Wind power generation system operation optimization method based on intelligent control

Abstract

The invention discloses a wind power generation system operation optimization method based on intelligent control, and the method comprises the following steps: S1, collecting the real-time data during the operation of a wind power generation system, including the wind speed, the wind direction, the rotating speed of a generator, the output power, and the environment temperature; s2, constructing a performance model of the wind power generation system according to the real-time data; s3, based on the performance model, adjusting the operation parameters of the generator in real time to optimize the power generation efficiency of the system; and S4, when detecting that the system is abnormal, automatically triggering a fault diagnosis module, generating a fault processing suggestion and executing an optimization strategy. Through real-time data acquisition, performance model construction, operation parameter dynamic adjustment and intelligent fault diagnosis, the problems that a traditional wind power generation system is low in efficiency and not timely in fault response are solved, and the system has the advantages of dynamically optimizing power generation efficiency, timely processing system abnormity and improving operation stability.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation, and more specifically, to an operation optimization method for a wind power generation system based on intelligent control. Background Art

[0002] Wind power generation, as a clean and renewable energy form, has been widely applied globally in recent years. However, the operation efficiency and stability of wind power generation systems still face many challenges. In the prior art, wind power generation systems usually have the following problems: low operation efficiency, the operation parameters of wind power generation systems (such as blade pitch angles, generator speeds, etc.) are usually set based on fixed control strategies and cannot be dynamically adjusted according to environmental conditions such as real-time wind speed and wind direction, resulting in the power generation efficiency not reaching the optimal state.

[0003] Lack of timely fault response, the existing systems rely on manual intervention or simple threshold judgment for fault detection and handling, lacking intelligent fault diagnosis and automatic repair capabilities, resulting in a long fault response time and affecting the system operation stability. The problem of power fluctuation, the randomness and volatility of wind speed will cause the output power of wind power generation systems to be unstable, affecting the stability of the power grid and the power quality.

[0004] Insufficient environmental adaptability, the operation performance of wind power generation systems is poor under extreme weather conditions (such as high temperature, low temperature, strong wind, etc.), lacking effective protection mechanisms, which may lead to equipment damage or shortened service life.

[0005] Insufficient remote monitoring and management capabilities, the existing systems have limited capabilities for monitoring and managing operation data, lacking functions of real-time data transmission and remote optimization adjustment, and it is difficult to achieve efficient centralized management.

[0006] Therefore, how to provide an operation optimization method for a wind power generation system based on intelligent control is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0007] In view of this, the present invention provides an operation optimization method for a wind power generation system based on intelligent control, aiming to solve the technical problems of low operation efficiency, lack of timely fault response, power fluctuation and insufficient environmental adaptability.

[0008] To achieve the above object, the present invention adopts the following technical solutions:

[0009] An operation optimization method for a wind power generation system based on intelligent control, comprising the following steps:

[0010] S1. Collect real-time data during the operation of the wind power generation system, including wind speed, wind direction, generator speed, output power, and environmental temperature;

[0011] S2. Construct a performance model of the wind power generation system based on the real-time data;

[0012] S3. Based on the performance model, adjust the operating parameters of the generator in real time to optimize the power generation efficiency of the system;

[0013] S4. When a system anomaly is detected, automatically trigger the fault diagnosis module to generate fault handling suggestions and execute optimization strategies.

[0014] Furthermore, the real-time data further includes blade angle, gearbox temperature, and grid load parameters.

[0015] Furthermore, the performance model is constructed based on machine learning algorithms, and the algorithms include but are not limited to support vector machines, neural networks, or decision trees.

[0016] Furthermore, the adjustment of the operating parameters includes:

[0017] Dynamically adjust the blade pitch angle to adapt to wind speed changes;

[0018] Optimize the generator speed to match the current wind speed and load demand;

[0019] Adjust the output power of the converter to improve the power quality.

[0020] Furthermore, the fault diagnosis module includes:

[0021] Real-time monitor abnormal fluctuations of key system parameters;

[0022] Generate fault types and handling suggestions based on historical fault data and a preset rule library;

[0023] Automatically adjust the system operating mode to avoid further damage.

[0024] Furthermore, it further includes a power prediction module:

[0025] Predict the wind speed and power output in the next period of time based on historical data and real-time data;

[0026] According to the prediction results, adjust the system operating parameters in advance to optimize the power generation efficiency.

[0027] Furthermore, the power prediction module adopts the Kalman filter algorithm or time series analysis method.

[0028] Furthermore, it further includes a remote monitoring module:

[0029] Transmit the system operating data to the monitoring center in real time through a wireless communication network;

[0030] The monitoring center generates optimization instructions based on the data and sends them to the wind power generation system.

[0031] Furthermore, it also includes an intelligent scheduling module:

[0032] Dynamically adjust the power generation according to the grid load demand and the real-time status of the wind power generation system;

[0033] Output the maximum power preferentially during the peak period of grid load, and reduce the output during the low period to extend the equipment life.

[0034] Furthermore, it also includes an environmental adaptability module:

[0035] Automatically adjust the operating parameters of the cooling system and the lubrication system according to the changes of environmental temperature, humidity and air pressure;

[0036] Trigger the protection mechanism under extreme weather conditions to ensure the safe operation of the system.

[0037] Compared with the prior art, the present invention solves the problems of low efficiency and untimely fault response of the traditional wind power generation system through real-time data acquisition, constructing a performance model, dynamically adjusting operating parameters and intelligent fault diagnosis, and has the advantages of dynamically optimizing the power generation efficiency, timely handling system anomalies and improving the operation stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0039] Figure 1 It is a schematic flowchart of an operation optimization method for a wind power generation system based on intelligent control of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] The following clearly and completely describes the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0041] In the operation control strategy of the traditional existing wind power generation system, the fixed parameter setting cannot adapt to the dynamic environmental changes, resulting in the generator frequently deviating from the optimal operation range when the wind speed fluctuates. The artificial-dependent fault response mechanism has a lag, and it is difficult to capture sudden situations such as abnormal mechanical stress of the blades or overheating of the gearbox in a timely manner. The lack of systematic compensation for the coupled influence of environmental parameters makes the lubrication failure or the cooling efficiency decline easily caused by extreme temperature and humidity changes.

[0042] For example, for a megawatt-class doubly-fed wind turbine deployed in a coastal area with high salt fog, its pitch system adjusts the pitch angle based on a preset wind speed threshold. When encountering the superposition of gusts and turbulence, the angle of attack of the blade deviates from the aerodynamic optimal range for more than 15 seconds, and the fluctuation of the generator speed causes the torque oscillation of the drive train to intensify. At the same time, the monitoring of the gearbox oil temperature relies on daily manual inspections at fixed times, and the abnormal temperature rise caused by sudden bearing wear can only be identified after 2 hours. During this period, the decrease in the oil viscosity causes an increase in the wear amount of the gear meshing surface by 0.12 mm. When the environmental humidity breaks through the 85% threshold, the existing system does not establish a humidity-cooling efficiency correlation model, and the temperature rise rate of the generator winding increases to 1.3 times that of the normal operating condition.

[0043] If the above problems are not solved, the continuous loss of aerodynamic efficiency will cause the annual power generation gap to exceed 18% of the design value, and the cumulative abnormal vibration of the drive train will lead to a 40% reduction in the service life of the main shaft bearing. The delay in the fault diagnosis window period may increase the single repair cost by 2.5 million yuan, and the environmental adaptability defect will cause the annual unplanned shutdown time of the unit in the tropical monsoon climate area to increase by 300 hours.

