An offshore wind turbine generator set and intelligent control method thereof

Abstract

The present invention discloses an offshore wind turbine and an intelligent control method thereof, comprising a height column, a height rod fixedly installed in the middle of the top of the height column, an assembly shell fixedly installed on the top of the height rod, one end of the assembly shell is rotatably connected with a rotating shaft, one end of the rotating shaft is fixedly installed with an assembly frame, and four sides of the assembly frame are fixedly installed with blades. The present invention sets a mounting mechanism, and a user slides a movable block along the height column to adjust the position of a positioning plate. The positioning plate and the first mounting plate are installed at the use location through ground piles in a tripod manner, thereby improving the stability of its installation and avoiding the phenomenon that moisture on the sea surface impacts the offshore wind turbine and causes the offshore wind turbine to tilt or shake. By setting a power generation mechanism, the offshore wind turbine completes wind power generation through a generator and blades, or completes solar power generation through a solar panel. The power generation method is flexible and changeable, thereby improving its practicality.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore wind turbines, and in particular to an offshore wind turbine and an intelligent control method thereof. Background Art

[0002] Offshore wind turbines are devices that use wind energy to generate electricity. They can effectively reduce carbon dioxide emissions. Compared with traditional power generation equipment such as thermal power plants and nuclear power plants, they have less impact on environmental pollution. In addition, the labor and materials required for the construction, operation and maintenance of offshore wind turbines all come from local areas, which is conducive to promoting local economic development.

[0003] However, traditional offshore wind turbines have the following disadvantages:

[0004] Traditional offshore wind turbines are installed on the sea surface. As the wind force increases, the moisture on the sea surface impacts the offshore wind turbines. Over time, the offshore wind turbines are prone to tilt or shake. Summary of the invention

[0005] The object of the present invention is to provide an offshore wind turbine and an intelligent control method thereof, so as to solve the problem proposed in the above background technology that a traditional offshore wind turbine is installed on the sea surface, and the moisture in the sea surface impacts the offshore wind turbine as the wind force increases. Over time, the offshore wind turbine is prone to tilt or shake.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an offshore wind turbine generator set, comprising a height column, a height rod is fixedly installed in the middle of the top of the height column, an assembly shell is fixedly installed on the top of the height rod, one end of the assembly shell is rotatably connected with a rotating shaft, one end of the rotating shaft is fixedly installed with an assembly frame, four sides of the assembly frame are fixedly installed with blades, one end of the assembly frame is fixedly installed with a blade calibrator, two positioning rings are slidably connected to the middle of the height rod, a control box body is provided on one side of the height rod, a power generation mechanism is fixedly installed on one side of the top of the assembly shell, a wind sensor is fixedly installed in the middle of the top of the assembly shell, a mounting mechanism is fixedly installed in the middle of the height column, and a generator is fixedly installed inside the assembly shell.

[0007] As a preferred technical solution of the present invention, the mounting mechanism includes a movable block and two positioning plates, both sides of the movable block are rotatably connected with angle rods, and the ends of the two angle rods away from the movable block are respectively rotatably connected to the top ends of the two positioning plates, and a plurality of first ground piles are fixedly installed on the bottom ends of the two positioning plates, and a first mounting plate is provided at the bottom end of the movable block, and shock-absorbing springs are fixedly installed on the four corners of the bottom end of the first mounting plate, and a second mounting plate is provided at the bottom end of the first mounting plate, and the ends of the four shock-absorbing springs away from the first mounting plate are respectively fixedly connected to the four corners of the top end of the second mounting plate, and a plurality of second ground piles are fixedly installed on the bottom end of the second mounting plate, and shock absorbers are installed on the opposite sides of the two angle rods. Users install them at the use place through the first ground pile and the second ground pile. After the water flow impacts the first mounting plate and the height column, the shock-absorbing spring undergoes elastic deformation to buffer the impact force, and the shock absorber buffers the deformation force generated by the shock-absorbing spring.

[0008] As a preferred technical solution of the present invention, the interior of the movable block is slidably connected to the height column, the top of the first mounting plate is fixedly connected to the height column, the opposite ends of the two shock absorbers are connected to the side directly opposite the height column, and the mounting mechanism is installed on the height column through the movable block and the first mounting plate.

[0009] As a preferred technical solution of the present invention, the power generation mechanism includes a vertical pole and a connecting block, the top end of the vertical pole is fixedly connected to the bottom end of the connecting block, one side of the connecting block is rotatably connected to a direction plate, the top end of the direction plate is fixedly installed with a solar panel, the middle part of one side of the vertical pole is fixedly installed with two limit plates, a screw rod is rotatably connected between the two limit plates, the middle part of the screw rod is threadedly connected to a height block slidably connected to the vertical pole, one side of the height block is rotatably connected to a push rod, and the push rod is rotatably connected to the side opposite to the direction plate at one end away from the height block The solar panel is fixedly installed on the top of the direction plate, and a stepper motor for driving the screw rod to rotate is fixedly installed on the bottom end of the bottom limit plate. The stepper motor starts after being energized, and the stepper motor drives the screw rod to rotate. The thread on the surface of the screw rod matches the thread on the inner wall of the height block. The height block is limited by the height rod, so the height block slides along the screw rod. During the sliding process of the height block, the push rod is pulled from one side, and the push rod pulls the direction plate, so that the direction plate is deflected along the connection block, the angle of the solar panel is adjusted, and the solar panel converts solar energy into electrical energy.

[0010] As a preferred technical solution of the present invention, the bottom end of the vertical pole is fixedly connected to the assembly shell, and the power generation mechanism is installed on the assembly shell through the vertical pole.

[0011] As a preferred technical solution of the present invention, a guardrail body located outside the height rod is fixedly installed on the top of the height column, one side of the two positioning rings is fixedly connected to one side of the control box body, and the control box body is installed on the height rod through the positioning ring.

