Self-adaptive MPPT control method and system

Abstract

The invention discloses a self-adaptive MPPT control method and system, and the method comprises the following steps: S1, data collection and processing: collecting operation wind speed data in real time, and carrying out the filtering processing; s2, wind speed feature extraction: calculating a real-time wind speed fluctuation amplitude based on the processed real-time wind speed data; multi-mode cooperative control can adapt to different wind conditions and power grid states, efficient power generation is achieved when a wind field is stable and a power grid is normal, power difference is rapidly compensated through energy storage when the wind conditions fluctuate or the power grid is out of tolerance, load power supply continuity is guaranteed when the power grid is interrupted or the wind speed is low, meanwhile, excessive charging and discharging of equipment are avoided through energy storage SOC constraint, and the energy storage life is prolonged; and the comprehensive requirements on power generation efficiency, power supply stability and equipment protection in small and medium-sized distributed scenes are integrally met.

Description

Technical Field

[0001] This application relates to the technical field of wind power generation, and in particular to an adaptive MPPT control method and system. Background Technology

[0002] Small and medium-sized distributed wind turbine generators are mostly used in distributed wind farms. Their operating environment is characterized by highly variable wind conditions, with wind speeds easily fluctuating frequently in the short term due to terrain and climate. They also need to adapt to dynamic changes in the distribution network voltage and often require collaboration with energy storage devices to meet power supply stability requirements. Existing traditional MPPT control methods for these generators struggle to balance power point tracking accuracy and response speed when facing wind speed fluctuations: when wind speeds are stable, the fixed tracking step size can lead to response lag, missing some of the maximum power output; when wind speeds fluctuate drastically, insufficient step size adaptation can cause power overshoot, affecting the generator's operational stability. Summary of the Invention

[0003] This application aims to at least partially address one of the technical problems in the related art.

[0004] Therefore, one objective of this application is to provide an adaptive MPPT control method and system. The multi-mode collaborative control can adapt to different wind conditions and grid conditions, achieve efficient power generation when the wind field is stable and the grid is normal, quickly compensate for power difference through energy storage when wind conditions fluctuate or the grid exceeds the tolerance, ensure the continuity of power supply to the load when the grid is interrupted or the wind speed is low, and avoid overcharging and discharging of equipment through energy storage SOC constraint, thereby extending the life of energy storage. Overall, it meets the comprehensive requirements of power generation efficiency, power supply stability and equipment protection in small and medium-sized distributed scenarios.

[0005] To achieve the above objectives, the first aspect of this application proposes an adaptive MPPT control method, comprising the following steps: S1. Data Acquisition and Processing: Real-time acquisition of operating wind speed data and filtering processing; S2. Wind speed feature extraction: Based on the processed real-time wind speed data, calculate the real-time wind speed fluctuation amplitude. ; S3. Short-term wind speed prediction: Using a pre-trained time series prediction model, the processed real-time wind speed data is rolled for prediction, and the predicted wind speed sequence for the next period is output. S4. Dynamic Decision-Making for Step Size Coefficients: A multi-level decision-making process is executed to determine the adaptive step size coefficients. This process includes: S41. Basic Mapping: Based on the real-time wind speed fluctuation amplitude. Query a preset mapping relationship to obtain the basic step size coefficient. The mapping relationship is predetermined based on the system stability analysis and wind condition simulation of the wind turbine generator set. S42. Trend Optimization: Perform quantitative trend analysis on the predicted wind speed sequence, and adjust the base step size coefficient according to the trend category obtained from the analysis. Optimization is performed to obtain an optimization step size coefficient. ; S43. Instantaneous Correction: Real-time calculation of generator output power change rate And when detected When the sign changes, the optimization step size coefficient is adjusted. The attenuation process is performed to obtain the final adaptive step size coefficient. ; S5.MPPT Tracking Control: Employs an improved hill-climbing method with the aforementioned adaptive step size coefficient. The determined dynamic step size is used for maximum power point tracking.

[0006] In addition, the adaptive MPPT control method proposed in this application may also have the following additional technical features: Furthermore, in step S41, the method for determining the preset mapping relationship includes: establishing a small-signal model of the unit and analyzing the stability margin of the system under different wind speed fluctuation intensities; Simulations were conducted to simulate different real-time wind speed fluctuation amplitudes. Tracking performance and overshoot when using different fixed step sizes; Based on the stability margin and simulation results, the real-time wind speed fluctuation range is configured to achieve optimal tracking efficiency for the system under stable conditions. With the basic step size coefficient The corresponding interval.

[0007] Furthermore, in step S42, the quantitative trend analysis specifically includes the following sub-steps: S421. Calculate at least one trend quantification index of the predicted wind speed sequence, the trend quantification index including sequence variance, gradient mean or range; S422. Compare the calculated trend quantification indicators with a set of preset trend judgment thresholds; S423. Based on the comparison results, the future wind speed trend is classified into one of the following: a stable trend, a gradual trend, or a dramatic trend. S424. Based on the identified trend categories, adopt a trend optimization factor corresponding to that category.

[0008] For the basic step size coefficient Perform multiplication optimization, where the corresponding trend is dramatic. Corresponding to a gradual trend Corresponding to a steady trend .

[0009] Furthermore, the method further includes a multi-mode cooperative control step S6: S6. State machine-based collaborative control: Define three modes: high-efficiency power generation, energy storage collaboration, and emergency support; Set the first threshold. Used to trigger energy storage coordination, setting a second threshold. Used to trigger emergency response; Wherein, the first threshold The wind speed fluctuation boundary value is calculated by inversion through a system model based on the grid-connected power fluctuation rate standard of the target power grid. Second threshold It is determined based on the cut-in wind speed or the lower limit wind speed for economic operation in the power curve of the wind turbine generator set; Based on real-time measurements , power grid status, measured wind speed and and The comparison results drive the state machine to switch between the three modes.

[0010] Furthermore, in step S6, the energy storage coordination mode also includes constraining the operation of the energy storage device according to the energy storage state of charge (SOC), wherein the upper and lower limits of the allowable operating range of the SOC are determined jointly based on the cycle life characteristics of the energy storage device and the system reserve capacity requirements.

[0011] A second aspect of this application provides an adaptive MPPT control system, comprising: Perception layer: used for data acquisition for execution method, including sensor group adapted to small and medium-sized wind turbine generator sets and data preprocessing unit, wherein the data preprocessing unit has at least filtering function; Control layer: Used for wind speed analysis, step size adjustment and MPPT control, including main control unit, algorithm storage unit and MPPT control logic module. The main control unit has data processing and instruction output capabilities. Collaboration layer: used for multi-device collaborative control in executing method step S5, including a communication module and a collaborative control unit. The communication module supports bidirectional communication with generators, energy storage devices and grid-connected inverters, and the collaborative control unit has a built-in state machine model. Hardware support layer: includes power management module and power drive module.

