Gravity assisted loop heat pipe driven thermal management system and method for wind turbines

Abstract

The application discloses a gravity loop heat pipe driven wind driven generator thermal management system and method, and belongs to the field of generator thermal management, wherein the method comprises: a collection module for collecting multi-source operation state data of the wind driven generator and the gravity loop heat pipe system; a processing module for fusing and feature extracting the multi-source operation state data, online correcting the physical heat transfer limit of the gravity loop heat pipe based on the extracted features, and evaluating the current thermal management demand according to the corrected physical heat transfer limit; and an execution module for generating corresponding thermal management action instructions according to the current thermal management demand, and executing the wind driven generator thermal management control. The application solves the technical problems that the existing wind driven generator thermal management technology is difficult to respond to the thermal shock caused by sudden increase of power generation power in advance, and is prone to cause overheating aging of core components, unplanned shutdown and energy waste.

Description

Technical Field

[0001] This invention relates to the field of generator thermal management, and more specifically to a gravity-driven loop heat pipe-driven wind turbine generator thermal management system and method. Background Technology

[0002] Wind turbines are mostly deployed in remote, extreme weather environments in onshore and offshore wind farms. The thermal management level of their core heat-generating components, such as generator windings, gearboxes, and converters, directly determines the operational reliability and total life-cycle cost of the unit. Traditional active liquid cooling and air cooling systems rely on high-power pumps and fans operating continuously, resulting in inherent drawbacks such as high energy consumption, easy wear of moving mechanical parts, high failure rate in salt spray and dust environments, and high maintenance frequency. Therefore, gravity-driven loop heat pipes without power drive have become an important development direction in the field of wind power thermal management due to their high heat transfer efficiency, high reliability, and zero passive operating energy consumption. However, existing wind power thermal management systems... The system generally adopts fixed heat transfer limit parameters calibrated at the factory and lag-driven pure feedback control logic, which does not fully consider the dynamic impact of nacelle yaw adjustment, uneven tower settlement, and dynamic wind load on the return head. It lacks quantitative monitoring and early warning of the hidden performance degradation such as slow leakage of heat pipe working fluid and gradual accumulation of non-condensable gases. It is difficult to cope in advance with the thermal shock caused by sudden increase in power generation and the gearbox lubrication preheating requirements under extreme low temperature. It is prone to overheating and aging of core components, unplanned shutdowns, and ineffective energy waste caused by excessive heat dissipation, which seriously restricts the large-scale application and safe operation level of gravity loop heat pipe in wind turbine thermal management. Summary of the Invention

[0003] This application provides a gravity-driven loop heat pipe-driven wind turbine thermal management system and method, aiming to solve the technical problems of existing wind turbine thermal management technologies being unable to cope with the thermal shock caused by sudden increases in power generation in advance, which easily leads to overheating and aging of core components, unplanned shutdowns and energy waste, and achieves the technical effect of improving the intelligence level and long-term operational reliability of the wind turbine thermal management system.

[0004] In view of the above problems, this application provides a gravity-driven loop heat pipe-driven wind turbine thermal management system and method.

[0005] The first aspect disclosed in this application provides a gravity-driven loop heat pipe-driven thermal management system for a wind turbine generator, the system comprising: The acquisition module is used to acquire multi-source operating status data of the wind turbine generator and gravity loop heat pipe system; the processing module is used to fuse and extract features from the multi-source operating status data, correct the physical heat transfer limit of the gravity loop heat pipe online based on the extracted features, and evaluate the current thermal management requirements based on the corrected physical heat transfer limit; the execution module is used to generate corresponding thermal management action commands based on the current thermal management requirements and execute the thermal management control of the wind turbine generator.

[0006] Another aspect of this application discloses a thermal management method for a gravity-driven loop heat pipe-driven wind turbine generator, the method comprising: Collect multi-source operating status data of the wind turbine generator and gravity loop heat pipe system; fuse and extract features from the multi-source operating status data; correct the physical heat transfer limit of the gravity loop heat pipe online based on the extracted features; assess the current thermal management requirements based on the corrected physical heat transfer limit; generate corresponding thermal management action commands based on the current thermal management requirements; and execute the thermal management control of the wind turbine generator.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: By constructing a multi-source heterogeneous data acquisition and fusion system covering the entire chain of generator heat generation, heat pipe heat transfer, and environmental heat dissipation, the physical heat transfer limit of the gravity loop heat pipe is corrected online based on the real-time tilt angle of the nacelle and the physical properties of the working fluid. The heat load margin is calculated by combining the corrected heat transfer limit with the temperature of core components to achieve a layered and accurate assessment of thermal management requirements. At the same time, a dual feedforward control logic of natural convection equivalent heat dissipation power and ultra-short-term power generation prediction is introduced. A heat pipe performance degradation index and predictive maintenance mechanism based on long-term temperature difference offset and rate of change are established. While ensuring the heat dissipation safety of the unit under all operating conditions and avoiding overheating damage and unplanned shutdowns of core components, the energy-saving heat dissipation advantage of the gravity loop heat pipe is maximized, improving the intelligence level and long-term operational reliability of the wind turbine thermal management system.

[0008] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0009] Figure 1 A schematic diagram of the structure of a gravity-driven loop heat pipe-driven wind turbine thermal management system is provided for embodiments of this application; Figure 2 A schematic flowchart of a gravity-driven loop heat pipe-driven thermal management method for wind turbine generators is provided for embodiments of this application. Figure 3 A typical operating condition schematic diagram of a gravity-driven loop heat pipe-driven wind turbine thermal management system is provided for embodiments of this application.

[0010] Explanation of reference numerals in the attached diagram: Acquisition module 10, Processing module 20, Execution module 30. Detailed Implementation

[0011] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0012] The overall concept of the technical solution provided in this application is as follows: This application provides a gravity-driven loop heat pipe-driven wind turbine thermal management system and method. By constructing a multi-source heterogeneous data acquisition and fusion system covering the entire chain of generator heat generation, heat pipe heat transfer, and environmental heat dissipation, the physical heat transfer limit of the gravity-driven loop heat pipe is corrected online based on the real-time tilt angle of the nacelle and the physical properties of the working fluid. The heat load margin is calculated by combining the corrected heat transfer limit with the temperature of core components to achieve a layered and accurate assessment of thermal management requirements. At the same time, a dual feedforward control logic of natural convection equivalent heat dissipation power and ultra-short-term power generation prediction is introduced, and a heat pipe performance degradation index and predictive maintenance based on long-term temperature difference offset and rate of change are established. This ensures the heat dissipation safety of the unit under all operating conditions, avoids overheating damage to core components and unplanned shutdowns, and maximizes the energy-free heat dissipation advantage of the gravity-driven loop heat pipe, thereby improving the intelligence level and long-term operational reliability of the wind turbine thermal management system.

[0013] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0014] Example 1, as Figure 1 As shown in the embodiment of this application, a gravity-driven loop heat pipe-driven wind turbine thermal management system is provided, the system comprising: The acquisition module 10 is used to acquire multi-source operating status data of the wind turbine generator and gravity loop heat pipe system.

[0015] Specifically, the system first collects multi-source operational status data. This data encompasses a complete set of nine heterogeneous physical quantities across four main categories: core heat-generating components of the wind turbine, the gravity-driven loop heat pipe body, nacelle attitude, and the external environment. These include generator winding temperature, gearbox oil temperature, converter module temperature, heat pipe evaporation section temperature, heat pipe condensation section temperature, nacelle tilt angle, ambient wind speed, and ambient temperature, covering key influencing factors across the entire system's heat generation, heat transfer, and heat dissipation chain. The modules employ a differentiated industrial-grade sensor deployment scheme. For core heat sources such as generator stator windings, pre-embedded platinum resistance sensors are used, which are directly embedded in the bottom of the winding slots and the end insulation layer to avoid temperature deviations and response delays caused by indirect measurement through the air layer.

[0016] The gearbox oil temperature is achieved by using an immersion-type armored temperature sensor installed in the main circulation loop of the lubricating oil in the oil pan, ensuring that the measured value accurately reflects the overall temperature of the oil involved in gear meshing and heat dissipation.

[0017] Temperature monitoring of the converter power module involves attaching a high-precision surface-mount thermistor between the module substrate and the heat sink to directly capture the junction temperature correlation value of the power device that determines the lifespan of the converter.

[0018] For gravity-type loop heat pipe systems, 3×3 array thermocouples are arranged in the evaporation section, which is the heat-absorbing end where the heat pipe absorbs heat from the generator and vaporizes the internal working fluid from a liquid state to a gaseous state, and is the only entry point for heat into the heat pipe system, and the condensation section, which is the heat-releasing end where the heat pipe releases the heat carried by the gaseous working fluid to the external environment and causes the working fluid to recondense into a liquid state, and is the only exit point for heat to be discharged from the system. Multi-point averaging eliminates measurement distortion caused by local hot spots.

[0019] The nacelle tilt angle is achieved by using a high-precision tilt sensor installed at the geometric center of the nacelle chassis to capture the wind turbine's yaw and pitch attitude changes in real time. This parameter directly determines the gravity recirculation capability of the heat pipe working fluid.

[0020] Ambient wind speed and direction are collected by an ultrasonic anemometer on the top of the cabin. Its characteristic of having no mechanical rotating parts can effectively resist the wear and tear of sand and salt spray commonly found in wind fields. The ambient temperature sensor is placed in a well-ventilated and shady place away from the cabin heat dissipation vents and avoids direct sunlight to ensure that the measured value is the true ambient temperature.

