Active anti-typhoon offshore wind power system based on fusion prediction and control method

Abstract

The present application relates to the technical field of offshore wind turbine, and particularly relates to an active anti-typhoon offshore wind power system based on fusion prediction and a control method thereof. The active anti-typhoon offshore wind power system based on fusion prediction comprises: a perception layer, which is used for collecting and preprocessing environmental meteorological data and wind turbine operation state data, and outputting a structured perception multi-source data set; a decision layer, which is used for receiving the structured perception multi-source data set, generating future wind field prediction and structural load risk early warning based on a WRF meteorological model and an LSTM load prediction model, and outputting a cooperative control instruction; meanwhile, the decision layer re-receives the preprocessed multi-source heterogeneous data, updates the WRF meteorological model and the LSTM load prediction model parameters, and optimizes the cooperative control instruction; a response layer, which is used for receiving the cooperative control instruction, executing active aerodynamic control and structural dynamic regulation, and feeding back the wind turbine state; and a communication network, which is used for realizing data interaction and instruction transmission among the perception layer, the decision layer and the response layer.

Description

Technical Field

[0001] This invention relates to the field of offshore wind turbine technology, specifically to an active typhoon-resistant offshore wind power system and control method based on fusion prediction. Background Technology

[0002] Offshore wind turbines deployed in typhoon-prone areas, extreme weather zones, or complex marine environments face significantly different operational challenges compared to conventional sea areas. These harsh environments drastically raise the standards for disaster resistance and continuous operational stability. When encountering extreme weather events such as typhoons, the turbines often face a sudden surge in load, posing potential risks to their structural safety and operational control. This is because typhoons are often accompanied by extremely high wind speeds, rapidly changing wind directions, and dramatic fluctuations in air pressure, which can easily damage the structural integrity and operational stability of the wind turbine, leading to mechanical failures or even turbine shutdowns, ultimately affecting its power generation efficiency and service life.

[0003] Currently, traditional wind turbine control methods are mainly based on fixed-parameter control strategies, relying on preset operating conditions and static models. These methods are ill-suited to rapidly changing wind conditions and complex meteorological environments. When typhoons arrive, they can only passively shut down, unable to actively resist the storm, leading to economic losses. Furthermore, these methods offer limited resilience and operational reliability for wind turbines in extreme weather conditions, increasing maintenance costs and operational risks. Summary of the Invention

[0004] The purpose of this invention is to provide an active typhoon-resistant offshore wind power system based on fusion prediction, so as to solve the problem that the control methods of traditional wind turbines in the prior art have weak disaster resistance and operational reliability under extreme weather conditions.

[0005] To address the aforementioned problems, this invention proposes an active typhoon-resistant offshore wind power system based on fusion prediction. The technical solution adopted is as follows: An active typhoon-resistant offshore wind power system based on fusion prediction includes: The sensing layer is used for real-time acquisition and preprocessing of environmental meteorological data and wind turbine operating status data, and outputs... Structured, perceptual multi-source datasets; The decision layer receives structured sensing multi-source datasets, generates future wind field forecasts and structural load risk warnings based on the WRF meteorological model and LSTM load prediction model, and outputs collaborative control commands and triggers graded response strategies. At the same time, it periodically receives structured sensing multi-source datasets, updates the parameters of the WRF meteorological model and LSTM load prediction model, and optimizes collaborative control commands. The response layer includes a plasma flow control unit integrated into the pressure surface of the trailing edge of the wind turbine blade and an active mass damping system located inside the wind turbine blade. It is used to receive cooperative control commands, perform active aerodynamic control and structural dynamic adjustment, and provide feedback on the wind turbine status. The communication system is used to enable data interaction and command transmission between the perception layer, decision-making layer, and response layer.

[0006] Furthermore, the perception layer includes: The meteorological satellite and external data interface module is used to receive meteorological satellite cloud images, GFS numerical forecast fields, and meteorological radar reflectivity data in real time to obtain macro-meteorological information; A fixed sensor module is used to collect environmental parameters at the height of the wind turbine hub in real time; The weather forecasting system is used to perform real-time analysis and preliminary diagnosis of macro meteorological information and environmental parameters, identify data anomalies and trigger quality markers to obtain abnormal data. The full-time operation data acquisition unit is used to synchronously collect real-time status data of the wind turbine body, including the structural dynamic response obtained by vibration acceleration sensors installed on the blades and tower, as well as the operating parameters of the pitch system, generator, and yaw system. The data acquisition and preprocessing module is used to preprocess macro-meteorological information, environmental parameters, abnormal data and real-time status data to obtain structured sensing multi-source datasets; The preprocessing includes: The system performs time synchronization, format standardization, invalid value removal, and missing data imputation on macro meteorological information, environmental parameters, abnormal data, and real-time status data. It also uses data fusion algorithms to calibrate and match point-based measured data and area-based remote sensing data within a unified spatiotemporal framework, outputting a structured sensing multi-source dataset.

