Intelligent control method for braking of wind power generator

Abstract

The application provides an intelligent control method for wind turbine braking, and belongs to the technical field of wind power generation. In the automatic braking control mode, the current operating parameters of the wind turbine are collected, and intelligent judgment is performed in combination with a self-learning model to link control of the electromagnetic unloading device and the mechanical braking device to perform braking operation; in the manual braking control mode, the judgment and execution of the automatic braking control mode are suspended, and the mechanical braking device is directly controlled by the operator, and the manual braking control mode has the highest priority; wherein the self-learning model is trained and iterated based on historical wind turbine operating parameters, historical meteorological data and current meteorological data, and is used for predicting the wind speed change trend and optimizing the braking decision. Through the fusion of the self-learning model and the dual-mode braking control mechanism, the intelligentization, precision and high reliability of the wind turbine braking process are realized, while effectively reducing the false braking, excessive braking and manual intervention while ensuring the safety of the equipment.

Description

Technical Field

[0001] This application belongs to the field of wind power generation technology, specifically relating to an intelligent control method for braking a wind turbine. Background Technology

[0002] Current wind power generation systems generally employ braking methods such as electromagnetic braking, short-circuit braking, and mechanical braking to ensure the safety of the unit under overspeed or extreme wind conditions.

[0003] However, existing braking control strategies mostly rely on fixed threshold triggering, lacking the ability to predict dynamic changes in wind speed. This easily leads to frequent start-stop or braking lag, which not only exacerbates mechanical wear on the transmission system and braking components but may also affect power generation efficiency and equipment lifespan. Especially in continuous extreme wind environments, traditional methods struggle to achieve intelligent matching of braking timing and intensity, often requiring manual intervention, which presents problems such as untimely response and high operational risks. Summary of the Invention

[0004] To address at least one of the technical problems existing in the background art, this application provides an intelligent control method for wind turbine braking. By integrating a self-learning model and a dual-mode braking control mechanism, the method achieves intelligent, precise, and highly reliable wind turbine braking process, effectively reducing false braking, excessive braking, and manual intervention while ensuring equipment safety.

[0005] The technical solution adopted in this application is as follows: This application provides an intelligent control method for wind turbine braking, including an automatic braking control mode and a manual braking control mode. In the automatic braking control mode, the current operating parameters of the wind turbine are collected and intelligent judgment is made in combination with the self-learning model, and the electromagnetic unloading device and the mechanical braking device are linked to perform braking operation. In the manual braking control mode, the judgment and execution of the automatic braking control mode are suspended, and the operator directly controls the mechanical braking device. The manual braking control mode has the highest priority. The self-learning model is trained and iterated based on historical wind turbine operating parameters, historical meteorological data, and current meteorological data to predict wind speed change trends and optimize braking decisions.

[0006] The intelligent control method for wind turbine braking provided in this application achieves a unified approach to safety, intelligence, and operational flexibility by constructing a dual-mode control architecture that combines automatic and manual braking control modes. In automatic braking control mode, the system collects key operating parameters such as wind speed, rotor speed, output voltage, and current in real time. It also integrates a self-learning model trained based on historical operating data, historical meteorological information, and current weather forecasts to dynamically predict and assess wind speed trends. This allows for intelligent decision-making on whether to initiate braking and what braking strategy to adopt. For example, when a brief exceedance of wind speed followed by a rapid decline is predicted, only the low-energy-consumption electromagnetic unloading device is activated for flexible deceleration. Conversely, when a sustained extreme wind is anticipated, electromagnetic unloading and mechanical braking are sequentially activated to ensure the wind turbine's safe shutdown and entry into a protection state. This mechanism significantly improves the foresight and accuracy of braking response, effectively avoiding the frequent start-stop, braking lag, or over-braking problems caused by traditional fixed threshold control, and extending the service life of the transmission system and braking components. Meanwhile, the manual braking control mode is given the highest priority. Once an operator intervenes, the system immediately suspends all automatic judgment logic and directly hands over manual control of the mechanical braking device, ensuring reliable manual intervention capability even in the event of communication interruptions, model failures, or sudden emergencies. Furthermore, the system automatically records operating status, environmental parameters, and operation logs during the switching between the two modes and uploads them to a remote platform, enhancing operational traceability and system transparency. In summary, this method not only improves the operational safety and stability of wind turbine generators under complex wind conditions but also continuously optimizes braking strategies through a data-driven self-learning mechanism, reducing operational costs and reliance on manual intervention, and promoting the intelligent and adaptive development of wind power control systems.