[0021] Subsequently, the module employs differentiated sampling frequencies based on the dynamic characteristics of different parameters: generator winding temperature, converter module temperature, and heat pipe evaporation / condensation section temperature are sampled at a high frequency of 1Hz to adapt to their thermal response characteristics that fluctuate rapidly with power generation; gearbox oil temperature, ambient temperature, and nacelle tilt angle are sampled at a low frequency of 0.1Hz to reduce redundant data; ambient wind speed and direction are sampled at 0.5Hz to balance wind speed pulsation characteristics and processing efficiency. After the acquired raw analog signals are converted into digital signals by the module's built-in 16-bit high-precision AD converter, three levels of preprocessing are performed sequentially: First, outlier removal is performed, based on the 3σ statistical criterion to identify and directly remove spike values ​​caused by sensor momentary poor contact and electromagnetic interference, avoiding abnormal data from contaminating subsequent analysis; second, sliding filtering is performed, using a 5-point moving average algorithm to suppress signal fluctuations caused by sensor random noise and industrial electromagnetic interference, making the data curve smoother.

[0022] Finally, time synchronization and alignment are performed, adding a unified GPS time stamp to each data frame. This maps heterogeneous data with different sampling frequencies and clock deviations onto the same time axis, ensuring that the time correspondence of all parameters is completely accurate during subsequent data fusion and completely avoiding feature extraction errors caused by sensor clock asynchrony. The standardized dataset after preprocessing is transmitted in real time to the system's edge computing unit via the industrial Ethernet bus for subsequent fusion analysis and control decision-making.

[0023] The processing module 20 is used to fuse and extract features from the multi-source operating status data, correct the physical heat transfer limit of the gravity loop heat pipe online based on the extracted features, and evaluate the current thermal management requirements based on the corrected physical heat transfer limit.

[0024] Specifically, the processing module 20 first receives standardized multi-source operating status data with a unified GPS timestamp transmitted by the acquisition module, and performs multi-source data fusion. Fusion is to linearly interpolate and resample the pre-processed heterogeneous time series data with different sampling frequencies according to a unified time granularity of 1 second, and construct a unified time series sample in which each time node fully includes 8 parameters: generator winding temperature, gearbox oil temperature, converter module temperature, heat pipe evaporation section temperature, heat pipe condensation section temperature, nacelle tilt angle, ambient wind speed, and ambient temperature.

[0025] Based on this, feature extraction is performed to mine nine key features from a unified time series sample that can reflect the entire chain of heat generation, heat transfer, and heat dissipation of the system. These include: the real-time temperature difference between the evaporation and condensation sections, which directly characterizes the heat transfer state of the heat pipe. That is, the temperature difference between the heat-absorbing end and the heat-releasing end of the heat pipe at the same time point, which is the most direct characteristic quantity reflecting the current heat load of the heat pipe. The larger the temperature difference, the more heat the heat pipe transfers. The rate of change of the real-time temperature difference relative to time, which reflects the fluctuation rate of the heat load. A sharp increase in the rate of change usually indicates a thermal shock caused by a sudden increase in power generation. The long-term deviation of the real-time temperature difference from the theoretical temperature difference under the design conditions of the heat pipe, that is, the average deviation of the measured temperature difference from the design value over an exponential period, can eliminate the interference of short-term heat load fluctuations and reflect the performance degradation trend of the heat pipe after long-term operation. At the same time, state features that can affect the system operation are extracted, namely, the real-time tilt angle of the nacelle, the real-time temperature of the generator windings, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time ambient temperature.

[0026] During feature extraction, the module simultaneously constructs a heat pipe performance degradation index based on the long-term offset and rate of change of the real-time temperature difference. This index quantifies the deviation between the measured heat transfer performance and the design performance, enabling sensitive identification of hidden faults such as slow leakage of the working fluid or gradual accumulation of non-condensable gases inside the heat pipe. When the degradation index exceeds the preset degradation threshold, it is directly marked as a performance degradation warning feature, providing a basis for subsequent demand assessment. After feature extraction is completed, the module enters the online correction stage of the physical heat transfer limit. First, based on the extracted real-time tilt angle of the nacelle, combined with the cold... The design vertical height difference between the condenser and evaporator, the working fluid density of the heat pipe, and the gravitational acceleration are used to calculate the current effective return head. This is the core driving force for the stable operation of the gravity-type loop heat pipe. It refers to the gravitational pressure generated by the actual vertical height difference between the condenser and evaporator sections under the current nacelle attitude. It is equal to the design vertical height difference multiplied by the cosine of the nacelle tilt angle, and then multiplied by the working fluid density and gravitational acceleration. When the nacelle tilt angle increases due to yaw or pitch, the effective vertical height difference will decrease, and the effective return head will decrease linearly, thereby reducing the maximum heat transfer capacity of the heat pipe.

[0027] The module then retrieves the standard heat transfer limit and corresponding standard return head under the heat pipe design conditions, i.e., the return head when the nacelle is horizontal. Based on the experimentally calibrated head correction relationship, a nonlinear physical mapping model between the return head and the heat transfer limit is established to reflect the attenuation law of the heat pipe heat transfer limit when the return head is insufficient. The ratio of the current effective return head to the standard return head is substituted into the model to calculate the corrected actual physical heat transfer limit.

[0028] The next module assesses the current thermal management requirements based on the corrected physical heat transfer limit. First, it calculates the actual heat load currently borne by the heat pipe based on the real-time temperature difference between the evaporator and condenser sections, combined with the total thermal resistance of the heat pipe (i.e., the total thermal resistance of heat transfer from the outer wall of the evaporator section to the outer wall of the condenser section, composed of the evaporator section thermal resistance, steam flow thermal resistance, condenser section thermal resistance, and pipe wall thermal resistance). This resistance is approximately constant when the heat pipe is operating normally and without performance degradation. Then, it uses the formula: (Modified physical heat transfer limit - current actual heat load) / Modified physical heat transfer limit; The calculated heat load margin is a core indicator that directly reflects the remaining heat dissipation capacity of the heat pipe system. A larger margin indicates more sufficient passive heat dissipation, while a smaller margin indicates that the heat pipe is close to its heat transfer limit and there is a risk of overheating.

[0029] The module then comprehensively compares the heat load margin with the real-time temperatures of the generator windings, gearbox oil, converter module, and ambient temperature, and integrates two feedforward correction logics to complete the hierarchical assessment of demand patterns: The first feedforward is to calculate the equivalent heat dissipation power of natural convection based on the angle between the ambient wind speed, wind direction, and the current azimuth angle of the condenser, combined with the condenser's windward area. This is used as the feedforward quantity to pre-correct the trigger threshold for the auxiliary enhanced heat dissipation mode. When the equivalent heat dissipation power is greater than the preset sufficient air cooling threshold, it indicates that natural wind can provide sufficient auxiliary heat dissipation. At this time, the trigger threshold is increased to reduce the number of fan starts and reduce energy consumption. When the equivalent heat dissipation power is less than the preset insufficient air cooling threshold, the trigger threshold is decreased to start the fan in advance. The second feedforward is to access the wind farm's ultra-short-term power generation forecast data. When it is predicted that the power generation will increase sharply within the next 15-30 minutes, exceeding the preset power change rate threshold, the judgment threshold of the passive sufficient heat dissipation mode is lowered in advance or the auxiliary heat dissipation is pre-started directly, so that the heat pipe system can establish a high heat transfer state in advance to cope with thermal shock. When the predicted power generation is continuously lower than the preset low power threshold, it enters the low temperature preheating preparation state.

[0030] The final module categorizes current thermal management requirements into four modes based on comprehensive evaluation results: When the heat load margin exceeds the first preset threshold (typically 30%) and the temperatures of all core components are below their respective warning limits, it is assessed as a passive sufficient cooling mode; when the heat load margin is less than the second preset threshold (typically 10%) or the temperature of any core component exceeds the warning limit, it is assessed as an auxiliary enhanced cooling requirement mode; when the ambient temperature is below the preset low-temperature threshold and the gearbox oil temperature is below the allowable start-up lower limit temperature, it is assessed as a low-temperature preheating requirement mode; and when the heat pipe performance degradation index exceeds the preset degradation threshold, it is assessed as a performance degradation alarm mode, such as... Figure 3 As shown, the decay index changes over time under three typical operating conditions: normal, working fluid leakage, and non-condensable gas accumulation. These are superimposed with a yellow warning threshold of 0.2 and a red alarm threshold of 0.5. The system verifies that the decay index constructed using long-term offsets and rates of change can distinguish between gradual leakage faults and step-accumulation faults, enabling predictive maintenance. Under normal operating conditions, the decay index remains stable below 0.1 for a long period, slowly increasing slightly. Under working fluid leakage conditions, the decay index increases gradually and linearly from day 50, with a significant slope, reaching the yellow warning threshold around day 110. Under non-condensable gas accumulation conditions, the decay index jumps abruptly from day 80, with a three-stage increasing slope, reaching the yellow warning threshold around day 100 and approaching the red alarm threshold around day 130. The threshold lines are represented by two dashed lines: yellow warning (0.2) and red alarm (0.5).

[0031] The execution module 30 is used to generate corresponding thermal management action instructions based on the current thermal management requirements and execute the thermal management control of the wind turbine generator.

[0032] Specifically, the module first receives the thermal management demand mode identifier and corresponding control parameters output by the processing module, generates differentiated action instructions for different modes and sends them to the corresponding actuators. When the demand is a passive heat dissipation sufficient mode, it generates a command to maintain a fully passive state. The command to maintain a fully passive state means that heat transfer is achieved entirely by the natural circulation of the working fluid in the gravity loop heat pipe, without the intervention of any active energy-consuming equipment, resulting in a zero-power heat dissipation state. It cuts off the power supply to all active actuators such as the auxiliary fan on the condenser side and the electric heating element of the gearbox, putting them into a low-power standby state. At the same time, it continuously collects the temperature difference between the two ends of the heat pipe and the temperature of the core components of the generator as feedback signals to verify the stability of the demand mode in real time. Once an abnormal temperature rise trend is detected, the mode switch is immediately triggered.