[0044] When facing the above problems, this application first analyzes the technical bottlenecks of traditional wind power generation systems in terms of dynamic environmental adaptability, fault response mechanisms, and environmental parameter coupling processing. Aiming at the contradiction between the fixed parameter control strategy and the dynamic wind speed change, it is considered to introduce a real-time data acquisition and dynamic adjustment mechanism to replace the preset threshold control. At the same time, to solve the response delay of the manual inspection mode to sudden faults, an attempt is made to establish a closed-loop control system for automatic monitoring and active diagnosis. For the chain reaction caused by environmental parameter changes, the possibility of multi-dimensional data fusion analysis is studied.

[0045] During the process of scheme construction, attempts are made to optimize through the following paths: First, a sensor network is used to obtain key parameters such as wind speed and temperature in real time, but it is found that single parameter acquisition cannot accurately reflect the overall operating state of the system; Second, an offline prediction model based on historical data is established, but tests show that it cannot adapt to real-time working condition changes; Third, an attempt is made to dynamically correlate and control the generator speed and output power, but the influence of environmental factors on the equipment performance is not solved. Through comparison, it is found that a real-time data system including environmental parameters needs to be constructed, and a system performance model that can be dynamically updated needs to be established to achieve precise control.

[0046] In response to this, this application proposes an operation optimization method for a wind power generation system based on intelligent control, such as Figure 1As shown, it includes the following steps: collecting real-time data during the operation of the wind power generation system, including wind speed, wind direction, generator speed, output power, and environmental temperature; constructing a performance model of the wind power generation system based on the real-time data; adjusting the operation parameters of the generator in real time based on the performance model to optimize the power generation efficiency of the system; when detecting system anomalies, automatically triggering a fault diagnosis module to generate fault handling suggestions and execute optimization strategies.

[0047] Among them, the real-time data refers to a set of dynamic parameters reflecting the operation status of the wind power generation system and environmental conditions. Specifically, it can be implemented by a sensor network composed of a wind speed sensor, a rotation speed encoder, a power transmitter, and a temperature sensor. Through multi-dimensional data fusion, it provides input support for performance modeling, and solves the problem of insufficient adjustment accuracy caused by single-dimensional data in traditional control strategies. Among them, the performance model refers to a mathematical expression describing the relationship between system input parameters and output efficiency. Specifically, it can be implemented by training with historical data using a support vector machine or a neural network. By updating the model parameters in real time, it realizes dynamic adaptation to working conditions and overcomes the problem that fixed control strategies cannot respond to environmental mutations. Among them, the adjustment of operation parameters refers to the dynamic correction of device control variables based on the model output. Specifically, it can be implemented through the coordinated operation of a variable pitch actuator, a frequency converter control unit, and a power regulator. By optimizing the closed-loop of pitch angle, speed, and power output, it improves the energy conversion efficiency. Among them, the fault diagnosis module refers to a system for abnormal state recognition and response strategy generation. Specifically, it can be implemented by combining a threshold comparator and a rule engine with a historical fault database. By automatically detecting abnormal fluctuations and executing processing suggestions, it shortens the fault downtime and solves the problem of low efficiency of manual diagnosis.

[0048] The core innovation of this application lies in upgrading the traditional static control strategy to an adaptive intelligent optimization system through a closed-loop control mechanism of real-time data-driven modeling and dynamic parameter adjustment. This solution realizes the coordinated operation of online update of the performance model and automatic execution of fault diagnosis in wind power generation control for the first time. It not only improves the model prediction accuracy through multi-dimensional data fusion, but also shortens the abnormal response time through the fault handling strategy library. At the same time, it dynamically adjusts the device operation status in combination with environmental parameters, forming a complete solution covering efficiency optimization, fault tolerance, and environmental adaptation.

[0049] The working process and principle of this application are as follows: First, key data during the operation of the wind power generation system, including wind speed, wind direction, generator speed, output power, and environmental temperature, are collected in real time through a sensor network. These multi-dimensional data provide comprehensive input for the construction of the system performance model.

[0050] Then, a performance model of the wind power generation system is constructed based on the collected real-time data. This model converts the complex system operation status into a quantifiable mathematical model, reflects the actual working conditions of the system in real time, and provides a basis for subsequent parameter optimization.

[0051] Then, based on the constructed performance model, the system adjusts the operating parameters of the generator in real time. For example, the blade pitch angle can be dynamically adjusted to match the wind speed fluctuations, and the generator speed can be optimized to balance the power generation efficiency and equipment load, thereby improving the energy conversion efficiency and reducing mechanical losses.

[0052] Finally, the system continuously monitors the operating status. When an anomaly is detected, the fault diagnosis module is automatically triggered. This module combines the real-time monitoring data with a preset rule base to quickly generate fault handling suggestions and execute optimization strategies, improving the system's self-recovery ability.

[0053] Through this closed-loop intelligent control process, the system can continuously adapt to dynamic environmental changes, respond promptly to potential faults, and maintain efficient and stable operation.

[0054] As a preferred embodiment, the solution of the present application is specifically implemented as follows:

[0055] First, install multiple sensors in the wind power generation system, including a wind speed sensor, a wind direction sensor, a rotational speed sensor, a power sensor, and a temperature sensor. These sensors are respectively installed at key positions such as the wind turbine, the generator, and the control box. The system collects the data of the above sensors once per second through the data acquisition unit.

[0056] Next, input the collected data into a pre-trained machine learning model. This model can be a performance prediction model constructed based on neural network or support vector machine algorithms. The model inputs include the historical data of the most recent 10 minutes and the current real-time data, and the output is the predicted system performance indicators.

[0057] Then, the control system calculates the optimal operating parameters according to the output of the performance model. For example, when a sudden increase in wind speed is detected, the system calculates the optimal blade pitch angle and generator speed to maximize energy capture. The control instructions are transmitted to the corresponding actuating elements, such as the pitch system and the frequency converter, through the actuator.

[0058] At the same time, the fault diagnosis module continuously monitors the change trends of key parameters. When a certain parameter exceeds the preset threshold or shows abnormal fluctuations, the system triggers the diagnostic program. The diagnostic program identifies possible fault types by comparing the characteristic patterns in the historical fault database. According to the identification results, the system automatically generates handling suggestions, such as adjusting the operating mode or issuing a maintenance alarm.

[0059] Finally, the system records the execution results and diagnostic information in the database for subsequent model optimization and fault prediction. The whole process forms a closed-loop adaptive control system, continuously optimizing the operating efficiency and reliability of the wind power generation system.

[0060] Through the above solutions, the present application can effectively improve the operating efficiency of the wind power generation system. Due to real-time data acquisition and dynamic model construction, the system can quickly adapt to changes in wind speed and environmental conditions, enabling the generator to always operate near the optimal efficiency point. This adaptive control strategy can significantly enhance the energy conversion efficiency and reduce energy losses caused by parameter mismatches.

[0061] In addition, through an intelligent fault diagnosis and handling mechanism, the present application significantly shortens the fault response time. The system can capture abnormal signals at the initial stage of a fault and quickly adopt corresponding optimization strategies. This not only reduces the equipment downtime but also prevents small faults from evolving into major problems, thereby improving the reliability and lifespan of the entire system.