[0007] Furthermore, the decision-making layer includes: The typhoon prediction unit is used to receive structured sensing multi-source datasets and generate future wind field forecasts and structural load risk warnings through WRF meteorological models and LSTM load prediction models. Collaborative Decision-Making Unit: Based on future wind field forecasts and structural load risk warnings, it outputs collaborative control commands and triggers hierarchical response strategies; at the same time, it periodically receives structured sensing multi-source datasets, updates the parameters of the WRF meteorological model and the LSTM load prediction model, and optimizes collaborative control commands. The WRF meteorological model adopts a triple nested grid and integrates pre-processed satellite cloud image data, meteorological radar reflectivity data and environmental parameters from structured sensing multi-source data through data assimilation technology to generate a high spatiotemporal resolution forecast wind field for the next 72 hours. The LSTM load prediction model is a three-layer neural network structure. It takes as input the forecast wind field data, environmental parameters and real-time status data of the wind turbine body, predicts the dynamic load spectrum and main excitation frequency of the blade root in the future period, and calculates the probability of resonance risk.

[0008] Furthermore, the collaborative decision-making unit includes a hierarchical response unit and a model predictive control framework. The graded response unit outputs coordinated control commands and triggers graded response strategies based on future wind field forecasts and structural load risk warnings. The model predictive control framework performs rolling optimization updates at fixed cycles: in each control cycle, it re-receives structured sensing multi-source data, updates the input parameters of the WRF meteorological model and the LSTM predictive model accordingly, and re-solves the optimization problem to generate a new round of optimal control commands, thereby achieving closed-loop adaptive control.

[0009] Furthermore, the hierarchical response strategy includes: When the predicted wind speed reaches the Level I threshold, it enters the Level I response state and generates a command to start the plasma flow control unit. When the predicted wind speed reaches the Level II threshold and the resonance risk probability exceeds the preset value, it is in the Level II response state. The command is generated to activate the plasma flow control unit and the active mass damping system at the same time, and to calculate the optimal target position of the counterweight. When the predicted wind speed reaches the Level III threshold or other uncontrollable risks are detected, the response status is Level III, an emergency shutdown command is generated, and the counterweight is locked.

[0010] Furthermore, the response layer also includes: Activate the emergency handling module, which is activated during a Level 3 response. The wind turbine main control coordination unit is used to perform yaw attitude adjustment, emergency shutdown, and counterweight locking; The emergency resource dispatch unit is used to monitor the status of each actuator in real time; at the same time, it coordinates mobile emergency power supplies and links with external rescue forces.

[0011] Furthermore, the response strategy of the response layer is as follows: In the first-level response state, the plasma flow control unit is activated, and by applying optimized high-voltage pulse excitation to the trailing edge electrode array of the blade, flow separation is actively suppressed to maintain aerodynamic stability. When the response is upgraded to the second level, the active mass damping system is activated simultaneously. Based on the optimal control solution, the target position of the counterweight of the active mass damping system is quickly adjusted to change the natural frequency of the blades to avoid resonance. At the same time, the emergency resource scheduling unit monitors the status of each actuator in real time. When the response level is upgraded to Level 3, the emergency handling module is activated, and the wind turbine main control coordination unit immediately performs an emergency shutdown and locks all moving parts; at the same time, the emergency resource dispatch unit coordinates mobile emergency power supplies and links with external rescue forces.

[0012] Furthermore, the plasma flow control unit includes an array of needle-ring electrodes and a high-voltage pulse power supply, with an electrode spacing of 15 mm, covering 15% of the chord length of the blade trailing edge; the electrode array is connected to a high-voltage pulse power supply module, the high-voltage pulse power supply module has a response time ≤5ms, and adjustable parameters include voltage amplitude of 5~15kV, pulse frequency of 1~10kHz, and duty cycle of 10~50%; the active mass damping system includes a carbon fiber guide rail arranged along the blade spanwise, a movable tungsten alloy counterweight slidably mounted on the carbon fiber guide rail, a linear motor mechanism for driving the tungsten alloy counterweight to move, and a high-frequency sensor array for monitoring position and vibration.

[0013] Furthermore, the communication system includes: The space-based satellite communication subnetwork connects with edge communication nodes and communicates with the WRF weather forecasting system and underwater equipment through the BeiDou satellite communication system. It provides command and control and status data feedback channels between the decision-making level and sea-based unmanned vessels, monitoring equipment in remote sea areas, and some mobile platforms. At the same time, it is responsible for the uplink transmission of meteorological data from the WRF weather forecasting system. The sea-based wireless communication subnetwork communicates with edge communication nodes and adopts a hybrid mode of "surface wireless + underwater acoustic". Surface devices communicate through wireless links, while underwater devices rely on acoustic channels to interact with surface relay nodes or sea-based unmanned vessels. The land-based wired communication subnetwork communicates with the edge communication nodes and connects with the fixed sensors of the equipment via single-mode optical fiber communication. It is used to collect the sensor data of the land-based fixed equipment and upload it to the edge communication nodes. Edge communication nodes, which have bidirectional communication connections with the decision-making layer, are deployed at hub locations corresponding to offshore platforms or key islands. They integrate multiple network interfaces and are responsible for protocol conversion, fusion, and intelligent routing of heterogeneous data from different sub-networks, and uploading it to the decision-making layer. At the same time, they receive decision data or control instructions from the decision-making layer and distribute them to various communication sub-networks.

