Optimal control system for adjusting pitch angle of wind generating set

Abstract

The invention discloses an optimal control system for adjusting a pitch angle of a wind generating set. The optimal control system comprises a data monitoring module, a strategy analysis module, a pitch angle control module, a safety margin evaluation module and a threshold iterative optimization module, according to the method, the clearance distance between the blades and the tower drum can be accurately monitored in real time, the control strategy is dynamically adjusted in combination with actual environmental parameters, the dynamic optimal power generation pitch angle special for the wind generating set is controlled and measured, and the optimal power generation efficiency is achieved. The problem that the optimal power generation pitch angle of a current mainstream blade is not matched with the clearance due to terrain, wind conditions and the like is solved, and the problems of frequent power limitation and shutdown caused by too small clearance of the blade of a unit at present are solved. And meanwhile, the control robustness is improved, and the problem of oscillation of the wind generating set caused by pitch angle adjustment is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine pitch angle control, and in particular to an optimization control system for adjusting the wind turbine pitch angle. Background Art

[0002] Currently, as the capacity of wind turbines continues to increase, blade lengths can reach hundreds of meters to meet the demand for a larger swept area. However, due to the weight of the blades themselves and material costs, blade stiffness is insufficient. In strong winds and turbulent weather, blade sweep accidents are prone to occur, causing damage to the blades and tower, and even causing serious consequences such as tower collapse.

[0003] To prevent blade sweeps, the current mainstream prevention and control approach involves simulation calculations based on IEC standards to determine the relationship between the minimum clearance and the optimal fine pitch angle for a designed wind turbine. However, due to limitations in simulation software, actual operating environments differ significantly from simulated ones. This results in frequent problems with low clearance, sweeps, blade breakage, or tower collapse when wind turbines are operated in different environments based on simulated fine pitch. Furthermore, after installing the fine pitch, the control strategy typically shifts from conservative to aggressive, but the blades often cannot support this aggressive power generation mode, resulting in a high volume of data from the protective logic's retracted blades. This results in a less significant increase in power generation compared to before the fine pitch was installed, leading to frequent power limiting and shutdowns due to insufficient blade clearance. Long-term analysis reveals that frequent pitch angle adjustments can cause wind turbines to experience power and clearance oscillations. Summary of the Invention

[0004] The present invention aims to overcome the shortcomings of existing technologies by proposing an optimized control system for adjusting the pitch angle of a wind turbine generator set. This system addresses the mismatch between simulated fine pitch and clearance of mainstream blades due to terrain, wind conditions, and other factors. It also addresses the frequent power limiting and shutdowns caused by insufficient blade clearance. It also improves control robustness and suppresses turbine oscillations caused by pitch angle adjustment.

[0005] The object of the present invention is achieved through the following technical solution: A wind turbine generator set pitch angle optimization control system, comprising:

[0006] Data monitoring module, used to collect SCADA data of wind turbine generator sets at the current optimal power generation pitch angle;

[0007] The strategy analysis module is used to divide power into bins and set dynamic thresholds. It independently collects statistics on the operating data in each power bin and determines whether the clearance distance of each power bin falls into the high-risk power range or the high-redundancy power range.

[0008] The pitch angle control module adjusts the pitch angle based on the judgment results provided by the strategy analysis module, and ultimately adjusts the clearance distance of each power bin to remove high-risk power segments or high-redundancy power segments;

[0009] The safety margin assessment module constructs an input feature set based on historical SCADA data and outputs a clearance distance prediction value. It uses machine learning and gradient descent optimization loss function to optimize the prediction results, and predicts the clearance change trend under different pitch angle adjustment strategies within a preset time period in real time, and selects an adjustment scheme that maintains the clearance distance within the preset range.

[0010] The threshold iterative optimization module defines the clearance threshold for triggering blade retraction and the shutdown protection threshold. When the wind turbine generator set has not triggered blade retraction for many consecutive days and the average clearance of each power bin is greater than the preset threshold, the margin compression strategy is executed, and iterative optimization is performed to converge the threshold. When the shutdown protection threshold is triggered, the wind turbine generator set's safety protection system is directly activated.

[0011] Furthermore, the SCADA data of the wind turbine generator set includes power, wind speed and blade clearance distance.

[0012] Furthermore, the strategy analysis module includes:

[0013] Refined power binning: With 100kW power bins as the interval, the operating data in each power bin is independently counted. The statistical data includes real-time clearance distance D, corresponding pitch angle β, wind speed V, and blade root load Fload.