[0062] Finally, by comprehensively considering factors such as environmental temperature, the present application enhances the environmental adaptability of the wind power generation system. The system can automatically adjust operating parameters according to temperature changes, for example, optimizing the cooling strategy in high-temperature environments and adjusting the lubrication system in low-temperature environments. This comprehensive environmental adaptability enables the wind power generation system to operate stably and efficiently under various climate conditions.

[0063] In some of the above solutions of the present application, since the blade angle, gearbox temperature, and grid load parameters are not covered, the model cannot accurately evaluate the matching degree between the mechanical component loss state and grid demand, restricting the comprehensiveness of operating parameter adjustment.

[0064] In response to this, the present application further proposes that the real-time data also includes the blade angle, gearbox temperature, and grid load parameters.

[0065] Among them, the acquisition of the blade angle is achieved through a pitch system encoder. For example, an absolute encoder is used to record the rotation angle of the blade root in real time, and its measurement accuracy can reach ±0.1 degrees. This parameter is input into the control algorithm synchronously with the wind speed data for calculating the optimal pitch angle adjustment amount; the acquisition of the gearbox temperature is achieved through an embedded temperature sensor. For example, a PT100 platinum resistance temperature probe is arranged on the inner wall of the gearbox oil sump, and the monitoring interval can be set to once per second, with the temperature monitoring range covering -20°C to 120°C; the acquisition of the grid load parameters is achieved through a grid-side communication module. For example, the IEC 61850 protocol is used to interact with the substation SCADA system to receive grid frequency, voltage amplitude, and load demand data in real time.

[0066] Specifically, during the operation of the system, the blade angle data is combined with the wind speed and wind direction parameters, and the pitch actuator is controlled through the proportional integral derivative algorithm. For example, when the wind speed suddenly changes, the pitch angle is dynamically adjusted based on the angle deviation value, increasing the blade capture efficiency by 5% - 8%; the gearbox temperature data is combined with the generator speed parameter, and a warning is triggered when the temperature exceeds the set threshold. For example, when the oil temperature reaches 85°C, the auxiliary cooling pump is automatically started to control the temperature within the safe range of 70°C ± 5°C, avoiding increased gear wear; the grid load parameter is combined with the environmental temperature and output power data, and the converter adjustment instruction is generated through the model predictive control algorithm. For example, during the low load period of the grid, the output current harmonic distortion rate is controlled below 3%, and at the same time, the voltage fluctuation range at the grid connection point is maintained within ±2%. Through the synergistic effect of the above parameters, the performance model can optimize the lifespan of mechanical components and grid adaptability simultaneously, reducing the failure rate of the drive system by 15% - 20% and increasing the power utilization rate to over 95%.

[0067] As a preferred embodiment, the solution of the present application is specifically implemented as follows: The real-time data acquisition system includes multiple sensors installed on the wind turbine generator set. These sensors include a wind speed sensor, a wind direction sensor, a generator speed sensor, an output power sensor, an environmental temperature sensor, a blade angle sensor, a gearbox temperature sensor, and a grid load parameter sensor. The blade angle sensor is installed at the root of the blade of the wind turbine generator set for measuring the angle of the blade relative to the horizontal plane. The gearbox temperature sensor is installed on the outer shell of the gearbox for measuring the operating temperature of the gearbox. The grid load parameter sensor is installed at the grid connection point for measuring the real-time load condition of the grid. These sensors collect data once per second and transmit the data to the central control system. The central control system receives this data and uses it to construct the performance model of the wind power system.

[0068] Through the above technical solution, the present application improves the real-time data acquisition dimension and provides more comprehensive input support for the performance model. The blade angle data can be used to accurately control the windward angle of the blade to maximize wind energy utilization. The gearbox temperature data can be used to monitor the mechanical loss state of the drive system and give an early warning of the overheating risk. The grid load parameter can be used to dynamically optimize the generator speed and converter output to ensure the balance between the generated power and the grid load, reduce power waste, and improve grid connection stability.

[0069] In some of the above solutions of the present application, during the process of constructing the performance model, traditional modeling methods may result in insufficient model accuracy due to complex data dimensions and difficult-to-analyze non-linear relationships, and cannot accurately reflect the dynamic characteristics of the system, thus affecting the parameter optimization effect.

[0070] In response to this, the present application further proposes that the performance model is constructed based on machine learning algorithms, including but not limited to support vector machines, neural networks, or decision trees.

[0071] Among them, the support vector machine algorithm transforms the high-dimensional non-linear relationship of wind speed and temperature data into linearly separable features through kernel function mapping. For example, when using the Gaussian kernel function, the kernel width parameter can be set to a value within the range of 0.1 to 1.5 to balance the model complexity and generalization ability; the neural network algorithm adopts a multi-layer perceptron structure with 3 to 5 hidden layers, and the number of nodes in each hidden layer is set to 1.2 to 2 times the dimension of the input layer. The dynamic association between the generator speed and output power is corrected by the backpropagation mechanism, and the learning rate is controlled within the range of 0.001 to 0.01; the decision tree algorithm uses information gain or Gini coefficient as the feature splitting criterion to generate a judgment logic with a branch depth not exceeding 8 layers for discrete decision-making behaviors such as blade angle adjustment and converter control. The minimum number of samples for a single split is set to 20 to 50 data points to avoid overfitting. These three algorithms process data feature extraction, dynamic association modeling, and decision logic generation respectively. Among them, the support vector machine processes short-term data samples during wind speed mutations, the neural network captures the temporal association between the generator speed and power output, and the decision tree generates discrete control instructions for the blade angle. The real-time fusion of the model output is achieved through a parallel computing framework.

[0072] Specifically, when constructing the performance model, the wind speed and temperature data in the real-time data are input into the support vector machine algorithm to generate high-dimensional feature vectors through kernel function mapping; the generator speed and output power data are input into the neural network algorithm, and the dynamic fitting results are output after multi-layer non-linear transformation; the blade angle and converter control signals generate hierarchical optimized decision-making paths through the decision tree algorithm. The output results of the three algorithms are integrated through a weighted fusion module, and the weight coefficients are dynamically adjusted according to environmental conditions. For example, when the wind speed fluctuates violently, the weight ratio of the support vector machine is increased to 60%-70%, and when the load demand is stable, the weight of the decision tree algorithm is enhanced to 50%-55%. The performance model formed in this way can simultaneously analyze the non-linear relationship, temporal association, and discrete control logic in the data, with the modeling accuracy improved by 15%-22% compared with a single algorithm, and the parameter optimization response time under conditions such as wind speed mutations and load jumps is shortened to within 0.8 seconds. Through the combined application of machine learning algorithms, the complete modeling requirements from data feature extraction to control strategy generation are covered, and the problem of insufficient model generalization ability caused by complex data dimensions in traditional methods is solved.