[0014] This invention also provides a control method for an active typhoon-resistant offshore wind power system based on fusion prediction, comprising the following steps: Step 1: The sensing layer collects and preprocesses environmental meteorological data and wind turbine operating status data in real time, and outputs a structured sensing multi-source dataset; Step 2: The decision-making layer receives the structured sensing multi-source dataset, generates future wind field forecasts and structural load risk warnings based on the WRF meteorological model and LSTM load prediction model, and outputs collaborative control commands and triggers graded response strategies; at the same time, it receives the structured sensing multi-source dataset at regular intervals, updates the parameters of the WRF meteorological model and LSTM load prediction model, and optimizes the collaborative control commands. Step 3: The response layer receives the cooperative control command and executes active aerodynamic control and structural dynamic adjustment.

[0015] Beneficial Effects: This invention is an improved version. By integrating the WRF meteorological model and the LSTM load prediction algorithm, it achieves early perception of typhoon path, intensity, and wind turbine load. Before extreme wind conditions arrive, it executes active aerodynamic control and structural dynamic adjustment, constructing an integrated intelligent control architecture of "perception-prediction-decision-execution." This architecture integrates macro-level meteorological forecasts and micro-level load perception, and actively adjusts the aerodynamic and structural dynamic characteristics of the wind turbine. While ensuring the safety of the wind turbine power generation system, it extends the wind turbine operating window during typhoons, improving power generation efficiency and equipment reliability. Simultaneously, it monitors extreme weather conditions in real time, reducing unit load, avoiding mechanical damage, and improving disaster resistance and operational stability.

[0016] (1) A cross-scale fusion prediction and early warning system of "meteorology-load" was constructed: by coupling physical driving and data driving models, the entire chain of advanced perception and prediction from the 100-kilometer-level typhoon path to the microenvironment at the wind turbine location level and then to the internal load of the structure was realized, and the risk response mode was transformed from passive response to active early warning and intervention, which significantly extended the emergency preparation time. (2) An innovative active typhoon resistance technology path combining aerodynamics and structure was developed: Plasma flow control technology and adjustable active mass damping technology were integrated and applied to wind turbine blades. Through the coordinated scheduling of the central intelligent decision-making unit, progressive defenses from aerodynamic stall suppression to structural resonance avoidance were dynamically implemented for wind speed and turbulence characteristics at different stages of typhoon development, effectively expanding the safe operation boundary of wind turbines under extreme wind conditions. (3) Adaptive closed-loop optimization control is realized: The closed-loop control architecture of "perception-decision-execution-verification" is adopted, and combined with model predictive control and online optimization algorithm, so that the system can dynamically adjust the control strategy and parameters according to the actual evolution of the typhoon situation and the real-time response status of the wind turbine, thereby improving the robustness and control accuracy of the system in dealing with complex and uncertain environments. (4) Comprehensive improvement in economy and safety: Under the core premise of ensuring the structural safety of the wind turbine, the unnecessary downtime during typhoons was reduced through active control technology, which improved the annual power generation and asset utilization of the wind farm. At the same time, actively suppressing extreme loads and vibrations helps to reduce fatigue damage to key components, extend the service life of the wind turbine, and achieve a dual improvement in equipment safety and reliability and system economic benefits. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the active typhoon-resistant offshore wind power system based on fusion prediction of the present invention; Figure 2 This is a schematic diagram of the sensing layer structure in the active typhoon-resistant offshore wind power system based on fusion prediction of the present invention; Figure 3 This is a schematic diagram of the decision-making layer in the active typhoon-resistant offshore wind power system based on fusion prediction of the present invention; Figure 4 This is a schematic diagram of the response layer structure in the active typhoon-resistant offshore wind power system based on fusion prediction of the present invention; Figure 5 This is a schematic diagram of the communication system in the active typhoon-resistant offshore wind power system based on fusion prediction of the present invention. Detailed Implementation

[0018] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0019] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0020] The active typhoon-resistant offshore wind power system and control method based on fusion prediction, according to embodiments of this application, are described below with reference to the accompanying drawings.

[0021] The following is combined Figure 1This application provides a detailed description of the active typhoon-resistant offshore wind power system based on fusion prediction.