[0012] As a preferred technical solution of the present invention, the control box body includes a communication module, a wind speed monitoring module, a wind direction monitoring module, an operation monitoring module, a blade monitoring module, a voltage monitoring module, a current monitoring module, a reactive power compensation module, a battery power supply module, a city grid power supply module, a power supply switching module, a signal encryption module, a signal isolation module and an environmental monitoring module. The communication module transmits the power generation information of the wind power group to the power dispatching center in real time, the wind speed monitoring module monitors the offshore wind speed in real time, the wind direction monitoring module monitors the offshore wind direction in real time, the operation monitoring module monitors the operation of the generator in real time, and the blade monitoring module monitors the rotation of the blade in real time. , the shape of the blades, the voltage monitoring module monitors the electric energy voltage generated by the wind turbine set, the current monitoring module monitors the electric energy current generated by the wind turbine set, the reactive power compensation module compensates the electric energy in real time, the battery power supply module transmits the electric energy generated by the wind turbine set to the battery for storage, the city grid power supply module transmits the electric energy generated by the wind turbine set to the city grid for power supply, the power supply switching module switches the electric energy in real time according to the power demand, the signal encryption module encrypts the information of the wind turbine set, the signal isolation module isolates the wind turbine set from the external environment information in real time for anti-interference protection, and the environmental monitoring module monitors the offshore environment in real time.

[0013] An intelligent control method for an offshore wind turbine generator system comprises the following steps:

[0014] Step 1: Establish a 3D model scene: Based on the B / S framework under the WebGL standard, the front-end adopts HTML+JavaScript technology as the front-end research and development basis. After entering the SovitD / SovitD visualization platform, combine the real world to establish a 3D scene or upload a background image, drag the wind turbines, booster stations, and equipment components in the resource library online, configure colors and animation effects, and intuitively and visually present the operating status of the offshore wind farm. By zooming in, out, and panning the scene, you can view the scene effects and details, and connect environmental parameters, real-time power generation indicators, and energy-saving and emission reduction information data to the D panel;

[0015] Step 2: Global information overview: The above panel also displays the overall production and operation information of the offshore wind farm operation and maintenance center and the wind farm, including environmental conditions, real-time indicators, unit status, environmental parameters, power generation statistics, and wind turbine quantity information;

[0016] Step 3: Intelligent equipment management: Install intelligent wind turbines, intelligent submarine cables and intelligent booster stations on offshore wind turbines.

[0017] As a preferred technical solution of the present invention, the intelligent wind turbine in step three specifically adopts advanced status monitoring, data analysis, and intelligent control technologies on the wind turbine, so that the unit can accurately perceive itself and the external environment, establish data monitoring and abnormal alarm mechanisms, and intelligently lock the fault location. The status monitoring of important components of the mechanical transmission chain, tower, foundation, blades, and important bolt loads is matched for the wind turbine, and intelligent fault warning and diagnosis of gradual and sudden changes in unit parameters are realized. The intelligent submarine cable in step three specifically monitors the temperature and stress changes of the submarine cable in real time, and uses the optical fiber core in the submarine cable to connect to the new optical fiber distributed sensing technology to determine whether the cable is damaged and whether the surrounding environment has changed, and alarm and locate submarine cable abnormalities. The intelligent booster station evaluates and determines whether the structure of the offshore booster station is within the safety limit by online monitoring and storing the key safety parameters of the main structural components.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] 1. By setting up the installation mechanism, the user slides the movable block along the height column to adjust the position of the positioning plate. The positioning plate and the first installation plate are installed at the place of use through the ground piles in a tripod manner, thereby improving the stability of its installation and preventing the moisture on the sea surface from impacting the offshore wind turbine, causing the offshore wind turbine to tilt or shake;

[0020] 2. By setting up a power generation mechanism, the offshore wind turbine unit can generate wind power through generators and blades, or generate solar power through solar panels. The power generation method is flexible and varied, which improves its practicality.

[0021] 3. Through online learning and simulation evaluation, optimize the operating parameters and strategies of the booster station to ensure structural safety and improve operating efficiency;

[0022] 4. Optimize the control strategy of the system through continuous trial and error and feedback through deep reinforcement learning algorithms; BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a side view of the present invention;

[0024] Figure 2 It is a side view of the power generation mechanism of the present invention;

[0025] Figure 3 is a side view of the mounting mechanism of the present invention;

[0026] Figure 4 It is a schematic diagram of the structure of the control box body of the present invention;

[0027] Figure 5 is a flow chart of the present invention;

[0028] Figure 6 The figure is a flow chart of the optimization control algorithm of the intelligent booster station of the present invention.

[0029] In the figure: 1. height column; 2. mounting mechanism; 21. movable block; 22. angle rod; 23. positioning plate; 24. first ground pile; 25. second ground pile; 26. second mounting plate; 27. shock absorbing spring; 28. first mounting plate; 29. ​​shock absorber; 3. guardrail body; 4. control box body; 401. communication module; 402. wind speed monitoring module; 403. wind direction monitoring module; 404. operation monitoring module; 405. blade monitoring module; 406. voltage monitoring module; 407. current monitoring module; 408. reactive power compensation module; 409. battery power supply module; 410, city grid power supply module; 411, power supply switching module; 412, signal encryption module; 413, signal isolation module; 414, environmental monitoring module; 5, positioning ring; 6, height rod; 7, blade; 8, blade calibrator; 9, assembly frame; 10, rotating shaft; 11, wind sensor; 12, power generation mechanism; 121, vertical pole; 122, stepping motor; 123, limit plate; 124, height block; 125, screw rod; 126, connecting block; 127, direction plate; 128, solar panel; 129, push rod; 13, assembly shell; 14, generator. DETAILED DESCRIPTION

[0030] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0031] See also Figure 1-5 The present invention provides an offshore wind turbine generator set, including a height column 1, a height rod 6 is fixedly installed in the middle of the top of the height column 1, an assembly shell 13 is fixedly installed on the top of the height rod 6, one end of the assembly shell 13 is rotatably connected with a rotating shaft 10, one end of the rotating shaft 10 is fixedly installed with an assembly frame 9, four sides of the assembly frame 9 are fixedly installed with blades 7, one end of the assembly frame 9 is fixedly installed with a blade calibrator 8, two positioning rings 5 ​​are slidably connected to the middle of the height rod 6, a control box body 4 is provided on one side of the height rod 6, a power generation mechanism 12 is fixedly installed on one side of the top of the assembly shell 13, a wind sensor 11 is fixedly installed in the middle of the top of the assembly shell 13, a mounting mechanism 2 is fixedly installed in the middle of the height column 1, and a generator 14 is fixedly installed inside the assembly shell 13.