[0073] As a preferred embodiment, the solution of the present application is specifically implemented as follows: The performance model is constructed based on machine learning algorithms, including the combined application of three algorithms: support vector machine, neural network, and decision tree. In the actual implementation process, the system first collects the operation data of the wind turbine generator set, including parameters such as wind speed, wind direction, generator speed, output power, and ambient temperature. These data are preprocessed and then input into the machine learning model. Specifically, the support vector machine algorithm is used to process the non-linear mapping relationship between wind speed and output power. The original data is mapped to a high-dimensional feature space through a radial basis kernel function to establish a wind speed-power curve model. The neural network algorithm adopts a three-layer structure. The input layer receives the processed operation parameters, the hidden layer contains multiple neuron nodes for feature extraction and non-linear transformation, and the output layer generates the optimal generator speed and blade angle adjustment values. The decision tree algorithm is used to construct a logical judgment framework for system fault diagnosis and operation mode selection, and classifies and makes decisions on operation strategies under different working conditions through the information gain criterion. The output results of these three algorithms are comprehensively evaluated through an ensemble learning method to form the final control decision signal. The system updates the model parameters with newly collected data at regular time intervals to ensure that the model can adapt to the dynamic changes of the wind farm environment and equipment status.

[0074] Through the above technical solution, the present application solves the problem of insufficient accuracy of traditional modeling methods in processing complex data of wind power generation systems. Machine learning algorithms can automatically identify and learn the complex non-linear relationships between environmental factors such as wind speed and temperature and power generation performance, without the need to manually establish a mathematical model. The support vector machine algorithm effectively processes high-dimensional feature data through kernel function technology, improving the generalization ability of the model in the case of small samples; the neural network algorithm captures the deep correlations between system parameters through a multi-layer structure, enhancing the adaptability of the model to time-varying characteristics; the decision tree algorithm provides an interpretable control logic framework, facilitating engineers to understand and adjust the system operation. This method of constructing a performance model based on machine learning significantly improves the accuracy and real-time performance of parameter optimization of wind power generation systems, enabling the system to quickly adjust operation parameters according to changes in environmental conditions and maintain the best power generation efficiency. At the same time, the self-learning characteristics of the model enable the system to continuously improve its own performance with the accumulation of operation data, reducing maintenance costs and the need for manual intervention.

[0075] In some of the above solutions of the present application, it is proposed to adjust the generator operation parameters in real time through the performance model to optimize the system power generation efficiency. However, in the specific implementation process, due to the lack of clear specific dimensions for adjusting the operation parameters, the adjustment strategy may lack pertinence. For example, it is impossible to accurately adapt to the dynamic demand of the blade angle for sudden changes in wind speed, difficult to match the real-time coupling relationship between the generator speed and load changes, and fail to synchronously optimize the impact of the converter output power on the power quality.

[0076] In this regard, the present application further proposes the adjustment of operating parameters, including: dynamically adjusting the blade pitch angle to adapt to wind speed changes; optimizing the generator speed to match the current wind speed and load demand; and adjusting the output power of the converter to improve the power quality.

[0077] Among them, the dynamic adjustment of the blade pitch angle is implemented as follows: the pitch angle adjustment range is set to 0 - 30 degrees, and the angle adjustment is achieved by driving the blade rotation mechanism with a pitch motor, where the adjustment response time is controlled within 200 milliseconds. The optimization of the generator speed is implemented as follows: based on the tip speed ratio corresponding to the current wind speed and the load power demand, the optimal speed range is dynamically matched through a proportional-integral-derivative control algorithm. For example, when the wind speed is 12 m / s and the grid load is 80% of the rated power, the speed is adjusted to the range of 1800 - 2000 rpm. The adjustment of the converter output power is implemented as follows: the inverter switching frequency is adjusted through pulse width modulation technology to suppress the total harmonic distortion rate of the output voltage below 3%, and at the same time, the grid connection frequency error is made less than 0.1 Hz through a phase-locked loop technology.

[0078] Specifically, in the scenario of sudden wind speed change, when the wind speed change rate exceeds 2 m / s 2 the pitch angle adjustment command is preferentially executed, and the pitch angle is quickly increased by 5 - 10 degrees to reduce the aerodynamic load on the blade. At this time, the generator speed is synchronously lowered by 100 - 150 rpm to avoid overload, and at the same time, the modulation ratio of the converter is increased to 0.95 to maintain the stability of the output voltage. In the steady-state operation stage, the pitch angle is maintained in the optimal angle of attack range, the generator speed is adjusted step by step in increments of 50 rpm according to the load demand, and the converter dynamically adjusts the switching timing to compensate for the phase difference by real-time detecting the phase deviation of the grid voltage. In the case of sudden wind speed drop, when it is detected that the speed deviates from the set value by more than 5% for 10 seconds, the active power command of the converter is automatically lowered by 20%, and at the same time, the pitch angle is increased by 3 - 5 degrees to maintain the dynamic balance between the mechanical energy input and the electrical energy output.

[0079] As a preferred embodiment, the solution of the present application is specifically implemented as follows:

[0080] The adjustment of operating parameters includes dynamically adjusting the blade pitch angle, optimizing the generator speed, and adjusting the output power of the converter. The dynamic adjustment of the blade pitch angle can be achieved through a hydraulic drive system, which consists of a pressure sensor, a solenoid valve, and a hydraulic cylinder. When the wind speed changes, the pressure sensor detects the change in the force on the blade, and then triggers the solenoid valve to open, driving the hydraulic cylinder to push the blade to rotate, realizing the real-time adjustment of the pitch angle. The optimization of the generator speed can adopt variable speed constant frequency technology, which is achieved through a doubly-fed induction generator and a full-power converter. The converter calculates the optimal speed according to the wind speed and load demand, and controls the generator speed by adjusting the frequency of the rotor current, so that it always operates at the optimal efficiency point. The adjustment of the converter output power can be achieved through the PWM control strategy. According to the fluctuations of the grid voltage and frequency, the switching frequency and duty cycle of the inverter are adjusted in real time, so as to optimize the output voltage waveform and reduce the harmonic content.

[0081] Through the above technical solutions, the present application realizes the multi-dimensional collaborative adjustment of the operating parameters of the wind power generation system. Dynamically adjusting the blade pitch angle can effectively cope with sudden changes in wind speed, maintain the optimal angle of attack, and reduce the fluctuation of aerodynamic loads; optimizing the generator speed can achieve real-time matching of wind speed and load demand, and improve the energy conversion efficiency; adjusting the converter output power can improve the power quality, reduce harmonic interference, and improve the voltage stability. This multi-level collaborative control mechanism significantly improves the system's adaptability to wind speed changes, enhances the coupling relationship between the generator and the load, and at the same time improves the grid-connected power quality, thus comprehensively improving the operating efficiency and stability of the wind power generation system.

[0082] In some of the above solutions of the present application, a fault diagnosis module is proposed to detect system anomalies in real time and generate treatment suggestions. However, in this process, if only relying on a single threshold judgment or manual analysis, it may lead to abnormal fluctuations not being captured in time, insufficient accuracy in identifying fault types, and inability to quickly execute the adjustment of the operating mode to prevent the expansion of faults. There are still problems of response lag and unsystematic treatment measures.

[0083] In response to this, the present application further proposes that the fault diagnosis module includes: real-time monitoring of abnormal fluctuations of key parameters of the system; generating fault types and treatment suggestions based on historical fault data and a preset rule library; automatically adjusting the system operating mode to avoid further damage.

