a higher sampling frequency than in normal mode. When the battery voltage is detected to be continuously higher than the lower limit of the safe voltage and remains so for a set period of time, it is determined that the voltage has recovered to the safe range.

[0031] When the voltage recovers to a safe range, the voltage recovery rate is calculated to determine whether to completely exit the protection mode. If the voltage recovery rate meets the preset conditions, the battery recovery capability is determined to be good, the wind turbine main controller control system exits the protection mode and resumes normal monitoring. Otherwise, the protection mode is maintained, monitoring continues, and the recovery status is periodically reassessed.

[0032] As a preferred embodiment of the safety control method for wind turbine generator sets described in this invention, the main controller of the wind turbine generator set simultaneously processes the status information of multiple battery packs through a multi-task scheduling mechanism, including:

[0033] Calculate priority parameters for each battery pack status monitoring task;

[0034] The wind turbine generator main controller sorts the monitoring tasks of multiple battery packs based on the priority parameters, and processes the voltage data, early warning information and protection decisions of each battery pack in order from high to low in each scheduling cycle.

[0035] The priority parameters of each battery pack are dynamically updated according to changes in system operating status to ensure that battery packs with high abnormality levels are continuously processed with priority; when protection mode is activated or deactivated, the priority of relevant battery packs is recalculated to adapt to changes in system status.

[0036] Secondly, the present invention provides a safety control system for wind turbine generator sets, comprising:

[0037] The trend analysis module is used to set up monitoring nodes and perform initial configuration on each battery pack of the pitch system. The monitoring nodes collect battery voltage data in real time and use a sliding time window algorithm to obtain the voltage change trend.

[0038] The undervoltage warning trigger module is used to compare the voltage change trend with a preset undervoltage threshold, trigger the undervoltage warning mechanism based on the comparison result, and send the undervoltage warning information to the wind turbine main controller.

[0039] A centralized risk prediction module is used by the main controller of the wind turbine generator to determine whether to activate protection measures after receiving the undervoltage warning information and combining data from multiple monitoring nodes through a voltage prediction model.

[0040] The undervoltage protection execution module is used to execute a graded protection strategy if the voltage prediction result of the voltage prediction model shows that the voltage will continue to drop; in the protection mode, the voltage recovery is monitored by increasing the sampling frequency, and the protection mode is exited when the voltage recovers to a safe range and the voltage recovery rate meets the conditions.

[0041] Thirdly, the present invention provides an electronic device, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor executes the computer-executable instructions to implement the steps of a safety control method applied to a wind turbine generator set.

[0042] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement steps of a safety control method applied to a wind turbine generator set.

[0043] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention achieves real-time monitoring and rapid response to the voltage status of the pitch battery through a combination of distributed monitoring and centralized control. The sliding time window algorithm effectively reduces the impact of data fluctuations on monitoring results and improves the accuracy of undervoltage detection; the two-level criterion design avoids false alarms caused by single-condition triggering, enhancing system reliability; the hierarchical protection strategy dynamically adjusts according to actual operating conditions, maximizing operational efficiency while ensuring system safety. Furthermore, the linear regression-based prediction model can identify potential undervoltage risks in advance, providing a basis for proactive protection. The combination of technical means provided by this invention significantly improves the stability and safety of the pitch system, extends the service life of the backup power supply, and provides strong protection for the safe operation of wind turbine generators. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of the overall flow logic of a safety control method for wind turbine generator sets provided in one embodiment of the present invention. Detailed Implementation

[0046] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0047] Example 1, referring to Figure 1 As one embodiment of the present invention, a safety control method for wind turbine generator sets is provided, such as... Figure 1 The specific steps shown are as follows:

[0048] S100: Set up monitoring nodes on each battery pack of the pitch system and perform initial configuration. The monitoring nodes collect battery voltage data in real time and use a sliding time window algorithm to obtain the voltage change trend.

[0049] S200: Compares the voltage change trend with the preset undervoltage threshold, triggers the undervoltage warning mechanism based on the comparison result, and sends the undervoltage warning information to the wind turbine main controller;

[0050] S300: After receiving the undervoltage warning information, the main controller of the wind turbine generator combines the data from multiple monitoring nodes and uses a voltage prediction model to determine whether to activate protection measures.

[0051] S400: If the voltage prediction model shows that the voltage will continue to drop, a graded protection strategy will be implemented. In protection mode, the voltage recovery will be monitored by increasing the sampling frequency. When the voltage recovers to a safe range and the voltage recovery rate meets the conditions, the protection mode will be exited.