[0012] In addition, the adaptive MPPT control system proposed in this application may also have the following additional technical features: Furthermore, the main control unit of the control layer has a built-in ARIMA prediction model execution module, which includes: Model parameter retrieval unit: used to read the ARIMA model parameters and model coefficients determined during the offline training phase from the algorithm storage unit; Real-time data receiving unit: used to receive filtered wind speed data output by the sensing layer data preprocessing unit, and sort it by timestamp to form a real-time wind speed sequence; Prediction calculation unit: Based on ARIMA model parameters and real-time wind speed sequence, it performs autoregressive calculation, moving average calculation and inverse difference operation, which can be adjusted to 3-10 seconds according to system configuration; Result Correction Unit: Performs boundary correction on the predicted wind speed output by the prediction calculation unit, and transmits the corrected predicted wind speed to the MPPT control logic module of the control layer for adaptive step size calculation.

[0013] Furthermore, the sensor group of the sensing layer includes: Ultrasonic anemometer: used to collect wind speed data. The signal output terminal of the ultrasonic anemometer is electrically connected to the wind speed signal input terminal of the data preprocessing unit through a shielded cable. The temperature compensation circuit built into the ultrasonic anemometer first transmits the collected raw wind speed data to the data preprocessing unit for moving average filtering, and then outputs it to the control layer for wind speed characteristic analysis. Current transformer: used to collect stator current data of generator. The primary side of the current transformer is connected in series with the generator output circuit. The secondary side of the current transformer is connected to the current signal input terminal of the data preprocessing unit through a signal conditioning circuit. The output current signal is denoised by the data preprocessing unit and used in conjunction with the output signal of the voltage sampling module to calculate the generator output power. Voltage sampling module: used to collect the generator terminal voltage data. The sampling end of the voltage sampling module is connected in parallel to the generator output end, and the signal output end is directly connected to the voltage signal input end of the data preprocessing unit. The collected voltage data and the current data output by the current transformer are synchronously transmitted to the data preprocessing unit. After synchronous calibration, the data is output to the control layer to provide power calculation for MPPT tracking control. Energy storage status acquisition unit: used to acquire the SOC value, charging and discharging current and terminal voltage data of energy storage device. The acquisition interface of the energy storage status acquisition unit is connected to the communication interface of energy storage device through CAN bus. The data output terminal of the energy storage status acquisition unit is connected to the energy storage signal input terminal of the data preprocessing unit. After the acquired energy storage status data is converted by the data preprocessing unit, it is output to the coordination layer for control mode switching judgment. The data preprocessing unit achieves synchronous acquisition of data from each sensor using the same clock signal.

[0014] A third aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the first aspect of an adaptive MPPT control method.

[0015] The fourth aspect of this application provides a wind turbine generator set, including the second aspect of an adaptive MPPT control system.

[0016] Beneficial effects: According to the adaptive MPPT control method and system of the embodiments of this application, multi-mode collaborative control can adapt to different wind conditions and grid conditions. It can achieve efficient power generation when the wind field is stable and the grid is normal. When the wind conditions fluctuate or the grid exceeds the tolerance, it can quickly compensate for the power difference through energy storage. When the grid is interrupted or the wind speed is low, it can ensure the continuity of power supply to the load. At the same time, it can avoid overcharging and discharging of equipment through energy storage SOC constraint and extend the life of energy storage. Overall, it meets the comprehensive needs of power generation efficiency, power supply stability and equipment protection in small and medium-sized distributed scenarios.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 The overall control flow diagram of the adaptive MPPT control method and system according to this application is shown below. Figure 2 The diagram below shows the multi-mode cooperative control block diagram of the adaptive MPPT control method and system according to this application. Detailed Implementation

[0019] Embodiments of this application are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. Rather, embodiments of this application include all variations, modifications, and equivalents falling within the spirit and scope of the appended claims.

[0020] The adaptive MPPT control method and system of this application are described below with reference to the accompanying drawings.

[0021] like Figures 1-2 As shown, Example 1: The adaptive MPPT control method of this application includes the following steps: S1. Data Acquisition and Processing: Real-time acquisition of operating wind speed data and filtering processing; S2. Wind speed feature extraction: Based on the processed real-time wind speed data, calculate the real-time wind speed fluctuation amplitude. ; S3. Short-term wind speed prediction: Using a pre-trained time series prediction model, the processed real-time wind speed data is rolled for prediction, and the predicted wind speed sequence for the next period is output. S4. Dynamic Decision-Making for Step Size Coefficients: A multi-level decision-making process is executed to determine the adaptive step size coefficients. This process includes: S41. Basic Mapping: Based on real-time wind speed fluctuations. Query a preset mapping relationship to obtain the basic step size coefficient. The mapping relationship is predetermined based on the system stability analysis and wind condition simulation of the wind turbine generator set; S42. Trend Optimization: Perform quantitative trend analysis on the predicted wind speed sequence, and adjust the base step size coefficient based on the trend category obtained from the analysis. Optimization is performed to obtain an optimization step size coefficient. ; S43. Instantaneous Correction: Real-time calculation of generator output power change rate And when detected When the sign changes, the optimization step size coefficient is affected. The attenuation process is performed to obtain the final adaptive step size coefficient. ; S5.MPPT tracking control: Employs an improved hill-climbing method with an adaptive step size coefficient. The determined dynamic step size is used for maximum power point tracking.

[0022] In this embodiment, in step S41, the method for determining the preset mapping relationship includes: establishing a small-signal model of the unit and analyzing the stability margin of the system under different wind speed fluctuation intensities; Simulations were conducted to simulate different real-time wind speed fluctuations. Tracking performance and overshoot when using different fixed step sizes; Based on stability margin and simulation results, the real-time wind speed fluctuation range that optimizes the system's tracking efficiency under stable conditions is determined. With the basic step size coefficient The corresponding interval.

[0023] In this embodiment, step S42, the quantitative trend analysis specifically includes the following sub-steps: S421. Calculate at least one trend quantification index for the predicted wind speed sequence, including sequence variance, gradient mean, or range. S422. Compare the calculated trend quantification indicators with a set of preset trend judgment thresholds; S423. Based on the comparison results, the future wind speed trend is classified into one of the following: a stable trend, a gradual trend, or a dramatic trend. S424. Based on the identified trend categories, adopt a trend optimization factor corresponding to that category.

[0024] For the basic step size coefficient Perform multiplication optimization, where the corresponding trend is dramatic. Corresponding to a gradual trend Corresponding to a steady trend .