[0084] Among them, for the abnormal fluctuations of the key parameters of the real-time monitoring system, the sliding window algorithm can be used to continuously calculate the parameter change rate. For example, the vibration amplitude of the gearbox and the change of the winding temperature are monitored with a period of 0.5 seconds. The detection of abnormal fluctuations can set a dynamic threshold range. For example, when the temperature change rate exceeds 2 times the standard deviation of the historical average in the same period, a warning is triggered. When generating treatment suggestions based on historical fault data and a preset rule base, the rule base can include multi-level judgment logic. For example, first match the fault feature waveform library, and then analyze the fault priority in combination with the decision tree model. The historical data can include the fault case library of the same type of unit in the past three years, where the fault mode and the treatment plan are associated through a Bayesian network. The automatic adjustment of the system operation mode can be specifically manifested as reducing the load operation or switching to the standby subsystem. For example, when the temperature of the generator bearing is abnormal, the speed is reduced to the range of 80%-90% of the rated value, and at the same time, the auxiliary cooling device is started.

[0085] Specifically, when it is detected that the oil temperature of the gearbox rises at a rate of 3°C per minute, by comparing the fault cases under similar working conditions in the historical database, the probability of lubrication failure is identified to reach 85%. At this time, the coping strategy preset in the rule base is activated, and the control unit immediately gradually reduces the generator output power from 2MW to 1.5MW and switches the lubrication system to the standby oil circuit for circulation. This process completes the whole process of parameter acquisition, diagnostic decision-making, and execution actions within 30 seconds, shortening the response time by 80% compared with the traditional manual processing. Through the linkage mechanism of dynamic monitoring and the rule base, the problem of misjudgment of the fixed threshold judgment under low wind speed conditions is avoided, and the fault recognition accuracy rate is increased to more than 92%. The operation mode adjustment strategy is combined with the performance model. For example, the blade pitch angle is synchronously optimized during the load reduction operation stage, which not only prevents the equipment from overheating and damage but also maintains the power generation efficiency in the range of 90%-95% of the optimal state. This closed-loop control mechanism realizes the seamless connection between fault diagnosis and operation optimization through the interaction of real-time data flow and historical knowledge base.

[0086] As a preferred embodiment, the solution of the present application is specifically implemented as follows:

[0087] The fault diagnosis module includes a real-time monitoring unit, a fault analysis unit, and an automatic adjustment unit. The real-time monitoring unit collects key parameters of the wind power generation system, such as wind speed, wind direction, generator speed, output power, ambient temperature, etc., and establishes a multi-dimensional parameter monitoring matrix. This matrix can be a 10×10 two-dimensional array, where each row represents a time point and each column represents a parameter. The fault analysis unit uses pre-trained machine learning models, such as support vector machines or random forest algorithms, to analyze the monitoring matrix. The machine learning model is trained based on historical fault data and can identify 20 common fault types. When an anomaly is detected, the fault analysis unit will call corresponding handling suggestions from a preset rule library. The rule library can contain more than 100 if-then rules, covering various possible fault scenarios. The automatic adjustment unit executes corresponding system adjustment instructions according to the fault type and handling suggestions. For example, when it detects that the generator is overheating, the automatic adjustment unit may reduce the generator load, increase the power of the cooling system, or switch to a standby generator.

[0088] Through the above technical solutions, this application realizes the rapid diagnosis and automatic processing of faults in the wind power generation system. The real-time monitoring unit can promptly capture the subtle changes in system parameters, avoiding the missed detection problems that may be caused by traditional fixed-threshold judgment methods. The fault analysis unit uses machine learning models and a preset rule library to improve the accuracy of fault type identification and give targeted handling suggestions. The automatic adjustment unit can quickly execute optimization strategies without manual intervention, preventing the further expansion of faults. This intelligent fault diagnosis and processing mechanism significantly improves the reliability and operating efficiency of the wind power generation system, reducing the downtime and maintenance costs caused by faults.

[0089] In some of the above solutions of this application, due to the randomness and volatility of wind speed having a time lag, relying solely on real-time data to adjust parameters will result in the system response speed being insufficient to cope with the upcoming wind speed changes, causing the output power fluctuations to intensify, making it difficult to optimize the operating parameters in advance to match the future wind speed trend, and thus affecting the power quality and grid stability.

[0090] In response to this, this application further proposes a power prediction module, which predicts the wind speed and power output in the next period based on historical data and real-time data, and adjusts the system operating parameters in advance according to the prediction results to optimize the power generation efficiency.

[0091] Among them, the power prediction module can integrate the Kalman filtering algorithm or the time series analysis method, and the prediction time window is set in the range of 5 - 15 minutes. The historical data includes the wind speed fluctuation pattern and the corresponding power generation curve in the past 72 hours, and the real-time data includes the current wind speed gradient, the wind direction deflection angle, and the generator speed change rate. The prediction model generates a prediction result with trend extrapolation ability by correlating the periodic characteristics of the historical data with the transient fluctuation amount of the real-time data. In the parameter adjustment process, the generator speed control instruction is sent to the converter 30 seconds in advance, and the blade pitch angle adjustment instruction is pre-judgmentally corrected based on the rising rate of the predicted wind speed. This module forms a data closed-loop with the real-time performance model. The predicted power curve is used as the input boundary condition of the performance model, and the optimized operating parameters are executed after being verified by the performance model.

[0092] Specifically, the power prediction module first collects the historical wind speed sequence stored in the database and synchronously receives the sampling value of the wind speed sensor at the current moment. Through the time series analysis method, the correlation between the standard deviation of the hourly wind speed fluctuation in the historical data and the real-time wind speed change rate is analyzed, and a prediction model including the lag effect is established. When it is predicted that the wind speed will increase by more than 10% in the next 5 minutes, the control unit raises the preset value of the generator speed to 80% of the corresponding interval 30 seconds before the actual increase in wind speed. At the same time, according to the duration of the predicted wind speed, the optimal adjustment amplitude of the blade pitch angle is dynamically calculated, so that the blades have completed 60% of the adjustment amount before the wind speed changes. Through this lead control, when the wind speed actually reaches the predicted value, the generator is already in the optimal speed interval, and the power conversion efficiency of the converter is increased by more than 15%. The prediction module updates the data input every 30 seconds to ensure the synchronization of the prediction result with the real-time working condition, and at the same time retains the periodic characteristics in the historical data for trend correction. When the deviation between the real-time wind speed and the predicted value exceeds the preset threshold, a dynamic weight adjustment mechanism is started to reduce the weight of the historical data and increase the update frequency of the real-time data to ensure the adaptability of the prediction model.

[0093] As a preferred embodiment, the solution of the present application is specifically implemented as follows:

[0094] In the method for optimizing the operation of a wind power generation system based on intelligent control, in addition to collecting real-time data during the operation of the wind power generation system, constructing a performance model, adjusting operation parameters in real time, and fault diagnosis, it also includes a power prediction module. This power prediction module combines historical data and real-time data to predict and analyze future wind conditions, and then optimizes system parameters in advance.

[0095] Specifically, the power prediction module first extracts historical data such as wind speed, wind direction, and output power in the past 72 hours from the system database, and at the same time obtains data such as the real-time wind speed, wind direction, and barometric pressure change trends within the current 10 minutes. The system inputs these data into a pre-trained prediction algorithm, which can be a recurrent neural network or a long short-term memory network model based on deep learning. The prediction algorithm generates prediction curves of wind speed and power output for the next 30 minutes to 4 hours by analyzing the periodic patterns, seasonal variations in the historical data, and the immediate trends reflected in the real-time data.