[0014] Perform normal distribution or Weibull distribution fitting on the headroom data in each power bin, calculate the mean μ and standard deviation σ, generate the headroom probability density curve f(D), and set the initial safety threshold to [μ-2σ, μ+2σ] or [μ-3σ, μ+3σ];

[0015] Set the lower limit of the clearance distance D min =μ-kσ, k=2 or 3, which is the critical value for triggering propeller retraction;

[0016] Set the clearance distance upper limit D max =μ+kσ is the optimized boundary value of propeller opening; if the clearance distance D in the corresponding power bin is <D min If the data of exceeds 5%, it is determined that the clearance distance of the power bin falls into the high-risk power range; if the clearance distance D>D max If the proportion exceeds 10%, it is determined that the clearance distance of the power bin falls into the high-redundancy power segment.

[0017] Furthermore, the pitch angle control module includes:

[0018] For the power compartment that falls into the high-risk power range, the pitch angle is increased by Δβ = +k*0.5° each time, where k is the adjustment coefficient, to reduce the lift of the blade and reduce the deformation until the clearance distance D in the power compartment is <D min The data accounted for ≤2%;

[0019] For the power bin that falls into the high-redundancy power range, the pitch angle Δβ is reduced by -k*0.5° each time to increase the blade lift and increase the power generation until the clearance distance D>D max The data account for ≤5%.

[0020] Furthermore, the safety margin assessment module includes:

[0021] Using historical SCADA data, construct the input feature set {wind speed V, pitch angle β, clearance temp, generator power}, and output the clearance distance prediction value Use LSTM neural network or random forest regression model and gradient descent to optimize the loss function Where N is the number of samples, is the model’s predicted value of the clearance distance for the i-th sample, D i is the true value of the clearance distance of the i-th sample; the square of the difference between the predicted value and the true value of each sample is calculated by this function, and then the average value is calculated. The prediction is optimized by minimizing the average value; at the same time, the model parameters are updated every 10 minutes, and the clearance change trend under different pitch angle adjustment schemes in the next 10 minutes is predicted in real time. The clearance is kept at [D min +0.5m,D max -0.5m] interval adjustment strategy.

[0022] Furthermore, the threshold iteration optimization module includes:

[0023] Define the clearance threshold for retracting the propellers A1 = D min And the shutdown protection threshold A2 = A1-2m; use the margin compression strategy, that is, when the wind turbine generator set does not trigger the blade retraction for 7 consecutive days and the average clearance value of each power bin μ> A1+0.2m, execute:

[0024] A. Lower A1 by 0.1m and A2 by 0.1m simultaneously;

[0025] B. Re-fit all power bins, return to the strategy analysis module and the pitch angle control module to adjust the optimal power generation pitch angle;

[0026] C. The wind turbine is in a normal power generation state and is monitored for 48 hours after each round of adjustment. If the number of blade retractions is ≤3 times / day and the load does not exceed the design value, it enters the next round of compression until two consecutive blade retraction triggers occur with an interval of less than 1 hour. The last valid threshold combination is retained to obtain the iteratively optimized threshold.

[0027] A method for optimizing the control of the pitch angle of a wind turbine generator set is provided. The method is implemented by a processor calling the data monitoring module, strategy analysis module, pitch angle control module, safety margin assessment module and threshold iteration optimization module in the above-mentioned control system for optimizing the pitch angle of a wind turbine generator set.

[0028] A non-transitory computer-readable medium storing instructions, when the instructions are executed by a processor, executes the above-mentioned method for optimizing the control of adjusting the pitch angle of a wind turbine generator set.

[0029] A computing device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the above-mentioned method for optimizing the control of adjusting the pitch angle of a wind turbine generator set is implemented.

[0030] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0031] 1. Increased power generation: Through power distribution and dynamic pitch angle optimization, aerodynamic efficiency is maximized and annual power generation is increased.

[0032] 2. Enhanced safety: By setting dual-threshold protection for A1 and A2, as well as machine learning predictive control, the risk of blade impact is reduced, sudden blade retraction is reduced, and system robustness is improved.

[0033] 3. Strong adaptability: Dynamic distribution fitting and threshold iterative optimization automatically adapt to different wind field turbulence characteristics and seasonal changes.

[0034] 4. Low cost and high compatibility: Pure software solution, adaptable to various models, no need for additional hardware, and low deployment cost.