[0007] According to one embodiment of this application, the current operating parameters of the wind turbine include one or more combinations of wind speed, rotor speed, output voltage, and output current.

[0008] According to one embodiment of this application, the self-learning model constructs a regional meteorological prediction model based on historical meteorological data with time scales of hour, day, month, and year, for day-ahead and intraday multi-scale wind speed and direction prediction.

[0009] According to one embodiment of this application, in automatic braking control mode: When the operating parameters meet the first cut-out condition but the wind speed is predicted to drop back to a safe range in the short term, electromagnetic unloading is activated only to reduce the speed, and the brake is automatically released after the operating conditions are met. When the operating parameters meet the second cut-out condition and it is predicted that the wind will continue to be in extreme wind conditions, the electromagnetic unloading will be activated first, and then the mechanical braking device will be activated to completely stop the wind turbine and put it into protection mode.

[0010] According to one embodiment of this application, during the protection state of the wind turbine, the self-learning model continuously determines whether the conditions for releasing the mechanical brake are met. If not, the combined braking state of electromagnetic unloading and mechanical braking is maintained.

[0011] According to one embodiment of this application, in automatic braking control mode, if a fault is detected in the electromagnetic unloading device or the mechanical braking device, the system immediately switches to safety degradation mode, prioritizes the activation of the backup braking mechanism, and issues a fault alarm signal.

[0012] According to one embodiment of this application, when switching to manual braking control mode, the operating parameters, weather conditions and operation log at the time of switching are automatically recorded and uploaded to the remote monitoring platform.

[0013] According to one embodiment of this application, the method further includes: In automatic braking control mode, when the self-learning model predicts that the wind speed will exceed the limit within a preset time window, the early warning mechanism will be activated in advance, and braking preparation prompts will be sent to the operation and maintenance personnel.

[0014] According to one embodiment of this application, the method further includes: After each braking operation is completed, a braking event report is automatically generated. The report includes the reason for braking triggering, the status of the actuator, the duration, environmental parameters, and the model prediction accuracy evaluation index.

[0015] According to one embodiment of this application, the method further includes: Regularly validate and update the self-learning model offline, retrain the model using newly added historical operational data and meteorological data, and decide whether to deploy the new version after comparing the performance of the old and new models through A / B testing. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating the intelligent control method for wind turbine braking provided in an embodiment of this application. Detailed Implementation

[0017] To more clearly illustrate the overall concept of this application, a detailed explanation is provided below with reference to the accompanying drawings.

[0018] Many specific details are set forth in the following description to provide a thorough understanding of this application. However, this application may also be implemented in other ways different from those described herein. Therefore, the scope of protection of this application is not limited to the specific embodiments disclosed below. It should be noted that, unless otherwise specified, the embodiments of this application and the features thereof can be combined with each other.

[0019] In this application, unless otherwise expressly specified and limited, the "above" or "below" of the second feature can mean that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediate medium. In the description of this specification, references to terms such as "an embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples.

[0020] like Figure 1 As shown in the figure, this application provides an intelligent control method for wind turbine braking, including an automatic braking control mode and a manual braking control mode. Step 100: In automatic braking control mode, by collecting the current operating parameters of the wind turbine generator and combining them with the self-learning model for intelligent judgment, the electromagnetic unloading device and the mechanical braking device are linked to perform braking operations.

[0021] Step 200: In manual braking control mode, the judgment and execution of automatic braking control mode are suspended, and the operator directly controls the mechanical braking device. Manual braking control mode has the highest priority.

[0022] The self-learning model is trained and iterated based on historical wind turbine operating parameters, historical meteorological data, and current meteorological data to predict wind speed change trends and optimize braking decisions.