[0096] Furthermore, the power prediction module calculates the optimal operating parameter combination according to the prediction results. For example, when it is predicted that the wind speed will increase from 8 m / s to 12 m / s in 15 minutes, the system will calculate in advance the optimal blade pitch angle and generator speed at this wind speed and start gradually adjusting these parameters 10 minutes before the actual change in wind speed. During the specific adjustment process, the system will adjust the parameters in a smooth transition manner according to the predicted wind speed change rate to avoid an increase in mechanical stress caused by sudden changes.

[0097] Thus, when it is predicted that the wind speed is about to decrease, the system can reduce the generator speed in advance to avoid energy loss caused by too high a speed; when it is predicted that the wind direction changes, the system can adjust the yaw angle in advance so that the wind turbine always faces the incoming wind direction. This predictive pre-adjustment enables the wind power generation system to always operate close to the optimal efficiency point.

[0098] As a preferred implementation, the power prediction module can also interact with the power grid dispatching system to adjust the power generation plan according to the power grid load prediction. For example, when it is predicted that the power grid is in a peak electricity consumption period and the wind resources are sufficient, the system will adjust to the maximum power output mode in advance; while when the power grid load is low and the wind speed fluctuates greatly, a stable output mode may be selected, sacrificing some power generation efficiency in exchange for more stable power quality.

[0099] Through the above technical solutions, this application solves the problem of lag caused by relying only on real-time data adjustment in the wind power generation system. By introducing the power prediction module, the system can sense the change trend of wind conditions in advance and realize pre-adjustment of parameters, significantly reducing the power fluctuations caused by sudden changes in wind speed. Specifically, in the case of large wind speed changes, the traditional system may have power fluctuations of more than 20%, while with predictive adjustment, the power fluctuations can be effectively controlled. At the same time, the predictive adjustment reduces the frequent operation of mechanical components, reduces equipment wear, and extends the life of key components. In addition, since the system always operates close to the optimal efficiency point, the overall power generation efficiency is improved, the power quality is more stable, which is beneficial to the safe operation of the power grid. In practical applications, this predictive adjustment strategy is particularly important for large wind farms, which can significantly improve the grid connection friendliness and economy of wind farms.

[0100] In some of the above solutions of the present application, a power prediction module is proposed to predict future wind speed and power output based on historical data and real-time data. However, in this process, the selection of the prediction algorithm directly affects the prediction accuracy and real-time performance. Traditional prediction methods may not be able to effectively handle the randomness of wind speed and data noise, resulting in a large deviation in the prediction results, thereby affecting the accuracy of system operation parameter adjustment and the effect of power generation efficiency optimization.

[0101] In response to this, the present application further proposes that the power prediction module adopts the Kalman filter algorithm or the time series analysis method.

[0102] Among them, the Kalman filter algorithm is configured to perform real-time state estimation on the dynamic system through recursive calculation, and its state vector can include the covariance matrix of wind speed, power deviation and system noise. For example, a four-dimensional state vector is adopted to correspond to the wind speed component, power output, noise variance and system error respectively; the time series analysis method is implemented to establish an autoregressive integrated moving average model based on historical data, where the time window length of the historical data can be set from 24 hours to 72 hours to capture daily and weekly periodic characteristics. The input data of both algorithms are associated with the real-time collected wind speed, generator speed and output power, and the historical data sequence is dynamically updated through the sliding window mechanism. In the prediction stage of the Kalman filter, the prior estimated value is calculated through the state transition matrix, and the estimation error is corrected by combining the real-time wind speed measurement value in the update stage; the time series analysis eliminates the data non-stationarity through the difference operation and determines the model order by using the partial autocorrelation function. The prediction results output by both algorithms are transmitted to the performance model as input parameters for generator speed adjustment and blade pitch angle optimization.

[0103] Specifically, after the wind speed data is collected, the Kalman filter algorithm first initializes the system state vector and the error covariance matrix. Subsequently, it predicts the wind speed and power output at the next moment through the state equation, and calculates the Kalman gain in combination with the real-time measurement values to update the state estimation. For example, when it is detected that the sudden change in wind speed causes the measurement noise variance to exceed the threshold, the weight coefficient of the covariance matrix is dynamically adjusted to 0.6 to 0.8 to balance the prediction stability. The time series analysis method performs seasonal decomposition on the historical power data in the data preprocessing stage. After extracting the trend term and the periodic term, it fits the model parameters by the least squares method, and the lag order can be set to 12 hours to match the semi-daily cycle characteristics of the wind speed change. After the prediction result is generated, the response time of the blade pitch angle control command is shortened to within 5 seconds, so that the phase difference between the generator speed adjustment and the predicted wind speed change is controlled within the range of ±10%. Thus, the amplitude of the power output fluctuation is suppressed below 3% of the rated power, and at the same time, the calculation delay of the prediction algorithm is controlled within 200 milliseconds, ensuring that the power generation efficiency of the system is increased by 8% to 12% under the random fluctuation of the wind speed. As a preferred embodiment, the solution of the present application is specifically implemented as follows: The power prediction module can use the Kalman filter algorithm to predict the wind speed and power output. The Kalman filter algorithm performs state estimation on the dynamic system through recursion, and can effectively handle the noise interference in the wind speed data. In specific implementation, first establish the wind speed state equation and the observation equation, where the state equation describes the change law of the wind speed over time, and the observation equation describes the relationship between the actual measurement value and the true wind speed. Then, through the alternating execution of the prediction step and the update step, the wind speed estimation value is continuously optimized. In the prediction step, based on the state estimation at the previous moment and the system model, predict the wind speed state at the current moment; in the update step, correct the prediction result in combination with the actual observation value to obtain the optimal estimation. In this way, the Kalman filter algorithm can track the wind speed change in real time and provide accurate short-term prediction results.

[0104] Furthermore, the power prediction module can also combine the time series analysis method, such as the autoregressive moving average model (ARMA), to perform modeling analysis on the historical wind speed and power data. The ARMA model captures the periodic and trend characteristics of the wind speed and power changes by analyzing the autocorrelation and moving average characteristics of the data. In specific implementation, first perform the stationarity test and differencing process on the historical data, then determine the model order, estimate the model parameters, and finally perform model diagnosis and prediction. This method can provide relatively accurate medium- and long-term power predictions based on long-term historical data.

[0105] Thus, by combining the short-term prediction advantage of the Kalman filter algorithm and the long-term prediction ability of the time series analysis method, the power prediction module can meet the requirements of real-time control and long-term planning at the same time, providing a reliable basis for the dynamic adjustment of the system operation parameters.

[0106] Through the above technical solutions, the present application can effectively improve the accuracy and real-time performance of wind speed and power output prediction. The Kalman filter algorithm can quickly respond to the random fluctuations of wind speed and reduce prediction errors, while the time series analysis method can capture long-term change trends and enhance the adaptability of the prediction model. This method of combining short-term and long-term predictions enables the power prediction module to provide a more reliable basis for adjusting the system operation parameters, thereby optimizing the power generation efficiency and reducing the impact of power fluctuations on the power grid. At the same time, accurate power prediction also helps to improve the economy of the system and provides support for power grid dispatching and energy management.

[0107] In some of the above solutions of the present application, a performance model and a fault diagnosis module are proposed to improve the operation efficiency and fault handling ability. However, during this process, there is a lack of remote centralized monitoring and real-time optimization adjustment capabilities for system operation data, resulting in the monitoring center being unable to generate optimization instructions based on global data and making it difficult to achieve coordinated management and dynamic response of cross-regional systems.

