Low-wind-speed environment wind power efficient power generation capturing control method

Abstract

The application discloses a kind of low wind speed environment wind power efficient power generation capture control methods, it is related to wind power generation technical field, first identification real-time incoming flow wind speed belongs to interval, when judging as low wind speed working condition, synchronous acquisition local unit operating parameter and the wake operating parameter of adjacent unit in upstream, dynamic matching optimal tip speed ratio interval corresponding to current working condition, synchronous adjustment impeller pitch angle and generator electromagnetic torque, maintain tip speed ratio stable in optimal interval, synchronous superposition transmission chain damping loss torque compensation, reduce the energy consumption of unit itself operation.According to the coverage ratio of upstream wake to the wind-swept surface of the local unit, adjust the yaw angle.When the wind speed is lower than the cut-in wind speed, maintain the low speed of the impeller standby.The method effectively improves the wind energy capture efficiency at low wind speed, reduces unnecessary variable pitch, yaw adjustment energy consumption, avoids the wind energy loss caused by upstream wake interference, solves the problem that the existing control scheme does not consider the influence of wake, the high proportion of energy consumption.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, and in particular to a method for efficient wind power generation capture and control in low wind speed environments. Background Technology

[0002] Low-wind-speed wind resources refer to exploitable wind resources with an annual average wind speed of 3 to 5 m / s, accounting for more than 60% of the total exploitable wind resources in my country's land area. This represents a core strategic direction for the industry to expand its development boundaries and reduce the cost per kilowatt-hour in the current era of grid parity for wind power. The wind energy capture efficiency of wind turbines under low-wind-speed conditions directly determines the investment return of these wind farms. The industry's core need is to maximize wind energy capture in low-wind-speed conditions while minimizing the turbine's own operating energy consumption and improving overall net power generation revenue.

[0003] Current mainstream low-wind-speed power generation control schemes primarily employ fixed optimal tip speed ratio control. The working principle involves pre-calibrating the optimal tip speed ratio range for the corresponding turbine model in the low-wind-speed range through bench tests. During operation, real-time incoming wind speeds are collected, and the impeller pitch angle and generator electromagnetic torque are adjusted accordingly to maintain the tip speed ratio within the calibrated range, achieving maximum wind energy capture. This scheme has simple control logic and high operational stability, and has been widely applied to mass-produced grid-connected wind turbine units. However, this scheme does not consider the attenuation effect of the wake of upstream adjacent units on incoming wind energy at low wind speeds, nor does it optimize the control logic for the characteristics of high transmission chain losses and high energy consumption of the pitch system at low wind speeds. The additional energy consumption offsets some of the power generation gain, and the actual capture efficiency only reaches about 70% of the theoretical optimal value.

[0004] Another mainstream auxiliary control scheme is active yaw wind control. Its working principle is based on data collected by the anemometer on the top of the nacelle, adjusting the unit's yaw angle in real time to ensure the rotor's swept surface faces the incoming flow direction, reducing wind energy loss due to yaw error. This scheme can reduce yaw losses under normal operating conditions, but under low wind speed conditions, it does not consider the dynamic offset characteristics of the upstream unit's wake, adjusting yaw only based on the local incoming wind direction. This easily causes the rotor's swept surface to fall into the low-speed loss zone of the upstream wake, further reducing wind energy capture. Furthermore, it does not optimize the energy consumption-benefit ratio of yaw adjustment under low wind speeds, resulting in situations where adjustment energy consumption exceeds power generation gain.

[0005] In addition, both existing solutions shut down the turbine when the incoming wind speed is lower than the turbine's cut-in wind speed, requiring it to overcome the static friction torque of the stationary impeller to restart when the wind speed increases. This not only consumes extra energy but also misses short-term wind power generation windows, further reducing the power generation duration under low wind speed conditions. To address these issues, the industry urgently needs a dedicated power capture and control method adapted to low wind speed environments to systematically resolve the shortcomings of existing solutions and improve the overall benefits of low-wind-speed wind farms. Summary of the Invention