[0022] This application discloses an active typhoon-resistant offshore wind power system based on fusion prediction, comprising: a perception layer for real-time acquisition and preprocessing of environmental meteorological data and wind turbine operating status data, outputting a structured perception multi-source dataset; a decision layer for receiving the structured perception multi-source dataset, generating future wind field forecasts and structural load risk warnings based on the WRF meteorological model and the LSTM load prediction model, and outputting collaborative control commands and triggering graded response strategies; simultaneously, periodically receiving the structured perception multi-source dataset, updating the parameters of the WRF meteorological model and the LSTM load prediction model, and optimizing the collaborative control commands; a response layer including a plasma flow control unit integrated into the pressure surface of the wind turbine blade trailing edge and an active mass damping system installed inside the wind turbine blade, for receiving collaborative control commands, executing active aerodynamic control and structural dynamic adjustment, and providing feedback on the wind turbine status; and a communication system for realizing data interaction and command transmission between the perception layer, decision layer, and response layer.

[0023] Specifically, the perception layer includes: a meteorological satellite and external data interface module for receiving meteorological satellite cloud images, GFS numerical forecast fields, and meteorological radar reflectivity data to acquire macro-meteorological information; fixed sensors for continuously collecting environmental parameters at the wind turbine hub height; a detection weather forecasting system for real-time analysis and preliminary diagnosis of macro-meteorological information and environmental parameters, identifying data anomalies and triggering quality markers to obtain abnormal data; a full-time operation data acquisition unit for synchronously collecting real-time status data of the wind turbine body, including real-time status data obtained through vibration acceleration sensors installed on the blades and tower, structural dynamic responses, and operating parameters of the pitch system, generator, and yaw system; and a data acquisition and preprocessing module for preprocessing macro-meteorological information, environmental parameters, abnormal data, and real-time status data to obtain a structured perception multi-source dataset.

[0024] The preprocessing includes: performing time synchronization, format standardization, invalid value removal and missing data imputation on macro meteorological information, environmental parameters, abnormal data and real-time status data; and calibrating and matching point-based measured data and area-based remote sensing data in a unified spatiotemporal framework through data fusion algorithms to output a structured sensing multi-source dataset.

[0025] Here, meteorological satellite cloud images are visible and infrared cloud images from meteorological satellites (such as Fengyun-4 and Himawari-8); macro-meteorological information includes typhoon paths, intensity, and cloud structure, which are large-scale and highly timely. The fixed sensor array consists of ultrasonic anemometers, temperature and humidity sensors, atmospheric pressure sensors, etc., deployed in and around the wind farm, continuously collecting (frequency ≥10Hz) fine-grained environmental parameters such as wind speed, wind direction, temperature, and air pressure at the height of the wind turbine hub.

[0026] In one specific embodiment, the weather forecasting system acts as the initial analysis node, performing real-time analysis and preliminary diagnosis of the acquired raw satellite and sensor data, identifying data anomalies and triggering quality marking. All multi-source, heterogeneous, and massive data streams from external interfaces, fixed sensors, and wind turbine systems ultimately converge into the data processing and preprocessing module. This module performs strict time synchronization, format standardization, invalid value removal, and missing data imputation. It also uses data fusion algorithms to calibrate and match point-based measured data and area-based remote sensing data within a unified spatiotemporal framework, ultimately outputting a high-quality, spatiotemporally consistent, structured sensing dataset that can directly drive prediction models.

[0027] Specifically, the decision-making layer includes: receiving structured sensing multi-source datasets and generating future wind field forecasts and structural load risk warnings through WRF meteorological models and LSTM load prediction models; Collaborative Decision-Making Unit: Based on future wind field forecasts and structural load risk warnings, it outputs collaborative control commands and triggers hierarchical response strategies; at the same time, it periodically receives structured sensing multi-source datasets, updates the parameters of the WRF meteorological model and the LSTM load prediction model, and optimizes the collaborative control commands.

[0028] In one specific embodiment, the WRF meteorological model employs a triple-nested grid, fusing input satellite cloud imagery data, meteorological radar reflectivity data, and real-time observation data uploaded by field sensors through data assimilation technology to generate a high spatiotemporal resolution forecast wind field for the next 72 hours. Specifically, in the triple-nested grid of the WRF meteorological model, the outermost grid has a resolution of 27km and is used to capture the large-scale circulation of typhoons; the middle grid has a resolution of 9km and is used to simulate the macroscopic structure of typhoons; and the innermost grid has a resolution of 3km, covering the wind farm area, and is used to output a refined wind field forecast at the wind turbine location scale.

[0029] The LSTM load prediction model is a three-layer neural network structure. It takes as input refined wind field data, real-time wind speed, turbulence intensity, and wind turbine operating status obtained from the aforementioned WRF meteorological model, predicts the dynamic load spectrum and main excitation frequency at the blade root for future periods, calculates the resonance risk probability, and outputs the load confidence interval. Here, data assimilation technology is used to fuse input satellite cloud imagery data, meteorological radar reflectivity data, and real-time observation data uploaded by on-site sensors to dynamically sense and predict the typhoon's path.