[0108] In response to this, the present application further proposes a remote monitoring module: real-time transmission of system operation data to the monitoring center through a wireless communication network; the monitoring center generates optimization instructions based on the data and sends them to the wind power generation system.

[0109] Among them, the wireless communication network can adopt LoRaWAN, NB-IoT or 5G technology to achieve low-latency data transmission, and the transmission interval can be set from once per second to once every 5 seconds to ensure data real-time performance. The monitoring center is configured with a distributed database for storing the historical operation data of multiple wind turbines, and the data storage capacity can be designed to be in the TB level to meet the long-term storage requirements. The optimization instruction generation algorithm can integrate a dynamic optimization model based on deep reinforcement learning. The model input includes parameters such as generator speed, output power, and ambient temperature, and the output is the adjustment amplitude of the blade pitch angle or the set value of the generator speed. The adjustment amplitude accuracy can be controlled within the range of ±0.5°. The data feedback mechanism adopts a two-way communication protocol, and the instruction transmission delay requirement is less than 500 milliseconds.

[0110] Specifically, after the sensors built into the wind turbine generator collect the data of the generator speed and output power, they are encapsulated into structured data packets through the wireless communication module. The data packets are encapsulated in JSON format and appended with timestamp information. After receiving the data, the monitoring center performs data cleaning, eliminates outliers, and then inputs the data into a pre-trained performance optimization model. This model calculates the combination of operating parameters that need to be adjusted by comparing the deviation of the current operating parameters from the historical optimal parameter set. The generated optimization instructions are encrypted and then transmitted back to the specified wind turbine generator through a dedicated communication channel. The execution result of the instructions is confirmed through a check code. During this process, the local control system retains the final execution authority for parameter adjustment. When the instructions from the monitoring center conflict with the local safety threshold, the local protection strategy is preferentially executed. Through the continuous operation of this module, coordinated power regulation of wind turbine generators across regions can be achieved. For example, when the load of a certain regional power grid suddenly increases, the monitoring center can synchronously increase the output power of 5 wind turbines within a range of 3 kilometers adjacent to it. The power adjustment step size is set to 2%-5% of the rated power, and the power balance is completed within 30 seconds.

[0111] As a preferred embodiment, the solution of the present application is specifically implemented as follows: A remote monitoring module is deployed in the wind power system, and a wireless communication unit is configured to establish a data channel with the monitoring center. The system operation data includes generator speed, output power, and gearbox temperature parameters, and data packet transmission is carried out at a cycle of 5 seconds through the LTE-M communication protocol. The monitoring center configures a data fusion platform, performs correlation analysis after receiving the real-time data streams of multiple wind turbines, and generates coordinated optimization instructions when it is detected that the output power difference between adjacent units exceeds the set threshold. Specifically, when implemented, the optimization instructions include pitch rate correction coefficients and power limit parameters, and are sent to the controller of the target wind turbine through the MQTT protocol to achieve balanced control of the output of multiple units. The data channel uses the AES-256 encryption algorithm to ensure transmission security, and enables a local caching mechanism to temporarily store operation data when the communication is interrupted.

[0112] Through the above technical solutions, the present application realizes centralized monitoring and coordinated optimization of the operating states of wind turbine generators across regions, and solves the problem of the lack of global optimization caused by data islands in traditional systems. Through the real-time two-way interaction of the wireless communication network, the monitoring center can generate system-level optimization strategies based on multi-source heterogeneous data analysis, enabling the adjustment of the operating parameters of a single wind turbine generator to dynamically match the load demand of the regional power grid, forming a closed-loop control from data collection to instruction feedback. This solution effectively improves the overall response speed and coordinated control accuracy of the distributed wind power system, and at the same time ensures the reliability and security of remote monitoring through the encryption transmission and disconnection resumption mechanism.

[0113] In some of the above solutions of this application, the existing solutions fail to perform coordinated scheduling according to the real-time demand of the power grid load and the state of the wind power generation system itself, resulting in the equipment still running at a high load continuously during the low-load period of the power grid, accelerating equipment wear, and at the same time being unable to effectively improve the power supply capacity during the peak period of the power grid.

[0114] In response to this, this application further proposes a technical solution including an intelligent scheduling module: dynamically adjusting the power generation according to the power grid load demand and the real-time state of the wind power generation system; giving priority to outputting the maximum power during the peak period of the power grid load, and reducing the output during the low-load period to extend the equipment life.

[0115] Among them, the power grid load demand obtains real-time load data by accessing the power grid dispatching system, such as the total load value of the regional power grid updated every 15 minutes. The real-time state of the wind power generation system includes the generator speed, the gearbox temperature, and the converter output current, where the gearbox temperature is collected by an embedded temperature sensor at a frequency of once per second. Dynamically adjusting the power generation is achieved through a power controller, which is built with a priority algorithm. When the power grid load exceeds a preset threshold (such as 85% of the total load of the regional power grid), the generator speed limit is automatically lifted and the maximum power tracking mode of the converter is enabled; when the power grid load is lower than the set threshold (such as 40% of the total load of the regional power grid), the generator speed is controlled to operate within the range of 60%-80% of the rated speed.

[0116] Specifically, during the implementation of giving priority to outputting the maximum power during the peak period of the power grid load, the power controller receives the load demand signal sent by the power grid dispatching system and simultaneously collects the temperature rise data of the generator winding. When it is detected that the power grid load is at the peak and the generator temperature does not exceed the safety threshold, the controller adjusts the converter switching frequency to make the output power reach 105% of the nameplate nominal value. At this time, the lubrication system of the generator bearing synchronously increases the oil supply pressure to 0.25 MPa - 0.35 MPa to reduce the friction loss. During the low-load period of the power grid, the controller dynamically adjusts the power output curve according to the feedback signal of the gearbox vibration sensor. For example, when the axial vibration amplitude of the gearbox exceeds 50 μm, the power generation is gradually reduced to 70% of the rated power, and at the same time, the pitch system is triggered to increase the blade angle by 2° - 5° to reduce the aerodynamic load. Through the differential control strategy in different time periods, the cumulative wear of the gearbox during the low-load period is reduced by about 30% compared with the traditional operation mode, and the power supply capacity of the power grid during the peak period is increased by 12% - 15%. This solution realizes the balance between equipment durability and power generation efficiency while ensuring the stability of the power grid through power distribution optimization in the time dimension.

[0117] As a preferred embodiment, the solution of the present application is specifically implemented as follows: An intelligent scheduling module is integrated in the main control unit of the wind power generation system. This module obtains real-time load data through the power grid monitoring interface and simultaneously receives the operation status signals from the nacelle sensor group. When it is detected that the power grid load exceeds the preset threshold and lasts for 15 minutes, it is determined as the peak period. At this time, the scheduling algorithm controls the pitch system to adjust the blade pitch angle to the optimal aerodynamic efficiency angle, simultaneously increases the speed of the doubly-fed generator to 102% of the rated speed, and increases the output power to the maximum output value under the current environmental conditions through the converter control system. When the power grid load is lower than 70% of the reference load and lasts for more than 30 minutes, the system enters the low-load operation mode. At this time, the generator speed is limited within 85% of the rated speed, and the pitch system gradually increases the pitch angle at a rate of 0.5 degrees per minute to reduce the wind energy capture amount. At the same time, the gearbox lubricating oil circulation system is activated to enter the energy-saving mode.