[0033] When the demand is for enhanced cooling, a command to start the auxiliary fan on the condenser side is generated. On-demand cooling is achieved based on closed-loop fan speed adjustment logic. Closed-loop fan speed adjustment refers to a continuous adjustment method that dynamically adjusts the fan output power with the core component temperature as the controlled variable and the heat load margin as the feedforward variable. First, the initial fan speed is calculated based on the current heat load margin. The smaller the heat load margin, the higher the initial speed. Then, a PID closed-loop control loop is established with the generator winding temperature as the controlled object. The deviation between the winding temperature feedback value and the target value is collected in real time, and the fan speed is dynamically adjusted. At the same time, the feedforward signal of the equivalent heat dissipation power of natural convection output by the processing module is received and the gain parameter of the PID controller is corrected in real time. When natural convection cooling is sufficient, the gain is reduced to avoid frequent speed fluctuations. When natural convection is insufficient, the gain is increased to speed up the response. If the cooling demand cannot be met by a single fan running at full speed, the module will start the backup auxiliary fan in sequence to form a multi-machine collaborative cooling mechanism.

[0034] When the demand is in the low-temperature preheating mode, an intermittent power-on command is generated, which adopts an energy-saving heating control method of power-on heating-power-off temperature equalization-temperature detection cycle. This command is sent to the electric heating element embedded in the gearbox oil pan or generator end cover. First, the initial power-on duty cycle is calculated based on the difference between the current gearbox oil temperature and the allowable lower limit temperature for starting. The larger the temperature difference, the higher the duty cycle. After each heating cycle, the oil temperature is detected. If it is still below the lower limit, the current duty cycle is maintained to continue heating. If it is close to the lower limit, the duty cycle is gradually reduced until the oil temperature stabilizes within the allowable starting range. At the same time, the working current and surface temperature of the heating element are monitored in real time to prevent short circuits or dry burning that could cause safety accidents. When the ambient temperature rises or the oil temperature rises naturally after the generator starts, the heating power supply is automatically cut off.

[0035] When the demand is in performance degradation alarm mode, a graded alarm and maintenance prompt information are first generated and uploaded to the wind farm's central monitoring system in real time via industrial Ethernet. The information includes the heat pipe performance degradation index, estimated remaining service life, recommended maintenance time, and fault cause, namely working fluid leakage or accumulation of non-condensable gases. At the same time, an audible and visual alarm is triggered in the local control cabinet of the nacelle. Then, the system immediately switches to the backup auxiliary cooling circuit, which is an emergency cooling channel independent of the main gravity loop heat pipe system. It consists of an independent radiator, fan, and coolant circulation pump. It temporarily takes over the cooling task when the main heat pipe fails. The control link of the main heat pipe system is cut off, the circulation pump and fan of the backup circuit are started, and the cooling power of the backup circuit is dynamically adjusted according to the temperature of the generator's core components to ensure that the wind turbine can operate safely at reduced power before maintenance personnel arrive.

[0036] In addition, the execution module has a built-in independent hardware watchdog circuit. When a failure of the processing module or communication link is detected, it will automatically switch to a safe operation mode, start all auxiliary fans to run at maximum speed and trigger a fault alarm to prevent the system from going out of control. At the same time, the module will feed back the operating status of all actuators, namely fan speed, heating element on / off status and backup circuit working status, to the processing module 20 in real time, forming a control loop of perception-decision-execution-feedback.

[0037] In one possible implementation, the processing module 20 performs fusion and feature extraction on the multi-source operating status data, including: performing time synchronization alignment, outlier removal, and sliding filtering on the multi-source operating status data, which includes generator winding temperature, gearbox oil temperature, converter module temperature, heat pipe evaporator section temperature, heat pipe condenser section temperature, nacelle tilt angle, ambient wind speed, and ambient temperature; the processed multi-source operating status data is fused into a unified time series sample, and state features are extracted, including the real-time temperature difference between the evaporator section and the condenser section, the rate of change of the real-time temperature difference relative to time, the long-term offset of the real-time temperature difference from the theoretical temperature difference under the heat pipe design conditions, the real-time tilt angle of the nacelle, the real-time temperature of the generator winding, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time ambient temperature.

[0038] Specifically, the system first receives multi-source heterogeneous raw data with a unified GPS time stamp transmitted from the acquisition module 10. This data comes from eight different types of sensors with significantly different sampling frequencies: generator winding, converter, and heat pipe temperatures at 1 Hz; ambient wind speed and direction at 0.5 Hz; and gearbox oil temperature, ambient temperature, and nacelle tilt angle at 0.1 Hz. The first step is to perform time synchronization alignment, which means mapping sensor data with different sampling frequencies and local clock deviations to the same time axis using GPS absolute time as a unified reference. This is achieved through a linear interpolation resampling algorithm, using 1 second as the unified time granularity. For parameters below this frequency, such as gearbox oil temperature at 0.1 Hz, a linear fitting value is inserted between two adjacent measured points. For parameters above this frequency, the average value within the corresponding time window is taken. Finally, an equally spaced time series containing all eight parameters is generated for each timestamp. The second step is to perform outlier removal. The first step involves identifying and removing distorted spikes and jumps in the raw data caused by factors such as momentary poor sensor contact, industrial electromagnetic interference, and lightning induction. This is achieved using the 3σ statistical criterion commonly used in the industrial field. The mean μ and standard deviation σ of each parameter are calculated based on the historical data of the most recent 10 minutes. Data exceeding the μ ± 3σ range are identified as outliers and directly removed. Simultaneously, linear interpolation of two adjacent valid values ​​is used to fill gaps, preventing outlier data from contaminating subsequent analysis results while ensuring the continuity of the time series. The third step is sliding filtering, which involves smoothing the data within a sliding window to suppress random noise and signal fluctuations from the sensor. A sliding average algorithm with a window size of 5 is selected, meaning the output value at each time step is the arithmetic mean of the data at that time and the previous four time steps. This window size has been experimentally calibrated to effectively filter out high-frequency noise from the complex electromagnetic environment of the wind field without excessive smoothing that could lead to signal response lag.

[0039] After completing the three-level preprocessing, multi-source data fusion is performed. Fusion integrates the eight cleaned independent time-series data into a structured multi-dimensional unified time-series sample. Each sample corresponds to a 1-second timestamp and contains values ​​for eight dimensions: generator winding temperature, gearbox oil temperature, converter module temperature, heat pipe evaporator section temperature, heat pipe condenser section temperature, nacelle tilt angle, ambient wind speed, and ambient temperature. This forms a snapshot that can completely depict the system's operating state at a specific moment. Based on this, key state features are extracted, resulting in eight features: The first is the real-time temperature difference between the evaporator and condenser sections, which refers to the difference between the average temperature of multiple points in the heat pipe evaporator section and the average temperature of multiple points in the condenser section at the same timestamp. Since the thermal resistance of steam flow is negligible during normal operation of the gravity loop heat pipe, this temperature difference is approximately linearly positively correlated with the actual heat load transferred by the heat pipe, making it the most direct characteristic reflecting the current heat transfer state of the heat pipe. The second is the rate of change of the real-time temperature difference relative to time, which refers to the rate of change of the temperature difference per unit hour. The change in real-time temperature difference, ΔT / Δt, is obtained by first-order difference calculation of the real-time temperature difference time series. It can sensitively capture the thermal shock signal caused by a sudden increase in power generation. A sharp increase in the rate of change often indicates an impending risk of overheating. The third item is the long-term deviation between the real-time temperature difference and the theoretical temperature difference under the heat pipe design conditions. It refers to the difference between the sliding average of the measured real-time temperature difference over more than 7 consecutive days and the theoretical temperature difference under the design conditions specified by the heat pipe at the factory. This feature eliminates the temperature difference disturbance caused by short-term power generation fluctuations and changes in ambient temperature. It can stably reflect the performance degradation trend of the heat pipe after long-term operation and is the basis for constructing the heat pipe performance degradation index. The remaining 5 features are the pre-processed real-time tilt angle of the nacelle, the real-time temperature of the generator windings, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time ambient temperature. These correspond to the attitude factors that affect the heat transfer capacity of the heat pipe, the temperature factors that reflect the overheating risk of the generator core components, and the environmental factors that affect the heat dissipation conditions, respectively.

[0040] In one possible implementation, the processing module 20 further includes: constructing a heat pipe performance degradation index based on the long-term offset and the rate of change of the real-time temperature difference relative to time; when the degradation index exceeds a preset threshold, determining that the working fluid is leaking or that non-condensable gas is accumulating, generating an early warning signal for predictive maintenance feedback.

[0041] Specifically, based on the extracted rate of change of real-time temperature difference relative to time, i.e., the fluctuation range of the temperature difference between the evaporation section and the condensation section of the heat pipe per unit time, it reflects the dynamic response speed of the heat pipe's heat transfer capacity to changes in heat load. A sudden increase usually indicates an abnormal deterioration in the heat transfer state and the long-term deviation between the real-time temperature difference and the theoretical temperature difference under the heat pipe's design conditions, i.e., the difference between the moving average of the measured temperature difference over 14 consecutive days and the theoretical temperature difference under the rated operating conditions specified by the heat pipe at the factory. By using a long-term moving average, the influence of temporary factors such as short-term power generation fluctuations, sudden changes in ambient temperature, and gust interference is completely eliminated, and it can stably reflect the overall trend of the heat pipe's long-term heat transfer performance. These two complementary features construct a quantitative heat pipe performance degradation index, a dimensionless comprehensive evaluation index with a value range of 0 to 1. The larger the value, the more severe the degradation of the heat pipe's heat transfer capacity relative to the design value. 0 represents the ideal performance state of a brand-new heat pipe, and 1 represents the heat pipe completely losing its heat transfer capacity.