[0030] In another specific embodiment, the collaborative decision-making unit includes a hierarchical response unit and a model predictive control framework. The hierarchical response unit outputs collaborative control commands and triggers a hierarchical response strategy based on future wind field forecasts and structural load risk warnings. The model predictive control framework performs rolling optimization updates at fixed intervals: within each control cycle, it re-receives structured sensing multi-source data, updates the input parameters of the WRF meteorological model and the LSTM prediction model accordingly, and re-solves the optimization problem to generate a new round of optimal control commands, thereby achieving closed-loop adaptive control. It should be noted that the entire decision-making process is embedded within a model predictive control framework. The fixed interval is 30 seconds.

[0031] Specifically, the tiered response strategy includes: When the predicted wind speed reaches the Level I threshold, it enters the Level I response state and generates a command to initiate plasma flow control; here, the Level I threshold is a predicted wind speed ≥ 22 m / s. When the predicted wind speed reaches the Level II threshold and the resonance risk probability exceeds the preset value, the response state is Level II. A command is generated to activate the plasma flow control and active mass damping system simultaneously. Here, the Level II threshold is a predicted wind speed ≥ 28 m / s; the preset value for the resonance risk probability is 0.7. When the predicted wind speed reaches the Level III threshold or other uncontrollable risks are detected, the response state is Level III, an emergency shutdown command is generated and the counterweight is locked; here, the Level III threshold is a predicted wind speed ≥ 35 m / s; In one specific embodiment, a plasma flow control unit, integrated into the trailing edge pressure surface of the wind turbine blade, includes an array of needle-ring electrodes and a high-voltage pulse power supply. The electrode spacing is 15 mm, covering 15% of the chord length of the trailing edge of the blade. The electrode array is connected to a high-voltage pulse power supply module with a response time ≤ 5 ms and adjustable parameters including voltage amplitude 5~15 kV, pulse frequency 1~10 kHz, and duty cycle 10~50%. An active mass damping system, located inside the blade, includes a carbon fiber guide rail arranged along the blade spanwise, a movable tungsten alloy counterweight slidably mounted on the carbon fiber guide rail, a linear motor mechanism driving the movement of the tungsten alloy counterweight, and a high-frequency sensor array for monitoring position and vibration. The mass of a single counterweight accounts for 0.8~1.2% of the total blade mass. The linear motor mechanism is a linear dual-motor drive mechanism with a maximum acceleration of 2.5 g. The counterweight position is fed back via a laser scale and a photoelectric encoder. The system also includes a vibration sensor array distributed at 20%, 50%, and 80% of the blade span, with a sampling frequency ≥ 200 Hz.

[0032] In another specific embodiment, the response layer further includes: an emergency processing module for activation during a Level 3 response; a wind turbine main control coordination unit for performing yaw attitude adjustment, emergency shutdown, and counterweight locking; an emergency resource scheduling unit for real-time monitoring of the status of each actuator; and simultaneously, coordinating mobile emergency power supplies and linking external rescue forces.

[0033] In another specific embodiment, the active typhoon-resistant offshore wind power system based on fusion prediction further includes: The decision-making layer also includes a central control unit, which is configured to execute the aforementioned hierarchical early warning and coordinated control strategies. When calculating the optimal target position of the counterweight, the central control unit uses an optimization algorithm with the objective function of "maximizing the distance between the first-order natural frequency of the blade after adjustment and the predicted excitation frequency" and the constraint of the drive mechanism capability.

[0034] Specifically, the response strategy of the response layer is as follows: In the first-level response state, the plasma flow control unit is activated, and by applying optimized high-voltage pulse excitation to the trailing edge electrode array of the blade, flow separation is actively suppressed to maintain aerodynamic stability. When the response is upgraded to the second level, the active mass damping system is activated simultaneously. Based on the optimal control solution, the target position of the counterweight of the active mass damping system is quickly adjusted to change the natural frequency of the blades to avoid resonance. At the same time, the emergency resource scheduling unit monitors the status of each actuator in real time. When the response level is upgraded to Level 3, the emergency handling module is activated, and the wind turbine main control coordination unit immediately performs an emergency shutdown and locks all moving parts; at the same time, the emergency resource dispatch unit coordinates mobile emergency power supplies and links with external rescue forces.

[0035] In one specific embodiment, during the first-level response state: Activate the plasma flow control unit to enter working status; Data on the blades is collected using sensors to obtain local flow conditions. Dynamically determine the optimal combination of excitation parameters; The local controller drives the high-voltage pulse power supply subsystem; Monitor the effectiveness of flow control and form a closed-loop verification.

[0036] If you need to stop or encounter an abnormality, enter safe mode or stop state.