[0025] The flow of the adaptive MPPT control method: Step S1: Data acquisition and processing.

[0026] Data source: An ultrasonic anemometer installed at the hub of the wind turbine generator, with a sampling frequency set to 10Hz.

[0027] Filtering: The raw wind speed data is filtered in real time using a first-order low-pass digital filter. The cutoff frequency fc of the filter is set to 1Hz to retain low-frequency wind speed changes that reflect the aerodynamic characteristics of the wind turbine, while filtering out high-frequency measurement noise and turbulent micro-pulsations.

[0028] Step S2: Wind speed feature extraction.

[0029] Calculate the real-time wind speed fluctuation amplitude based on the filtered wind speed data. Specifically, taking the current time t as the center, the wind speed data sequence within the time window T (e.g., T=5 seconds) is taken. Defined as the current instantaneous wind speed Compared with the average wind speed within that window The absolute value of the difference, i.e.: in, The unit is meters per second (m / s). This indicator directly reflects the degree of deviation of instantaneous wind speed from its recent average.

[0030] Step S3: Short-term wind speed forecast.

[0031] A pre-trained ARIMA(2,1,1) model was used as the time series prediction model.

[0032] The filtered wind speed data from the latest 200 sampling points (i.e., the past 20 seconds) is input into the model for rolling prediction, and the predicted wind speed sequence for the next 3 seconds (30 sampling points in total, with an interval of 100ms) is output.

[0033] Step S4: Dynamic decision-making for step size coefficient (three-level decision-making process).

[0034] S41. Basic Mapping: The pre-defined mapping relationships exist as a lookup table stored in the controller's memory. This table is generated as follows: Modeling and stability analysis: For the 1.5MW permanent magnet direct-drive wind turbine generator set in this embodiment, a small-signal model including wind turbine aerodynamics, transmission chain, generator and converter control was established in Matlab / Simulink.

[0035] Through frequency domain analysis (such as Nyquist plots and Bode plots), the system is calculated at different wind speed reference points, corresponding to different... The phase margin and gain margin (when used as the disturbance input strength) are used to determine the maximum allowable step size boundary to ensure system stability.

[0036] Time-domain performance simulation: In the simulation platform, turbulent winds of different intensities (corresponding to different...) are set up. The range is determined, and different fixed hill-climbing step sizes are tested. The maximum power point tracking efficiency and power overshoot are recorded for each combination.

[0037] Comprehensive configuration partitioning: Combining the stability margin boundary (safety red line) from step 1 and the simulation performance results (efficiency and overshoot) from step 2, for each Configure a base step size coefficient within the interval to maximize tracking efficiency, controllable overshoot, and within the stability boundary. Range of values.

[0038] Example of mapping representation: when (With stable wind speed), the mapping is obtained At this point, the system has a large stability margin, allowing for the use of larger step sizes for rapid tracking.

[0039] when (Medium fluctuation), mapped to obtain The step size is reduced accordingly to suppress fluctuations.

[0040] when (Surging fluctuations), mapped to obtain Using small step sizes ensures stability first. The midpoint of the interval is usually taken.

[0041] At runtime, based on real-time calculations If the value falls within the specified interval, select one value from that interval as the current value. (For example, take the midpoint of the interval) as the basic step size coefficient for this control cycle. .

[0042] S42. Trend Optimization: S421. Calculate the trend quantification index: Calculate the variance of the predicted wind speed sequence output by S3. As a core indicator.

[0043] S422. Comparison with preset thresholds: Two preset trend determination thresholds are: , These thresholds are derived from statistical analysis of historical wind data and are used to distinguish different patterns of change. This is the "low volatility" threshold. When the trend quantification indicator (such as variance) of the predicted sequence is lower than this value, the future wind speed trend is determined to be a "stable trend".

[0044] This is the "high volatility" threshold. When the trend quantification index of the predicted sequence is higher than or equal to this value, the future wind speed trend is determined to be a "drastic change trend".

[0045] It is a core indicator used to measure the degree of fluctuation in the predicted wind speed sequence. Its calculation is based on the predicted wind speed sequence output by the short-term forecasting module.

[0046] S423. Trend Classification: like Classified as a stable trend (The wind speed will change gradually in the future).

[0047] like Classified as a gradual trend (The wind speed is expected to change in a clear but gradual direction in the future).

[0048] like Classified as a trend of dramatic change (Future wind speeds may change abruptly or fluctuate significantly).

[0049] For a containing Predicted wind speed sequence at each sampling point Its variance The calculation formula is: Alternatively, a correction formula using unbiased estimation can be used: ,when If the condition is met, a biased estimation formula is used; otherwise, an unbiased estimation formula is used.

[0050] in: Indicates the first in the sequence Predicted wind speed values ​​for each point (unit: m / s).

[0051] This represents the arithmetic mean of the predicted wind speed sequence, i.e. .

[0052] This represents the total number of sampling points in the predicted sequence, determined by the prediction duration and sampling interval (e.g., predicting the next 5 seconds, one point every 100ms). ).

[0053] This indicates a summation operation.

[0054] S424. Based on the identified trend categories: Trend optimization factor The range of values ​​is a guiding principle determined based on a trade-off between system stability and tracking efficiency. In this embodiment, to illustrate its application, a set of typical values ​​is selected for implementation. However, those skilled in the art should understand that in actual systems, the range of values ​​can be adjusted within the stated principle range based on wind condition statistical characteristics, unit inertia, and control performance requirements. The specific values ​​will be adjusted for adaptability.

[0055] Regarding the trend of dramatic changes ,set up Because drastic changes are anticipated, a conservative approach is needed in advance, reducing the step size to prevent instability. .

[0056] For gradual trends ,set up The trend is clear but gradual; therefore, we will maintain the basic step size for adaptive tracking. .

[0057] For a stable trend ,set up If the wind conditions are predicted to be stable, the step size can be appropriately increased to improve search efficiency. .

[0058] Calculate the optimization step size coefficient: .

[0059] S43. Instantaneous Correction:

[0060] Real-time calculation of generator output power change rate When detected When the sign changes, the optimization step size coefficient is affected. The attenuation process is performed to obtain the final adaptive step size coefficient. .

[0061] The specific implementation is as follows: System continuous monitoring The sign of the curve. Once a change in sign is detected (e.g., from positive to negative), it indicates that the power curve may have passed its peak and the system is approaching its maximum power point. At this time, immediate action should be taken. Perform attenuation correction to obtain the final step size coefficient. Step size coefficient The method of obtaining the result is as follows: like (Sign change): The attenuation factor ; like (Symbols remain unchanged): , in, This is a sign function used to determine the sign of a numerical value: when the input value... (here) represent When ) is greater than 0, When less than 0, .