[0052] It should be noted that, to address the lack of an effective dynamic monitoring and graded response mechanism in existing technologies for undervoltage fault protection of pitch batteries, which makes it difficult to accurately identify actual undervoltage risks and achieve rapid and appropriate protection, steps S100-S400 above achieve real-time monitoring and rapid response of the pitch battery voltage status through a combination of distributed monitoring and centralized control. The sliding time window algorithm effectively reduces the impact of data fluctuations on monitoring results, improving the accuracy of undervoltage detection; the two-level criterion design avoids false alarms caused by single-condition triggering, enhancing system reliability; the graded protection strategy dynamically adjusts according to actual operating conditions, maximizing operational efficiency while ensuring system safety. Furthermore, the linear regression-based predictive model can identify potential undervoltage risks in advance, providing a basis for proactive protection. The combination of technical means provided by this invention significantly improves the stability and safety of the pitch system, extends the service life of the backup power supply, and provides strong support for the safe operation of wind turbine generators.

[0053] It should be noted that, in this embodiment of the invention, the pitch system includes a wind turbine main controller, multiple monitoring nodes, and a communication network. The wind turbine main controller is located at the center of a star topology. The monitoring nodes are connected to the wind turbine main controller via wired or wireless means. The wired connection uses the Controller Area Network (CLAN) bus protocol with a transmission rate of 500 kbps; the wireless connection uses a long-range radio communication module with a transmission power of 20 dBm and a communication distance of up to 5 km. Each monitoring node is directly installed inside the electrical cabinet of the pitch system using a DIN rail mounting method and is fixed to the power line with bolts to ensure a contact resistance of less than 10 milliohms. The monitoring nodes maintain a connection with the wind turbine main controller through a heartbeat mechanism, with a heartbeat transmission cycle of 1 second.

[0054] It should be noted that the monitoring node's hardware components include a high-precision analog-to-digital converter (ADC), a microcontroller unit (MCU), and a communication interface. The ADC has a 16-bit resolution and a sampling rate of 1kHz, used to acquire voltage, current, and temperature data from the pitch battery. The MCU uses an ARM Cortex-M4 processor architecture with a clock speed of 168MHz, responsible for running a Kalman filter algorithm to suppress noise interference in the acquired signals and improve data reliability. The communication interface supports Universal Asynchronous Receiver / Transmitter (UART), Serial Peripheral Interface (SPEI) protocol, and Integrated Circuit Bus (ICB) protocol, used to transmit processed data to the wind turbine generator's main controller. During initial configuration, the monitoring node requires hardware parameter settings and software algorithm loading. Hardware parameter settings include calibrating the sampling circuit to ensure the accuracy of voltage and current signal acquisition.

[0055] In this embodiment of the invention, step S100 includes the following sub-steps A1 and A2:

[0056] In A1: Set up a monitoring node on each battery pack of the pitch system and perform initial configuration;

[0057] Specifically, in the actual operation of wind turbine generators, the backup power supply for the pitch system consists of multiple battery packs, each equipped with a monitoring node. The monitoring node is fixed inside the electrical cabinet via a DIN rail mounting method and directly connected to the power line, ensuring a contact resistance of less than 10 milliohms to reduce errors during signal acquisition. After the monitoring node completes its initial configuration, a high-precision analog-to-digital converter begins acquiring battery voltage, current, and temperature data at a sampling frequency of 1 kHz. This data is processed by the Kalman filter algorithm of the microcontroller unit and then transmitted to the wind turbine generator's main controller for global analysis.

[0058] Specifically, each monitoring node undergoes initial configuration, including setting the undervoltage threshold, sampling frequency, and data transmission protocol. The undervoltage threshold is dynamically adjusted based on the rated voltage range of the pitch battery, the sampling frequency is set to 1kHz, and the data transmission protocol adopts the User Datagram Protocol (UDP) based on timestamp synchronization to reduce data transmission latency and ensure data consistency.

[0059] In A2: The monitoring node collects battery voltage data in real time and uses a sliding time window algorithm to obtain the voltage change trend;

[0060] Specifically, the real-time acquired voltage data is stored in a fixed-length buffer in chronological order. The length of the buffer is determined by both the sampling frequency and the window time. Assuming the sampling frequency is... f The window time is T Then the buffer length L = f × T ;

[0061] For example, when the sampling frequency is 1 kHz and the window time is 1 second, the length of the buffer is 1000 data points.