[0037] When upgraded to the second-level response state: the active mass damping system is activated synchronously, and the target position of the counterweight of the active mass damping system is quickly adjusted according to the solved optimal control solution to change the natural frequency of the blades to avoid resonance; at the same time, the emergency resource scheduling unit monitors the status of each actuator in real time. Excitation frequency perception and prediction, identifying external excitation characteristics; Online identification of the structure's natural frequencies enables real-time acquisition of the structure's dynamic characteristics. Resonance risk assessment to evaluate safety under current operating conditions; The optimal target position of the counterweight in an active mass damping system is determined by solving the target position optimization problem and identifying the best moving target of the counterweight. Motion trajectory planning and collaborative driving, followed by effect verification and adaptive fine-tuning, and adjustment of control strategy based on feedback; If the process is completed or an error occurs, the system will enter safe mode and reset state.

[0038] When the response level is upgraded to Level 3, the regional collaborative protection mechanism is fully activated: the emergency handling module is activated, the wind turbine main control coordination unit immediately executes an emergency shutdown and locks all moving parts, and the emergency resource dispatch unit coordinates and mobilizes wide-area resources such as mobile emergency power supplies for support in accordance with the plan, and links with external rescue forces, thereby forming an integrated closed-loop response system from local active control to overall safety protection and cross-regional emergency support.

[0039] In one specific embodiment, such as Figure 5As shown, the communication system includes: a space-based satellite communication subnetwork, which communicates with edge communication nodes and connects with the WRF weather forecasting system and underwater equipment via the BeiDou satellite communication system. Primarily relying on the BeiDou satellite communication system, it provides a command and control and status data feedback channel between the sea-based unmanned surface vessels (USVs), monitoring equipment in remote sea areas, and some mobile platforms and the decision-making level; it also handles the uplink transmission of meteorological data for the WRF weather forecasting system. A sea-based wireless communication subnetwork, which communicates with edge communication nodes, employs a hybrid mode of "surface wireless + underwater acoustic," where surface equipment communicates via wireless links, while underwater equipment relies on acoustic channels to interact with surface relay nodes or sea-based USVs. A land-based wired communication subnetwork, which communicates with edge communication nodes... The system connects to fixed sensors on land-based equipment via single-mode fiber optic communication to collect sensor data and upload it to edge communication nodes. This provides a high-bandwidth, low-latency, reliable data link for WRF weather forecasting systems and fixed sensors deployed on land or in fixed facilities. Edge communication nodes, bidirectionally connected to the decision-making layer, are deployed at key locations on offshore platforms or critical islands. They integrate multiple network interfaces and are responsible for protocol conversion, fusion, and intelligent routing of heterogeneous data from different sub-networks, uploading it to the decision-making layer. Simultaneously, they receive and distribute decision data or control commands from the decision-making layer to various communication sub-networks. Furthermore, the system performs local preprocessing on the collected raw data to reduce the load on the core network, thereby ensuring the real-time performance and reliability of the transmission of sensing data and control commands in complex marine environments. Here, local preprocessing includes anomaly filtering and image compression.

[0040] This application also provides a control method for an active typhoon-resistant offshore wind power system based on fusion prediction, including the following steps: Step 1: The sensing layer collects and preprocesses environmental meteorological data and wind turbine operating status data in real time, and outputs a structured sensing multi-source dataset; Step 2: The decision-making layer receives the structured sensing multi-source dataset, generates future wind field forecasts and structural load risk warnings based on the WRF meteorological model and LSTM load prediction model, and outputs collaborative control commands and triggers graded response strategies; at the same time, it receives the structured sensing multi-source dataset at regular intervals, updates the parameters of the WRF meteorological model and LSTM load prediction model, and optimizes the collaborative control commands. Step 3: The response layer receives the cooperative control command and executes active aerodynamic control and structural dynamic adjustment.

[0041] In one specific implementation, the control method for an active typhoon-resistant offshore wind power system based on fusion prediction includes the following steps: Step 1: Integrate perception and prediction, assimilate satellite, radar and field sensor data, and drive the typhoon prediction unit to continuously update the refined wind field forecast for the next 72 hours; simultaneously input the predicted wind field and real-time wind turbine data into the load prediction unit to obtain the short-term load risk prediction; with a time resolution of 10 minutes, the LSTM load prediction model calculates the blade root bending moment probability density function. Step 2, Risk Assessment and Decision-Making: The central control unit determines the current risk level based on the predicted wind speed and resonance risk assessment results, and generates the corresponding collaborative control instruction set; when the resonance risk probability is ≥0.7, for cases where the active mass damping unit needs to be activated, the optimization algorithm is called to solve the target position of the counterweight block online; Step 3, Collaborative Execution and Closed-Loop Update: Control commands are synchronously sent to each response unit for execution via industrial Ethernet, and the rolling time-domain parameters of model predictive control are updated every 30 seconds. The system adopts a model predictive control framework to update predictive information and optimize control parameters at fixed intervals.