[0062] Real-time calculation of generator output power change rate , Using the central difference method: ,in, Indicates the current sampling time The generator output power; This indicates the historical power value traversed back two sampling periods; This indicates the system's fixed control and sampling period.

[0063] The above can be summarized as continuous monitoring. The symbol, when detected When the sign changes from positive to negative or from negative to positive, it indicates that the power curve may have passed its peak value, and the system is at or near its maximum power point. At this time, immediate action should be taken... The attenuation process is performed to obtain the final adaptive step size coefficient. Among them, attenuation factor This value is based on experience and is chosen to strike a balance between suppressing oscillations and maintaining tracking sensitivity; it is generally a median value.

[0064] like If the sign remains unchanged, then .

[0065] Step S5: MPPT Tracking Control An improved hill-climbing method is adopted. In each control cycle, the direction of the disturbance is determined based on the current power P and the power of the previous cycle.

[0066] Dynamic step size calculation: The actual power or speed disturbance step size value is: , in, It is a preset reference step size value for a power unit, used to adjust the step size coefficient. Converted into power regulation commands with actual physical units.

[0067] controller with The step size is used to adjust the generator torque command or speed command to achieve dynamic step size maximum power point tracking.

[0068] In this embodiment, it should be noted that: the reference power perturbation value The determination, The value must simultaneously satisfy stability constraints, tracking accuracy requirements, and actuator resolution limitations, and is determined through the following steps: Step 1: Upper limit analysis based on system stability.

[0069] Establish a small-signal model or detailed nonlinear model of the unit, including the generator, converter, and control loop. Linearize the system near rated operating conditions and analyze the bandwidth and stability margin of the system's open-loop transfer function.

[0070] Through simulation, the maximum permissible single power step change that will not cause system divergence or sustained oscillation under worst-case wind disturbances was determined, and half to one-third of this value was taken as the maximum permissible single power step change. Theoretical upper limit .

[0071] The calculation formula can be expressed as: .

[0072] in: This represents the closed-loop bandwidth of the control system, measured in rad / s. This represents the small-signal model phase margin of the system at its rated operating point; This represents the maximum design wind speed disturbance intensity used for stability analysis, which can be quantified as turbulence intensity or wind speed change rate. It represents a functional relationship determined by a system model, which is usually determined by frequency domain analysis (such as Bode plots) combined with time domain simulation; It is a design upper limit value, representing the maximum step change allowed in a single power command while ensuring the stability of small signals of the wind turbine generator.

[0073] therefore, Must meet: .

[0074] Step 2: Lower limit analysis based on tracking accuracy and response speed.

[0075] Consider the system's requirement to track wind speed changes. Assume the wind speed changes at a certain rate. Changes require the unit to be in control cycle The system responds internally to follow the maximum power point, based on the wind turbine power characteristic curve. The required minimum power tracking rate can be derived. .

[0076] It should be noted that, This represents the power characteristic curve of a wind turbine, i.e., the generator's output power. With wind speed The functional relationship between these factors is determined by the aerodynamic characteristics of the wind turbine, transmission efficiency, and generator efficiency. This curve is typically provided by the wind turbine manufacturer and calibrated experimentally. In actual control, this curve is usually stored in the controller as a lookup table, determined by wind speed. Directly query the corresponding maximum capture power .

[0077] To ensure effective tracking, a single disturbance should produce an observable power change, therefore It should be several times greater than the system power measurement noise, and satisfy the following: ( ),in, This represents the noise amplitude in the generator output power measurement system (which can be estimated by collecting power data over a long period under steady state and calculating its standard deviation). For safety, a factor of 3 to 5 is typically used to ensure that the disturbance energy significantly exceeds the noise level; The minimum power change rate required to meet the tracking requirements can be derived from the wind speed change range and the wind turbine power characteristic curve. This refers to the system's control cycle; The function represents taking the larger of the two values.

[0078] Step 3: Comprehensive determination and verification.

[0079] In the interval Choose an intermediate value as The initial selected values ​​were then used. Subsequently, in a simulation model incorporating the complete MPPT adaptive control algorithm, these initial values ​​and other parameters were substituted, and tests were conducted under various typical wind conditions, including turbulent wind, step wind, and gradual wind. The average capture efficiency and power fluctuation standard deviation were used as evaluation metrics. Fine-tuning was performed to ultimately determine an optimal value that balances stability and speed across various scenarios.

[0080] Typical value range: For small and medium-sized wind turbine generators, experience shows that... Typically, it is 0.2% to 1% of the unit's rated power. For example, for a 100kW unit, its... It can be determined to be a specific value between 200W and 1000W.

[0081] Example 2: In one embodiment of this application, the method further includes a multi-mode cooperative control step S6: S6. State machine-based collaborative control: Define three modes: high-efficiency power generation, energy storage collaboration, and emergency support; Set the first threshold. Used to trigger energy storage coordination, setting a second threshold. Used to trigger emergency response; Among them, the first threshold The wind speed fluctuation boundary value is calculated by inversion through a system model based on the grid-connected power fluctuation rate standard of the target power grid. Second threshold It is determined based on the cut-in wind speed or the lower limit wind speed for economic operation in the power curve of the wind turbine generator set; Based on real-time measurements , power grid status, measured wind speed and and The comparison results drive the state machine to switch between three modes.

[0082] In this embodiment, step S6 further includes constraining the operation of the energy storage device according to its state of charge (SOC), wherein the upper and lower limits of the allowable operating range of the SOC are determined jointly based on the cycle life characteristics of the energy storage device and the system's reserve capacity requirements.

[0083] Specifically, step S6 is based on state machine-based cooperative control.

[0084] Pattern and threshold definitions: High-efficiency power generation mode: This mode has the highest priority and aims to maximize power generation revenue.

[0085] Energy storage collaborative mode: used to smooth power fluctuations or respond to grid dispatch.

[0086] Emergency backup mode: Ensure power supply to local critical loads in the event of off-grid or weak grid conditions.

[0087] First threshold : Set to 1 m / s. This value is calculated by inversion based on the requirements of the target power grid (such as national standard GB / T19963-2021) for wind power grid-connected power fluctuation rate, through the establishment of a transfer model of "wind speed fluctuation - unit power fluctuation - grid connection point power fluctuation". When Exceeding this value may cause grid-connected power fluctuations to exceed the limit due to fluctuations in the unit's own power. The first threshold... The preset fluctuation threshold is 1-2 m / s.