[0062] Specifically, whenever a new voltage sampling data is acquired and stored in the buffer, the historical voltage data that has been stored in the buffer for the longest time is removed to maintain a constant total amount of data in the buffer.

[0063] Specifically, after each data update, the average voltage and standard deviation within the time window are calculated based on all voltage data in the current buffer. The average voltage and standard deviation are then used as the output of the voltage change trend for subsequent undervoltage judgment. The average voltage is used to characterize the overall level of the current voltage, and the standard deviation is used to characterize the degree of fluctuation of the current voltage.

[0064] It should be noted that a significant increase in the standard deviation indicates severe voltage fluctuations, potentially suggesting abnormal operating conditions; conversely, a small standard deviation but a continuously decreasing average voltage value may indicate a gradual accumulation of undervoltage risk. In this way, the sliding time window algorithm effectively reduces the impact of random data fluctuations on monitoring results, thereby improving the accuracy of undervoltage detection.

[0065] It should be noted that the above step S100 realizes high-frequency sampling and dynamic trend analysis of battery voltage, effectively avoiding misjudgment caused by instantaneous voltage fluctuations. At the same time, through real-time data buffering and calculation, it improves the continuity and stability of voltage status monitoring.

[0066] In this embodiment of the invention, step S200 compares the voltage change trend with a preset undervoltage threshold, triggers an undervoltage warning mechanism based on the comparison result, and sends the undervoltage warning information to the wind turbine main controller, including the following sub-steps B1~B3:

[0067] In B1: If the average voltage is less than the preset undervoltage threshold, the voltage drop rate is calculated and compared with the preset rate threshold.

[0068] Specifically, the preset undervoltage threshold mainly considers the rated operating voltage of the pitch battery, its chemical characteristics, and the system's safety margin. It is typically set within 80%-85% of the battery's rated voltage. For example, for a battery with a rated voltage of 24V, the undervoltage threshold is set to 20V to ensure timely triggering of protection before the battery voltage enters a dangerous range that could lead to performance degradation or irreversible damage, while also avoiding malfunctions caused by normal fluctuations. The preset undervoltage threshold can also be dynamically calibrated based on actual operating conditions such as battery aging and ambient temperature to balance the accuracy and adaptability of the protection.

[0069] Specifically, voltage drop rate r The calculation formula is:

[0070]

[0071] in, This represents the voltage value at the current moment. This represents the voltage values ​​at the first n sampling points. n Indicates the number of sampling points. Indicates the sampling interval time;

[0072] Specifically, the preset rate threshold is mainly based on the voltage drop characteristics of the pitch battery under typical abnormal operating conditions and the maximum safe rate of voltage drop that the system can tolerate. It is determined by analyzing the voltage drop patterns in historical fault data, for example, by setting it to -50V / s. This aims to identify abnormal rapid voltage drops while avoiding false triggers caused by normal load switching or instantaneous fluctuations. The threshold value must ensure sufficient time for protection to activate before the battery voltage enters the dangerous range and match the system response delay, thereby achieving a balance between rapid early warning and reliable protection.

[0073] In B2: When the average voltage is lower than the preset undervoltage threshold and the voltage drop rate exceeds the preset rate threshold at the same time, the undervoltage warning mechanism is triggered.

[0074] In B3: After determining that the undervoltage warning mechanism has been triggered, an undervoltage warning message containing the battery pack identifier, current voltage value, voltage drop rate and timestamp is generated, and the undervoltage warning message is sent to the wind turbine main controller based on the communication network.

[0075] It should be noted that the triggering condition of the undervoltage warning mechanism is determined by two levels of criteria. First, the first-level criterion determines whether there is an undervoltage risk by comparing the current voltage value with a preset undervoltage threshold. For example, if the rated voltage range of a battery pack is 24V and the undervoltage threshold is set to 20V, then the first-level criterion is met when the monitoring node detects a voltage below 20V. Second, the second-level criterion determines the severity of the voltage change by calculating the voltage drop rate. Assuming a sampling frequency of 1kHz, a sampling interval Δt of 1ms, a voltage value of 21V for the first 10 sampling points, and a current voltage value of 20V, then the voltage drop rate is -100V / s. If the voltage drop rate exceeds a set threshold (e.g., -50V / s), then the second-level criterion is met. When both criteria are met simultaneously, the monitoring node will trigger the undervoltage warning mechanism and transmit the relevant information to the wind turbine generator main controller via the communication network.