[0118] Through the above technical solution, the present application realizes the dynamic matching between the operation mode of the power generation equipment and the power grid supply and demand status. During the peak load period of the power grid, the power supply pressure of the power grid is effectively relieved by maximizing the power output. While during the low-load period of the power grid, the wear of the gearbox bearings and the temperature rise of the generator windings are significantly reduced by reducing the operation intensity of mechanical components. Experimental data shows that the fatigue life of the key components of the gearbox can be extended by about 23%, and at the same time, the power grid peak shaving capacity is increased by 19.7%. This two-way adjustment mechanism realizes the coordinated optimization of equipment durability and power generation economy while ensuring the power supply reliability.

[0119] In some of the above solutions of the present application, the existing solutions fail to perform coordinated scheduling according to the real-time demand of the power grid load and the status of the wind power generation system itself, resulting in the equipment still running at a high load continuously during the low-load period of the power grid, accelerating the equipment loss, and at the same time being unable to effectively improve the power supply capacity during the peak period of the power grid.

[0120] In response to this, the present application further proposes a technical solution including an intelligent scheduling module: Dynamically adjust the power generation power according to the power grid load demand and the real-time status of the wind power generation system; Give priority to outputting the maximum power during the peak load period of the power grid, and reduce the output during the low-load period to extend the equipment life.

[0121] Among them, the real-time load data of the grid load demand is obtained by accessing the grid dispatching system, such as the total load value of the regional power grid updated every 15 minutes. The real-time status of the wind power generation system includes the generator speed, the gearbox temperature, and the converter output current, where the gearbox temperature is collected by an embedded temperature sensor at a frequency of once per second. The dynamic adjustment of the power generation is achieved through a power controller, which is built with a priority algorithm. When the grid load exceeds a preset threshold (such as 85% of the total load of the regional power grid), the generator speed limit is automatically lifted and the maximum power tracking mode of the converter is enabled; when the grid load is lower than the set threshold (such as 40% of the total load of the regional power grid), the generator speed is controlled to operate within the range of 60%-80% of the rated speed.

[0122] Specifically, during the implementation of giving priority to outputting the maximum power during the peak period of the grid load, the power controller receives the load demand signal sent by the grid dispatching system and simultaneously collects the temperature rise data of the generator winding. When it is detected that the grid load is at a peak and the generator temperature does not exceed the safety threshold, the controller adjusts the switching frequency of the converter to make the output power reach 105% of the nameplate nominal value. At this time, the lubrication system of the generator bearing synchronously increases the oil supply pressure to 0.25 MPa - 0.35 MPa to reduce the friction loss. During the low valley period of the grid load, the controller dynamically adjusts the power output curve according to the feedback signal of the gearbox vibration sensor. For example, when the axial vibration amplitude of the gearbox exceeds 50 μm, the power generation is gradually reduced to 70% of the rated power, and at the same time, the pitch system is triggered to increase the blade angle by 2° - 5° to reduce the aerodynamic load. Through the differential control strategy in different time periods, the cumulative wear of the gearbox during the low valley period is reduced by about 30% compared with the traditional operation mode, while the power supply capacity of the power grid during the peak period is increased by 12% - 15%. This solution realizes the balance between equipment durability and power generation efficiency while ensuring the stability of the power grid through the optimization of power distribution in the time dimension.

[0123] As a preferred embodiment, the solution of the present application is specifically implemented as follows: An intelligent scheduling module is integrated in the main control unit of the wind power generation system. This module obtains real-time load data through the power grid monitoring interface and simultaneously receives the operation status signals from the nacelle sensor group. When it is detected that the power grid load exceeds the preset threshold and lasts for 15 minutes, it is determined as the peak period. At this time, the scheduling algorithm controls the pitch system to adjust the blade pitch angle to the optimal aerodynamic efficiency angle, simultaneously increases the speed of the doubly-fed generator to 102% of the rated speed, and increases the output power to the maximum output value under the current environmental conditions through the converter control system. When the power grid load is lower than 70% of the reference load and lasts for more than 30 minutes, the system enters the low valley operation mode. At this time, the generator speed is restricted within the range of 85% of the rated speed, and the pitch system gradually increases the pitch angle at a rate of 0.5 degrees per minute to reduce the wind energy capture amount. At the same time, the gearbox lubricating oil circulation system is activated to enter the energy-saving mode.

[0124] Through the above technical solution, the present application realizes the dynamic matching between the operation mode of the power generation equipment and the power supply and demand state of the power grid. During the peak load period, the power supply pressure of the power grid is effectively relieved by maximizing the power output. While during the low valley load period, the wear of the gearbox bearings and the temperature rise of the generator windings are significantly reduced by reducing the operation intensity of mechanical components. Experimental data shows that the fatigue life of the key components of the gearbox can be extended by about 23%, and at the same time, the power grid peak shaving capacity is increased by 19.7%. This two-way adjustment mechanism realizes the coordinated optimization of equipment durability and power generation economy while ensuring the power supply reliability.

[0125] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other.

[0126] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An operation optimization method for a wind power generation system based on intelligent control, characterized in that It includes the following steps: S1. Collect real-time data during the operation of the wind power generation system, including wind speed, wind direction, generator speed, output power, and ambient temperature; S2. Build a performance model of the wind power generation system based on the real-time data; S3. Based on the performance model, adjust the operating parameters of the generator in real time to optimize the system power generation efficiency; S4. When a system anomaly is detected, automatically trigger the fault diagnosis module to generate fault handling suggestions and execute optimization strategies.

2. The operation optimization method of a wind power generation system based on intelligent control according to claim 1, characterized in that The real-time data also includes blade angle, gearbox temperature, and grid load parameters.

3. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, The performance model is built based on machine learning algorithms, and the algorithms include but are not limited to support vector machines, neural networks, or decision trees.

4. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, The adjustment of the operating parameters includes: Dynamically adjust the blade pitch angle to adapt to wind speed changes; Optimize the generator speed to match the current wind speed and load demand; Adjust the output power of the converter to improve power quality.

5. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, The fault diagnosis module includes: Real-time monitor the abnormal fluctuations of key system parameters; Generate fault types and handling suggestions based on historical fault data and a preset rule library; Automatically adjust the system operating mode to avoid further damage.

6. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, It also includes a power prediction module: Based on historical data and real-time data, predict the wind speed and power output in the next period of time; According to the prediction results, adjust the system operating parameters in advance to optimize the power generation efficiency.

7. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, The power prediction module adopts the Kalman filtering algorithm or time series analysis method.

8. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, It also includes a remote monitoring module: Transmit the system operating data to the monitoring center in real time through a wireless communication network; The monitoring center generates optimization instructions based on the data and sends them to the wind power generation system.

9. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, It also includes an intelligent scheduling module: Dynamically adjust the power generation according to the grid load demand and the real-time status of the wind power generation system; Give priority to outputting the maximum power during peak grid load periods and reduce the output during low load periods to extend the equipment life.

10. A method for optimizing the operation of a wind power generation system based on intelligent control according to claim 1, characterized in that, It also includes an environmental adaptability module: Automatically adjust the operating parameters of the cooling system and lubrication system according to changes in environmental temperature, humidity, and air pressure; Trigger a protection mechanism under extreme weather conditions to ensure the safe operation of the system.

Images