[0032] The mounting mechanism 2 includes a movable block 21 and two positioning plates 23. Both sides of the movable block 21 are rotatably connected with angle rods 22. The ends of the two angle rods 22 away from the movable block 21 are rotatably connected with the top ends of the two positioning plates 23. The bottom ends of the two positioning plates 23 are fixedly installed with a plurality of first ground piles 24. The bottom end of the movable block 21 is provided with a first mounting plate 28. The four corners of the bottom end of the first mounting plate 28 are fixedly installed with shock absorbing springs 27. The bottom end of the first mounting plate 28 is provided with a second mounting plate 26. The four shock absorbing springs 27 are fixedly installed with the first mounting plate 28. One end of the spring 27 away from the first mounting plate 28 is fixedly connected to the four corners of the top of the second mounting plate 26, and a plurality of second ground piles 25 are fixedly installed on the bottom end of the second mounting plate 26. Shock absorbers 29 are installed on the opposite sides of the two angle rods 22. The user installs the first ground piles 24 and the second ground piles 25 at the place of use. After the water flow hits the first mounting plate 28 and the height column 1, the shock-absorbing spring 27 undergoes elastic deformation to buffer the impact force, and the shock absorber 29 buffers the deformation force generated by the shock-absorbing spring 27.

[0033] The interior of the movable block 21 is slidably connected to the height column 1, the top of the first mounting plate 28 is fixedly connected to the height column 1, the opposite ends of the two shock absorbers 29 are connected to the side directly opposite the height column 1, and the mounting mechanism 2 is installed on the height column 1 through the movable block 21 and the first mounting plate 28.

[0034] The power generation mechanism 12 includes a vertical pole 121 and a connecting block 126, the top of the vertical pole 121 is fixedly connected to the bottom of the connecting block 126, one side of the connecting block 126 is rotatably connected to a direction plate 127, the top of the direction plate 127 is fixedly installed with a solar panel 128, two limit plates 123 are fixedly installed in the middle of one side of the vertical pole 121, a screw rod 125 is rotatably connected between the two limit plates 123, the middle part of the screw rod 125 is threadedly connected to a height block 124 slidably connected to the vertical pole 121, one side of the height block 124 is rotatably connected to a push rod 129, one end of the push rod 129 away from the height block 124 is rotatably connected to the side opposite to the direction plate 127, and the top of the direction plate 127 is fixedly installed There is a solar panel 128, and a stepper motor 122 for driving the screw rod 125 to rotate is fixedly installed at the bottom end of the bottom limit plate 123. The stepper motor 122 is started after being energized, and the stepper motor 122 drives the screw rod 125 to rotate. The thread on the surface of the screw rod 125 matches the thread on the inner wall of the height block 124. The height block 124 is limited by the height rod 6, so the height block 124 slides along the screw rod 125. During the sliding process of the height block 124, the push rod 129 is pulled from one side, and the push rod 129 pulls the direction plate 127, so that the direction plate 127 is deflected along the connecting block 126, so that the angle of the solar panel 128 is adjusted, and the solar panel 128 converts solar energy into electrical energy.

[0035] The bottom end of the vertical rod 121 is fixedly connected to the assembly shell 13 , and the power generation mechanism 12 is installed on the assembly shell 13 through the vertical rod 121 .

[0036] A guardrail body 3 located outside the height rod 6 is fixedly mounted on the top of the height column 1 , one side of two positioning rings 5 ​​are fixedly connected to one side of the control box body 4 , and the control box body 4 is mounted on the height rod 6 via the positioning rings 5 ​​.

[0037] The control box body 4 includes a communication module 401, a wind speed monitoring module 402, a wind direction monitoring module 403, an operation monitoring module 404, a blade monitoring module 405, a voltage monitoring module 406, a current monitoring module 407, a reactive power compensation module 408, a battery power supply module 409, a city grid power supply module 410, a power supply switching module 411, a signal encryption module 412, a signal isolation module 413 and an environmental monitoring module 414. The communication module 401 transmits the power generation information of the wind power group to the power dispatching center in real time, the wind speed monitoring module 402 monitors the offshore wind speed in real time, the wind direction monitoring module 403 monitors the offshore wind direction in real time, the operation monitoring module 404 monitors the operation of the generator 14 in real time, and the blade monitoring module 405 monitors the rotation and shape of blade 7 in real time, the voltage monitoring module 406 monitors the electric energy voltage generated by the wind turbine, the current monitoring module 407 monitors the electric energy current generated by the wind turbine, the reactive power compensation module 408 compensates the electric energy in real time, the battery power supply module 409 transmits the electric energy generated by the wind turbine to the battery for storage, the city grid power supply module 410 transmits the electric energy generated by the wind turbine to the city grid for power supply, the power supply switching module 411 switches the electric energy in real time according to the power demand, the signal encryption module 412 encrypts the information of the wind turbine, the signal isolation module 413 isolates the wind turbine from the external environment information in real time for anti-interference protection, and the environmental monitoring module 414 monitors the offshore environment in real time.

[0038] Example 1

[0039] An intelligent control method for an offshore wind turbine generator system comprises the following steps:

[0040] Step 1: Establish a 3D model scene: Based on the B / S framework under the WebGL standard, the front-end adopts HTML5+JavaScript technology as the front-end research and development basis. After entering the Sovit2D / Sovit3D visualization platform, combine the real world to establish a 3D scene or upload a background image, drag the wind turbines, booster stations, and equipment components in the resource library online, configure colors and animation effects, and intuitively and visually present the operating status of the offshore wind farm. By zooming in, out, and panning the scene, you can view the scene effects and details, and connect environmental parameters, real-time power generation indicators, and energy-saving and emission reduction information data to the 2D panel;

[0041] Step 2: Global information overview: The above panel also displays the overall production and operation information of the offshore wind farm operation and maintenance center and the wind farm, including environmental conditions, real-time indicators, unit status, environmental parameters, power generation statistics, and wind turbine quantity information;

[0042] Step 3: Intelligent equipment management: Install intelligent wind turbines, intelligent submarine cables and intelligent booster stations on offshore wind turbines.