[0023] In step 100, the system operates in automatic braking control mode. In this mode, the control system collects key operating parameters of the wind turbine in real time, including but not limited to wind speed, rotor speed, output voltage, and output current, and simultaneously acquires current meteorological data (such as real-time wind speed, wind direction, and air pressure). This data is input into a trained and continuously iterated self-learning model. Based on a large amount of historical wind turbine operating data, regional historical meteorological records, and real-time weather forecast information, this model has built a multi-timescale (e.g., hourly, daily, and monthly) wind speed change prediction capability. By analyzing the current state and future trends, the model can intelligently determine whether braking needs to be initiated, when to initiate it, and what braking strategy to adopt: for example, when the predicted wind speed briefly exceeds the cut-out threshold but is about to fall back, only the electromagnetic unloading device is triggered for flexible deceleration to avoid unnecessary mechanical wear; when it is determined that a continuous extreme gale will be encountered, the electromagnetic unloading is activated first to buffer kinetic energy, and then the mechanical braking device is activated to completely stop the wind turbine and put it into a protection state. The entire process has transformed from "passive response" to "active prediction," significantly improving the accuracy, safety, and energy efficiency of braking.

[0024] In step 200, the system switches to manual braking control mode, which serves as a safety redundancy mechanism and has the highest priority. Once the operator triggers a manual braking command via a physical switch or human-machine interface, the system immediately suspends all data acquisition, model inference, and control output logic in automatic braking control mode, ensuring that automatic judgment no longer interferes with manual operation. At this time, the operator can directly control the mechanical braking device to perform emergency braking, regardless of the current operating state of the wind turbine or whether the automatic system is functioning normally. This fully considers the actual needs of on-site operation and maintenance, ensuring the operator's absolute control over the equipment in abnormal scenarios such as communication interruptions, sensor failures, model misjudgments, or sudden emergencies, greatly enhancing the system's reliability and emergency response capabilities.

[0025] The intelligent control method for wind turbine braking provided in this application achieves a unified approach to safety, intelligence, and operational flexibility by constructing a dual-mode control architecture that combines automatic and manual braking control modes. In automatic braking control mode, the system collects key operating parameters such as wind speed, rotor speed, output voltage, and current in real time. It also integrates a self-learning model trained based on historical operating data, historical meteorological information, and current weather forecasts to dynamically predict and assess wind speed trends. This allows for intelligent decision-making on whether to initiate braking and what braking strategy to adopt. For example, when a brief exceedance of wind speed followed by a rapid decline is predicted, only the low-energy-consumption electromagnetic unloading device is activated for flexible deceleration. Conversely, when a sustained extreme wind is anticipated, electromagnetic unloading and mechanical braking are sequentially activated to ensure the wind turbine's safe shutdown and entry into a protection state. This mechanism significantly improves the foresight and accuracy of braking response, effectively avoiding the frequent start-stop, braking lag, or over-braking problems caused by traditional fixed threshold control, and extending the service life of the transmission system and braking components. Meanwhile, the manual braking control mode is given the highest priority. Once an operator intervenes, the system immediately suspends all automatic judgment logic and directly hands over manual control of the mechanical braking device, ensuring reliable manual intervention capability even in the event of communication interruptions, model failures, or sudden emergencies. Furthermore, the system automatically records operating status, environmental parameters, and operation logs during the switching between the two modes and uploads them to a remote platform, enhancing operational traceability and system transparency. In summary, this method not only improves the operational safety and stability of wind turbine generators under complex wind conditions but also continuously optimizes braking strategies through a data-driven self-learning mechanism, reducing operational costs and reliance on manual intervention, and promoting the intelligent and adaptive development of wind power control systems.

[0026] In some embodiments of this application, the current operating parameters of the wind turbine include one or more combinations of wind speed, rotor speed, output voltage, and output current.

[0027] Specifically, wind speed directly reflects the external environmental load and is a key indicator for judging whether it is close to or exceeds the safe operating threshold; the rotor speed reflects the actual response of the wind turbine's mechanical system and can effectively identify overspeed risks, avoiding the lag or deviation that may be caused by relying solely on wind speed prediction; the output voltage and output current reflect the generator's load status and energy output level from the electrical side. When the power grid is abnormal or an internal fault causes the power to be unable to be transmitted normally, even if the wind speed does not reach the cut-out value, the rotational speed may rise abnormally. At this time, electrical parameters can serve as an important supplementary criterion.

[0028] By integrating the aforementioned multi-source parameters, the system can not only more accurately identify actual braking needs but also distinguish between different types of abnormal operating conditions (such as gusts of wind, grid disconnection, and sensor drift), thereby avoiding false braking or missed braking caused by misjudgment of a single parameter. This multi-parameter collaborative sensing mechanism significantly improves the robustness and adaptability of braking decisions, ensuring the structural safety of the wind turbine while reducing unnecessary shutdowns, improving power generation efficiency, and extending the service life of key components such as electromagnetic unloading devices, brake pads, and gearboxes.