[0088] Second threshold Set to 3 m / s. This value represents the cut-in wind speed in the unit's power curve. Generating power below this speed is uneconomical and can easily lead to repeated start-ups and shutdowns of the unit. The second threshold... The preset wind speed threshold is 2-3 m / s.

[0089] State machine transition logic: Initial state: High-efficiency power generation mode.

[0090] Switch to energy storage collaborative mode: when Or when a grid voltage deviation is detected to exceed ±5% of the rated value for more than 2 seconds.

[0091] In this mode, calculation , Indicates power difference; Indicates the optimal power for MPPT; Indicates the allowable grid connection power; controls the energy storage system to Charge / discharge commands with a response time of <0.2s.

[0092] SOC Constraint: The SOC operating range of the energy storage battery is set at [25%, 85%]. The lower limit of 25% is based on the cycle life characteristics of lithium batteries (to prevent over-discharge), while the upper limit of 85% is to reserve at least 15% of backup charging capacity for the system to cope with sudden power shortages. When the SOC reaches the boundary, coordination commands will be restricted.

[0093] Switch to emergency backup mode: When a complete power grid outage or measured average wind speed is detected. When it lasts for 60 seconds.

[0094] MPPT tracking is immediately stopped, and the generator switches to power control mode, with power commands derived from local load demand.

[0095] Priority is given to the energy storage discharge, which works together with the generator to ensure the power supply to the load. The load power outage time during the switching process is less than 100ms.

[0096] Switch back to high-efficiency power generation mode: When The power grid voltage has returned to normal, and the wind speed has also returned to normal. When switching back from other modes.

[0097] Smooth transition: At the moment of mode switching, a first-order inertial element is used to smooth the generator power command and energy storage power command, limit the rate of power change, and avoid impact on equipment and power grid.

[0098] in, This refers to the control system taking a short-time calculation window from the real-time wind speed data collected and filtered in step S1 at the current moment. ,in To calculate the total number of sampling points within the window, These are the wind speed values ​​at each sampling point. This refers to the cut-in wind speed in the unit's power curve. Generating electricity at wind speeds below this is uneconomical and can easily lead to repeated start-ups and shutdowns of the unit. To preset the wind speed threshold, The value is directly derived from the "Technical Specifications" or power characteristic curve of the wind turbine generator set.

[0099] Example 3: In one embodiment of this application, the execution process of the ARIMA prediction model includes an offline training phase and an online prediction phase. In Example 1, the optimized training is as follows: Specifically, the ARIMA prediction model is executed in two phases: the offline training phase, which uses historical data covering the wind speed range of 3-15 m / s to complete data preprocessing, model parameter determination and verification, and form a usable model; and the online prediction phase, which inputs processed real-time wind speed data and calculates the time through the model, which can be adjusted to 3-10 seconds according to the system configuration. The two phases are connected to achieve short-term wind speed prediction.

[0100] In one embodiment of this scheme, the offline training phase includes: S31. Training Data Acquisition and Preprocessing: Collect historical data on wind speeds in the 3-15 m / s range for small and medium-sized distributed wind turbine installation scenarios, using the same sampling frequency as the wind speed data sampling frequency in step S1; preprocess the collected historical data, including: Outlier removal: Remove extreme wind speed data that exceed the 3-15 m / s range, as well as abnormal fluctuation data where the wind speed change rate is greater than 5 m / s for 10 consecutive sampling points; Stationarity processing: The non-stationary wind speed time series is transformed into a stationary series through differencing operations, ensuring that the series meets the ADF stationarity test. The order of differencing is [not specified]. ; S32. Model Parameter Determination: Based on the preprocessed stationary series, the autoregressive order of the ARIMA model is determined using the AIC information criterion. Moving average order : Traversal , Calculate the AIC value corresponding to each parameter combination; The parameter combination with the lowest AIC value is selected as the final model parameters. These parameter combinations include... , , ,in The difference order determined in step S31; Determining the model order Then, based on the preprocessed stationary training sequence, the specific coefficients of the model are estimated using the maximum likelihood estimation method or the least squares method.

[0101] S321. Establish the model equations: The ARIMA(2,1,1) model can be represented as: To obtain the optimal model order Then, the specific coefficients of the ARIMA model are solved by the maximum likelihood estimation method.

[0102] Specifically, the implementation involves: processing the pre-processed stable training sequence... As input, construct and solve for the coefficients , , The likelihood equations are used to obtain the coefficient values ​​that maximize the probability of the model fitting the sequence. This calculation is performed by calling a standard scientific computing library.

[0103] Coefficient estimation steps: After determining the model order Subsequently, based on the preprocessed stationary training sequences, the maximum likelihood estimation method was used, and standard time series analysis tools were employed to fit the model, thereby obtaining the specific autoregressive coefficients. , and moving average coefficient The numerical solution is obtained and stored.

[0104] S33. Model Validation and Optimization: Divide the preprocessed historical data into a training set and a validation set in a 7:3 ratio. Use the training set to train the ARIMA model, and use the validation set to verify the prediction accuracy. If the prediction error on the validation set is greater than ±0.3 m / s, adjust... , Retrain after each order until the prediction error is less than or equal to ±0.3m / s; The optimized ARIMA model parameters and the trained model coefficients are stored in the algorithm storage unit of the control layer for online prediction.

[0105] Specifically, the training data acquisition and preprocessing in S31 starts from the actual installation scenario of small and medium-sized distributed wind turbine units, selects historical data in the wind speed range of 3-15m / s, and maintains the same sampling frequency as the real-time acquisition (step S1) to ensure data scenario adaptability; outlier removal is to exclude extreme data that exceeds the target wind speed range and abnormal data with sudden changes in wind speed in a short period of time, so as to avoid such data interfering with model training.

[0106] The model parameters for S32 were determined based on the preprocessed stationary sequence, using the AIC information criterion. , Among the parameter combinations, the combination with the smallest AIC value is selected, and then combined with the difference order determined by S31. , forming the final ( , , Adjust the model parameters to ensure they are adapted to the wind speed data characteristics of the current scenario.

[0107] Model validation and optimization for S33 involves dividing the training and validation sets in a 7:3 ratio. The model is trained on the training set, and then the prediction accuracy is tested on the validation set. If the accuracy does not meet the requirement of error ≤ ±0.3 m / s, adjustments are made. , The model is retrained at different orders until the required accuracy is achieved; finally, the optimized model parameters and coefficients are stored in the control layer algorithm storage unit.