[0006] This invention proposes a high-efficiency wind power capture and control method for low-wind-speed environments to solve the problems mentioned in the prior art, such as low wind energy capture efficiency, lack of adaptation to wake interference, and high standby and auxiliary energy consumption.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a method for capturing and controlling wind power for high-efficiency generation in low-wind-speed environments, comprising the following steps: S1. The incoming wind speed and direction of the target wind turbine are collected in real time by the redundant laser wind radar and cup-type anemometer deployed on the top of the nacelle. The real-time rotational speed of the rotor is collected by the main shaft encoder and the real-time output power of the generator is collected by the grid-connected power transmitter. Based on the collected operating parameters, it is determined whether the target wind turbine has entered the preset low wind speed condition. The preset low wind speed condition corresponds to an incoming wind speed between 30% of the wind turbine's cut-in wind speed and the rated wind speed. S2. If it is determined that a preset low wind speed condition has been entered, retrieve the pre-stored correspondence between the incoming wind speed and the optimal tip speed ratio, which has been calibrated through wind tunnel testing and on-site operation. Dynamically adjust the blade pitch angle of the target wind turbine to the optimal blade pitch angle range of -1° to 2° for low wind speed. At the same time, adjust the electromagnetic torque of the generator to the optimal value that matches the current incoming wind speed and blade speed. Maintain the tip speed ratio within the preset optimal range throughout the process. The preset optimal range is calibrated based on the aerodynamic characteristics of the turbine blade, with an error not exceeding ±5%. S3. Obtain wake offset data of adjacent wind turbines within a range of 3 to 7 times the rotor diameter upstream of the target wind turbine in the prevailing wind direction. The wake offset data includes at least three types of parameters: wake lateral offset distance, axial velocity loss rate, and swept wind coverage range. Based on the wake offset data, adjust the yaw angle of the target wind turbine in real time to counteract the interference of the wake of the upstream adjacent wind turbines on the current turbine's low-wind-speed downstream wind energy. S4. At a preset verification cycle of 10 to 15 minutes, calculate the deviation between the actual average output power of the generator and the theoretical optimal output power under the same incoming wind speed within the current verification cycle. The theoretical optimal output power is based on the Bates limit combined with the unit's energy conversion efficiency calibration. If the deviation exceeds the preset allowable deviation range, iteratively correct the corresponding control parameters of the pitch angle, electromagnetic torque, and yaw angle.

[0008] Preferably, the specific rule for determining the entry into the preset low wind speed condition in step S1 is as follows: the incoming wind speed is continuously collected for no less than three preset collection cycles of 1 to 3 seconds and is within the preset low wind speed threshold range. Abnormal data with a single cycle wind speed deviation of more than 20% from the average value of the preceding and following cycles are removed during the collection process. When no warning signals for extreme turbulence, thunderstorms, or strong gusts are received from the wind farm SCADA system, the target wind turbine is determined to have entered the preset low wind speed condition.

[0009] Preferably, in step S2, during the dynamic adjustment of the impeller pitch angle, the real-time rate of change of the incoming wind speed is calculated based on the average slope of the wind speed change over 10 consecutive acquisition cycles. The preset rate of change threshold is set to 0.2 m / s / second. When the rate of change of the incoming wind speed is higher than the preset rate of change threshold, the pitch angle adjustment step size is reduced from the default 0.5° to 0.1° each time. After each pitch angle adjustment is completed, the next adjustment is performed after an interval of 2 acquisition cycles to avoid overshooting of the pitch angle adjustment.

[0010] Preferably, in step S2, during the process of adjusting the generator electromagnetic torque, a pre-established mapping table of incoming wind speed, impeller speed and optimal electromagnetic torque is called to match the corresponding electromagnetic torque. The mapping table is calibrated by combining the actual operating data of the unit at the factory bench test with more than 12 months of field operation data. At the same time, a preset torque compensation coefficient of 1.02 to 1.05 is superimposed to offset the damping loss of the unit's transmission chain and the friction loss of the bearings, ensuring that the electromagnetic torque adjustment error does not exceed 1%.

[0011] Preferably, the wake offset data in step S3 is obtained by retrieving the real-time pitch angle, speed, output power, and yaw angle operation data stored in the SCADA system of the upstream adjacent wind turbine, and combining it with the digital twin wake simulation model pre-constructed by the wind farm and trained based on the Lagrangian particle tracking method and terrain roughness parameters, to simulate the wake offset direction and wake influence range of the upstream adjacent wind turbine, with the wake data update frequency not less than 1 time / minute.

[0012] Preferably, the specific rule for adjusting the yaw angle in step S3 is as follows: when the proportion of the target wind turbine rotor swept area covered by the upstream wake exceeds a preset threshold of 20%, the yaw angle is adjusted so that the rotor swept surface deviates from the wake influence area. The adjustment range is between 0.5° and 5°. During the adjustment process, the tip speed ratio is monitored in real time. If the tip speed ratio deviates from the preset optimal range by more than 5%, the yaw adjustment is paused, and the electromagnetic torque is adjusted first to bring the tip speed ratio back to the optimal range before continuing to complete the yaw adjustment.

[0013] Preferably, the specific rule for triggering the iterative correction of control parameters in step S4 is as follows: when the average deviation value calculated for three consecutive preset verification cycles exceeds the preset allowable deviation range, and external interference factors such as sudden changes in wind speed and abnormal fluctuations in wake are excluded, the iterative correction process of control parameters is triggered. In each verification cycle, no less than 100 sets of valid operating data are collected to calculate the average output power, and invalid data during unit faults and power-limited operation periods are eliminated.