[0042] Specifically, the module first calibrates the weighting coefficients of two features using long-term historical operating data of the wind farm and accelerated aging experiments of heat pipes in the laboratory. Among them, the long-term offset is the core dominant factor reflecting the essence of performance degradation, and the weighting coefficient is... The real-time temperature difference change rate is set to 0.7, serving as an auxiliary dynamic factor reflecting the rate of fault deterioration, with a weighting coefficient of 0.7. Set to 0.3, the formula for calculating the decay index I is: ; in This is the long-term offset calculated so far. This represents the theoretical temperature difference of the corresponding heat pipe model under rated heat load and standard orientation. This represents the current 1-hour moving average rate of change of real-time temperature difference. The maximum allowable rate of temperature change under normal operating conditions will be calculated first. A ±5℃ / h limiting process is implemented to eliminate instantaneous abnormal change rates caused by sudden increases or decreases in power generation due to grid dispatch, thus avoiding misjudgment.

[0043] After the index calculation is completed, the module further distinguishes the two most common heat pipe failure modes by combining the index's trend with the correlation of operating conditions: If the decay index shows a monotonically gradual increase, that is, the daily growth rate is stable between 0.001 and 0.005 without obvious jumps, and the temperature difference deviation remains basically consistent under high and low power generation load conditions, it is determined to be a working fluid leakage. That is, due to long-term vibration fatigue and salt spray corrosion, micron-level gaps appear on the heat pipe weld sealing surface, causing the internal heat transfer working fluids such as anhydrous ethanol and acetone to slowly evaporate and be lost, reducing the effective vapor volume in the heat pipe and the heat transfer capacity. The decay index decreases linearly with the remaining amount of working fluid. If the decay index shows a step-like increase in stages, and the temperature difference deviation increases significantly under low power generation load (i.e., below 30% of rated power) and decreases under high load, it is determined to be the accumulation of non-condensable gases. This means that non-condensable gases such as air and nitrogen are left over from the heat pipe manufacturing process or seep in through sealing defects during operation. These gases cannot participate in phase change heat transfer and will accumulate at the end of the condensing section to form a gas plug, which hinders the contact between the gaseous working fluid and the condensing wall. At low load, the steam pressure is low, and the gas plug occupies a larger proportion of the condensing area, which has a more significant impact on heat transfer.

[0044] Subsequently, the module pre-calibrated two levels of warning thresholds through failure experiments. The yellow warning threshold was set at 0.2, corresponding to a heat pipe heat transfer capacity degradation of approximately 20%, still safe to operate but with performance deterioration. The red alarm threshold was set at 0.5, corresponding to a heat pipe heat transfer capacity degradation of approximately 50%, approaching the lower limit of safe operation and with the risk of overheating at any time. When the degradation index exceeds the yellow warning threshold, the module generates a first-level warning signal, including the current precise value of the degradation index, a preliminary determination of the fault type, an estimated remaining service life based on a linear extrapolation of the daily growth rate of the index, and a recommended maintenance cycle. When the index exceeds the red alarm threshold, a second-level emergency alarm signal is generated, simultaneously triggering the local audible and visual alarms in the cabin and immediately sending instructions to the execution module. The thermal management system is forcibly switched to an auxiliary enhanced cooling mode to increase the condenser fan speed to compensate for the insufficient heat transfer capacity of the heat pipes. All early warning signals are uploaded in real time to the equipment health management platform of the wind farm's central monitoring system via industrial Ethernet, and simultaneously pushed to the mobile terminals of maintenance personnel. Standardized maintenance work orders are automatically generated, clearly indicating the fault type, required tools and spare parts. For example, for working fluid leakage, the corresponding model of heat transfer working fluid and leak detector are required; for non-condensable gas accumulation, a vacuum pump is required, guiding maintenance personnel to carry out precise maintenance. After maintenance is completed, the module automatically collects 7 days of normal operation data, recalculates the baseline theoretical temperature difference and normal change rate of the heat pipes, updates the calculation baseline of the decay index, and completes system self-calibration.

[0045] In one possible implementation, the processing module 20 performs online correction of the physical heat transfer limit of the gravity-type loop heat pipe based on extracted features, including: calculating the current effective return head based on the real-time tilt angle of the nacelle, combined with the vertical height difference between the condenser and evaporator, the working fluid density, and the gravitational acceleration; obtaining the standard heat transfer limit and the corresponding standard return head under the heat pipe design conditions, and constructing a head correction relationship; and performing online correction of the physical heat transfer limit based on the current effective return head and the head correction relationship, wherein the corrected physical heat transfer limit is calculated using the current effective return head and the head correction relationship.

[0046] Specifically, firstly, the module obtains the real-time tilt angle of the nacelle after a 10-second moving average processing from the extracted state features. This is the tilt angle of the nacelle chassis relative to the horizontal plane, collected by a high-precision MEMS tilt sensor installed at the geometric center of the nacelle. The tilt angle primarily originates from the yaw adjustment of the wind turbine to track the optimal wind direction and the long-term settlement of the tower. The design allowable range is ±25°. The moving average is used to eliminate instantaneous tilt angle fluctuations caused by mechanical vibration during yaw. The module then calculates the current effective return head. The effective return head is the sole driving force for the condensed liquid working fluid to flow back to the evaporation section along the return pipeline. It is generated solely by the vertical height difference between the condensation and evaporation sections. Nacelle tilt reduces the projection of the designed vertical height difference onto the direction of gravity, thus decreasing the return power. The specific calculation is as follows: First, based on the synchronously extracted real-time temperature of the heat pipe condensation section, the module looks up the saturated liquid density at the corresponding temperature from a pre-stored working fluid property parameter table. For example, when anhydrous ethanol is used as the working fluid, the density at 25°C is 789 kg / m³. 3 30℃ corresponds to 785kg / m 3 To avoid introducing temperature errors by using a fixed density value; the second step is to determine the vertical height difference between the condenser and evaporator as specified by the heat pipe manufacturer. Such as 1.2m and local gravitational acceleration =9.81m / s 2 The third step is to substitute the formula: ; Where θ is the real-time cabin tilt angle after moving average. The projection coefficient representing the vertical elevation difference of the design along the direction of gravity; for example, when When =15°, 15°≈0.9659, the effective vertical height difference is 1.2×0.9659≈1.159m, substituting... =789kg / m 3 achievable ≈789×9.81×1.159≈8980Pa, approximately 9kPa.

[0047] Next, the module retrieves the design operating condition standard parameters for this model of heat pipe from the non-volatile memory, that is, the baseline operating conditions calibrated in the laboratory before the heat pipe leaves the factory, i.e., the nacelle level. Ideal conditions under standard ambient temperatures of 0°C and 25°C, and rated working fluid filling capacity, including standard reflux head. The backflow head under design conditions is calculated using the following formula: ; in, Let be the density of the saturated liquid working fluid at 25℃. Substituting the values, we can obtain... ≈789×9.81×1.2≈9296Pa, approximately 9.3kPa, and the standard physical heat transfer limit. This refers to the maximum heat power that the heat pipe can stably transfer under design operating conditions. Exceeding this value will cause the working fluid in the evaporation section to dry out and the temperature to rise sharply, resulting in failure phenomena, such as this type of heat pipe. =12kW, then the pre-calibrated head correction relationship is invoked. This is a set of continuous curves reflecting the nonlinear mapping relationship between the return head and the heat transfer limit. Specifically, it is constructed as follows: an adjustable-angle heat pipe performance test bench is built in the laboratory to simulate all operating angles of the nacelle from 0° to 25°, corresponding to the dimensionless head ratio: ; The input heat load was gradually increased from 1 to approximately 0.9063 cos25°C. At each head ratio, the input heat load was gradually increased until the temperature in the evaporation section experienced a sudden increase exceeding 5°C / min, at which point it was determined to be a drying failure. The actual heat transfer limit under this condition was recorded. ; Calculate the corresponding correction factor: ; Multiple discrete ( , Calibration data, for example =1 =1、 =0.95 =0.88、 =0.9 =0.75、 =0.85 =0.58, which intuitively reflects the accelerated decay characteristics of the heat transfer limit under low head; finally, a cubic spline interpolation algorithm is used to fit the discrete data into a continuously differentiable correction curve, i.e., the head correction relationship. The data is then stored in the system memory, and each heat pipe is individually calibrated to eliminate the effects of manufacturing tolerances.

[0048] Finally, the module performs the corrected physical heat transfer limit calculation, the specific process of which is as follows: The first step is to calculate the current dimensionless pressure head ratio: ; The second step is to substitute r into the pressure head correction relationship. The corresponding correction coefficient is calculated through cubic spline interpolation. For example, when r=0.966, the interpolation yields f≈0.93. The third step is to substitute the values ​​into the formula: ; The final correction value is calculated as 12kW × 0.93 ≈ 11.16kW. This value is updated every second to keep pace with the data acquisition and feature extraction frequency. At the same time, the correction values ​​of the most recent 10 minutes are stored for smooth calculation of subsequent heat load margin. If the nacelle tilt angle exceeds the design limit of 25°, the system will automatically trigger a power limit warning to avoid overheating.