[0043] The smart wind turbine in step three specifically adopts advanced status monitoring, data analysis, and intelligent control technologies on the wind turbine, so that the unit can accurately perceive itself and the external environment, establish data monitoring and abnormal alarm mechanisms, intelligently lock the fault location, and match the status monitoring of mechanical transmission chains, towers, foundations, blades, and important bolt load components for the wind turbine, so as to realize intelligent fault warning and diagnosis of gradual and sudden changes in unit parameters. The smart submarine cable in step three specifically monitors the temperature and stress changes of the submarine cable in real time, and uses the optical fiber core in the submarine cable to connect to the new optical fiber distributed sensing technology to determine whether the cable is damaged and whether the surrounding environment has changed, and alarm and locate submarine cable anomalies. The smart booster station monitors and stores the key safety parameters of the main structural components online to evaluate and determine whether the structure of the offshore booster station is within the safety limit.

[0044] In the present invention, based on the B / S framework under the WebGL standard, the front end adopts HTML5+JavaScript technology as the front-end research and development basis. After entering the Sovit2D / Sovit3D visualization platform, a three-dimensional scene is established in combination with the real world or a background image is uploaded. The wind turbines, booster stations, and equipment components in the resource library are dragged online, and colors and animation effects are configured to intuitively and visually present the operating status of the offshore wind farm. The scene effects and details are viewed by zooming in, out, and translating the scene, and environmental parameters, real-time power generation indicators, and energy-saving and emission reduction information data are connected to the 2D panel; the above panel also displays the offshore wind farm operation and maintenance center and the overall production and operation information of the wind farm, including environmental conditions, real-time indicators, unit status, environmental parameters, power generation statistics, and wind turbine quantity information; smart wind turbines, smart submarine cables, and smart booster stations are installed on the offshore wind turbines.

[0045] Example 2

[0046] Introduce an optimization control algorithm based on reinforcement learning in the intelligent booster station; optimize the operating parameters and strategies of the booster station through online learning and simulation evaluation to ensure structural safety while improving operating efficiency; optimize the control strategy of the system through continuous trial and error and feedback through the deep reinforcement learning (DRL) algorithm;

[0047] Step 1: Data Collection

[0048] Obtain historical data from the sensors and monitoring systems of the smart booster station, including key structural parameters (temperature, stress, current, voltage) and operating status data;

[0049] Acquiring historical data from the sensors and monitoring system of the smart booster station is the first step in implementing the intelligent control algorithm; the following are the detailed data collection steps:

[0050] 1.1 Data Source

[0051] Temperature sensor: used to measure the temperature of key components of the booster station, including transformers, switchgear, and cooling systems;

[0052] Stress sensor: used to monitor stress changes in structural parts to prevent structural overload;

[0053] Current sensor: used to measure current value and monitor the stability of power transmission;

[0054] Voltage sensor: used to measure voltage value and ensure the stability of power supply;

[0055] Vibration sensor: used to monitor the vibration of equipment and prevent mechanical failure;

[0056] SCADA system: used for real-time monitoring and data collection, providing system operation status and alarm information;

[0057] Historical database: stores long-term operating data and provides historical data required for data analysis and model training;

[0058] 1.2 Frequency of Data Collection

[0059] Sensor data is usually collected at a frequency of seconds. Here, temperature, current, and voltage data are recorded every 5 seconds. For historical data, it is extracted from the historical database, covering months or even years of data, for model training and verification.

[0060] 1.3 Data variables

[0061] Temperature data: the unit is Celsius (℃); stress data: the unit is megapascal (MPa); current data: the unit is ampere (A); voltage data: the unit is volt (V);

[0062] Step 2: Environment Modeling

[0063] Using the collected data and physical models, build a simulation environment for the intelligent booster station, including the physical structure, control system, and external environment parameters of the booster station; define the state space and action space of the reinforcement learning algorithm;

[0064] 2.1 Simulation environment construction

[0065] The purpose of setting up the simulation environment is to test and optimize the control strategy of the intelligent booster station in a virtual environment to ensure the reliability and adaptability of the model. The following are the detailed simulation environment construction steps and parameter variable descriptions: The data obtained from the data collection include temperature T, stress σ, current I, and voltage V, which are used as input parameters of the simulation environment; the simulation environment contains the physical model of the intelligent booster station to simulate its physical properties and dynamic behavior; for the booster station components: including the thermodynamic and electromagnetic properties of the transformer, the mechanical and electrical properties of the simulated switch, and the heat exchange and cooling efficiency of the simulated cooling system;

[0066] The simulation environment also includes the control system of the intelligent booster station, simulating its control strategy and response behavior; including temperature control, load management, stress monitoring

[0067] 2.2 State and Action Space Definition

[0068] In the reinforcement learning algorithm, the definition of state space and action space is crucial; the following are the detailed state and action space definition steps and parameter variable descriptions:

[0069] The state space consists of the current state information of the intelligent booster station, reflecting the operating state of the system; the variables in the state space include:

[0070] Temperature status: Transformer temperature T transformer , switchgear temperature T switch

[0071] Stress state: structural stress σ structure

[0072] Current state: output current I output

[0073] Voltage state: output voltage V output

[0074] The state space vector is defined as:

[0075] S=[T transformer , T switch , σ structure , I output , V output ]

[0076] The action space consists of the operations performed by the intelligent booster station to adjust the system operating parameters; the variables in the action space include: cooling fan speed A cooling , load distribution factor A load , the stress relief measure is to adjust the structural support A support

[0077] The action space vector is defined as:

[0078] A=[A cooling , A load , A support ]

[0079] Where: Cooling fan speed is A cooling , the load sharing factor is A load , the structural support is A support ;

[0080] 2.3 Specific implementation cases

[0081] To test and optimize the following control strategy in a simulation environment: By adjusting the cooling fan speed A cooling , maintain the transformer temperature T transformer Below 60℃; by adjusting the load distribution factor A load , optimize the current output I output , ensure that the rated current range (145A to 155A); by adjusting the structure to support A support , maintaining the structural stress σ structure Below 15MPa;

[0082] Simulation process

[0083] 2.3.1 Initial state: transformer temperature is 65.2℃, switchgear temperature is 64.8℃, structural stress is 12.5MPa, output current is 150A, and output voltage is 33.5V;

[0084] 2.3.2 State Update: At each time step, the state in the simulation environment is updated according to the current state S and the executed action A:

[0085] S new =f(S, A)

[0086] Among them, f is the dynamic model of the simulation environment;

[0087] 2.3.3 Action selection: Based on the current state S, the reinforcement learning algorithm is used to select the optimal action A to adjust the cooling fan speed to reduce the temperature:

[0088] A = arg maxQ(S, A)