[0042] In step 1, the ensemble Kalman filter algorithm is used to assimilate the measured data within a 200km radius of the wind farm, and the initial field of the WRF model is corrected in real time to generate an ensemble forecast containing multiple disturbance members, generating 50 sets of disturbance ensemble forecast members; the typhoon path prediction error is controlled within 30km / 24h to reduce prediction uncertainty.

[0043] In step 2, the online identification of the first-order natural frequency of the blade adopts a random subspace identification method based on the vibration response signal; the optimization algorithm adopts a genetic algorithm to optimize the position of the counterweight and perform multi-objective optimization.

[0044] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the description and drawings of the present invention shall also be included within the scope of protection of the present invention.

Claims

1. A proactive typhoon-resistant offshore wind power system based on fusion prediction, characterized in that, include: The sensing layer is used for real-time acquisition and preprocessing of environmental meteorological data and wind turbine operating status data, and outputs... Structured, perceptual multi-source datasets; The decision layer receives structured sensing multi-source datasets, generates future wind field forecasts and structural load risk warnings based on the WRF meteorological model and LSTM load prediction model, and outputs collaborative control commands and triggers graded response strategies. At the same time, it periodically receives structured sensing multi-source datasets, updates the parameters of the WRF meteorological model and LSTM load prediction model, and optimizes collaborative control commands. The response layer includes a plasma flow control unit integrated into the pressure surface of the trailing edge of the wind turbine blade and an active mass damping system located inside the wind turbine blade. It is used to receive cooperative control commands, perform active aerodynamic control and structural dynamic adjustment, and provide feedback on the wind turbine status. The communication system is used to enable data interaction and command transmission between the perception layer, decision-making layer, and response layer.

2. The active typhoon-resistant offshore wind power system based on fusion prediction as described in claim 1, characterized in that, The sensing layer includes: The meteorological satellite and external data interface module is used to receive meteorological satellite cloud images, GFS numerical forecast fields, and meteorological radar reflectivity data in real time to obtain macro-meteorological information; A fixed sensor module is used to collect environmental parameters at the height of the wind turbine hub in real time; The weather forecasting system is used to perform real-time analysis and preliminary diagnosis of macro meteorological information and environmental parameters, identify data anomalies and trigger quality markers to obtain abnormal data. The full-time operation data acquisition unit is used to synchronously collect real-time status data of the wind turbine body. The real-time data includes the structural dynamic response obtained by vibration acceleration sensors installed on the blades and tower, as well as the operating parameters of the pitch system, generator, and yaw system. The data acquisition and preprocessing module is used to preprocess macro-meteorological information, environmental parameters, abnormal data and real-time status data to obtain structured sensing multi-source datasets; The preprocessing includes: The system performs time synchronization, format standardization, invalid value removal, and missing data imputation on macro meteorological information, environmental parameters, abnormal data, and real-time status data. It also uses data fusion algorithms to calibrate and match point-based measured data and area-based remote sensing data within a unified spatiotemporal framework, outputting a structured sensing multi-source dataset.

3. The active typhoon-resistant offshore wind power system based on fusion prediction according to claim 1, characterized in that, The decision-making body includes: The typhoon prediction unit is used to receive structured sensing multi-source datasets and generate future wind field forecasts and structural load risk warnings through WRF meteorological models and LSTM load prediction models. Collaborative Decision-Making Unit: Based on future wind field forecasts and structural load risk warnings, it outputs collaborative control commands and triggers hierarchical response strategies; at the same time, it periodically receives structured sensing multi-source datasets, updates the parameters of the WRF meteorological model and the LSTM load prediction model, and optimizes collaborative control commands. The WRF meteorological model adopts a triple nested grid and integrates pre-processed satellite cloud image data, meteorological radar reflectivity data and environmental parameters from structured sensing multi-source data through data assimilation technology to generate a high spatiotemporal resolution forecast wind field for the next 72 hours. The LSTM load prediction model is a three-layer neural network structure. It takes as input the forecast wind field data, environmental parameters and real-time status data of the wind turbine body, predicts the dynamic load spectrum and main excitation frequency of the blade root in the future period, and calculates the probability of resonance risk.

4. The active typhoon-resistant offshore wind power system based on fusion prediction according to claim 3, characterized in that, The collaborative decision-making unit includes a hierarchical response unit and a model prediction control framework. The hierarchical response unit outputs collaborative control commands and triggers hierarchical response strategies based on future wind field forecasts and structural load risk warnings. The model predictive control framework performs rolling optimization updates at fixed cycles: in each control cycle, it re-receives structured sensing multi-source data, updates the input parameters of the WRF meteorological model and the LSTM predictive model accordingly, and re-solves the optimization problem to generate a new round of optimal control commands, thereby achieving closed-loop adaptive control.