[0029] In some embodiments of this application, the self-learning model constructs a regional meteorological prediction model based on historical meteorological data with time scales of hour, day, month, and year, for day-ahead and intraday multi-scale wind speed and direction prediction.

[0030] Traditional braking strategies typically rely solely on instantaneous or short-term wind speed thresholds for passive responses, making them ill-suited to rapidly changing wind conditions or persistent extreme weather. This application, however, incorporates historical meteorological data across multiple timescales (such as wind speed distribution in the same season over the past few years, typical diurnal variation patterns, and monthly climate trends) to train a self-learning model that comprehensively understands the long-term statistical characteristics and short-term dynamic features of regional wind resources. This model can simultaneously output high-precision wind speed and direction predictions for the next few minutes to several days. For example, on an "intra-day" scale, it can capture sudden strong winds caused by gusts, turbulence, or the passage of fronts; on a "day-ahead" scale, it can predict sustained strong winds caused by southward movement of cold air or the outer influence of typhoons. Based on this multi-scale forecast information, the braking control system can plan a graded response strategy in advance—for example, if it predicts that the wind speed will exceed the limit within 6 hours, it can enter a standby state in advance and notify the operation and maintenance personnel; if it predicts that the wind speed will briefly exceed the limit within 30 minutes but quickly drop, it can only use low-loss electromagnetic unloading for flexible adjustment to avoid frequent start-stop of mechanical braking.

[0031] This "prediction-decision-execution" closed loop not only significantly reduces the risk of equipment impact caused by braking lag, but also reduces unnecessary complete shutdowns, improving the availability and power generation revenue of wind turbine units. At the same time, multi-scale meteorological modeling gives the system good regional adaptability, allowing it to fine-tune the model using local historical data in different wind farms, achieving personalized and adaptive intelligent braking control.

[0032] In some embodiments of this application, under automatic braking control mode: When the operating parameters meet the first cut-out condition but the wind speed is predicted to drop back to a safe range in the short term, electromagnetic unloading is activated only to reduce the speed, and the brake is automatically released after the operating conditions are met. When the operating parameters meet the second cut-out condition and it is predicted that the wind will continue to be in extreme wind conditions, the electromagnetic unloading will be activated first, and then the mechanical braking device will be activated to completely stop the wind turbine and put it into protection mode.

[0033] Specifically, when the operating parameters of the wind turbine (such as wind speed or rotational speed) reach the preset first cut-off condition (i.e., the initial over-limit threshold), but the self-learning model predicts that the wind speed will drop back to the safe operating range within a short period of time (e.g., 10–30 minutes), the system only activates the electromagnetic unloading device. By short-circuiting the generator output or connecting an unloading resistor, excess kinetic energy is quickly consumed, thereby effectively suppressing the increase in rotor speed and preventing it from entering a dangerous state. During this process, the mechanical braking device remains inactive. Once the weather and operating parameters return to the normal range, the system automatically releases the electromagnetic unloading, and the wind turbine reconnects to the grid for power generation. This strategy avoids complete shutdown due to brief gusts of wind, reducing mechanical wear and power generation losses.

[0034] On the other hand, when the operating parameters meet the more stringent second cut-out conditions (such as wind speed continuously exceeding the rated cut-out wind speed and the trend intensifying), and the self-learning model further predicts that the wind turbine will be in extreme wind conditions for a long time (such as typhoons, strong cold air passage, etc.), the system will initiate a step-by-step joint braking process: first, the electromagnetic unloading device is triggered to perform preliminary deceleration and reduce the impact load on the transmission chain; after the speed drops below the safety threshold, the mechanical braking device is activated to completely lock the wind turbine, and the wind turbine enters a deep protection state.

[0035] This "electricity first, then mechanical" sequential control not only ensures absolute safety under extreme working conditions, but also effectively mitigates the high kinetic energy impact that mechanical braking is subjected to instantaneously, extending the service life of key components such as brake pads and brake discs.

[0036] In summary, this tiered braking mechanism, guided by prediction, based on risk level, and centered on equipment protection, achieves an intelligent leap from "one-size-fits-all shutdown" to "precise braking on demand," maximizing operational efficiency and system reliability while ensuring the structural safety of the wind turbine.