[0089] Q(S, A) represents the expected reward after performing action A in state S plus the sum of the discounted cumulative rewards of future states. This value is estimated by the Q network;

[0090] 2.3.4 Reward calculation: Calculate the reward R obtained for the current state S and action A. The lower the transformer temperature, the higher the reward:

[0091] R=-|T transformer-60|-penalty for high stress

[0092] "penalty for high stress" refers to the penalty term for high stress conditions; specifically, this penalty term is used to reduce the total reward when the system stress is too large, so as to encourage the control strategy to avoid excessive stress and thus ensure the safety of the system;

[0093] 2.3.5 Training and optimization: Through multiple simulations and strategy updates, the reinforcement learning model is trained and the control strategy is optimized to enable the booster station to operate efficiently and safely;

[0094] Step 3: Reinforcement learning algorithm design and training

[0095] Select appropriate deep reinforcement learning algorithms, including Deep Q Network (DQN) and Deep Deterministic Policy Gradient (DDPG); design a reasonable reward function to reflect the operating efficiency and safety of the booster station;

[0096] 3.1 Algorithm Selection

[0097] In combination with the offshore wind turbine intelligent control method of the present invention, a deep Q network (DQN) and a deep deterministic policy gradient (DDPG) algorithm are selected to optimize the operation control strategy of the booster station;

[0098] DQN: Applicable to discrete action space, selecting the optimal action by approximating the Q value function;

[0099] DDPG: Applicable to continuous action space, generates optimal actions through policy network, and uses Q-value network to evaluate action value;

[0100] 3.2 Reward Function Design

[0101] In order to effectively reflect the operating efficiency and safety of the booster station, the following reward function is designed: the higher the system operating efficiency, the higher the reward; the higher the system operates within the safety limit, the higher the reward; once the limit is exceeded, the reward is reduced or punished;

[0102] The specific reward function formula is as follows:

[0103] R=R efficiency +R security

[0104] in,

[0105]

[0106] Where: E output : Current output power, E max : Maximum output power, T transformer: Transformer temperature, T limit : Transformer temperature limit, σ structure : Structural stress, σ limit : structural stress limit, α, β, γ: bonus coefficients;

[0107] 3.3 Model Training

[0108] 3.3.1DQN

[0109] Initialize the Q network Q(s, a; θ) and the target Q network Q′(s, a; θ - ) parameters θ and θ - ; Initialize the experience playback buffer D;

[0110] Q value update:

[0111]

[0112] Where: s is the current state, a is the current action, r is the current reward, s′ is the next state, α is the learning rate, γ is the discount factor, θ is the Q network parameter, θ - is the target Q network parameter;

[0113] Experience Replay:

[0114] In each step, a mini-batch (s, a, r, s′) is randomly drawn from the buffer D for training to reduce data correlation and improve training stability;

[0115] Periodically copy the Q network parameters to the target Q network:

[0116] θ - ←θ

[0117] 3.3.2DDPG

[0118] Initialize the policy network μ(s;θ μ ) and Q network Q(s, a; θ Q ) parameter θ μ and θ Q Initialize the target policy network μ′(s;θ μ ) and the target Q network Q′(s, a; θ Q ) parameter θ μ and θ Q ; Initialize the experience playback buffer D;

[0119] Q value update:

[0120]

[0121] in: represents the expected value of a sample extracted from the experience replay buffer D; the experience replay buffer D stores the past state s, action a, reward r, and next state s′, where r is the current reward and s′ is the next state. Q is the Q network parameter, μ′(s′; θ μ ) is the optimal action output by the target policy network;

[0122] Policy Update:

[0123]

[0124] in:

[0125] represents the expectation of the state s extracted from the experience replay buffer D; that is, E represents the expected value, and s~D indicates that the state s is extracted from the experience replay buffer D, J is the objective function of the policy network, and θ μ are policy network parameters, is the gradient of the Q network under the current strategy; μ(s; θ μ ) represents the action output by the policy network μ in state s; this policy network is composed of parameters θ μ definition; represents the output (i.e., action) of the policy network μ relative to its parameters θ μ The gradient of

[0126] Experience Replay:

[0127] At each step, a mini-batch (s, a, r, s′) is randomly drawn from the buffer DD for training;

[0128] Soft update target network parameters:

[0129] θ μ ←τθ μ +(1-τ)θ μ

[0130] θ Q ←τθ Q +(1-τ)θ Q

[0131] Where τ is the soft update rate;

[0132] 3.4. Specific implementation cases

[0133] Initialize the simulation environment, set the initial state and external environment parameters;

[0134] Run DQN and DDPG algorithms in a simulation environment, use experience replay and target network technology, and continuously conduct trial and error and feedback to optimize control strategies;

[0135] DQN selects the optimal discrete action by updating the Q-value function; DDPG generates the optimal continuous action through the policy network, and the Q network evaluates the action value;

[0136] Deploy the trained model to the actual smart booster station and continuously optimize it through online learning to ensure the efficiency and safety of the system in actual operation;

[0137] Step 4: Model deployment and online learning

[0138] Deploy the trained reinforcement learning model to the actual intelligent booster station control system; use online learning technology to continuously update and optimize the model to adapt to changes in actual operation and improve the adaptability and effectiveness of the control strategy;

[0139] 4.1 Model Deployment

[0140] Deploying the trained reinforcement learning model to the actual intelligent booster station control system is a key step; the following are detailed model deployment steps and parameter variable descriptions: Export the trained reinforcement learning model from the training environment and convert it into a format suitable for actual deployment (ONNX);

[0141] In the DQN model, the Q network parameters θ are derived; in the DDPG model, the policy network parameters θ are derived μ and Q network parameters θ Q ;

[0142] Save the model files: DQN model file: dqn_model.pb, DDPG policy network file: ddpg_actor.pb, DDPGQ network file: ddpg_critic.pb;

[0143] Integrate a deep learning inference framework (ONNX Runtime) in the intelligent booster station control system to load and run trained models;

[0144] Develop the interface between the control system and the model to ensure real-time data is passed to the model and receive the control strategy output by the model;

[0145] Before actual deployment, conduct offline testing to verify the performance and stability of the model under real data; gradually enable the model in the actual system and conduct online testing to ensure the reliability of the model in actual operation;