[0035] 5. Extended life: Reduce blade load fluctuations, reduce maintenance frequency, and significantly extend the life of key components of wind turbines. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Architecture diagram of the control system for adjusting the pitch angle optimization of wind turbine generator sets.

[0037] Figure 2 This is a comparison chart of the test prototype of a wind farm experiment. DETAILED DESCRIPTION

[0038] The present invention will be further described below with reference to specific embodiments.

[0039] Example 1

[0040] See also Figure 1 As shown, the wind turbine generator set pitch angle optimization control system provided by this embodiment includes:

[0041] 1) Data monitoring module, used to collect SCADA data of wind turbine generators at the current optimal power generation pitch angle, including power, wind speed and blade clearance distance.

[0042] 2) Strategy Analysis Module: This module is used to perform power binning and dynamic threshold setting. It independently collects statistics on the operating data within each power bin and determines whether the clearance distance of each power bin falls into the high-risk power segment or the high-redundancy power segment. This module includes:

[0043] Refined power binning, with 100kW power segments as intervals, and independent statistics of the operating data in each power bin. The statistical data includes real-time clearance distance D in meters, corresponding pitch angle β in degrees, wind speed V in meters per second, and blade root load F. load , unit is kN.

[0044] Perform normal distribution fitting or Weibull distribution fitting on the clearance data in each power bin, calculate the mean μ and standard deviation σ, generate the clearance probability density curve f(D), and set the initial safety threshold to [μ-2σ,μ+2σ] that can cover 95% of the data or [μ-3σ,μ+3σ] that can cover 99.7% of the data.

[0045] Set the lower limit of the clearance distance D min =μ-kσ, k=2 or 3, which is the critical value for triggering the retraction of the propeller.

[0046] Set the clearance distance upper limit D max =μ+kσ is the optimized boundary value of propeller opening; if the clearance distance D in the corresponding power bin is <D min If the data of exceeds 5%, it is determined that the clearance distance of the power bin falls into the high-risk power range; if the clearance distance D>D max If the proportion exceeds 10%, it is determined that the clearance distance of the power bin falls into the high-redundancy power segment.

[0047] 3) The pitch angle control module adjusts the pitch angle according to the judgment results provided by the strategy analysis module, and finally adjusts the clearance distance of each power bin to remove high-risk power segments or high-redundancy power segments.

[0048] For the power compartment that falls into the high-risk power range, each adjustment increases the pitch angle Δβ=+k*0.5°, where k is the adjustment coefficient and is adjusted according to the risk level; this is to reduce the blade lift and deformation until the clearance distance D in the power compartment is <D min The data accounts for ≤2%.

[0049] For the power bin that falls into the high-redundancy power range, the pitch angle Δβ is reduced by -k*0.5° each time to increase the blade lift and increase the power generation until the clearance distance D>D max The data account for ≤5%.

[0050] 4) The safety margin assessment module constructs an input feature set based on historical SCADA data and outputs a clearance distance prediction value. It uses machine learning and gradient descent optimization loss function to optimize the prediction results, and predicts the clearance change trend under different pitch angle adjustment strategies within a preset time period in the future in real time. It selects an adjustment scheme that maintains the clearance distance within the preset range.

[0051] Using historical SCADA data, construct the input feature set {wind speed V, pitch angle β, clearance temp, generator power}, and output the clearance distance prediction value Use LSTM neural network or random forest regression model and gradient descent to optimize the loss function Where N is the number of samples, is the model’s predicted value of the clearance distance for the i-th sample, D i is the true value of the clearance distance of the i-th sample; the square of the difference between the predicted value and the true value of each sample is calculated by this function, and then the average value is calculated. The prediction is optimized by minimizing the average value; at the same time, the model parameters are updated every 10 minutes, and the clearance change trend under different pitch angle adjustment schemes in the next 10 minutes is predicted in real time. The clearance is kept at [D min +0.5m,D max -0.5m] interval adjustment strategy.

[0052] 5) Threshold iterative optimization module, which defines the clearance threshold for triggering blade retraction and the shutdown protection threshold. When the wind turbine generator set has not triggered blade retraction for many consecutive days and the average clearance of each power bin is greater than the preset threshold, the margin compression strategy is implemented, and iterative optimization is performed to converge the threshold. When the shutdown protection threshold is triggered, the wind turbine generator set's safety protection system is directly activated.