[0108] Furthermore, the online prediction phase includes: S34. Real-time data input: Input the real-time wind speed data collected in step S1 into the trained ARIMA model in time series order; S35. Short-term forecast calculation: The ARIMA model is based on the input real-time wind speed sequence, which can be adjusted to 3-10 seconds according to the system configuration through the following steps: Autoregressive calculation: based on autoregressive order , before Use the wind speed data at each moment and the model's autoregressive coefficients to calculate the autoregressive term at the current moment; Moving average calculation: based on the moving average order , before The prediction error at each time point and the moving average coefficient of the model are used to calculate the moving average term at the current time point. Predicted output: The autoregressive term and the moving average term are superimposed, and after inverse difference operation, the predicted wind speed is output at a sampling point every 100ms for the next 5-10 seconds. S36. Prediction result correction: If the predicted wind speed exceeds the range of 3-15m / s, it shall be corrected according to the boundary value.

[0109] Specifically, S34. Real-time data input: The real-time wind speed sequence (e.g., the latest 200 sampling points, corresponding to the past 20 seconds, sampling frequency 10Hz), after being filtered in step S1, is input into the trained ARIMA(2,1,1) model in timestamp order. The parameters of this model ( , , ) and the corresponding autoregressive coefficients ( , ), moving average coefficient ( The values ​​have been determined and stored in the algorithm storage unit through the aforementioned offline training phase.

[0110] S35. Short-term forecast calculation: The model performs rolling forecasts as follows, outputting a predicted wind speed sequence for the next 3 seconds (30 points in total, with an interval of 100ms): : Autoregressive term calculation: based on autoregressive order Using the wind speed observed in the first two moments , and the stored autoregressive coefficients , Calculate the autoregressive term at the current time. .

[0111] Moving average calculation: based on the moving average order Utilizing the prediction error from the previous moment and the stored moving average coefficient Calculate the moving average term at the current time. .

[0112] Predicted output: The autoregressive term is superimposed with the moving average term, and the difference order is adjusted. Perform inverse difference operations to obtain the final predicted value. Repeat this process iteratively to generate predicted values ​​for the next 29 points.

[0113] in, for Forecast wind speed at any time The error between the predicted value and the actual value, Current moment The autoregressive term, For the moving average term, Defined as , for The actual observed wind speed at that moment for The model predicts wind speed at any given time.

[0114] S36. Prediction result correction and output: Check the output predicted wind speed sequence. If any predicted value exceeds the effective operating wind speed range of the unit [3,15] m / s, clamp it to the boundary value (i.e., if it is less than 3 m / s, correct it to 3 m / s, and if it is greater than 15 m / s, correct it to 15 m / s).

[0115] The corrected predicted wind speed sequence will be directly output to step S42 (trend optimization module) for calculating variance. And determine the future wind speed trend category.

[0116] In this embodiment, the prediction duration is set to 3 seconds, which is based on the response cycle (e.g., 100ms) of the control system of the present invention and the typical time scale of wind speed change. It aims to provide sufficient forward-looking trend information for subsequent adaptive step size adjustment while maintaining high prediction accuracy.

[0117] Example 4: An adaptive MPPT control system according to an embodiment of this application includes: The perception layer is used to perform data acquisition, including sensor groups adapted to small and medium-sized wind turbine generators and a data preprocessing unit. The data preprocessing unit has at least a filtering function. Control layer: Used to perform wind speed analysis, step size adjustment and MPPT control, including main control unit, algorithm storage unit and MPPT control logic module. The main control unit has data processing and instruction output capabilities. Collaboration Layer: Used for multi-device collaborative control in step S5, including a communication module and a collaborative control unit. The communication module supports bidirectional communication with generators, energy storage devices and grid-connected inverters, and the collaborative control unit has a built-in state machine model. Hardware support layer: includes power management module and power drive module.

[0118] Specifically, the perception layer provides raw data for subsequent control, corresponding to step S1. Its sensor array is adapted to the characteristics of small and medium-sized units, and can accurately collect wind speed data, generator operating data (stator current, rotor speed, etc.), and energy storage device status data (SOC value, terminal voltage, etc.). The data preprocessing unit, with filtering function as its core, can filter noise from the collected raw data (such as filtering instantaneous fluctuations in wind speed data) to avoid abnormal data interfering with subsequent analysis and provide stable data input for the control layer.

[0119] The control layer is responsible for the computation and execution of control logic. The main control unit, in particular, possesses data processing and command output capabilities, and can perform wind speed characteristic analysis (such as...). Calculation, ARIMA short-term forecast), adaptive step size dynamic adjustment (such as based on and The calculation of the step size coefficient is performed; the algorithm storage unit is used to store core algorithm data such as ARIMA prediction model parameters and improved hill-climbing logic, ensuring fast access during calculation; the MPPT control logic module generates instructions to adjust the generator power output according to the determined adaptive step size, realizing maximum power point tracking.

[0120] Multi-device collaborative control at the collaborative layer is the core of the联动 control of generators, energy storage devices, and grid-connected inverters. The communication module supports two-way communication and can transmit the status data of each device (such as generator power, energy storage SOC, grid voltage) and control instructions in real time to ensure smooth data interaction between devices; the collaborative control unit has a built-in state machine model, which can be based on the wind speed fluctuation characteristics ( ), grid operation status (voltage deviation), and energy storage SOC critical values (20%, 80%) to execute the switching of three modes: efficient power generation, energy storage collaboration, and emergency support. At the same time, it controls the power change rate during mode switching through a smooth transition algorithm to avoid sudden power changes.

[0121] In the hardware support layer, the power management module is responsible for providing stable power supply for each module in the sensing layer, control layer, and collaborative layer to avoid affecting data acquisition or control instruction execution due to power supply fluctuations; the power drive module receives the control instructions output by the control layer and collaborative layer and converts them into actual drive signals, such as driving the generator to adjust the power output and driving the energy storage device to perform charge and discharge actions.

[0122] In an embodiment of the present application, the main control unit of the control layer has a built-in ARIMA prediction model execution module, and the execution module includes: Model parameter call unit: used to read the ARIMA model parameters and model coefficients determined in the offline training stage from the algorithm storage unit; Real-time data reception unit: used to receive the filtered wind speed data output by the data preprocessing unit in the sensing layer and form a real-time wind speed sequence according to the time stamp; Prediction calculation unit: Based on the ARIMA model parameters and the real-time wind speed sequence, perform autoregressive calculation, moving average calculation, and inverse difference operation, and can be adjusted to 3 - 10 seconds according to the system configuration; Result correction unit: Perform boundary correction on the predicted wind speed output by the prediction calculation unit and transmit the corrected predicted wind speed to the MPPT control logic module in the control layer for adaptive step size calculation.