[0014] Preferably, in step S4, when iteratively correcting the control parameters, a particle swarm optimization algorithm is used to optimize three types of control parameters: pitch angle adjustment coefficient, electromagnetic torque compensation coefficient, and yaw adjustment step size. The particle swarm size is set to 20, the upper limit of the number of iterations is 50, and the iteration termination condition is that the deviation between the actual output power of the generator and the theoretical optimal output power is within the preset allowable deviation range, or the upper limit of the number of iterations is reached. The corrected control parameters are automatically updated to the stored mapping table.

[0015] Preferably, the method further includes: when it is determined that the target wind turbine has entered a preset low wind speed operating condition, the three-way pitch drive redundancy backup function activated by the turbine in the operating condition above the rated wind speed is turned off, only one pitch drive path is kept in normal operation, and the other two pitch drive paths are switched to a low power standby state. At the same time, the gust prediction redundancy adjustment function of the pitch system is turned off, and only the pitch angle adjustment command output by this control method is responded to, thereby reducing the standby energy consumption of the pitch system.

[0016] Preferably, the method further includes: when the incoming wind speed is detected to be lower than the wind turbine cut-in wind speed threshold of 2.5 m / s to 3 m / s, controlling the generator to output a very small excitation torque to keep the impeller in an idling state, and setting the preset standby speed range to 2% to 5% of the rated speed to avoid the impeller from standing still. When the wind speed rises to above the cut-in wind speed, it can quickly increase to the speed required for power generation without overcoming the static friction torque of the stationary impeller, thereby reducing energy loss during the start-up process.

[0017] Compared with existing technologies, the beneficial effects of this invention are: This invention effectively solves the problems of low wind energy capture efficiency and lack of optimization control for unit losses under low wind speed conditions by dynamically matching the optimal pitch angle and electromagnetic torque to maintain the tip speed ratio in the optimal range, while superimposing transmission chain damping loss torque compensation technology. It significantly improves the energy conversion efficiency of the unit under low wind speed conditions. At the same time, it avoids overshoot by dynamically adjusting the pitch adjustment step size according to the wind speed change rate, reducing unnecessary pitch action energy consumption, and achieving dual optimization of wind energy capture and energy consumption control.

[0018] This invention effectively solves the problem that existing active yaw control does not consider the dynamic offset of the upstream wake and easily causes the impeller to fall into the low-speed loss zone of the wake by introducing wake offset data from adjacent upstream units to dynamically adjust the yaw angle. At the same time, it prioritizes maintaining the stability of the tip speed ratio during yaw adjustment and only triggers yaw adjustment when the wake coverage area exceeds the threshold, avoiding meaningless yaw energy consumption, maximizing the benefits of yaw adjustment, and further reducing wind energy loss caused by wake interference.

[0019] This invention effectively solves the problems of high start-up losses and missed small wind power generation windows caused by stopping the turbine below the wind speed in the existing control scheme, as well as the high proportion of standby energy consumption of the pitch redundancy system, by maintaining the rotor at low speed standby when the wind speed is below the cut-off speed and closing the pitch redundancy adjustment path under low wind speed conditions. It effectively extends the power generation time under low wind speed and reduces the operating energy consumption of the unit's auxiliary system.

[0020] This method is compatible with the control architecture of existing mainstream onshore low-wind-speed wind turbines. It requires no additional hardware equipment and can be deployed simply by iterating the control logic. It can be adapted to the factory configuration of newly built low-wind-speed wind farms, and can also be used to upgrade existing units through remote updates of the wind farm's SCADA system. It is suitable for various low-wind-speed wind farm application scenarios in plains and hilly terrains, and can effectively improve the overall power generation revenue of low-wind-speed wind farms. It has extremely high value for promotion and application. Attached Figure Description