[0053] Define the clearance threshold for retracting the propellers A1 = D min And the shutdown protection threshold A2 = A1-2m; use the margin compression strategy, that is, when the wind turbine generator set does not trigger the blade retraction for 7 consecutive days and the average clearance value of each power bin μ> A1+0.2m, execute:

[0054] A. Lower A1 by 0.1m and A2 by 0.1m simultaneously;

[0055] B. Re-fit all power bins, return to the strategy analysis module and the pitch angle control module to adjust the optimal power generation pitch angle;

[0056] C. The wind turbine is in a normal power generation state and is monitored for 48 hours after each round of adjustment. If the number of times the blades are retracted is ≤3 times / day and the load does not exceed the design value, it enters the next round of compression until the interval between two consecutive blade retraction triggers is less than 1 hour. The last valid threshold combination is retained to obtain the iteratively optimized threshold.

[0057] Based on the wind turbine generator set pitch angle optimization control system provided in this embodiment, a closed-loop control process is formed as follows:

[0058] a. Data collection-analysis closed loop:

[0059] SCADA data - power binning - distribution fitting - dynamic threshold generation - pitch angle adjustment - new data feedback.

[0060] b. Safety control dual circuit:

[0061] Main circuit: real-time clearance and D min / D max Compare - adjust the pitch angle immediately.

[0062] Prediction loop: LSTM predicts future clearance and proactively adjusts the pitch angle.

[0063] c. Threshold optimization process:

[0064] Start condition detection - threshold compression - safety verification period - convergence judgment - final threshold locking.

[0065] d. Exception handling path:

[0066] Emergency stop threshold A2 is triggered - bypassing all optimization modules - directly connected to the safety protection system.

[0067] See also Figure 2 As shown in the comparison chart of a wind farm experimental prototype test, compared with the traditional solution, under the same working conditions, the power generation rate of this embodiment is increased by 13.636%, the damage of the variable pitch drive is reduced to 0, and the grid dispatch qualification rate is increased by 5 percentage points.

[0068] Example 2

[0069] This embodiment discloses a method for optimizing the control of adjusting the pitch angle of a wind turbine generator set. The method is implemented by a processor calling the data monitoring module, strategy analysis module, pitch angle control module, safety margin assessment module, and threshold iterative optimization module in the control system for optimizing the pitch angle of a wind turbine generator set described in Example 1.

[0070] Example 3

[0071] This embodiment discloses a non-transitory computer-readable medium storing instructions. When the instructions are executed by a processor, the method for optimizing the control of adjusting the pitch angle of a wind turbine generator set according to embodiment 2 is executed.

[0072] The non-transitory computer-readable medium in this embodiment can be a disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), a USB flash drive, a mobile hard disk, or other media.

[0073] Example 4

[0074] This embodiment discloses a computing device, including a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the method for optimizing the control of adjusting the pitch angle of a wind turbine generator set described in Example 2 is implemented.

[0075] The computing device described in this embodiment may be a desktop computer, a laptop computer, a smart phone, a PDA handheld terminal, a tablet computer, a programmable logic controller (PLC), or other terminal devices with a processor function.

[0076] The embodiments described above are only preferred embodiments of the present invention and are not intended to limit the scope of implementation of the present invention. Therefore, any changes made based on the shape and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A wind turbine generator set pitch angle optimization control system, characterized in that: include: Data monitoring module, used to collect SCADA data of wind turbine generator sets at the current optimal power generation pitch angle; The strategy analysis module is used to divide power into bins and set dynamic thresholds. It independently collects statistics on the operating data in each power bin and determines whether the clearance distance of each power bin falls into the high-risk power range or the high-redundancy power range. The pitch angle control module adjusts the pitch angle based on the judgment results provided by the strategy analysis module, and ultimately adjusts the clearance distance of each power bin to remove high-risk power segments or high-redundancy power segments; The safety margin assessment module constructs an input feature set based on historical SCADA data and outputs a clearance distance prediction value. It uses machine learning and gradient descent optimization loss function to optimize the prediction results, and predicts the clearance change trend under different pitch angle adjustment strategies within a preset time period in real time, and selects an adjustment scheme that maintains the clearance distance within the preset range. The threshold iterative optimization module defines the clearance threshold for triggering blade retraction and the shutdown protection threshold. When the wind turbine generator set has not triggered blade retraction for many consecutive days and the average clearance of each power bin is greater than the preset threshold, the margin compression strategy is executed, and iterative optimization is performed to converge the threshold. When the shutdown protection threshold is triggered, the wind turbine generator set's safety protection system is directly activated.