5. The active typhoon-resistant offshore wind power system based on fusion prediction according to claim 4, characterized in that, The hierarchical response strategy includes: When the predicted wind speed reaches the Level I threshold, it enters the Level I response state and generates a command to start the plasma flow control unit. When the predicted wind speed reaches the Level II threshold and the resonance risk probability exceeds the preset value, it is in the Level II response state. The command is generated to activate the plasma flow control unit and the active mass damping system at the same time, and to calculate the optimal target position of the counterweight. When the predicted wind speed reaches the Level III threshold or other uncontrollable risks are detected, the response status is Level III, an emergency shutdown command is generated, and the counterweight is locked.

6. The active typhoon-resistant offshore wind power system based on fusion prediction according to claim 5, characterized in that, The response layer also includes: Activate the emergency handling module, which is activated during a Level 3 response. The wind turbine main control coordination unit is used to perform yaw attitude adjustment, emergency shutdown, and counterweight locking; The emergency resource dispatch unit is used to monitor the status of each actuator in real time; at the same time, it coordinates mobile emergency power supplies and links with external rescue forces.

7. The active typhoon-resistant offshore wind power system based on fusion prediction according to claim 6, characterized in that, The response strategy of the response layer is as follows: In the first-level response state, the plasma flow control unit is activated, and by applying optimized high-voltage pulse excitation to the trailing edge electrode array of the blade, flow separation is actively suppressed to maintain aerodynamic stability. When the response is upgraded to the second level, the active mass damping system is activated simultaneously. Based on the optimal control solution, the target position of the counterweight of the active mass damping system is quickly adjusted to change the natural frequency of the blades to avoid resonance. At the same time, the emergency resource scheduling unit monitors the status of each actuator in real time. When the response level is upgraded to Level 3, the emergency handling module is activated, and the wind turbine main control coordination unit immediately performs an emergency shutdown and locks all moving parts; at the same time, the emergency resource dispatch unit coordinates mobile emergency power supplies and links with external rescue forces.

8. The active typhoon-resistant offshore wind power system based on fusion prediction according to claim 1, characterized in that, The plasma flow control unit includes an array of needle-ring electrodes and a high-voltage pulse power supply, with an electrode spacing of 15 mm, covering 15% of the chord length of the blade trailing edge. The electrode array is connected to a high-voltage pulse power supply module, which has a response time of ≤5 ms and adjustable parameters including voltage amplitude of 5~15 kV, pulse frequency of 1~10 kHz, and duty cycle of 10~50%. The active mass damping system includes a carbon fiber guide rail arranged along the blade spanwise, a movable tungsten alloy counterweight slidably mounted on the carbon fiber guide rail, a linear motor mechanism for driving the tungsten alloy counterweight, and a high-frequency sensor array for monitoring position and vibration.

9. The active typhoon-resistant offshore wind power system based on fusion prediction according to claim 1, characterized in that, The communication system includes: The space-based satellite communication subnetwork is connected to the edge communication nodes and communicates with the WRF weather forecasting system and underwater equipment through the BeiDou satellite communication system. It provides command and control and status data feedback channels between the decision-making level and the sea-based unmanned vessels, monitoring equipment in remote sea areas, and some mobile platforms. At the same time, it is responsible for the uplink transmission of meteorological data of the WRF weather forecasting system. The sea-based wireless communication subnetwork communicates with edge communication nodes and adopts a hybrid mode of "surface wireless + underwater acoustic". Surface devices communicate through wireless links, while underwater devices rely on acoustic channels to interact with surface relay nodes or sea-based unmanned vessels. The land-based wired communication subnetwork communicates with the edge communication nodes and connects with the fixed sensors of the equipment via single-mode optical fiber communication. It is used to collect the sensor data of the land-based fixed equipment and upload it to the edge communication nodes. Edge communication nodes, which have bidirectional communication connections with the decision-making layer, are deployed at hub locations corresponding to offshore platforms or key islands. They integrate multiple network interfaces and are responsible for protocol conversion, fusion, and intelligent routing of heterogeneous data from different sub-networks, and uploading it to the decision-making layer. At the same time, they receive decision data or control instructions from the decision-making layer and distribute them to various communication sub-networks.

10. A control method for an active typhoon-resistant offshore wind power system based on fusion prediction, characterized in that, The active typhoon-resistant offshore wind power system based on fusion prediction as described in any one of claims 1-9 includes the following steps: Step 1: The sensing layer collects and preprocesses environmental meteorological data and wind turbine operating status data in real time, and outputs a structured sensing multi-source dataset; Step 2: The decision-making layer receives the structured sensing multi-source dataset, generates future wind field forecasts and structural load risk warnings based on the WRF meteorological model and LSTM load prediction model, and outputs collaborative control commands and triggers graded response strategies; at the same time, it receives the structured sensing multi-source dataset at regular intervals, updates the parameters of the WRF meteorological model and LSTM load prediction model, and optimizes the collaborative control commands. Step 3: The response layer receives the cooperative control command and executes active aerodynamic control and structural dynamic adjustment.

Images