[0049] In one possible implementation, the processing module 20 assesses the current thermal management requirements based on the corrected physical heat transfer limit, including: assessing the current actual heat load based on the real-time temperature difference between the evaporation section and the condensation section combined with the total thermal resistance of the heat pipe; calculating the heat load margin between the current actual heat load and the corrected physical heat transfer limit; and comprehensively comparing the real-time temperature of the generator winding, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time ambient temperature with the heat load margin to assess the current thermal management requirement mode.

[0050] Specifically, the first step is to quantitatively assess the current actual heat load. Here, actual heat load refers to the total heat power currently being transferred by the gravity-loop heat pipe, consisting of the total heat generated by core components such as the generator windings, gearbox, and converter. The assessment first utilizes the real-time temperature difference between the evaporator and condenser sections obtained from the feature extraction process. Specifically, it's the difference between the average temperature of multiple points on the evaporator section's 3×3 array thermocouples and the average temperature of multiple points on the condenser section's 3×3 array thermocouples at the same time stamp. Multi-point averaging eliminates measurement errors caused by localized hot spots. This is combined with the total thermal resistance of the heat pipe, which refers to the total thermal resistance from the outer wall of the evaporator section to the outer wall of the condenser section. This resistance is composed of the evaporator section wall thermal resistance, the working fluid evaporation thermal resistance, the steam flow thermal resistance, the working fluid condensation thermal resistance, and the condenser section wall thermal resistance connected in series. When the heat pipe is operating normally and without significant performance degradation, this resistance is approximately constant and is calibrated by the factory test. =0.008℃ / W, substituting into the basic formula for heat conduction: ; The actual heat load is calculated, for example, when the real-time temperature difference... At 64℃, =64℃ / 0.008℃ / W=8000W=8kW, meaning that there is no need to install expensive heat flow sensors on the heating components. High-precision measurement of heat load can be achieved indirectly through the temperature characteristics of the heat pipe itself, which greatly reduces the system cost.

[0051] Next, we calculate the heat load margin, a dimensionless core indicator that directly reflects the remaining heat dissipation capacity of the heat pipe system. A larger margin indicates greater unused heat transfer potential of the heat pipe and a lower likelihood of overheating. The specific calculation process is as follows: First, retrieve the corrected physical heat transfer limit obtained from the online correction in the previous step: For example, the correction value is 11.16kW when the nacelle tilt angle is 15°; the second step is to substitute into the formula: ; To calculate the margin, substitute the values ​​above to obtain: =(11.16kW-8kW) / 11.16kW×100%≈28.3%; If the calculation result is negative, it means that the current heat load has exceeded the maximum heat transfer capacity of the heat pipe, and there is a risk of failure due to the working fluid in the evaporation section drying up and the temperature rising sharply.

[0052] Finally, a multi-parameter comprehensive thermal management demand mode assessment is conducted. This assessment does not rely solely on the single indicator of thermal load margin, but cross-validates it with the real-time temperature of the generator winding, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time temperature of the environment. It also integrates two feedforward correction logics and a continuous confirmation mechanism, ultimately classifying the demand modes into four mutually exclusive and fully covered categories: The first category is the passive heat dissipation sufficient mode. The judgment condition is that the thermal load margin is greater than the first preset proportional threshold, usually 30%, and the real-time temperature of the generator winding is lower than the warning upper limit corresponding to its F-class insulation, usually 120℃; the real-time temperature of the gearbox oil is lower than the warning upper limit of lubricating oil oxidation, usually 80℃; and the real-time temperature of the converter IGBT module is lower than the warning upper limit of junction temperature, usually 75℃. For example, when the margin is 35% and all core temperatures are lower than the threshold, it is judged as this mode.

[0053] The second type is the auxiliary enhanced heat dissipation demand mode, which adopts the OR logic trigger. That is, it is triggered immediately when the heat load margin is less than the second preset proportional threshold, usually 10%, or when the temperature of any core component exceeds its warning limit. At the same time, this mode will introduce a natural convection equivalent heat dissipation power feedforward correction threshold: when the natural convection equivalent heat dissipation power calculated based on the ambient wind speed and wind direction is greater than 5kW, the second threshold is increased from 10% to 15% to use natural wind to assist heat dissipation and reduce fan start-up; when the equivalent heat dissipation power is less than 2kW, the second threshold is reduced to 5% to intervene in active heat dissipation in advance to avoid thermal shock.

[0054] The third type is the low-temperature preheating demand mode. The determination condition is that the real-time ambient temperature is lower than the preset low-temperature threshold, usually -10℃, and the real-time gearbox oil temperature is lower than the allowable starting lower limit, usually -5℃. It is triggered only when both conditions are met at the same time to avoid unnecessary preheating energy consumption.

[0055] The fourth type is the performance degradation alarm mode. The judgment condition is that the heat pipe performance degradation index exceeds the preset degradation threshold, which is 0.2 and can be adjusted according to the actual situation. This mode is judged in parallel with the other three types. Once triggered, the corresponding control strategy will be superimposed. In order to avoid frequent mode switching caused by instantaneous wind speed fluctuations and power generation jumps, all judgment results must be confirmed by three consecutive 1-second cycles. Only when the same mode condition is met three times in a row will the mode be officially switched.

[0056] In one possible implementation, the processing module 20 evaluates the current thermal management demand mode, including: when the heat load margin is greater than a first preset proportional threshold, and the real-time temperature of the generator winding, the real-time temperature of the gearbox oil, and the real-time temperature of the converter module are all lower than their respective corresponding warning upper limit temperatures, it is evaluated as a passive heat dissipation sufficient mode; when the heat load margin is less than a second preset proportional threshold, or when any of the real-time temperatures of the generator winding, the gearbox oil, and the converter module exceed their respective corresponding warning upper limit temperatures, it is evaluated as an auxiliary enhanced heat dissipation demand mode; when the real-time ambient temperature is lower than a preset low temperature threshold, and the real-time temperature of the gearbox oil is lower than the allowable start-up lower limit temperature, it is evaluated as a low temperature preheating demand mode; when the heat pipe performance degradation index exceeds a preset degradation threshold, it is evaluated as a performance degradation alarm mode.

[0057] Specifically, the module first retrieves a full-scale judgment threshold system from non-volatile memory, which has been calibrated in the laboratory and verified in the field. The first preset proportional threshold is set at 30%, not an empirical value, but rather a safety margin for heat dissipation to cope with instantaneous thermal shocks, based on the characteristic that wind turbine power generation fluctuations can reach 80% of rated power. This ensures that even if power generation suddenly increases, the heat pipe has sufficient capacity to handle the additional heat load without triggering overheating. The generator winding warning upper limit temperature is set at 120℃, corresponding to the maximum allowable temperature of 155℃ for F-class insulation materials, reserving a 35℃ thermal shock buffer zone to prevent accelerated aging of the winding insulation due to short-term overheating. The gearbox oil warning upper limit temperature is set at 80℃, lower than the oxidation initiation temperature of mineral gear oil (90℃), which can extend the service life of the lubricating oil by more than 2 times. The converter module warning upper limit... The temperature is set to 75℃, corresponding to the safe threshold of 150℃ junction temperature for IGBT power devices, to prevent thermal breakdown of semiconductor devices due to excessive junction temperature; the second preset proportional threshold is set to 10%, which is the critical margin for gravity-type loop heat pipes to approach dry-out failure. Below this value, the internal working fluid reflux rate of the heat pipe is insufficient to replenish the consumption of the evaporation section, and the heat transfer stability will drop sharply; the preset low temperature threshold is set to -10℃, which is the dividing point of continuous low temperature conditions commonly seen in northern wind farms during winter; the allowable start-up lower limit temperature is set to -5℃, corresponding to 5℃ above the pour point of No. 46 wind turbine gear oil. Below this temperature, the viscosity of the lubricating oil will increase sharply, leading to poor gear meshing lubrication and aggravated wear; the preset degradation threshold is set to 0.2, corresponding to the maintenance critical point where the heat pipe's heat transfer capacity decreases by 20% relative to the design value. At this point, the heat pipe can still operate for a short period of time, but has entered the irreversible degradation stage.

[0058] Subsequently, the module initiates independent judgment logic for four modes in parallel: The first type, passive heat dissipation sufficient mode, is triggered by the AND logic of satisfying all conditions, that is, the heat load margin must be greater than 30% at the same time, and the real-time temperatures of the generator winding, gearbox oil, and converter module must all be lower than their respective warning limits. For example, when the heat load margin calculated based on the corrected physical heat transfer limit is 34%, and the generator winding temperature is 106℃, the gearbox oil temperature is 73℃, and the converter module temperature is 67℃, it is initially judged to be in this mode. The goal of this mode is to maximize the advantage of energy-saving heat dissipation of gravity loop heat pipe.

[0059] The second type of auxiliary enhanced heat dissipation demand mode is triggered by an OR logic that satisfies any condition. That is, as long as the heat load margin is less than 10%, or the temperature of any core component exceeds the warning limit, this mode is initially determined. For example, when the heat load margin drops to 9.2%, even if all core temperatures are still below the warning level, or when the heat load margin is still 24% but the generator winding temperature rises to 123°C, this mode will be triggered. At the same time, this mode will introduce a feedforward correction of the second threshold for the equivalent heat dissipation power of natural convection: when the equivalent heat dissipation power of natural convection calculated based on the ambient wind speed, wind direction and condenser azimuth angle is greater than 5kW, the second threshold is increased from 10% to 15% to make full use of natural wind to assist heat dissipation and reduce the number of fan starts; when the equivalent heat dissipation power is less than 2kW, the second threshold is reduced to 5% to intervene in active heat dissipation in advance to cope with thermal shock.