[0146] Specific parameters include: ambient temperature is T env , the transformer temperature is T transformer , the structural stress is σ structure , the output current is I output , the output voltage is V output, cooling fan speed is A cooling , the load sharing factor is A load , the structural support is adjusted to A support ;

[0147] 4.2 Online Learning

[0148] Online learning is to use real-time data to continuously update and optimize the model to adapt to changes in actual operation and improve the adaptability and effectiveness of the control strategy; the specific steps are:

[0149] Real-time data collection: To collect real-time operating data from sensors and monitoring systems, including temperature, stress, current, and voltage; collect data once a minute to build a time series data set;

[0150] Data preprocessing: cleaning, normalizing and denoising real-time data to ensure data quality;

[0151] Incremental learning: In order to use real-time data for incremental learning, update model parameters and adapt to system changes; DQN incremental learning uses real-time data stored in the experience replay buffer to update Q network parameters; parameters are updated every 1000 steps; DDPG incremental learning uses real-time data to update policy network and Q network parameters; policy network parameters are updated every 100 steps; Q network parameters are updated every 500 steps;

[0152] Regularly update the parameters of the target Q network and the target policy network to ensure the stability and convergence of the model; update the target network parameters every 10,000 steps;

[0153] Online evaluation: To monitor the performance of the model in real time and evaluate its performance in actual operation;

[0154] The performance indicators are operating efficiency, which is the power generation efficiency of the real-time monitoring system, and safety, which is whether the key parameters of the monitoring system are within the safety range. The operating effect of the model is evaluated every 30 minutes;

[0155] Strategy adjustment: To adjust the control strategy of the model based on the evaluation results to ensure efficient and safe operation of the system; if the temperature is too high, increase the cooling fan speed; if the stress is too large, adjust the structural support;