[0021] Figure 1 This is the overall flowchart of the low-wind-speed power generation capture and control proposed in this invention; Figure 2 This is the logic diagram for low wind speed condition determination and optimal capture control proposed in this invention; Figure 3 This is a flowchart of the wake drift analysis and yaw coordinated adjustment proposed in this invention; Figure 4 This is a block diagram of power deviation verification and control parameter optimization proposed in this invention; Figure 5 This is the logic diagram for low wind speed energy-saving standby and fast start-up proposed in this invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Reference Figures 1 to 5This invention discloses a method for capturing and controlling wind power for high-efficiency generation in low-wind-speed environments. The specific implementation process is as follows: First, perform parameter acquisition and condition determination in step S1: The laser wind radar is redundantly deployed at the front end of the top of the nacelle, 0.8 meters away from the windward side. It adopts pulse coherent radar with a sampling frequency of 10Hz and can detect the incoming wind speed field within a range of 100 meters in front. A cup-type anemometer is positioned 1.5 meters behind the radar to avoid interference from the radar bracket affecting measurement accuracy, with a sampling frequency of 1Hz. A 17-bit absolute spindle encoder is installed on the bearing housing at the spindle drive end, outputting impeller angle signals with an accuracy of 0.01°, which can be converted into the real-time impeller speed. A 0.2-class high-precision power transmitter is installed on the incoming side of the unit's grid-connected cabinet to collect real-time active power data from the generator. All collected data is transmitted to the main PLC controller via the unit's internal Profinet bus. The controller pre-stores the current unit's cut-in wind speed and rated wind speed parameters. The preset low-wind-speed operating range is 30% of the cut-in wind speed to the rated wind speed. For example, for a 1.5MW doubly-fed generator unit with a cut-in wind speed of 3m / s and a rated wind speed of 12m / s, the corresponding low-wind-speed operating range is 3m / s to 3.6m / s. The parameters can be flexibly configured according to the actual factory calibration values ​​of the unit. Step S2 involves stabilizing the tip speed ratio: the pre-stored correspondence between the incoming wind speed and the optimal tip speed ratio is initially calibrated through a boundary layer wind tunnel test with a 1:20 scale impeller. The tip speed ratio sweep test is simulated under different low wind speeds in the wind tunnel, and the change curve of the wind energy utilization coefficient Cp is recorded. The tip speed ratio with a fluctuation of 5% above or below the maximum value of Cp is taken as the preset optimal range. Then, a second calibration is completed using 6 months of trial operation data from the unit to ensure that the range is adapted to the actual wind conditions on site.

[0024] The optimal pitch angle range for low wind speeds is set to -1° to 2°. This range corresponds to the pitch angle range where the Cp value is maintained above 95% of its peak value under low wind speeds during wind tunnel testing. When adjusting the electromagnetic torque, maintaining the tip speed ratio within the optimal range is the core objective. The real-time calculation formula for the tip speed ratio is: in This is the real-time tip speed ratio. This represents the real-time angular velocity of the impeller, in rad / s. The impeller radius is in meters (m). The real-time incoming air velocity is expressed in m / s.

[0025] When the tip speed ratio is below the lower limit of the optimal range, the generator electromagnetic torque is reduced to increase the impeller speed. When the tip speed ratio is above the upper limit of the optimal range, the generator electromagnetic torque is increased to decrease the impeller speed. The control error is kept within ±5% throughout the process. Step S3, wake interference cancellation control, is then implemented: the upstream adjacent unit is determined within a range of 3 to 7 times the impeller diameter in the prevailing wind direction. For example, a unit with an impeller diameter of 120m corresponds to units in the same row upstream within a range of 360m to 840m. The wake offset data includes three parameters: the wake lateral offset distance (the vertical distance from the wake center to the prevailing wind direction), the axial velocity loss rate (the ratio of the difference between the average wind speed in the wake region and the free-flow wind speed to the free-flow wind speed), and the swept coverage area (the percentage of overlap between the wake region and the swept surface of the target unit's impeller). The yaw angle is adjusted according to the wake coverage percentage to ensure the impeller swept surface avoids the low-speed region of the wake, reducing wind energy loss caused by the wake. Step S4 involves iterative calibration of control parameters: the preset calibration period is set to 12 minutes, which meets the requirement of 10-15 minutes. After each calibration period, the deviation between the actual average output power of the generator and the theoretical optimal output power is calculated. The formula for calculating the theoretical optimal output power is: ;in The theoretically optimal output power, in watts (W). The real-time air density at the site can be calculated using temperature, pressure, and humidity sensors installed in the cabin, with units of kg / m³. 3 ; The impeller swept area is expressed in m². 2 ; The maximum wind energy utilization factor calibrated for the unit is taken as 0.47 to 0.49; The average incoming air velocity during the verification period, in m / s; The total energy conversion efficiency of the drive train, generator, and converter is set to 0.88–0.92. The preset allowable deviation range is 8%. If the deviation exceeds this range, the parameter iterative correction process will be initiated.

[0026] This invention also discloses specific rules for determining low wind speed conditions, with the following specific implementation details: The preset acquisition period is set to 2 seconds, which meets the requirement of 1 to 3 seconds. The incoming wind speed is continuously acquired for 3 acquisition periods. If all effective wind speeds are within the preset low wind speed threshold range, and no level 2 or above extreme turbulence, thunderstorm, or strong gust warning signals are received from the wind field SCADA system, the low wind speed condition is determined to be in effect.