[0076] It should be noted that step S200 significantly improves the accuracy of undervoltage identification by introducing a two-level criterion-triggered early warning mechanism. The analysis method combining the average voltage value and standard deviation can distinguish between normal fluctuations and real faults, while the rate-based criterion can detect the deterioration trend of voltage in advance, thereby issuing an early warning before the battery enters deep undervoltage and enhancing the system's early response capability to potential risks.

[0077] In this embodiment of the invention, after the wind turbine main controller receives the undervoltage warning information in step S300, it combines the data from multiple monitoring nodes and uses a voltage prediction model to determine whether to activate protection measures, including the following sub-steps C1~C3:

[0078] In C1: The wind turbine generator main controller receives undervoltage warning information from the monitoring nodes and simultaneously collects voltage data from all other monitoring nodes within the same time window. The overall voltage level of the entire pitch battery system is calculated using a weighted average method.

[0079] In C2: Based on the comprehensive voltage level and historical voltage data time series, a voltage prediction model is constructed using a linear regression algorithm. The voltage prediction model is used to predict the voltage change trend in the future and output the slope of the voltage change trend.

[0080] In C3: If the slope of the voltage change trend is less than the slope threshold, it is determined that the voltage has a continuous downward trend and protection measures need to be activated; otherwise, the voltage trend is determined to be stable, protection measures are not activated for the time being, and the system status is monitored continuously.

[0081] In this embodiment of the invention, the voltage prediction model can be divided into five core modules: input layer, preprocessing layer, feature engineering layer, linear regression layer, and output layer.

[0082] Specifically, the input data for the input layer consists of historical time-series voltage data from multi-node monitoring, in the format {timestamp, voltage value}, for example (t0, V0), (t1, V1), ..., (t n V n ), where t n V is the sampling time. n The voltage monitoring value corresponds to the current time. The model needs to extract continuous voltage data segments as input features using a sliding time window. The size of the sliding window needs to be preset according to the voltage fluctuation period, for example, taking the voltage values ​​of the past 5 sampling times as a set of input features.

[0083] Specifically, the preprocessing layer cleans and standardizes the input time-series voltage data to ensure data quality. This includes: ① Filling missing values ​​within the window using interpolation between adjacent time points or historical averages. ② Identifying outliers using the 3σ criterion and replacing them with the median of adjacent time points or valid historical values. ③ Normalizing voltage values ​​to the [0,1] or [-1,1] interval to avoid the impact of voltage range differences at different nodes on the linear regression coefficients. ④ Converting the preprocessed time window data into feature vectors directly usable by the linear regression model. The core of this is constructing lag features and time features. Lag features use voltage values ​​within the time window as the basic feature, while time features, if strongly correlated with time, can supplement with time-derived features, such as converting timestamps into numerical features like seconds or hours within a day, enhancing the model's ability to capture periodic fluctuations.

[0084] Specifically, the core of the linear regression layer is to construct a linear regression equation, learning the linear relationship between input features and output trends through historical data; let the input feature vector be X=[ x 1, x 2,..., x W ],in x W Let the voltage value be the voltage value within the window. The model output is the slope k of the future voltage change trend, where the slope represents the voltage change per unit time, reflecting the steepness of the trend. The linear regression equation is:

[0085] k = w 1 x 1+ w 2 x 2+...+ w W x W + b

[0086] in, w 1 ,w 2 ,..., w W Here, b represents the weights of each input feature, and b is the bias term.

[0087] Specifically, the output layer converts the slope k from the linear regression layer into a voltage change trend that can be directly used for protection strategies. The sign of the slope k represents the trend direction (k<0 for a downward trend, k>0 for an upward trend, and k≈0 for a stable trend), and the absolute value represents the steepness of the trend (the larger |k| is, the more significant the trend). The final output of the model is (trend direction, rate of change), for example (decreasing, -0.5V / s), which directly serves as the basis for the wind turbine generator's main controller to determine whether to activate protection measures.

[0088] In this embodiment of the invention, the voltage prediction model processing can be divided into three stages: data preparation, model training, and real-time prediction. The entire process is automatically executed by the wind turbine generator main controller.

[0089] Specifically, the data preparation phase includes:

[0090] Historical voltage data is collected from multi-node sensors and stored in a time-series database. The data volume needs to cover at least three complete voltage fluctuation cycles, where a complete voltage fluctuation cycle is a complete discharge cycle of the battery from full charge to undervoltage.

[0091] The sliding window size W is preset based on the voltage sampling frequency and fluctuation period.