[0156] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent control method for an offshore wind turbine, the turbine comprising a height column (1), characterized in that: A height rod (6) is fixedly mounted in the middle of the top of the height column (1), an assembly shell (13) is fixedly mounted on the top of the height rod (6), one end of the assembly shell (13) is rotatably connected to a rotating shaft (10), one end of the rotating shaft (10) is fixedly mounted to an assembly frame (9), four sides of the assembly frame (9) are fixedly mounted with blades (7), one end of the assembly frame (9) is fixedly mounted with a blade calibrator (8), the middle of the height rod (6) is slidably connected to two positioning rings (5), one side of the height rod (6) is provided with a control box body (4), one side of the top of the assembly shell (13) is fixedly mounted with a power generation mechanism (12), the middle of the top of the assembly shell (13) is fixedly mounted with a wind sensor (11), the middle of the height column (1) is mounted with a mounting mechanism (2), and a generator (14) is fixedly mounted inside the assembly shell (13); The method comprises the following steps: Step 1: Establish a 3D model scene: Based on the B / S framework under the WebGL standard, the front-end adopts HTML5+JavaScript technology as the front-end research and development basis. After entering the Sovit2D / Sovit3D visualization platform, combine the real world to establish a 3D scene or upload a background image, drag the wind turbines, booster stations, and equipment components in the resource library online, configure colors and animation effects, and intuitively and visually present the operating status of the offshore wind farm. By zooming in, out, and panning the scene, you can view the scene effects and details, and connect environmental parameters, real-time power generation indicators, and energy-saving and emission reduction information data to the 2D panel; Step 2: Global information overview: The above panel also displays the overall production and operation information of the offshore wind farm operation and maintenance center and the wind farm, including environmental conditions, real-time indicators, unit status, environmental parameters, power generation statistics, and wind turbine quantity information; Step 3: Intelligent equipment management: Install intelligent wind turbines, intelligent submarine cables and intelligent booster stations on offshore wind turbines; In the step three, the smart wind turbine specifically adopts advanced status monitoring, data analysis, and intelligent control technologies on the wind turbine, so that the unit can accurately perceive itself and the external environment, establish data monitoring and abnormal alarm mechanisms, intelligently lock the fault location, and match the status monitoring of the mechanical transmission chain, tower, foundation, blades, and important bolt load components of the wind turbine to achieve intelligent fault warning and diagnosis of gradual and sudden changes in unit parameters. In the step three, the smart submarine cable specifically monitors the temperature and stress changes of the submarine cable in real time, and uses the optical fiber core in the submarine cable to connect to the new optical fiber distributed sensing technology to determine whether the cable is damaged and whether the surrounding environment has changed, and alarm and locate the submarine cable anomalies. The smart booster station monitors and stores the key safety parameters of the main structural parts online to evaluate and determine whether the structure of the offshore booster station is within the safety limit. Introduce an optimization control algorithm based on reinforcement learning in the intelligent booster station; optimize the operating parameters and strategies of the booster station through online learning and simulation evaluation to ensure structural safety while improving operating efficiency; optimize the control strategy of the system through continuous trial and error and feedback through the deep reinforcement learning DRL algorithm; Step 1: Data Collection Obtain historical data from the sensors and monitoring systems of the smart booster station, including key structural parameters and operating status data of temperature, stress, current, and voltage; Acquiring historical data from the sensors and monitoring system of the smart booster station is the first step in implementing the intelligent control algorithm; the following are the detailed data collection steps: 1.1 Data Source Temperature sensor: used to measure the temperature of key components of the booster station, including transformers, switchgear, and cooling systems; Stress sensor: used to monitor stress changes in structural parts to prevent structural overload; Current sensor: used to measure current value and monitor the stability of power transmission; Voltage sensor: used to measure voltage value and ensure the stability of power supply; Vibration sensor: used to monitor the vibration of equipment and prevent mechanical failure; SCADA system: used for real-time monitoring and data collection, providing system operation status and alarm information; Historical database: stores long-term operating data and provides historical data required for data analysis and model training; 1.2 Frequency of Data Collection Sensor data is usually collected at a frequency of seconds. Here, temperature, current, and voltage data are recorded every 5 seconds. For historical data, it is extracted from the historical database, covering months or even years of data, for model training and verification. 1.3 Data variables Temperature data: the unit is Celsius; stress data: the unit is MPa; current data: the unit is ampere; voltage data: the unit is volt; Step 2: Environment Modeling Using the collected data and physical models, build a simulation environment for the intelligent booster station, including the physical structure, control system, and external environment parameters of the booster station; define the state space and action space of the reinforcement learning algorithm; 2.1 Simulation environment construction The purpose of setting up the simulation environment is to test and optimize the control strategy of the intelligent booster station in a virtual environment to ensure the reliability and adaptability of the model. The following are the detailed simulation environment setting steps and parameter variable descriptions: The data obtained from data collection include temperature ,stress , Current ,Voltage , as the input parameters of the simulation environment; the simulation environment contains the physical model of the intelligent booster station, simulating its physical characteristics and dynamic behavior; for the booster station components: including the thermodynamic and electromagnetic characteristics of the transformer, simulating the mechanical and electrical characteristics of the switch, simulating the heat exchange and cooling efficiency of the cooling system; The simulation environment also includes the control system of the intelligent booster station, simulating its control strategy and response behavior; including temperature control, load management, and stress monitoring; 2.2 State and Action Space Definition In the reinforcement learning algorithm, the definition of state space and action space is crucial; the following are the detailed state and action space definition steps and parameter variable descriptions: The state space consists of the current state information of the intelligent booster station, reflecting the operating state of the system; the variables in the state space include: Temperature status: Transformer temperature , switchgear temperature ; Stress state: structural stress ; Current status: output current ; Voltage status: output voltage ; The state space vector is defined as: ； The action space consists of the operations performed by the smart booster station to adjust the system operating parameters; the variables in the action space include: cooling fan speed , load sharing factor , stress relief measures include adjusting the structural support ; The action space vector is defined as: ； Where: Cooling fan speed is , the load sharing factor is , the structural support is ; 2.3 Specific implementation process To test and optimize the following control strategy in a simulation environment: , maintain the transformer temperature Below 60℃; by adjusting the load distribution coefficient , optimize current output , ensure that it is within the rated current range; by adjusting the structural support , maintain structural stress Below 15MPa; Simulation process 2.3.1 Initial state: transformer temperature is 65.2℃, switchgear temperature is 64.8℃, structural stress is 12.5MPa, output current is 150A, and output voltage is 33.5V; 2.3.2 State Update: At each time step, according to the current state and actions performed , update the status in the simulation environment: ； in, A dynamic model of the simulation environment; 2.3.3 Action selection: based on current state , using reinforcement learning algorithm to select the optimal action , adjust the cooling fan speed to reduce temperature: ； Indicates in status Next, execute the action The expected reward after the event plus the sum of the discounted cumulative rewards for each future state, which is estimated by the Q network; 2.3.4 Reward calculation: Calculate the current state and actions Rewards received , the lower the transformer temperature, the higher the reward: ； "penalty for high stress" refers to the penalty term for high stress conditions; specifically, this penalty term is used to reduce the total reward when the system stress is too large, so as to encourage the control strategy to avoid excessive stress and thus ensure the safety of the system; 2.3.5 Training and optimization: Through multiple simulations and strategy updates, the reinforcement learning model is trained and the control strategy is optimized to enable the booster station to operate efficiently and safely; Step 3: Reinforcement learning algorithm design and training Select appropriate deep reinforcement learning algorithms, including deep Q-network DQN and deep deterministic policy gradient DDPG; design a reasonable reward function to reflect the operating efficiency and safety of the booster station; 3.1 Algorithm Selection In combination with the offshore wind turbine intelligent control method of the present invention, the deep Q network DQN and the deep deterministic policy gradient DDPG algorithm are selected to optimize the operation control strategy of the booster station; DQN: Applicable to discrete action space, selecting the optimal action by approximating the Q value function; DDPG: Applicable to continuous action space, generates optimal actions through policy network, and uses Q-value network to evaluate action value; 3.2 Reward Function Design In order to effectively reflect the operating efficiency and safety of the booster station, the following reward function is designed: the higher the system operating efficiency, the higher the reward; the higher the system operates within the safety limit, the higher the reward; once the limit is exceeded, the reward is reduced or punished; The specific reward function formula is as follows: ； in, ； ； in: , is the intermediate variable of the reward function; : Current output power, : Maximum output power, : Transformer temperature, : Transformer temperature limit, : Structural stress, : Structural stress limit, : Reward coefficient; 3.3 Model Training 3.3.1DQN Initialize the Q network and the target Q network Parameters and ; Initialize the experience replay buffer ; Q value update: ； in: is the current state, For the current action, For the current reward, For the next state, For the next action, is the learning rate, is the discount factor, is the Q network parameter, is the target Q network parameter; Experience Replay: At each step, from the buffer Randomly draw a small batch from Conduct training to reduce data correlation and improve training stability; Periodically copy the Q network parameters to the target Q network: ； 3.3.2DDPG Initialize the policy network and Q Network Parameters and , initialize the target policy network and the target Q network Parameters and ; Initialize the experience replay buffer ; Q value update: ； in: Represents the experience replay buffer The expected value of the sample drawn from the experience replay buffer Stores past status ,action ,award , and the next state , For the current reward, For the next state, is the Q network parameter, The optimal action output by the target policy network; Policy Update: ； in: : Indicates the playback buffer from the experience The state extracted The expected value of Indicates status From the experience replay buffer Extracted from is the objective function of the policy network, are policy network parameters, is the gradient of the Q network under the current strategy; Indicates in status Next, the policy network The output action; this policy network consists of parameters definition; Representation Policy Network The output (i.e., action) of a relative to its parameters The gradient of Experience Replay: At each step, from the buffer Randomly draw a small batch from Conduct training; Soft update target network parameters: ； in is the rate of soft updates; 3.