[0037] In some embodiments of this application, during the period when the wind turbine is in a protected state, the self-learning model continuously determines whether the conditions for releasing the mechanical brake are met. If not, the combined braking state of electromagnetic unloading and mechanical braking is maintained.

[0038] Specifically, even when the wind turbine has been completely shut down and the mechanical braking device has been activated, the control system continues to collect real-time meteorological data such as wind speed, wind direction, and air pressure. It then uses a self-learning model to make rolling predictions about subsequent wind conditions—for example, determining whether extreme winds have weakened, whether wind speed has stabilized below the cut-off threshold, and whether there is a risk of secondary strong winds. Simultaneously, the system also comprehensively considers factors such as the equipment's own condition (e.g., brake temperature, transmission system stress recovery) to construct multi-dimensional braking release criteria. Only when the model continuously confirms that the wind speed will remain within a safe operating range within a reasonable time window, and there is no risk of exceeding limits again, will the mechanical braking and electromagnetic unloading be released sequentially, allowing the wind turbine to restart and connect to the grid for power generation. If the prediction results indicate that the wind conditions are still unstable or there is a possibility of repeated exceedances, the system continues to maintain a combined braking state of electromagnetic unloading and mechanical braking to prevent premature brake release from causing unexpected rotor rotation, equipment damage, or safety accidents.

[0039] This closed-loop control mechanism of "continuous assessment - dynamic maintenance - condition triggering" effectively avoids the blindness and lag caused by relying on fixed delays or manual resets in traditional protection modes. It not only ensures long-term safety under extreme weather conditions, but also improves the wind turbine's autonomous recovery capability after wind conditions improve, significantly enhancing the intelligence level and operational resilience of the entire braking system.

[0040] In some embodiments of this application, in automatic braking control mode, if a fault is detected in the electromagnetic unloading device or the mechanical braking device, the system immediately switches to safety degradation mode, prioritizes the activation of the backup braking mechanism, and issues a fault alarm signal.

[0041] In automatic braking control mode, if the system detects a malfunction in the electromagnetic unloading device or mechanical braking device in real time through self-checking mechanisms or operational status monitoring (such as an open circuit in the electromagnetic unloading circuit, brake motor failure, excessive wear of brake pads, or abnormal pressure in the hydraulic / pneumatic system), it will immediately trigger a safety degradation strategy and actively switch to the preset safety degradation mode. The core objective of this mode is to maximize the structural safety and operational controllability of the wind turbine generator set in the event of partial failure of the main braking function.

[0042] Specifically, the system prioritizes the use of backup braking mechanisms. For example, if the electromagnetic unloading device malfunctions, the mechanical braking response logic is strengthened, and mechanical braking is activated directly when the cut-out conditions are met. If the mechanical braking device malfunctions, the electromagnetic unloading time is extended, and emergency feathering of the pitch system (if it has this function) is used in conjunction to achieve deceleration and shutdown. Simultaneously, the control system immediately generates and issues multi-level fault alarm signals, including local audible and visual alarms, remote monitoring platform push notifications, and SMS / APP notifications to maintenance personnel, detailing the fault type, location, and recommended handling measures. Furthermore, the system automatically records operating parameters and model decision logs before and after the fault occurs, providing data support for subsequent diagnosis and maintenance. Through this integrated safety mechanism of "fault perception—rapid degradation—redundant braking—real-time alarm," the risk of loss of braking capability due to the failure of a single braking component is effectively avoided, significantly improving the fault tolerance and inherent safety of the entire system.

[0043] In some embodiments of this application, when switching to manual braking control mode, the operating parameters, weather conditions and operation log at the time of switching are automatically recorded and uploaded to the remote monitoring platform.

[0044] When switching to manual braking control mode, the system automatically triggers the data solidification and log recording mechanism to accurately capture and save key information at the moment of switching, including the wind turbine's operating parameters (such as wind speed, rotor speed, output voltage, and output current), real-time meteorological conditions (such as wind direction, air pressure, ambient temperature, and short-term weather forecast results), and detailed operation logs (such as switching time, operator identification, triggering method, and current control mode status).