[0027] The rule for removing abnormal data is as follows: if the wind speed in a single cycle deviates from the average wind speed of the preceding and following two cycles by more than 20%, it is considered abnormal data and removed. Three consecutive cycles of valid data are then collected to complete the judgment, avoiding misjudgments caused by gusts or sensor fluctuations. The code snippet for the core judgment logic is as follows: def low_wind_judge(wind_seq, scada_warning, cut_in_wind, rated_wind):low_wind_thr = rated_wind * 0.3 # Eliminate abnormal data valid_wind = [] for i inrange(1, len(wind_seq)-1): mean_around = (wind_seq[i-1]+ wind_seq[i+1]) / 2 ifabs(wind_seq[i]- mean_around) / mean_around<= 0.2: valid_wind.append(wind_seq[i]) # Determination condition if len(valid_wind)>=3 and all([cut_in_wind<= w<= low_wind_thrfor w in valid_wind]) and not scada_warning: return True else: return False This invention also discloses a method for dynamically adjusting the pitch angle under low wind speeds, with the specific implementation details as follows: When dynamically adjusting the pitch angle, wind speed data is collected for 10 consecutive 2-second collection cycles, the slope of wind speed change between every two adjacent cycles is calculated, and the average value of the 10 slopes is taken as the real-time rate of change of the incoming wind speed. The preset rate of change threshold is 0.2 m / s / second.

[0028] When the real-time rate of change is higher than the threshold, the pitch angle adjustment step size is reduced from the default 0.5° / time to 0.1° / time. After each adjustment is completed, there is an interval of 2 acquisition cycles, or 4 seconds, before the next adjustment is executed. This avoids overshooting that could cause the pitch angle to deviate from the optimal range, while also reducing the losses from frequent start-stop of the pitch motor.

[0029] This invention also discloses a precise matching control method for electromagnetic torque, with the specific implementation details as follows: When adjusting the electromagnetic torque, a pre-stored three-dimensional mapping table is called to match the optimal electromagnetic torque corresponding to the current incoming wind speed and impeller speed. This mapping table is initially calibrated through a 1000-hour continuous operation test on the generator set at the factory. Then, combined with more than 12 months of actual operating data from the field, invalid data from faults, power limitations, and extreme operating conditions are eliminated, and a second calibration is completed by fitting using the least squares method.

[0030] After obtaining the basic electromagnetic torque, a preset torque compensation coefficient of 1.02 to 1.05 is superimposed. In plain areas where the transmission chain wear is relatively low, a coefficient of 1.02 is used; in hilly areas where vibration is greater and wear is higher, a coefficient of 1.05 is used. The compensation torque is used to offset transmission chain damping losses, bearing friction losses, and gearbox meshing losses. The final adjustment error of the electromagnetic torque is controlled within 1%. The formula for calculating the compensated target electromagnetic torque is as follows: ; in The target electromagnetic torque is the final output, and the base electromagnetic torque is obtained by matching the mapping table. This is the preset torque compensation coefficient.

[0031] This invention also discloses a method for obtaining upstream wake offset data. The specific implementation details are as follows: The wake offset data is calculated by a digital twin wake simulation model pre-constructed by the wind field. This model is built based on the Lagrangian particle tracking method. The input parameters include the real-time pitch angle, rotational speed, output power, and yaw angle of the upstream adjacent units, as well as the real-time terrain roughness, incoming wind speed and direction. The output parameters are the wake lateral offset distance, axial velocity deficit rate, and swept wind coverage.

[0032] The model was pre-trained and labeled using 1,000 sets of wake data measured by LiDAR, achieving a simulation accuracy of over 92%. The wake data update frequency was set to 45 seconds per update, meeting the requirement of no less than once per minute. The data was transmitted to the target unit controller via the industrial Ethernet bus of the wind farm SCADA system, with a transmission delay of no more than 100ms.

[0033] In this invention, a dynamic adjustment rule for the yaw angle against the wind is also disclosed, and the specific implementation details are as follows: when the swept coverage ratio of the upstream wake exceeds a preset threshold of 20%, yaw adjustment is initiated, and the adjustment range is linearly matched with the coverage ratio. When the coverage ratio is 20%, the adjustment range is 0.5°, and when the coverage ratio is 100%, the adjustment range is 5°.

[0034] During the adjustment process, the real-time tip speed ratio is checked every 0.5° adjustment. If the tip speed ratio deviates from the preset optimal range by more than 5%, the yaw adjustment is immediately paused. The generator electromagnetic torque is adjusted first to bring the tip speed ratio back to the optimal range before the remaining yaw adjustment operations are continued. This avoids the problem of reduced wind energy capture efficiency during the yaw adjustment process.

[0035] This invention also discloses the triggering rules for control parameter correction, and the specific implementation details are as follows: When the average deviation value calculated by three consecutive 12-minute verification cycles exceeds the preset allowable deviation range of 8%, the external interference investigation process is initiated. If the investigation confirms that there are no external interference factors such as sudden wind speed changes exceeding 1 m / s within 10 seconds or abnormal wake fluctuations caused by upstream unit cut-out, the iterative correction process of the control parameters is formally triggered.