[0092] Specifically, the model training phase includes:

[0093] Input features are extracted from historical data using a sliding window. For each window, a corresponding future trend label is calculated: voltage values ​​at M time points after the window's end are taken, and the slope k is obtained through linear fitting of the voltage values. The data is then divided into training and test sets in a 7:3 ratio. For the training set features and labels, a linear regression equation is solved using the least squares method to obtain weights w and biases b. The model accuracy is verified using the test set; if the accuracy is insufficient, the sliding window size W is readjusted or additional time features are added before retrying. The trained weights w and biases b are stored locally on the wind turbine generator's main controller as parameters for real-time prediction.

[0094] Specifically, the real-time prediction stage includes: inputting the preprocessed input features into a linear regression model, calculating the output slope, judging the trend based on the slope, and judging again whether to trigger the subsequent protection strategy evaluation based on the judgment result.

[0095] It should be noted that the main controller of the wind turbine generator automatically retrains the model with the latest collected voltage data at preset intervals, and updates the weights w and bias b to adapt to changes in voltage characteristics caused by long-term factors such as battery aging and load changes.

[0096] It should be noted that step S300 above, through weighted comprehensive voltage and historical trend prediction, can distinguish between local anomalies and systemic problems, avoid false alarms caused by single nodes, and provide a scientific basis for whether to activate protection measures, thereby improving the overall nature and reliability of decision-making.

[0097] In this embodiment of the invention, step S400 includes the following sub-steps D1 to D3:

[0098] In D1: If the voltage prediction model predicts that the voltage will continue to drop, then a graded protection strategy is executed.

[0099] Specifically, Level 1 response reduces the battery burden by adjusting load distribution, triggered by a single battery pack being continuously undervoltage and the predicted downward trend; Level 2 response activates backup power to supply power to critical equipment, triggered by multiple battery packs being undervoltage simultaneously or the voltage failing to recover after Level 1 response; Level 3 response disconnects unnecessary loads and forces shutdown, triggered by a continuous deterioration of system voltage and a prediction that it will endanger system safety.

[0100] Specifically, the wind turbine main controller determines the protection level that should be activated. If a Level 1 response is met, the power supply distribution of non-critical equipment in the pitch system is adjusted to reduce the load on the undervoltage battery pack. If the voltage does not recover to a safe range within a set time after the Level 1 response is executed, the response is upgraded to Level 2, and the backup power supply is activated to power the critical control circuits of the pitch system. If the system voltage still does not improve after the Level 2 response or the prediction results show that the voltage will continue to drop to the dangerous threshold, a Level 3 response is activated, non-essential loads are disconnected, and a system shutdown command is executed.

[0101] It should be noted that the key equipment in the wind turbine pitch system mainly includes core components that directly affect blade angle adjustment and system safety, such as the pitch controller, pitch motor drive unit, safety chain loop, and communication module. Non-critical equipment includes nacelle lighting, auxiliary heaters, status indicator lights, and some non-real-time monitoring sensors. Non-critical equipment can be appropriately deloaded or temporarily disconnected when the system power supply is tight to prioritize the normal operation of critical control functions. The key control circuits of the pitch system mainly include the pitch controller, pitch servo driver, position / speed feedback sensor, safety chain loop, and emergency feathering execution circuit.

[0102] It is important to note that the dangerous threshold is set to ensure that the pitch system has sufficient time to safely perform a shutdown operation before the battery voltage drops to a level that could cause irreversible damage. This threshold is typically determined based on a comprehensive analysis of the battery's chemical characteristics, discharge profile, historical fault data, and the system's minimum operating voltage requirements. Sufficient voltage margin must be provided to cover any further voltage drop during the entire process from threshold triggering to safe shutdown completion, thereby absolutely avoiding the risk of the system going out of control due to complete voltage loss.

[0103] Specifically, during the execution of any level of protection response, the wind turbine main controller continuously monitors battery voltage changes at an increased sampling frequency; if the voltage shows an upward trend after the protection action, the current protection level is maintained and the system continuously assesses whether the protection can be downgraded or deactivated; if the voltage continues to deteriorate, the protection level is gradually upgraded according to the conditions.

[0104] In D2: In protection mode, the voltage recovery is monitored by increasing the sampling frequency. When the voltage recovers to a safe range and the voltage recovery rate meets the conditions, the protection mode is exited.

[0105] Specifically, in protection mode, the main controller of the wind turbine monitors the battery voltage at a higher sampling frequency than in normal mode. When the battery voltage is detected to be continuously higher than the lower limit of the safe voltage and remains so for a set period of time, it is determined that the voltage has recovered to the safe range.