[0045] This data not only fully reconstructs the system context before and after manual intervention, providing highly reliable evidence for subsequent accident backtracking, responsibility determination, or operation and maintenance analysis, but also uploads it in real time to a remote monitoring platform or wind farm control center via a secure communication protocol. The remote platform can use this data for abnormal operation warnings, human-machine collaboration efficiency assessments, or supplementing model training samples, further optimizing the robustness of the self-learning algorithm.

[0046] More importantly, this mechanism enhances the auditability and transparency of operations, ensuring the priority of human intervention in emergency situations while avoiding equipment risks caused by misoperation or unnecessary manual intervention. It achieves an organic unity of "safety and controllability" and "intelligent traceability," significantly improving the standardization and intelligent management level of wind power system operation and maintenance.

[0047] In some embodiments of this application, the method further includes: In automatic braking control mode, when the self-learning model predicts that the wind speed will exceed the limit within a preset time window, the early warning mechanism will be activated in advance, and braking preparation prompts will be sent to the operation and maintenance personnel.

[0048] In automatic braking control mode, when the self-learning model determines, based on real-time operating data and multi-scale meteorological forecast results, that the wind speed will exceed the safe operating threshold of the wind turbine within a preset time window (e.g., 10 minutes, 30 minutes, or 2 hours), the system will activate the intelligent early warning mechanism in advance.

[0049] This mechanism not only displays warning information on the local human-machine interface, but also proactively sends braking preparation prompts to operation and maintenance personnel via SMS, mobile application push, SCADA system alarms or emails, including predicted wind speed curves, estimated over-limit time, suggested countermeasures and current equipment status.

[0050] This enables the operations and maintenance team to prepare for emergencies in advance, such as checking the status of the braking system, confirming that the communication link is unobstructed, arranging on-site duty, or adjusting the power grid dispatch plan, thereby building a proactive defense chain of "prediction-early warning-pre-control" before extreme wind conditions actually arrive.

[0051] Compared to the traditional passive response mode that only triggers braking when the wind speed exceeds the limit, this function significantly improves the risk prediction capability and operation and maintenance coordination efficiency of the wind power system, and effectively reduces equipment impact, unplanned shutdowns and even safety accidents caused by sudden strong winds.

[0052] In some embodiments of this application, the method further includes: After each braking operation is completed, a braking event report is automatically generated. The report includes the reason for braking triggering, the status of the actuator, the duration, environmental parameters, and the model prediction accuracy evaluation index.

[0053] After each braking operation (whether it's intelligent braking in automatic mode or emergency intervention in manual mode), the system automatically triggers the event archiving process to generate a structured braking event report. This report comprehensively records key information from the entire braking process, including the cause of braking (such as excessive wind speed, abnormal speed, manual command, or equipment failure), the status of the actuators (whether the electromagnetic unloading device is properly engaged, whether the mechanical brake is successfully locked, the response time and action feedback of each actuator), the duration of braking (the time from start to complete stop or release of the brake), environmental parameters (meteorological data such as wind speed, wind direction, temperature, and air pressure at the time of braking and in the preceding and following periods), and, most importantly, the model prediction accuracy evaluation indicators (such as the deviation between predicted and measured wind speed, the error between predicted over-limit time and actual occurrence time, and the matching degree between the braking strategy and actual needs).

[0054] This report not only provides maintenance personnel with clear and traceable operational review data, facilitating quick identification of potential issues such as incorrect or missed braking, or equipment malfunctions, but also serves as crucial feedback data for the iterative optimization of the self-learning model. By continuously accumulating comparison samples of real braking events and predicted results, the system can continuously correct model parameters, improving future prediction accuracy and decision reliability. Furthermore, the report can be automatically stored in a local database and simultaneously uploaded to a remote monitoring platform, supporting multi-dimensional retrieval and statistical analysis by time, turbine number, event type, and other dimensions. This provides data-driven support for intelligent operation and maintenance, reliability assessment, and preventative maintenance of wind farms, significantly enhancing the full lifecycle management capabilities of wind power systems.

[0055] In some embodiments of this application, the method further includes: Regularly validate and update the self-learning model offline, retrain the model using newly added historical operational data and meteorological data, and decide whether to deploy the new version after comparing the performance of the old and new models through A / B testing.