[0060] The third type of low-temperature preheating demand mode also uses the AND logic trigger, that is, the ambient real-time temperature must be below -10℃ and the gearbox oil real-time temperature must be below -5℃ at the same time. For example, when the ambient temperature is -13℃ but the gearbox oil temperature is still -2℃, preheating will not be triggered to avoid unnecessary energy waste. The fourth type of performance degradation alarm mode is the highest priority mode for independent parallel judgment. As long as the heat pipe performance degradation index constructed based on the long-term temperature difference offset and change rate exceeds 0.2, regardless of whether other parameters are normal, it is initially judged to be this mode to ensure that the heat pipe hidden faults can be identified first.

[0061] Subsequently, to prevent control jitter caused by large fluctuations in wind power generation and sensor noise, i.e. frequent start-stop of the actuator between different modes, which not only increases energy consumption but also significantly shortens equipment life, all preliminary judgment results must be confirmed for consistency over three consecutive 1-second cycles. Only when the data from three consecutive samplings meet the judgment conditions of the same mode will the mode identifier be officially output to the execution module. For example, if a sudden gust of wind causes the heat load margin to drop to 9% briefly, but then rises back to 11% within 1 second, the auxiliary cooling mode switching will not be triggered.

[0062] In one possible implementation, the execution module 30 generates corresponding thermal management action instructions based on the current thermal management requirements, including: when the current thermal management requirement mode is a passive heat dissipation sufficient mode, generating an instruction to maintain a fully passive state, with all active actuators on standby; when the current thermal management requirement mode is an auxiliary enhanced heat dissipation requirement mode, generating an instruction to start the auxiliary fan on the condenser side, and adjusting the fan speed in a closed loop according to the size of the heat load margin; when the current thermal management requirement mode is a low temperature preheating requirement mode, generating an intermittent power-on instruction to the electric heating element embedded in the gearbox oil pan or generator end cover; when the current thermal management requirement mode is a performance degradation alarm mode, generating alarm and maintenance prompt information to the wind farm central monitoring system, and switching to the backup auxiliary heat dissipation circuit.

[0063] Specifically, the system first receives the thermal management demand mode identifier and corresponding control parameters output by the industrial Ethernet bus processing module 20, including thermal load margin, core component temperature deviation, and natural convection equivalent heat dissipation power. It also collects real-time operating status feedback signals from all actuators. Subsequently, it generates and executes differentiated, precise action commands for four demand modes: when the demand is a passive heat dissipation sufficient mode, it generates a command to maintain a fully passive state, meaning heat transfer is achieved entirely through the natural gravity circulation of the working fluid inside the gravity-type loop heat pipe, resulting in zero-power heat dissipation without the intervention of any active energy-consuming equipment. The core energy-saving advantage of the gravity-type heat pipe thermal management system lies in its state of cutting off the main power supply to all active actuators, such as the auxiliary fan on the condenser side, the electric heating element of the gearbox, and the standby loop circulation pump, so that they enter a low-power state with only a standby current of less than 1mA. At the same time, it continuously collects the real-time temperature of the generator winding, gearbox oil, converter module, and the temperature difference between the two ends of the heat pipe at a frequency of 1Hz. Once the judgment conditions of other modes are met for three consecutive 1-second cycles, the corresponding actuator is immediately awakened to complete the mode switching without delay, maximizing the natural advantage of the gravity-type loop heat pipe's energy-saving heat dissipation.

[0064] When the demand is for enhanced heat dissipation, the command to start the main and auxiliary fans on the condenser side is first generated. Then, the fan speed is adjusted based on a closed loop. This refers to a continuous dynamic adjustment logic that uses the temperature of the generator core components as the controlled variable and the heat load margin as the feedforward variable to achieve fine temperature control. The initial fan speed is first calculated based on the current heat load margin M. ; For example, when the heat load margin is 5%, the initial speed is set to 95% of the rated speed for the fan. Then, a PID closed-loop control loop is established with the generator winding temperature as the controlled object. The deviation between the winding temperature feedback value and the target value (typically 110℃) is collected in real time, and the fan output power is dynamically adjusted. Simultaneously, the equivalent heat dissipation power from natural convection output by the processing module is used as a feedforward to correct the PID gain parameter. When the equivalent heat dissipation power is greater than 5kW, the proportional gain is reduced to avoid frequent speed fluctuations; when the equivalent heat dissipation power is less than 2kW, the proportional gain is increased to accelerate the thermal shock response speed. If the main fan is running at full capacity... If the core temperature does not decrease after 10 seconds of rapid operation, one or two backup auxiliary fans will be activated sequentially to form a multi-machine collaborative cooling mechanism. Simultaneously, the fan speed will be limited to 30%-100% of the rated speed to prevent fan blockage at low speeds or excessive noise and vibration at high speeds. When the demand is for low-temperature preheating, an intermittent power-on command is generated, employing an energy-saving heating control method of power-on heating-power-off temperature equalization-temperature detection cycle. This utilizes heat conduction during the power-off phase to achieve uniform overall oil temperature rise, sending the command to the armored electric heating element embedded at the bottom of the gearbox oil pan. The command first determines the temperature based on the current gearbox oil temperature. The initial duty cycle is calculated based on the difference between the allowable lower start-up temperature, typically -5°C. ; For example, when the oil temperature is -15℃, the initial duty cycle is set to 50%, which means adopting a cycle mode of 30 seconds of power on and 30 seconds of power off. After each heating cycle, the overall oil temperature is detected. If the oil temperature is still below -5℃, the current duty cycle is maintained to continue heating. If the oil temperature rises above -7℃, the duty cycle is gradually reduced to 20% for heat preservation. At the same time, the working current and surface temperature of the heating element are monitored in real time. If the current exceeds 1.2 times the rated value or the surface temperature exceeds 120℃, the heating power supply is immediately cut off and a local overheat protection alarm is triggered. When the oil temperature naturally rises above 0℃ after the generator starts, the preheating process is automatically terminated and the power supply to the heating element is completely cut off.

[0065] When the demand is in performance degradation alarm mode, a tiered alarm and standardized maintenance prompts are first generated. This information includes the precise value of the current heat pipe performance degradation index, the estimated remaining service life based on linear extrapolation of the daily growth rate of the index, the preliminary fault type (working fluid leakage or non-condensable gas accumulation), and the corresponding maintenance procedures and spare parts list. This information is uploaded in real-time via industrial Ethernet to the equipment health management platform of the wind farm's central monitoring system and pushed to the mobile terminals of maintenance personnel. Simultaneously, the corresponding level of audible and visual alarm is triggered in the local control cabinet of the nacelle. Then, a seamless switch to the backup auxiliary cooling circuit, independent of... The emergency liquid cooling channel of the main gravity loop heat pipe system has an independent radiator, circulation pump and fan, which is physically isolated from the main heat pipe system. It seamlessly takes over the heat dissipation task when the performance of the main heat pipe deteriorates, ensuring the safe operation of the unit. First, the coolant circulation pump of the backup circuit is started. After the circulation stabilizes for 3 seconds, the backup radiator fan is started. At the same time, the heat load ratio of the main heat pipe system is gradually reduced. If the backup circuit is running at full load and still cannot meet the heat dissipation demand, the generator power limiting command is triggered, and the power generation is gradually reduced to 70% of the rated power to ensure that the unit can operate safely and stably before the maintenance personnel arrive.

[0066] In addition, the execution module has a built-in independent hardware watchdog circuit. When a fault is detected in the processing module or a communication interruption lasts for more than 5 seconds, it automatically switches to a safe operation mode, starts all auxiliary fans to run at rated speed and triggers an emergency fault alarm to prevent the system from going out of control and causing a major accident. At the same time, the module will feed back the operating status of all actuators, namely the real-time fan speed, the on / off status of the heating element, and the working current and pressure of the backup circuit, to the processing module 20 in real time at a frequency of 1Hz, forming a control loop of perception-decision-execution-feedback, and continuously iterating and optimizing the accuracy and response speed of the control strategy.

[0067] In one possible implementation, the execution module 30 further includes: calculating the equivalent heat dissipation power of natural convection based on the ambient wind speed and direction, combined with the windward area of ​​the condenser and the current azimuth angle; using the equivalent heat dissipation power as a feedforward quantity to pre-correct the trigger threshold for starting the auxiliary fan in the auxiliary enhanced heat dissipation demand mode, wherein when the equivalent heat dissipation power is greater than the preset sufficient air cooling threshold, the trigger threshold is increased, and when the equivalent heat dissipation power is less than the preset insufficient air cooling threshold, the trigger threshold is decreased.

[0068] Specifically, the module first obtains the real-time ambient wind speed (in m / s) processed by a 5-second moving average from the feature extraction output of the processing module 20. This wind speed is collected by an ultrasonic anemometer on the top of the cabin. The moving average is used to eliminate the peak fluctuations of instantaneous gusts. The real-time ambient wind direction refers to the angle between the direction of the wind and due north, ranging from 0° to 360°. The average temperature of the heat pipe condensing section is also obtained. and real-time ambient temperature Simultaneously, the current azimuth angle of the condenser is obtained from the nacelle yaw system, which is the angle between the normal direction of the condenser's heat dissipation surface and true north. Since the condenser is fixedly installed at the stern of the nacelle, this angle changes in real time with the nacelle yaw. This angle is used to calculate the relative incident angle between the wind and the condenser, and to retrieve pre-stored condenser inherent parameters: design windward area. That is, the empirical formula for the total projected area of ​​the condenser heat dissipation fins and the experimentally calibrated convective heat transfer coefficient.