[0106] It should be noted that the lower limit of the safe voltage refers to the lowest voltage threshold at which the pitch battery pack can stably maintain the normal operation of all critical control circuits in the pitch system. Its setting must comprehensively consider the battery's chemical characteristics, discharge curve, the circuit's minimum operating voltage, and safety margin to ensure that the voltage will not trigger the warning again due to normal fluctuations after protection is deactivated. The set duration is used to judge the stability of voltage recovery. This duration must be sufficient to filter out brief voltage rises caused by instantaneous load changes or the activation of backup power, ensuring that the battery has truly escaped the undervoltage state and entered a sustainable stable operating range, thereby avoiding frequent switching of protection modes near the critical point.

[0107] Specifically, when the voltage recovers to a safe range, the system calculates the voltage recovery rate to determine whether to completely exit the protection mode. The voltage recovery rate... The calculation formula is:

[0108]

[0109] in, Indicates the lower limit of the safe voltage range. This indicates the minimum voltage value required to activate the protection mode. This indicates the time required for the voltage to recover to a safe range after the protection mode is activated.

[0110] Specifically, if the voltage recovery rate meets the preset conditions, the battery recovery capability is determined to be good, the wind turbine main controller control system exits the protection mode and resumes normal monitoring; otherwise, the protection mode is maintained, monitoring continues, and the recovery status is periodically reassessed.

[0111] It should be noted that the preset conditions are set based primarily on the normal charging characteristics of the battery and the requirements for stable system operation. They must be higher than the battery's natural recovery rate under typical operating conditions to ensure that the voltage can remain stable within a safe range after exiting the protection mode, avoiding frequent re-triggering of protection due to insufficient recovery. At the same time, the preset conditions should also match the expected effect of the system protection action, reflecting that the battery has effectively recovered from the abnormal state to a self-sustaining level.

[0112] In D3: The wind turbine main controller processes the status information of multiple battery packs simultaneously through a multi-task scheduling mechanism;

[0113] Specifically, the multi-task scheduling mechanism is implemented based on a priority queue, calculating priority parameters for each battery pack status monitoring task. P The calculation formula is:

[0114]

[0115] in, , , All represent weighting coefficients. Vratio Indicates the degree of voltage deviation. Tratio Indicates the degree of temperature deviation. Iratio This indicates the degree of deviation of the current.

[0116] Specifically, the main controller of the wind turbine generator sorts the monitoring tasks of multiple battery packs based on priority parameters, and processes the voltage data, early warning information and protection decisions of each battery pack in order from high to low in each scheduling cycle.

[0117] Specifically, the priority parameters of each battery pack are dynamically updated based on changes in the system's operating status to ensure that battery packs with high abnormality levels are continuously processed with priority; when the protection mode is activated or deactivated, the priority of the relevant battery packs is recalculated to adapt to changes in the system status.

[0118] For example, in practical applications, suppose the backup power supply of a wind turbine's pitch system consists of four battery packs, each equipped with monitoring nodes 2A, 2B, 2C, and 2D. When monitoring node 2A detects that the voltage is below the undervoltage threshold and the voltage drop rate exceeds a preset rate threshold, it triggers an undervoltage warning mechanism and transmits the relevant information to the wind turbine's main controller. The wind turbine's main controller analyzes the data from other monitoring nodes. If it finds that the voltages of monitoring nodes 2B and 2C are also showing a downward trend, it determines that there is a systemic undervoltage risk and initiates a Level 1 response. For example, the wind turbine's main controller reduces the overall load by decreasing the power supply demand of non-critical equipment. If the Level 1 response cannot effectively alleviate the undervoltage risk, it escalates to a Level 2 response, activating the backup power supply to power critical equipment. If the Level 2 response still cannot solve the problem, it further escalates to a Level 3 response, cutting off unnecessary loads and forcibly shutting down the system to protect its safety.

[0119] It should be noted that step S400 above achieves adaptive matching between protection intensity and system state through a graded protection strategy and a dynamic recovery monitoring mechanism. The graded response can take differentiated measures under different risk levels, which avoids over-protection from affecting operating efficiency and ensures timely stoppage of losses in the event of a serious fault; while the exit mechanism based on recovery rate can determine whether the battery has truly recovered to a stable state, prevent frequent switching of protection modes, and improve the stability and safety of system operation.