[0036] In each verification cycle, no fewer than 100 sets of valid operating data are collected to calculate the average output power. Invalid data during periods when the unit fault code is non-zero or the power limit command is greater than 0 are removed to ensure the accuracy of the deviation calculation.

[0037] This invention also discloses an iterative optimization method for control parameters. The specific implementation details are as follows: When iteratively correcting the control parameters, a particle swarm optimization algorithm is used to optimize three types of control parameters: pitch angle adjustment coefficient, electromagnetic torque compensation coefficient, and yaw adjustment step size. The particle swarm size is set to 20, with each particle being a three-dimensional vector corresponding to the three types of parameters to be optimized. The inertia weight is 0.7, and both the individual learning factor and the global learning factor are 1.49. The maximum number of iterations is 50, and the fitness function is set to maximize the ratio of actual output power to theoretically optimal output power.

[0038] The iteration terminates when the deviation value corresponding to the fitness function falls within the allowable deviation range of 8%, or when the maximum number of iterations (50) is reached. The corrected optimal parameters are automatically updated to the stored mapping table, and the new parameters are directly called to execute control in the next verification cycle. The code snippet of the core optimization logic is as follows: def pso_optimize(init_params, wind_data, power_data, P_opt, max_iter=50, pop_size=20): # Initialize the particle swarm pop = np.random.uniform(low=[0.8, 1.0, 0.1], high=[1.2, 1.1, 0.5], size=(pop_size, 3)) vel = np.zeros((pop_size, 3)) p_best = pop.copy() p_best_fit = np.array([fitness(p, wind_data, power_data, P_opt) for p in pop]) g_best = p_best[np.argmax(p_best_fit)] g_best_fit = np.max(p_best_fit) # Iterative optimization for i in range(max_iter): r1, r2 = np.random.rand(2) vel = 0.7 * vel + 1.49 * r1 * (p_best - pop) + 1.49 * r2 * (g_best - pop) pop = pop + vel # Boundary constraint pop = np.clip(pop, [0.8, 1.0, 0.1], [1.2, 1.1, 0.5]) # Update the optimal value current_fit = np.array([fitness(p, wind_data, power_data, P_opt) for p in pop]) update_idx = current_fit > p_best_fit p_best[update_idx] = pop[update_idx] p_best_fit[update_idx] = current_fit[update_idx] if np.max(p_best_fit) > g_best_fit: g_best = p_best[np.argmax(p_best_fit)] g_best_fit = np.max(p_best_fit) if 1 - g_best_fit <= 0.08: break return g_best This invention also discloses a method for optimizing the energy consumption of a pitch system under low wind speeds, with the specific implementation details as follows: When it is determined that the low wind speed condition has been entered, the redundant backup function of the three pitch drive that is enabled when the unit is operating above the rated wind speed is turned off. This redundancy function is that the three blade pitch drives are hot backups of each other. If any one of them fails, the other two can take over immediately. Under low wind speeds, the pitch angle adjustment range is small and the load is low, so the risk of failure is extremely low. Therefore, only one pitch drive path is kept in normal operation, and the power supply of the other two pitch drive paths is cut off. Only the control power supply is kept in a low-power standby state, and the standby power consumption of a single driver is reduced from 150W to 10W.

[0039] At the same time, the gust prediction redundancy adjustment function of the pitch system is turned off. This function is a pre-adjustment function that responds to gusts in advance under rated operating conditions. At low wind speeds, the energy density of gusts is low and there is no need for advance adjustment. It only responds to the pitch angle adjustment command output by this control method, which can ultimately reduce the overall energy consumption of the pitch system by more than 40%.

[0040] This invention also discloses a method for impeller standby control at ultra-low speeds, with specific implementation details as follows: When the incoming wind speed is detected to be below the cut-in wind speed threshold of 3 m / s, the generator is controlled to output a minimum excitation torque of 1% to 2% of its rated torque. This torque is output through the converter excitation control circuit and can be achieved without grid connection. It is only used to counteract the static friction of the bearings, keeping the impeller in an idling state. The preset standby speed range is 2% to 5% of the rated speed. For example, the rated speed of a 2MW direct-drive unit is 12 rpm, corresponding to a standby speed of 0.24 rpm to 0.6 rpm, to prevent the impeller from coming to a complete stop. When the wind speed rises above the cut-in wind speed, it is not necessary to overcome the maximum static friction torque of the impeller when it is stationary. Only the electromagnetic torque needs to be adjusted to increase the speed to the required power generation speed within 3 seconds, reducing energy loss during startup. At the same time, it can capture short-term small wind windows of less than 30 seconds, increasing the power generation duration.