[0069] Next, we will calculate the equivalent heat dissipation power of natural convection. The equivalent heat dissipation power of natural convection refers to the heat power carried away from the condenser surface by natural wind through forced convection, which is equivalent to the additional heat dissipation capacity provided by the active fan. This is done in three steps: The first step is to calculate the wind incidence angle θ, which is the angle between the wind direction and the condenser normal direction. The formula is: ; For the ambient wind direction angle, Given the current azimuth angle of the condenser, if the calculation result... If the angle is greater than 90°, then take 180°. Because when wind enters from the back of the condenser, the effective heat dissipation area decreases linearly with increasing angle. For example, when the ambient wind direction angle is 180° and the azimuth angle of the condenser is 175°, =5°; Step 2: Calculate the effective windward area: ; Substituting the values, we can obtain =2.5㎡×cos5°≈2.49㎡; The third step is to calculate the equivalent heat dissipation power based on Newton's law of cooling: ; in The convective heat transfer coefficient is calculated using the empirical formula h = 12.12 + 1.16v, calibrated for outdoor wind turbine condensers, where v is the real-time ambient wind speed in m / s, and ΔT is the difference between the average temperature of the condensing section and the ambient temperature. ; For example, when v = 6 m / s, ,like =45℃ =25℃ =20℃, substituting into the equation yields... =19.08×2.49×20≈950W, approximately 0.95kW. The entire calculation process is performed once every second, and the result is output after a 3-second sliding average to avoid power surges caused by gusts.

[0070] The module then calculates... The activation trigger threshold for the auxiliary enhanced heat dissipation demand mode in the dynamic correction processing module 20 is the second preset proportional threshold for evaluating auxiliary heat dissipation demand. The default value is 10% heat load margin. When it is lower than this value, the system will start the auxiliary fan. The module pre-stores two field-verified critical thresholds: a preset sufficient air cooling threshold, usually 5kW, corresponding to sufficient auxiliary heat dissipation provided by natural wind, without the need for active fan intervention; and a preset insufficient air cooling threshold, usually 2kW, corresponding to weak natural wind heat dissipation capacity, requiring active cooling to be activated in advance. The specific correction logic is as follows: when When the power is >5kW, it indicates that the natural wind cooling capacity is sufficient. At this point, the trigger threshold is increased from the default 10% to 15%, meaning the fan will only start when the heat load margin drops to 15%, maximizing the duration of passive cooling. When the power is less than 2kW, it indicates that the natural wind cooling capacity is insufficient. In this case, the trigger threshold will be lowered from 10% to 5%, and the auxiliary fan will be started in advance to deal with potential thermal shock. When 2kW ≤ When the power is ≤5kW, the default 10% trigger threshold is kept unchanged. At the same time, in order to avoid frequent threshold switching caused by gusts, i.e. to control jitter, all threshold switching is set to a 10-second hold time. The threshold will only be updated if the same correction condition is met for 10 consecutive seconds. The corrected threshold will be synchronized to the demand evaluation unit of the processing module 20 in real time through the internal bus, overriding its original fixed threshold parameter.

[0071] In one possible implementation, the execution module 30 further includes: acquiring power generation prediction data of the wind turbine within a future time window; when the power generation prediction data exceeds a preset power change rate threshold by a sudden increase, initiating an auxiliary enhanced heat dissipation demand mode or lowering the judgment threshold of the passive heat dissipation sufficient mode before the power surge occurs; when the power generation prediction data is continuously lower than a preset low power threshold, entering a low temperature preheating preparation state, and predicting whether preheating is needed based on the ambient temperature.

[0072] Specifically, the module first obtains the power generation prediction curve within the future time window from the ultra-short-term power prediction subsystem of the SCADA system via the wind farm industrial Ethernet bus. The future time window is the 15-30 minute time interval with the highest prediction accuracy of the wind farm. The average absolute error of power prediction within this window is less than 8%, which can meet the time requirements for advance control of thermal management. The data resolution is 1 minute, and the complete curve is updated every 1 minute. The module performs a 3-point moving average processing on the obtained raw prediction data to eliminate random spike noise introduced by the prediction algorithm.

[0073] The module then determines a power surge based on the processed prediction curve. The preset power change rate threshold is a relative value based on the generator winding thermal time constant calibration. The thermal time constant is typically about 2 minutes and is set to 10% of the rated power per minute. For example, a 2MW rated power wind turbine corresponds to 200kW / minute. The specific determination process is as follows: iterate through the predicted power points for the next 15 minutes and calculate the maximum power change rate within any adjacent 5-minute window. ,like If the power surge exceeds 10% of the rated power per minute, it is considered a power surge event, and the start time and peak power of the surge are recorded. When a power surge is detected, the module will perform hierarchical feedforward control based on the magnitude of the surge: if the surge magnitude is less than 30% of the rated power, a threshold correction command is sent to the processing module 20 to reduce the judgment threshold of the passive heat dissipation sufficient mode, i.e., the first preset ratio threshold, from the default 30% to 20%, so that the processing module can determine that auxiliary heat dissipation is needed when the heat load margin decreases earlier, and reserve time for heat transfer capacity to build up in advance.

[0074] If the power surge is greater than or equal to 30% of the rated power, a pre-start command will be generated 5 minutes before the start of the power surge, starting the condenser-side auxiliary fan to operate at 30% of the rated speed at low load. This allows the heat pipe system to establish a high heat transfer state in advance, preventing overshoot of more than 10°C in the winding temperature after the power surge. If there is no power surge within the predicted future time window, and the predicted power generation data remains below the preset low power threshold for more than 2 hours (the preset low power threshold is set at 10% of the rated power, i.e., 200kW for a 2MW wind turbine), then the unit will be determined to enter a long-term low-load standby state. In this state, the module enters a low-temperature preheating preparation state. First, based on the current gearbox oil temperature, ambient temperature, and the natural heat dissipation coefficient of the oil, the estimated time for the oil temperature to drop to the allowable starting lower limit is calculated. The lower limit is usually -5℃. If the estimated time is less than 12 hours and the real-time ambient temperature is below -5℃, the intermittent heating preparation parameters are set in advance. When the oil temperature drops to -3℃, the low duty cycle 20% preheating is automatically started to avoid poor gear meshing lubrication due to excessively low oil temperature. If it is predicted that the power will rise to more than 30% of the rated power within the next 4 hours, the preheating preparation is canceled to avoid unnecessary energy waste.

[0075] The module also has a confidence level verification mechanism. Full feedforward action is only executed when the official confidence level of the power prediction data is ≥85%. When the confidence level is between 60% and 85%, only threshold correction is performed and the fan is not started directly. When the confidence level is <60%, feedback control is relied upon completely. At the same time, all feedforward actions are set with a 10-minute timeout exit mechanism. If the actual power does not increase as predicted, the original control parameters are automatically restored.

[0076] In summary, the gravity-driven loop heat pipe-driven wind turbine thermal management system provided in this application has the following technical effects: By constructing a multi-source heterogeneous data acquisition and fusion system covering the entire chain of generator heat generation, heat pipe heat transfer, and environmental heat dissipation, the physical heat transfer limit of the gravity loop heat pipe is corrected online based on the real-time tilt angle of the nacelle and the physical properties of the working fluid. The heat load margin is calculated by combining the corrected heat transfer limit with the temperature of core components to achieve a layered and accurate assessment of thermal management requirements. At the same time, a dual feedforward control logic of natural convection equivalent heat dissipation power and ultra-short-term power generation prediction is introduced. A heat pipe performance degradation index and predictive maintenance mechanism based on long-term temperature difference offset and rate of change are established. While ensuring the heat dissipation safety of the unit under all operating conditions and avoiding overheating damage and unplanned shutdowns of core components, the energy-saving heat dissipation advantage of the gravity loop heat pipe is maximized, improving the intelligence level and long-term operational reliability of the wind turbine thermal management system.

[0077] Example 2, based on the same inventive concept as the gravity-driven loop heat pipe-driven wind turbine thermal management system in the previous examples, such as... Figure 2 As shown in the embodiment of this application, a thermal management method for a gravity-driven loop heat pipe-driven wind turbine generator is provided. The method includes: Collect multi-source operating status data of the wind turbine generator and gravity loop heat pipe system; fuse and extract features from the multi-source operating status data; correct the physical heat transfer limit of the gravity loop heat pipe online based on the extracted features; assess the current thermal management requirements based on the corrected physical heat transfer limit; generate corresponding thermal management action commands based on the current thermal management requirements; and execute the thermal management control of the wind turbine generator.

[0078] The process of fusing and extracting features from the multi-source operating status data includes: performing time synchronization alignment, outlier removal, and sliding filtering on the multi-source operating status data, which includes generator winding temperature, gearbox oil temperature, converter module temperature, heat pipe evaporator section temperature, heat pipe condenser section temperature, nacelle tilt angle, ambient wind speed, and ambient temperature. The processed multi-source operating status data is then fused into a unified time series sample, and status features are extracted, including the real-time temperature difference between the evaporator and condenser sections, the rate of change of the real-time temperature difference relative to time, the long-term offset of the real-time temperature difference from the theoretical temperature difference under the heat pipe design conditions, the real-time nacelle tilt angle, the real-time generator winding temperature, the real-time gearbox oil temperature, the real-time converter module temperature, and the real-time ambient temperature.

[0079] The method may further include: constructing a heat pipe performance degradation index based on the long-term offset and the rate of change of the real-time temperature difference relative to time; when the degradation index exceeds a preset threshold, determining that the working fluid is leaking or that non-condensable gas is accumulating, generating an early warning signal for predictive maintenance feedback.

[0080] The online correction of the physical heat transfer limit of the gravity-type loop heat pipe based on extracted features includes: calculating the current effective return head based on the real-time tilt angle of the nacelle, combined with the vertical height difference between the condenser and evaporator, the working fluid density, and the gravitational acceleration; obtaining the standard heat transfer limit and the corresponding standard return head under the heat pipe design conditions, and constructing a head correction relationship; and performing online correction of the physical heat transfer limit based on the current effective return head and the head correction relationship, wherein the corrected physical heat transfer limit is calculated using the current effective return head and the head correction relationship.