[0120] Example 2: This example provides a safety control system for wind turbine generator sets, including:

[0121] The trend analysis module is used to set up monitoring nodes and perform initial configuration on each battery pack of the pitch system. The monitoring nodes collect battery voltage data in real time and use a sliding time window algorithm to obtain the voltage change trend.

[0122] The undervoltage warning trigger module is used to compare the voltage change trend with the preset undervoltage threshold, trigger the undervoltage warning mechanism based on the comparison result, and send the undervoltage warning information to the wind turbine generator main controller.

[0123] The centralized risk prediction module is used by the main controller of the wind turbine generator to determine whether to activate protection measures after receiving undervoltage warning information and combining data from multiple monitoring nodes through a voltage prediction model.

[0124] The undervoltage protection execution module is used to execute a graded protection strategy if the voltage prediction model shows that the voltage will continue to drop. In protection mode, the voltage recovery is monitored by increasing the sampling frequency. When the voltage recovers to a safe range and the voltage recovery rate meets the conditions, the protection mode is exited.

[0125] It should be noted that the technical solution for the safety control system applied to wind turbine generator sets is based on the same concept as the technical solution for the safety control method applied to wind turbine generator sets described above. For details not described in detail in the technical solution for the safety control system applied to wind turbine generator sets in this embodiment, please refer to the description of the technical solution for the safety control method applied to wind turbine generator sets described above.

[0126] The above-mentioned unit modules can be embedded in the processor of the electronic device in hardware form or independent of it, or they can be stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of the above modules.

[0127] This embodiment also provides an electronic device, which includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a safety control method applied to wind turbine generator sets. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.

[0128] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method proposed in the above embodiments.

[0129] The storage medium proposed in this embodiment belongs to the same inventive concept as the method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0130] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory, random access memory, flash memory, hard disk, or optical disk, and includes several instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) to execute the method of the embodiments of the present invention.

[0131] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A safety control method applied to wind turbine generator sets, characterized in that, include: A monitoring node is set up and initialized on each battery pack of the pitch system. The monitoring node collects the battery voltage data in real time and uses a sliding time window algorithm to obtain the voltage change trend. The voltage change trend is compared with a preset undervoltage threshold. Based on the comparison result, an undervoltage warning mechanism is triggered, and the undervoltage warning information is sent to the wind turbine main controller. After receiving the undervoltage warning information, the main controller of the wind turbine generator set combines data from multiple monitoring nodes and uses a voltage prediction model to determine whether to activate protection measures. If the voltage prediction model predicts that the voltage will continue to drop, a graded protection strategy is implemented. In protection mode, the voltage recovery is monitored by increasing the sampling frequency. When the voltage recovers to a safe range and the voltage recovery rate meets the conditions, the protection mode is exited. The method of using a sliding time window algorithm to obtain voltage change trends includes: The real-time voltage data is stored in a fixed-length buffer in chronological order. The length of the buffer is determined by the sampling frequency and the window time. Whenever a new voltage sampling data is acquired and stored in the buffer, the historical voltage data with the longest storage time in the buffer is removed to maintain a constant total amount of data in the buffer. After each data update, the average voltage and standard deviation within the time window are calculated based on all voltage data in the current buffer. The average voltage and standard deviation are then used as the output of the voltage change trend for subsequent undervoltage judgment. The average voltage is used to characterize the overall level of the current voltage, and the standard deviation is used to characterize the degree of fluctuation of the current voltage. The method of determining whether to activate protection measures using a voltage prediction model includes: The wind turbine generator main controller receives undervoltage warning information from the monitoring nodes and simultaneously collects voltage data from all other monitoring nodes within the same time window. It then calculates the overall voltage level of the entire pitch battery system using a weighted average method. Based on the comprehensive voltage level, historical voltage data, and time series, a voltage prediction model is constructed using a linear regression algorithm. The voltage prediction model is then used to predict the voltage change trend in the future and output the slope of the voltage change trend. If the slope of the voltage change trend is less than the slope threshold, it is determined that the voltage has a continuous downward trend and protection measures need to be activated; otherwise, the voltage trend is determined to be stable, protection measures are not activated for the time being, and the system status is monitored continuously. The implementation of the hierarchical protection strategy includes: The trigger condition for a Level 1 response is a single battery pack that is continuously undervoltage and the predicted trend is downward; the trigger condition for a Level 2 response is multiple battery packs that are simultaneously undervoltage or the voltage does not recover after the Level 1 response; the trigger condition for a Level 3 response is a continuous deterioration of the system voltage and a prediction that it will endanger system safety. The main controller of the wind turbine generator set determines the protection level that should be activated at the current time; If the Level 1 response is met, the load on the undervoltage battery pack is reduced by adjusting the power distribution of non-critical equipment in the pitch system. If the voltage does not recover to a safe range within a set time after the Level 1 response is executed, the response is upgraded to Level 2, and the backup power supply is activated to power the critical control circuits of the pitch system. If the system voltage still does not improve after the Level 2 response or the prediction results show that the voltage will continue to drop to a dangerous threshold, the Level 3 response is activated, non-essential loads are disconnected, and a system shutdown command is executed. During any level of protection response, the wind turbine main controller continuously monitors battery voltage changes at an increased sampling frequency. If the voltage shows an upward trend after the protection action, the current protection level is maintained and the system continuously assesses whether the protection can be downgraded or deactivated. If the voltage continues to deteriorate, the protection level is gradually upgraded according to the conditions.