[0041] Scenario Example 1: Application of a 1.5MW Doubly Fed Generator in a Low-Wind-Speed ​​Wind Farm in a Plain The scenario is located in the northern plains of Anhui Province. The wind farm has 32 1.5MW doubly-fed wind turbine units with a rotor diameter of 121m. The cut-in wind speed is 3m / s, the rated wind speed is 12m / s, and the annual average wind speed is 4.2m / s. 75% of the power generation time is in the low wind speed range. The spacing between the units in the same row is 5 times the rotor diameter. The upstream wake interference is significant. Under the traditional control scheme, the wind energy utilization coefficient is only 0.42 at most. The energy consumption of the pitch system accounts for 8%, and the power generation loss caused by the wake is 11%.

[0042] No new hardware is required. Updates are completed remotely via the wind farm's SCADA system, which sends control programs to the main PLCs of each turbine. The terrain roughness parameter of the wake simulation model is set to 0.03m (typical value for plains), the torque compensation coefficient to 1.02, the yaw trigger coverage threshold to 20%, and the calibration cycle to 12 minutes. During turbine operation, data is first collected from multiple sensors to determine low-wind-speed conditions. The optimal tip speed ratio range is matched to adjust the pitch angle and electromagnetic torque, and transmission chain loss compensation is added. The digital twin model is used to obtain upstream wake data to adjust the yaw angle. Control parameters are calibrated every 12 minutes. Pitch redundancy is disabled at low wind speeds, and rotor idling is maintained at ultra-low speeds. After implementation, the wind energy utilization coefficient can reach a maximum of 0.48, pitch system energy consumption is reduced by 42%, wake power generation loss is reduced to 4.3%, and annual power generation is increased by approximately 8.7%.

[0043] Scenario Example 2: Application of 2MW direct-drive turbine in hilly low-wind-speed wind farms This scenario is located in the hilly region of northern Fujian. The wind farm has 24 2MW direct-drive wind turbines with a rotor diameter of 131m. The cut-in wind speed is 3m / s, the rated wind speed is 11m / s, and the annual average wind speed is 3.9m / s. 82% of the power generation time is in the low wind speed range. The significant terrain undulations result in a 22% higher wake offset compared to plain areas. Under traditional control schemes, the power generation rate during small wind windows is less than 14%, and the yaw misoperation rate reaches 17%. The control program is remotely updated, with the terrain roughness parameter of the wake simulation model set to 0.12m (typical value for hilly areas), the torque compensation coefficient set to 1.05, the yaw trigger coverage threshold set to 25%, and the verification cycle set to 10 minutes to avoid erroneous adjustments caused by wind direction fluctuations due to terrain. The operation process adds wake correction logic based on terrain roughness to the plain scenario. During parameter optimization, the yaw adjustment frequency is reduced first, and the ultra-low speed standby speed is set to 3% to 5% of the rated speed. After implementation, the yaw error rate dropped to 2.1%, the power generation time during small wind windows increased by 23%, the energy consumption of the pitch system decreased by 38%, and the annual power generation increased by approximately 9.4%.

[0044] refer to Figure 1 This diagram illustrates the entire lifecycle control path from environmental perception to closed-loop optimization. The process begins with real-time acquisition of multiple operating parameters such as wind speed, wind direction, engine speed, and power. Through logical judgment, the system identifies whether the unit has entered a preset low-wind-speed operating condition (entering a wind speed range of 30% of the rated wind speed). If the condition is met, it enters a three-dimensional integrated adjustment phase: first, it achieves maximum wind energy capture (TSR control) by optimizing the pitch angle and electromagnetic torque; second, it performs yaw-coordinated wind control by analyzing the wake deviation data of upstream units; and third, it performs periodic performance verification and parameter iterative optimization. Finally, in conjunction with low-power energy-saving mode and idling standby control, it ensures that the system maintains rapid response and high-efficiency output even at extremely low wind speeds.

[0045] refer to Figure 2 This diagram illustrates the specific logic behind low-wind speed identification and optimal wind energy capture. The system averages wind speed data from multiple consecutive acquisition cycles and automatically removes abnormal interference items with deviations exceeding 20%. It also performs secondary verification using extreme weather warnings from the SCADA system to ensure the accuracy of work order decisions. During adjustments, the system calculates the wind speed change rate in real time. If wind speed fluctuations are severe, it automatically switches to a fine-grained small-step adjustment mode (0.1-degree increments) to avoid mechanical overshoot. By coordinating the adjustment of the blade pitch angle and electromagnetic torque, the system can offset losses such as transmission chain damping, precisely locking the tip speed ratio within the optimal aerodynamic performance range.

[0046] refer to Figure 3 This figure highlights the shielding strategy for wake interference between adjacent turbine units. The system utilizes a digital twin simulation model, combined with the operating status of the upstream unit and terrain parameters, to calculate the wake offset distance and swept coverage area in real time. When the wake coverage area exceeds 20%, the system triggers active yaw adjustment to guide the rotor to avoid low-speed loss areas. During the adjustment process, a priority constraint mechanism is established: the tip speed ratio is constantly monitored; if yaw causes a rapid decrease in capture efficiency, electromagnetic torque compensation is adjusted first, and wind angle correction is completed only after energy capture stabilizes, achieving maximum wind energy utilization through multi-objective collaboration.