[0081] The assessment of current thermal management requirements based on the corrected physical heat transfer limit includes: assessing the current actual heat load based on the real-time temperature difference between the evaporation and condensation sections and the total thermal resistance of the heat pipe; calculating the heat load margin between the current actual heat load and the corrected physical heat transfer limit; and comprehensively comparing the real-time temperature of the generator windings, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time ambient temperature with the heat load margin to assess the current thermal management requirement mode.

[0082] The assessment of the current thermal management demand mode includes: when the heat load margin is greater than a first preset proportional threshold, and the real-time temperatures of the generator winding, gearbox oil, and converter module are all lower than their respective corresponding warning upper limit temperatures, it is assessed as a passive heat dissipation sufficient mode; when the heat load margin is less than a second preset proportional threshold, or when any of the real-time temperatures of the generator winding, gearbox oil, and converter module exceed their respective corresponding warning upper limit temperatures, it is assessed as an auxiliary enhanced heat dissipation demand mode; when the real-time ambient temperature is lower than a preset low temperature threshold, and the real-time gearbox oil temperature is lower than the allowable start-up lower limit temperature, it is assessed as a low temperature preheating demand mode; and when the heat pipe performance degradation index exceeds a preset degradation threshold, it is assessed as a performance degradation alarm mode.

[0083] Specifically, based on the current thermal management requirements, corresponding thermal management action commands are generated, including: when the current thermal management requirement mode is passive heat dissipation sufficient mode, a command to maintain a fully passive state is generated, and all active actuators are put on standby; when the current thermal management requirement mode is auxiliary enhanced heat dissipation requirement mode, a command to start the auxiliary fan on the condenser side is generated, and the fan speed is adjusted in a closed loop according to the size of the heat load margin; when the current thermal management requirement mode is low temperature preheating requirement mode, an intermittent power-on command is generated to the electric heating element embedded in the gearbox oil pan or generator end cover; when the current thermal management requirement mode is performance degradation alarm mode, alarm and maintenance prompt information is generated to the wind farm central monitoring system, and the system switches to the backup auxiliary heat dissipation circuit.

[0084] The method may further include: calculating the equivalent heat dissipation power of natural convection based on the ambient wind speed and direction, combined with the windward area of ​​the condenser and the current azimuth angle; using the equivalent heat dissipation power as a feedforward quantity to pre-correct the trigger threshold for starting the auxiliary fan in the auxiliary enhanced heat dissipation demand mode, wherein when the equivalent heat dissipation power is greater than the preset sufficient air cooling threshold, the trigger threshold is increased, and when the equivalent heat dissipation power is less than the preset insufficient air cooling threshold, the trigger threshold is decreased.

[0085] The method may further include: acquiring power generation prediction data of the wind turbine within a future time window; when the power generation prediction data increases by a sudden increase exceeding a preset power change rate threshold, activating the auxiliary enhanced heat dissipation demand mode or lowering the judgment threshold of the passive heat dissipation sufficient mode in advance before the power increase occurs; when the power generation prediction data remains below a preset low power threshold, entering a low temperature preheating preparation state, and predicting whether preheating is needed based on the ambient temperature.

[0086] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A gravity-driven loop heat pipe-driven wind turbine thermal management system, characterized in that, include: The data acquisition module is used to collect multi-source operating status data of wind turbine generators and gravity loop heat pipe systems. The processing module is used to fuse and extract features from the multi-source operating status data, correct the physical heat transfer limit of the gravity loop heat pipe online based on the extracted features, and evaluate the current thermal management requirements based on the corrected physical heat transfer limit. The execution module is used to generate corresponding thermal management action instructions based on the current thermal management requirements and execute the thermal management control of the wind turbine generator.

2. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 1, characterized in that, The processing module performs fusion and feature extraction on the multi-source operating status data, including: The multi-source operating status data is processed by time synchronization alignment, outlier removal and sliding filtering. The multi-source operating status data includes generator winding temperature, gearbox oil temperature, converter module temperature, heat pipe evaporation section temperature, heat pipe condensation section temperature, nacelle tilt angle, ambient wind speed and ambient temperature. The processed multi-source operating status data are fused into a unified time series sample, and status features are extracted, including the real-time temperature difference between the evaporation section and the condensation section, the rate of change of the real-time temperature difference relative to time, the long-term offset of the real-time temperature difference from the theoretical temperature difference under the heat pipe design conditions, the real-time tilt angle of the nacelle, the real-time temperature of the generator windings, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time ambient temperature.

3. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 2, characterized in that, The processing module also includes: Based on the long-term offset and the rate of change of real-time temperature difference relative to time, a heat pipe performance degradation index is constructed. When the decay index exceeds a preset threshold, it is determined that there is a working fluid leak or non-condensable gas accumulation, and an early warning signal is generated for predictive maintenance feedback.

4. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 2, characterized in that, In the processing module, the physical heat transfer limit of the gravity-type loop heat pipe is corrected online based on the extracted features, including: Based on the real-time tilt angle of the nacelle, combined with the vertical height difference between the condenser and the evaporator, the working fluid density, and the gravitational acceleration, the current effective backflow head is calculated. Obtain the standard heat transfer limit and the corresponding standard reflux head under the heat pipe design conditions, and construct the head correction relationship; Based on the current effective return head and the head correction relationship, the physical heat transfer limit is corrected online, wherein the corrected physical heat transfer limit is calculated using the current effective return head and the head correction relationship.

5. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 2, characterized in that, In the processing module, the current thermal management requirements are assessed based on the corrected physical heat transfer limit, including: The current actual heat load is assessed based on the real-time temperature difference between the evaporation and condensation sections and the total thermal resistance of the heat pipe. Calculate the heat load margin between the current actual heat load and the corrected physical heat transfer limit; The real-time temperature of the generator windings, the real-time temperature of the gearbox oil, the real-time temperature of the converter module, and the real-time ambient temperature are comprehensively compared with the heat load margin to assess the current thermal management requirements.

6. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 3 or 5, characterized in that, The processing module assesses the current thermal management requirement pattern, including: When the heat load margin is greater than the first preset proportional threshold, and the real-time temperature of the generator winding, the real-time temperature of the gearbox oil, and the real-time temperature of the converter module are all lower than their respective corresponding warning upper limit temperatures, it is evaluated as a passive heat dissipation sufficient mode. When the heat load margin is less than the second preset proportional threshold, or when any of the real-time temperature of the generator winding, the real-time temperature of the gearbox oil, or the real-time temperature of the converter module exceeds the corresponding warning upper limit temperature, it is evaluated as an auxiliary enhanced heat dissipation demand mode. When the real-time ambient temperature is lower than the preset low temperature threshold and the real-time gearbox oil temperature is lower than the allowable starting lower limit temperature, it is evaluated as a low temperature preheating demand mode. When the heat pipe performance degradation index exceeds the preset degradation threshold, it is evaluated as a performance degradation alarm mode.

7. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 6, characterized in that, In the execution module, based on the current thermal management requirements, corresponding thermal management action instructions are generated, including: When the current thermal management requirement mode is passive heat dissipation sufficient mode, a command is generated to maintain a fully passive state, and all active actuators are in standby mode; When the current thermal management demand mode is the auxiliary enhanced heat dissipation demand mode, a command to start the auxiliary fan on the condenser side is generated, and the fan speed is adjusted in a closed loop according to the size of the heat load margin. When the current thermal management demand mode is low temperature preheating demand mode, intermittent power-on commands are generated to the electric heating elements embedded in the gearbox oil pan or generator end cover; When the current thermal management requirement mode is the performance degradation alarm mode, an alarm and maintenance prompt message is generated and sent to the wind farm central monitoring system, and the system is switched to the backup auxiliary heat dissipation circuit.

8. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 6, characterized in that, The execution module also includes: Based on the ambient wind speed and direction, combined with the condenser's frontal area and current azimuth angle, calculate the equivalent heat dissipation power of natural convection. The equivalent heat dissipation power is used as a feedforward quantity to pre-correct the trigger threshold for starting the auxiliary fan in the auxiliary enhanced heat dissipation demand mode. Specifically, when the equivalent heat dissipation power is greater than the preset sufficient air cooling threshold, the trigger threshold is increased, and when the equivalent heat dissipation power is less than the preset insufficient air cooling threshold, the trigger threshold is decreased.

9. The gravity-driven loop heat pipe-driven wind turbine thermal management system according to claim 6, characterized in that, The execution module also includes: Obtain forecast data of power generation from wind turbines within a future time window; When the predicted power generation data exceeds the preset power change rate threshold by a sudden increase, the judgment threshold of the auxiliary enhanced heat dissipation demand mode or the passive heat dissipation sufficient mode is activated in advance before the power surge occurs. When the predicted power generation data is consistently lower than the preset low power threshold, the system enters a low-temperature preheating preparation state, and determines whether preheating is required based on the ambient temperature.

10. A thermal management method for a wind turbine generator driven by a gravity-driven loop heat pipe, characterized in that, The method is implemented by the gravity-driven loop heat pipe-driven wind turbine thermal management system according to any one of claims 1-9, and the method includes: Collect multi-source operational status data of wind turbine generators and gravity-type loop heat pipe systems; The multi-source operating status data is fused and features are extracted. Based on the extracted features, the physical heat transfer limit of the gravity loop heat pipe is corrected online, and the current thermal management requirements are evaluated based on the corrected physical heat transfer limit. Based on the current thermal management requirements, corresponding thermal management action instructions are generated and executed to control the thermal management of the wind turbine generator.

Images