2. The safety control method for wind turbine generator sets as described in claim 1, characterized in that, The undervoltage warning mechanism triggered based on the comparison result includes: If the average voltage is less than a preset undervoltage threshold, the voltage drop rate is calculated and compared with a preset rate threshold. When the average voltage is lower than the preset undervoltage threshold and the voltage drop rate exceeds the preset rate threshold, the undervoltage warning mechanism is triggered. After determining that the undervoltage warning mechanism has been triggered, an undervoltage warning message containing the battery pack identifier, current voltage value, voltage drop rate and timestamp is generated, and the undervoltage warning message is sent to the main controller of the wind turbine generator set based on the communication network.

3. The safety control method for wind turbine generator sets as described in claim 1, characterized in that, When the voltage recovers to a safe range and the voltage recovery rate meets the conditions, the protection mode exit includes: In protection mode, the main controller of the wind turbine generator monitors the battery voltage at a higher sampling frequency than in normal mode. When the battery voltage is detected to be continuously higher than the lower limit of the safe voltage and remains so for a set period of time, it is determined that the voltage has recovered to the safe range. When the voltage recovers to a safe range, the voltage recovery rate is calculated to determine whether to completely exit the protection mode. If the voltage recovery rate meets the preset conditions, the battery recovery capability is determined to be good, the wind turbine main controller control system exits the protection mode and resumes normal monitoring. Otherwise, the protection mode is maintained, monitoring continues, and the recovery status is periodically reassessed.

4. The safety control method for wind turbine generator sets as described in claim 1, characterized in that, The wind turbine main controller processes the status information of multiple battery packs simultaneously through a multi-task scheduling mechanism, including: Calculate priority parameters for each battery pack status monitoring task; The wind turbine generator main controller sorts the monitoring tasks of multiple battery packs based on the priority parameters, and processes the voltage data, early warning information and protection decisions of each battery pack in order from high to low in each scheduling cycle. The priority parameters of each battery pack are dynamically updated according to changes in system operating status to ensure that battery packs with high abnormality levels are continuously processed with priority; when protection mode is activated or deactivated, the priority of relevant battery packs is recalculated to adapt to changes in system status.

5. A safety control system for wind turbine generator sets, employing the safety control method for wind turbine generator sets as described in any one of claims 1 to 4, characterized in that, include: The trend analysis module is used to set up monitoring nodes and perform initial configuration on each battery pack of the pitch system. The monitoring nodes collect battery voltage data in real time and use a sliding time window algorithm to obtain the voltage change trend. The undervoltage warning trigger module is used to compare the voltage change trend with a preset undervoltage threshold, trigger the undervoltage warning mechanism based on the comparison result, and send the undervoltage warning information to the wind turbine main controller. A centralized risk prediction module is used by the main controller of the wind turbine generator to determine whether to activate protection measures after receiving the undervoltage warning information and combining data from multiple monitoring nodes through a voltage prediction model. The undervoltage protection execution module is used to execute a graded protection strategy if the voltage prediction result of the voltage prediction model shows that the voltage will continue to drop; in the protection mode, the voltage recovery is monitored by increasing the sampling frequency; when the voltage recovers to a safe range and the voltage recovery rate meets the conditions, the protection mode is exited.

6. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store computer-executable instructions, and when the processor executes the computer-executable instructions, it implements the steps of the safety control method for wind turbine generator sets as described in any one of claims 1 to 4.

7. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: When the computer-executable instructions are executed by the processor, they implement the steps of the safety control method for wind turbine generator sets as described in any one of claims 1 to 4.

Images