[0056] Specifically, the system periodically (e.g., weekly, monthly, or after accumulating sufficient new data) collects newly added historical operational data (including turbine speed, voltage, current, braking event records, etc.) and corresponding meteorological data (e.g., real wind speed and direction sequences recorded by authoritative weather stations or anemometer towers) to construct updated training and validation datasets. Based on this, the existing self-learning model is retrained to generate candidate new models. To scientifically evaluate the performance of the new models, the system employs an A / B testing mechanism: the old and new models are run in parallel under the same historical scenarios, comparing their performance on key indicators such as wind speed prediction accuracy, early warning time for exceeding limits, braking decision accuracy, and false alarm / missed alarm rate. Only when the new model significantly outperforms the current deployment version in multiple core indicators and passes security and stability reviews is it officially deployed to the online control system to replace the old model.

[0057] This closed-loop model management mechanism not only effectively prevents models from "degrading" due to data drift or concept drift, but also ensures the controllability and reliability of algorithm upgrades, avoiding operational risks caused by directly deploying unverified models. Through continuous data-driven optimization, the system can continuously improve the intelligence level of braking decisions, enabling wind turbine generators to maintain high safety, high availability, and high power generation efficiency in complex and ever-changing natural environments.

[0058] For any parts not mentioned in this application, existing technologies may be used or referenced.

[0059] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0060] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. An intelligent control method for braking a wind turbine, comprising an automatic braking control mode and a manual braking control mode, characterized in that, In the automatic braking control mode, the current operating parameters of the wind turbine are collected and intelligent judgment is made in combination with the self-learning model, and the electromagnetic unloading device and the mechanical braking device are linked to perform braking operation. In the manual braking control mode, the judgment and execution of the automatic braking control mode are suspended, and the operator directly controls the mechanical braking device. The manual braking control mode has the highest priority. The self-learning model is trained and iterated based on historical wind turbine operating parameters, historical meteorological data, and current meteorological data to predict wind speed change trends and optimize braking decisions.

2. The intelligent control method for wind turbine braking according to claim 1, characterized in that, The current operating parameters of the wind turbine include one or more combinations of wind speed, rotor speed, output voltage, and output current.

3. The intelligent control method for wind turbine braking according to claim 1, characterized in that, The self-learning model is based on historical meteorological data to construct a regional meteorological prediction model with time scales of hour, day, month, and year, which is used for day-ahead and intraday multi-scale wind speed and direction prediction.

4. The intelligent control method for wind turbine braking according to claim 1, characterized in that, In automatic braking control mode: When the operating parameters meet the first cut-out condition but the wind speed is predicted to drop back to a safe range in the short term, electromagnetic unloading is activated only to reduce the speed, and the brake is automatically released after the operating conditions are met. When the operating parameters meet the second cut-out condition and it is predicted that the wind will continue to be in extreme wind conditions, the electromagnetic unloading will be activated first, and then the mechanical braking device will be activated to completely stop the wind turbine and put it into protection mode.

5. The intelligent control method for wind turbine braking according to claim 4, characterized in that, During the protection state of the wind turbine, the self-learning model continuously determines whether the conditions for releasing the mechanical brake are met. If not, the combined braking state of electromagnetic unloading and mechanical braking is maintained.

6. The intelligent control method for wind turbine braking according to claim 4, characterized in that, In automatic braking control mode, if a malfunction is detected in the electromagnetic unloading device or the mechanical braking device, the system will immediately switch to safety degradation mode, prioritize the activation of the backup braking mechanism, and issue a fault alarm signal.

7. The intelligent control method for braking of a wind turbine generator according to any one of claims 1 to 6, characterized in that, When switching to manual braking control mode, the system automatically records the operating parameters, weather conditions, and operation log at the time of the switch and uploads them to the remote monitoring platform.

8. The intelligent control method for wind turbine braking according to claim 1, characterized in that, The method also includes: In automatic braking control mode, when the self-learning model predicts that the wind speed will exceed the limit within a preset time window, the early warning mechanism will be activated in advance, and braking preparation prompts will be sent to the operation and maintenance personnel.

9. The intelligent control method for wind turbine braking according to claim 1, characterized in that, The method also includes: After each braking operation is completed, a braking event report is automatically generated. The report includes the reason for braking triggering, the status of the actuator, the duration, environmental parameters, and the model prediction accuracy evaluation index.

10. The intelligent control method for wind turbine braking according to claim 1, characterized in that, The method also includes: Regularly validate and update the self-learning model offline, retrain the model using newly added historical operational data and meteorological data, and decide whether to deploy the new version after comparing the performance of the old and new models through A / B testing.

Images