[0123] Specifically, the ARIMA prediction model execution module of the main control unit in the control layer includes four core units: the model parameter call unit reads the well-trained model parameters and coefficients in the algorithm storage unit; the real-time data reception unit receives the filtered wind speed data in the sensing layer and sorts them into a real-time sequence according to the time stamp; the prediction calculation unit outputs a 5 - 10s predicted wind speed based on the parameters and the sequence through autoregressive, moving average, and inverse difference operations; the result correction unit performs boundary correction on the predicted wind speed and then transmits it to the MPPT control logic module for S3 step size calculation.

[0124] In an embodiment of the present application, the sensor group in the sensing layer includes: Ultrasonic anemometer: Used to collect wind speed data. The signal output terminal of the ultrasonic anemometer is electrically connected to the wind speed signal input terminal of the data preprocessing unit through a shielded cable. The built-in temperature compensation circuit of the ultrasonic anemometer first transmits the collected raw wind speed data to the data preprocessing unit for moving average filtering, and then outputs it to the control layer for wind speed characteristic analysis. Current transformer: Used to collect stator current data of generator. The primary side of the current transformer is connected in series with the generator output circuit. The secondary side of the current transformer is connected to the current signal input terminal of the data preprocessing unit through the signal conditioning circuit. The output current signal is denoised by the data preprocessing unit and used in conjunction with the output signal of the voltage sampling module to calculate the generator output power. Voltage sampling module: used to collect the generator's terminal voltage data. The sampling end of the voltage sampling module is connected in parallel to the generator's output end, and the signal output end is directly connected to the voltage signal input end of the data preprocessing unit. The collected voltage data and the current data output by the current transformer are synchronously transmitted to the data preprocessing unit. After synchronous calibration, the data is output to the control layer to provide power calculation for MPPT tracking control. Energy storage status acquisition unit: used to acquire the SOC value, charging and discharging current and terminal voltage data of energy storage device. The acquisition interface of the energy storage status acquisition unit is connected to the communication interface of energy storage device through CAN bus. The data output terminal of the energy storage status acquisition unit is connected to the energy storage signal input terminal of the data preprocessing unit. After the acquired energy storage status data is converted by the data preprocessing unit, it is output to the coordination layer for control mode switching judgment. The data preprocessing unit achieves synchronous acquisition of data from all sensors using the same clock signal.

[0125] Specifically, in the perception layer, four types of sensors—ultrasonic anemometer, current transformer, voltage sampling module, and energy storage status acquisition unit—collect corresponding data and establish connections with the data preprocessing unit (the anemometer is connected via shielded cable, the current transformer via signal conditioning circuit, the voltage sampling module directly connected in parallel, and the energy storage acquisition unit via CAN bus). The data preprocessing unit achieves synchronous acquisition of data from the four types of sensors through the same clock signal, and performs filtering, noise reduction, synchronization calibration, and format conversion on the data from each sensor. The processed wind speed, current, and voltage data are sent to the control layer to support wind speed characteristic analysis and MPPT power calculation, while the energy storage status data is sent to the coordination layer for control mode switching judgment.

[0126] This system achieves MPPT adaptive control through a process of "data acquisition - analysis and calculation - control execution - collaborative adaptation," the specific process of which is as follows: The perception layer first initiates multi-dimensional data acquisition: an ultrasonic anemometer (with temperature compensation) transmits wind speed data via shielded cable; a current transformer (connected to the generator output circuit in series) collects stator current; a voltage sampling module (connected to the generator output terminal) collects terminal voltage; and the energy storage status acquisition unit collects energy storage SOC, charging / discharging current, and terminal voltage via a CAN bus. All sensor data is accessed by the data preprocessing unit, which achieves synchronous acquisition through the same clock signal. After performing moving average filtering of wind speed data, noise reduction of current signal, synchronous calibration of voltage and current data, and conversion of energy storage data format, the wind speed and generator operation data are transmitted to the control layer, and the energy storage status data is transmitted to the coordination layer.

[0127] After receiving data from the perception layer, the control layer first performs wind speed characteristic analysis: the ARIMA prediction model execution module of the main control unit reads the offline trained model parameters and coefficients from the algorithm storage unit, receives the filtered wind speed data and sorts it by timestamp, and outputs the predicted wind speed for 5-10 seconds after autoregression, moving average and inverse difference operations. The predicted values ​​in the range of 3-15 m / s are corrected by the result correction unit according to the boundary; at the same time, the real-time wind speed fluctuation amplitude is calculated. Next, adaptive step size adjustment is performed: based on the improved hill-climbing method, according to... The matching step size coefficient is set for the corresponding interval (<1m / s: 0.8-1.0; 1-3m / s: 0.4-0.8; >3m / s: 0.1-0.4). If the generator power change rate is detected... The sign changes, the current step size coefficient is corrected, and the adjusted step size coefficient is transmitted to the MPPT control logic module.

[0128] Finally, MPPT power point tracking is executed: the MPPT control logic module generates adjustment commands based on the adaptive step size, which are then converted into actions by the power drive module to control the generator output power to approach the maximum power point corresponding to the current wind speed.

[0129] The collaborative layer transmits data such as generator power, grid voltage, and energy storage SOC in real time through the communication module. The state machine model of the collaborative control unit is combined with wind speed characteristics. Switch control modes based on predicted wind speed, grid status (voltage deviation ≤ ±5% of rated value), and energy storage status (SOC between 20% and 80%). When the speed is ≤1-2m / s and the grid voltage is normal, it enters the high-efficiency power generation mode, the generator operates at the MPPT optimal power, and the energy storage maintains 20%-80% SOC standby. When the wind speed is >1-2 m / s or the grid voltage exceeds the tolerance, the system enters the energy storage collaborative mode, calculates the power difference between the "MPPT optimal power and the grid-connected allowable power", transmits it to the energy storage converter within 100 ms, and completes the charging and discharging compensation within 0.2 s; when the grid is interrupted or the wind speed is <2-3 m / s, the system enters the emergency protection mode, stops MPPT control, the generator operates according to the load demand power, prioritizes the control of energy storage discharge, and completes the power supply connection within 0.1 s.

[0130] When switching modes, the power change rate is controlled by a smooth transition algorithm, and energy storage discharge is prohibited when SOC < 20% and energy storage charging is prohibited when SOC > 80%.

[0131] The application embodiment discloses a computer-readable storage medium, specifically, on which a computer program is stored, which, when executed by a processor, implements the steps of any of the above methods.

[0132] The application embodiment discloses a wind turbine generator set, including the aforementioned adaptive MPPT control system.