2. The wind turbine generator set pitch angle optimization control system according to claim 1, characterized in that: The SCADA data of the wind turbine generator set includes power, wind speed and blade clearance distance.

3. The wind turbine generator set pitch angle optimization control system according to claim 1, characterized in that: The strategy analysis module includes: Refined power binning: With 100kW power bins as the interval, the operating data in each power bin is independently counted. The statistical data includes real-time clearance distance D, corresponding pitch angle β, wind speed V, and blade root load Fload. Perform normal distribution or Weibull distribution fitting on the headroom data in each power bin, calculate the mean μ and standard deviation σ, generate the headroom probability density curve f(D), and set the initial safety threshold to [μ-2σ, μ+2σ] or [μ-3σ, μ+3σ]; Set the lower limit of the clearance distance D min =μ-kσ, k=2 or 3, which is the critical value for triggering propeller retraction; Set the clearance distance upper limit D max =μ+kσ is the optimized boundary value of propeller opening; if the clearance distance D in the corresponding power bin is <D min If the data of exceeds 5%, it is determined that the clearance distance of the power bin falls into the high-risk power range; if the clearance distance D>D max If the proportion exceeds 10%, it is determined that the clearance distance of the power bin falls into the high-redundancy power segment.

4. The wind turbine generator set pitch angle optimization control system according to claim 1, characterized in that: The pitch angle control module includes: For the power compartment that falls into the high-risk power range, the pitch angle is increased by Δβ = +k*0.5° each time, where k is the adjustment coefficient, to reduce the lift of the blade and reduce the deformation until the clearance distance D in the power compartment is <D min The data accounted for ≤2%; For the power bin that falls into the high-redundancy power range, the pitch angle Δβ is reduced by -k*0.5° each time to increase the blade lift and increase the power generation until the clearance distance D>D max The data account for ≤5%.

5. The wind turbine generator set pitch angle optimization control system according to claim 1, characterized in that: The safety margin assessment module includes: Using historical SCADA data, construct the input feature set {wind speed V, pitch angle β, clearance temp, generator power}, and output the clearance distance prediction value Use LSTM neural network or random forest regression model and gradient descent to optimize the loss function Where N is the number of samples, is the model’s predicted value of the clearance distance for the i-th sample, D i is the true value of the clearance distance of the i-th sample; the square of the difference between the predicted value and the true value of each sample is calculated by this function, and then the average value is calculated. The prediction is optimized by minimizing the average value; at the same time, the model parameters are updated every 10 minutes, and the clearance change trend under different pitch angle adjustment schemes in the next 10 minutes is predicted in real time. The clearance is kept at [D min +0.5m,D max -0.5m] interval adjustment strategy.

6. The wind turbine generator set pitch angle optimization control system according to claim 1, characterized in that: The threshold iteration optimization module includes: Define the clearance threshold for retracting the propellers A1 = D min And the shutdown protection threshold A2 = A1-2m; use the margin compression strategy, that is, when the wind turbine generator set does not trigger the blade retraction for 7 consecutive days and the average clearance value of each power bin μ> A1+0.2m, execute: A. Lower A1 by 0.1m and A2 by 0.1m simultaneously; B. Re-fit all power bins, return to the strategy analysis module and the pitch angle control module to adjust the optimal power generation pitch angle; C. The wind turbine is in a normal power generation state and is monitored for 48 hours after each round of adjustment. If the number of blade retractions is ≤3 times / day and the load does not exceed the design value, it enters the next round of compression until two consecutive blade retraction triggers occur with an interval of less than 1 hour. The last valid threshold combination is retained to obtain the iteratively optimized threshold.

7. A method for optimizing the pitch angle of a wind turbine generator set, characterized in that: The method is implemented by a processor calling the data monitoring module, strategy analysis module, pitch angle control module, safety margin assessment module and threshold iteration optimization module in the wind turbine pitch angle optimization control system described in any one of claims 1-6.

8. A non-transitory computer-readable medium storing instructions, characterized in that: When the instructions are executed by a processor, the method for optimizing the control of adjusting the pitch angle of a wind turbine generator set according to claim 7 is executed.

9. A computing device comprising a processor and a memory for storing a program executable by the processor, characterized in that: When the processor executes the program stored in the memory, the method for optimizing the control of adjusting the pitch angle of the wind turbine generator set as claimed in claim 7 is implemented.

Images