4. Specific implementation Initialize the simulation environment, set the initial state and external environment parameters; Run DQN and DDPG algorithms in a simulation environment, use experience replay and target network technology, and continuously conduct trial and error and feedback to optimize control strategies; DQN selects the optimal discrete action by updating the Q-value function; DDPG generates the optimal continuous action through the policy network, and the Q network evaluates the action value; Deploy the trained model to the actual smart booster station and continuously optimize it through online learning to ensure the efficiency and safety of the system in actual operation; Step 4: Model deployment and online learning Deploy the trained reinforcement learning model to the actual intelligent booster station control system; use online learning technology to continuously update and optimize the model to adapt to changes in actual operation and improve the adaptability and effectiveness of the control strategy; 4.1 Model Deployment Deploying the trained reinforcement learning model to the actual intelligent booster station control system is a key step; the following are detailed model deployment steps and parameter variable descriptions: Export the trained reinforcement learning model from the training environment and convert it to the ONNX format suitable for actual deployment; Among them, the DQN model derives the Q network parameters ;DDPG model, export policy network parameters and Q network parameters ; Save the model files: DQN model file: dqn_model.pb, DDPG policy network file: ddpg_actor.pb, DDPGQ network file: ddpg_critic.pb; Integrate the deep learning inference framework ONNX Runtime into the intelligent booster station control system to load and run the trained model; Develop the interface between the control system and the model to ensure real-time data is passed to the model and receive the control strategy output by the model; Before actual deployment, conduct offline testing to verify the performance and stability of the model under real data; gradually enable the model in the actual system and conduct online testing to ensure the reliability of the model in actual operation; Specific parameters include: ambient temperature , the transformer temperature is , the structural stress is , the output current is , the output voltage is , cooling fan speed is , the load sharing factor is , the structural support is adjusted to ; 4.2 Online Learning Online learning is to use real-time data to continuously update and optimize the model to adapt to changes in actual operation and improve the adaptability and effectiveness of the control strategy; the specific steps are: Real-time data collection: To collect real-time operating data from sensors and monitoring systems, including temperature, stress, current, and voltage; collect data once a minute to build a time series data set; Data preprocessing: cleaning, normalizing and denoising real-time data to ensure data quality; Incremental learning: In order to use real-time data for incremental learning, update model parameters and adapt to system changes; DQN incremental learning uses real-time data stored in the experience replay buffer to update Q network parameters; parameters are updated every 1000 steps; DDPG incremental learning uses real-time data to update policy network and Q network parameters; policy network parameters are updated every 100 steps; Q network parameters are updated every 500 steps; Regularly update the parameters of the target Q network and the target policy network to ensure the stability and convergence of the model; update the target network parameters every 10,000 steps; Online evaluation: To monitor the performance of the model in real time and evaluate its performance in actual operation; The performance indicators are operating efficiency, which is the power generation efficiency of the real-time monitoring system, and safety, which is whether the key parameters of the monitoring system are within the safety range. The operating effect of the model is evaluated every 30 minutes; Strategy adjustment: To adjust the control strategy of the model based on the evaluation results to ensure efficient and safe operation of the system; if the temperature is too high, increase the cooling fan speed; if the stress is too large, adjust the structural support.

2. The method according to claim 1, characterized in that: The mounting mechanism (2) comprises a movable block (21) and two positioning plates (23). Both sides of the movable block (21) are rotatably connected to angle rods (22). One end of the two angle rods (22) away from the movable block (21) is rotatably connected to the top of the two positioning plates (23). A plurality of first ground piles (24) are fixedly mounted on the bottom ends of the two positioning plates (23). A first mounting plate (28) is provided at the bottom end of the movable block (21). Shock-absorbing springs (27) are fixedly mounted on four corners of the bottom end of the first mounting plate (28). A second mounting plate (26) is provided at the bottom end of the first mounting plate (28). One end of the four shock-absorbing springs (27) away from the first mounting plate (28) is fixedly connected to four corners of the top end of the second mounting plate (26). A plurality of second ground piles (25) are fixedly mounted on the bottom end of the second mounting plate (26). Shock absorbers (29) are installed on opposite sides of the two angle rods (22).

3. The method according to claim 2, characterized in that: The interior of the movable block (21) is slidably connected to the height column (1), the top end of the first mounting plate (28) is fixedly connected to the height column (1), and opposite ends of the two shock absorbers (29) are connected to the side directly facing the height column (1).

4. The method according to claim 1, characterized in that: The power generation mechanism (12) comprises a vertical pole (121) and a connecting block (126); the top end of the vertical pole (121) is fixedly connected to the bottom end of the connecting block (126); one side of the connecting block (126) is rotatably connected to a direction plate (127); the top end of the direction plate (127) is fixedly mounted with a solar panel (128); two limit plates (123) are fixedly mounted in the middle of one side of the vertical pole (121); a screw rod (125) is rotatably connected between the two limit plates (123); the screw rod (125) is rotatably connected to the bottom end of the connecting block (126); A height block (124) slidably connected to the vertical rod (121) is threadedly connected to the middle of the rod (125); a push rod (129) is rotatably connected to one side of the height block (124); an end of the push rod (129) away from the height block (124) is rotatably connected to a side directly opposite to a direction plate (127); a solar panel (128) is fixedly mounted on the top of the direction plate (127); and a stepping motor (122) for driving the screw rod (125) to rotate is fixedly mounted on the bottom end of the bottom limit plate (123).

5. The method according to claim 4, characterized in that: The bottom end of the vertical rod (121) is fixedly connected to the assembly shell (13).

6. The method according to claim 1, characterized in that: A guardrail body (3) located outside the height rod (6) is fixedly mounted on the top of the height column (1), and one side of the two positioning rings (5) is fixedly connected to one side of the control box body (4).

7. The method according to claim 1, characterized in that: The control box body (4) comprises a communication module (401), a wind speed monitoring module (402), a wind direction monitoring module (403), an operation monitoring module (404), a blade monitoring module (405), a voltage monitoring module (406), a current monitoring module (407), a reactive power compensation module (408), a battery power supply module (409), a city grid power supply module (410), a power supply switching module (411), a signal encryption module (412), a signal isolation module (413) and an environment monitoring module (414). The communication module (401) transmits the power generation information of the wind power group to the power dispatching center in real time. The wind speed monitoring module (402) monitors the offshore wind speed in real time. The wind direction monitoring module (403) monitors the offshore wind direction in real time. The operation monitoring module (404) monitors the operation of the generator (14) in real time. The blade monitoring module (405) monitors the rotation of the blade (7) and the shape of the blade (7) in real time, the voltage monitoring module (406) monitors the voltage of the electric energy generated by the wind turbine set, the current monitoring module (407) monitors the current of the electric energy generated by the wind turbine set, the reactive power compensation module (408) compensates the electric energy in real time, the battery power supply module (409) transmits the electric energy generated by the wind turbine set to the battery for storage, the city grid power supply module (410) transmits the electric energy generated by the wind turbine set to the city grid for power supply, the power supply switching module (411) switches the electric energy in real time according to the power demand, the signal encryption module (412) encrypts the information of the wind turbine set, the signal isolation module (413) isolates the wind turbine set from the external environment information in real time for anti-interference protection, and the environment monitoring module (414) monitors the offshore environment in real time.

Images