[0133] In summary, the adaptive MPPT control method and system of this application can adapt to different wind conditions and grid states through multi-mode collaborative control. It can achieve efficient power generation when the wind farm is stable and the grid is normal. When the wind conditions fluctuate or the grid exceeds the tolerance, it can quickly compensate for the power difference through energy storage. When the grid is interrupted or the wind speed is low, it can ensure the continuity of power supply to the load. At the same time, it can avoid overcharging and discharging of the equipment through energy storage SOC constraint and extend the life of energy storage. Overall, it meets the comprehensive requirements of power generation efficiency, power supply stability and equipment protection in small and medium-sized distributed scenarios.

[0134] In the description of this specification, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0135] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0136] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. An adaptive MPPT control method, characterized in that, Includes the following steps: S1. Data Acquisition and Processing: Real-time acquisition of operating wind speed data and filtering processing; S2. Wind speed feature extraction: Based on the processed real-time wind speed data, calculate the real-time wind speed fluctuation amplitude. ; S3. Short-term wind speed prediction: Using a pre-trained time series prediction model, the processed real-time wind speed data is rolled for prediction, and the predicted wind speed sequence for the next period is output. S4. Dynamic Decision-Making for Step Size Coefficients: A multi-level decision-making process is executed to determine the adaptive step size coefficients. This process includes: S41. Basic Mapping: Based on the real-time wind speed fluctuation amplitude. Query a preset mapping relationship to obtain the basic step size coefficient. The mapping relationship is predetermined based on the system stability analysis and wind condition simulation of the wind turbine generator set. S42. Trend Optimization: Perform quantitative trend analysis on the predicted wind speed sequence, and adjust the base step size coefficient according to the trend category obtained from the analysis. Optimization is performed to obtain an optimization step size coefficient. ; S43. Instantaneous Correction: Real-time calculation of generator output power change rate And when detected When the sign changes, the optimization step size coefficient is adjusted. The attenuation process is performed to obtain the final adaptive step size coefficient. ; S5.MPPT Tracking Control: Employs an improved hill-climbing method with the aforementioned adaptive step size coefficient. The determined dynamic step size is used for maximum power point tracking.

2. The method according to claim 1, characterized in that, In step S41, the method for determining the preset mapping relationship includes: establishing a small-signal model of the unit and analyzing the stability margin of the system under different wind speed fluctuation intensities; Simulations were conducted to simulate different real-time wind speed fluctuation amplitudes. Tracking performance and overshoot when using different fixed step sizes; Based on the stability margin and simulation results, the real-time wind speed fluctuation range is configured to achieve optimal tracking efficiency for the system under stable conditions. With the basic step size coefficient The corresponding interval.

3. The method according to claim 1, characterized in that, In step S42, the quantitative trend analysis specifically includes the following sub-steps: S421. Calculate at least one trend quantification index of the predicted wind speed sequence, the trend quantification index including sequence variance, gradient mean or range; S422. Compare the calculated trend quantification indicators with a set of preset trend judgment thresholds; S423. Based on the comparison results, the future wind speed trend is classified into one of the following: a stable trend, a gradual trend, or a dramatic trend. S424. Based on the identified trend categories, adopt a trend optimization factor corresponding to that category. For the basic step size coefficient Perform multiplication optimization, where the corresponding trend is dramatic. Corresponding to a gradual trend Corresponding to a steady trend .

4. The method according to claim 1, characterized in that, The method further includes a multi-mode cooperative control step S6: S6. State machine-based collaborative control: Define three modes: high-efficiency power generation, energy storage collaboration, and emergency support; Set the first threshold. Used to trigger energy storage coordination, setting a second threshold. Used to trigger emergency response; Wherein, the first threshold The wind speed fluctuation boundary value is calculated by inversion through a system model based on the grid-connected power fluctuation rate standard of the target power grid. Second threshold It is determined based on the cut-in wind speed or the lower limit wind speed for economic operation in the power curve of the wind turbine generator set; Based on real-time measurements , power grid status, measured wind speed and and The comparison results drive the state machine to switch between the three modes.

5. The method according to claim 4, characterized in that, In step S6, the energy storage coordination mode further includes constraining the operation of the energy storage device according to the energy storage state of charge (SOC), wherein the upper and lower limits of the allowable operating range of the SOC are determined jointly based on the cycle life characteristics of the energy storage device and the system reserve capacity requirements.

6. An adaptive MPPT control system implementing the method of any one of claims 1-5, characterized in that, include: Perception layer: Used to perform data acquisition; Control layer: Used to perform wind speed analysis, step size adjustment and MPPT control, including main control unit, algorithm storage unit and MPPT control logic module; Collaboration layer: used to perform multi-device collaborative control, including a communication module and a collaborative control unit. The communication module supports bidirectional communication with generators, energy storage devices and grid-connected inverters, and the collaborative control unit has a built-in state machine model. Hardware support layer: includes power management module and power drive module.

7. The system according to claim 6, characterized in that, The main control unit of the control layer has a built-in ARIMA prediction model execution module, which includes: Model parameter retrieval unit: used to read the ARIMA model parameters and model coefficients determined during the offline training phase from the algorithm storage unit; Real-time data receiving unit: used to receive filtered wind speed data output by the sensing layer data preprocessing unit, and sort it by timestamp to form a real-time wind speed sequence; Prediction calculation unit: Based on ARIMA model parameters and real-time wind speed sequence, it performs autoregressive calculation, moving average calculation and inverse difference operation, which can be adjusted to 3-10 seconds according to system configuration; Result Correction Unit: Performs boundary correction on the predicted wind speed output by the prediction calculation unit, and transmits the corrected predicted wind speed to the MPPT control logic module of the control layer for adaptive step size calculation.

8. The system according to claim 6, characterized in that, The sensor group of the sensing layer includes: Ultrasonic anemometer: used to collect wind speed data. The signal output terminal of the ultrasonic anemometer is electrically connected to the wind speed signal input terminal of the data preprocessing unit. The ultrasonic anemometer has a built-in temperature compensation circuit. Current transformer: used to collect stator current data of generator. The primary side of the current transformer is connected in series with the generator output circuit, and the secondary side of the current transformer is connected to the current signal input terminal of the data preprocessing unit through a signal conditioning circuit. Voltage sampling module: used to collect the terminal voltage data of the generator, and the sampling terminal of the voltage sampling module is connected in parallel to the generator output terminal; Energy storage status acquisition unit: used to acquire the SOC value, charging and discharging current and terminal voltage data of the energy storage device. The acquisition interface of the energy storage status acquisition unit is connected to the communication interface of the energy storage device, and the data output terminal of the energy storage status acquisition unit is connected to the energy storage signal input terminal of the data preprocessing unit. The data preprocessing unit achieves synchronous acquisition of data from each sensor using the same clock signal.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 5.

10. A wind turbine generator set, characterized in that, Includes the MPPT adaptive control system as described in any one of claims 6 to 8.

Images