[0047] refer to Figure 4 This diagram illustrates the evolutionary path of the system's self-correction. Within each 10-15 minute verification cycle, the system compares the generator's actual average power with the theoretically optimal power calculated based on the Bates limit. To eliminate random interference, the system requires three consecutive cycles of abnormal deviation before triggering the optimization process. The core optimization stage employs a particle swarm optimization algorithm, using pitch angle coefficients, torque compensation coefficients, and other parameters as optimization variables to find the optimal solution that minimizes the power deviation within a 50-iteration limit. The optimization results automatically overwrite the original mapping table data, enabling the control strategy to adaptively adjust based on performance degradation due to the unit's service life.

[0048] refer to Figure 5 This diagram illustrates the system's auxiliary operation strategy. Upon confirming entry into low-wind-speed mode, the system employs a hardware load reduction strategy, shutting down unnecessary redundant pitch drive paths and retaining only a single path in operation, thereby directly reducing system self-power consumption. For extreme cases below the cut-in wind speed, the system does not perform full braking but instead utilizes minimal excitation torque to maintain the impeller at a micro-rotation state of 2% to 5% of its rated speed. This controlled idling mode allows the impeller to transition from static friction to dynamic friction, enabling the unit to quickly overcome the starting resistance torque and rapidly enter power generation mode once the incoming wind speed recovers, minimizing time and energy loss during start-up and shutdown.

[0049] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A low wind speed environment wind power efficient generation capture control method, characterized in that, Includes the following steps: S1. Real-time acquisition of the target wind turbine's incoming wind speed, incoming wind direction, real-time rotor speed, and real-time generator output power, and determination of whether the target wind turbine has entered the preset low wind speed condition; S2. If it is determined that the preset low wind speed condition has been entered, based on the real-time collected incoming wind speed and the pre-stored optimal tip speed ratio correspondence, the blade pitch angle of the target wind turbine is dynamically adjusted to the optimal pitch angle range for low wind speed. At the same time, the electromagnetic torque of the generator is adjusted to the optimal value that matches the current incoming wind speed and blade speed, so as to maintain the tip speed ratio in the preset optimal range. S3. Obtain wake offset data of adjacent wind turbines within a preset range upstream of the target wind turbine, and adjust the yaw angle of the target wind turbine in real time based on the wake offset data to counteract the interference of the wake of the upstream adjacent wind turbines on the downflow wind energy of the current turbine at low wind speed. S4. At each preset verification cycle, calculate the deviation between the actual output power of the generator and the theoretical optimal output power under the same incoming wind speed. If the deviation exceeds the preset allowable deviation range, iteratively correct the corresponding control parameters of the pitch angle, electromagnetic torque and yaw angle.

2. The low wind speed environment wind power efficient generation capturing control method according to claim 1, characterized in that, The specific rule for determining the entry into the preset low wind speed condition in step S1 is as follows: when the incoming wind speed is continuously collected for no less than three preset collection cycles and is within the preset low wind speed threshold range, and no extreme turbulence warning signal is received, the target wind turbine is determined to have entered the preset low wind speed condition.

3. The low wind speed environment wind power efficient generation capturing control method according to claim 1, characterized in that, In step S2, during the process of dynamically adjusting the impeller pitch angle, the adjustment step size of the pitch angle is adjusted based on the real-time rate of change of the incoming wind speed. When the rate of change of the incoming wind speed is higher than the preset rate of change threshold, the adjustment step size of the pitch angle is reduced.

4. The method for capturing and controlling wind power in low-wind-speed environments according to claim 1, characterized in that, In step S2, during the process of adjusting the generator electromagnetic torque, a pre-established mapping table of incoming wind speed, impeller speed and optimal electromagnetic torque is called to match the corresponding electromagnetic torque, and a preset torque compensation coefficient is superimposed to offset the damping loss of the unit's transmission chain.

5. The low wind speed environment wind power efficient generation capture control method according to claim 1, characterized in that, The method for obtaining wake offset data in step S3 is as follows: retrieve the SCADA operation data of the upstream adjacent wind turbine, and combine it with the pre-constructed digital twin wake simulation model of the wind farm to simulate the wake offset direction and wake influence range of the upstream adjacent wind turbine.

6. The low wind speed environment wind power efficient generation capturing control method according to claim 5, characterized in that, The specific rule for adjusting the yaw angle in step S3 is as follows: when the proportion of the target wind turbine rotor sweep area covered by the influence range of the upstream wake exceeds the preset proportion threshold, the yaw angle is adjusted so that the rotor sweep surface deviates from the wake influence area, and the tip speed ratio is maintained in the preset optimal range during the adjustment process.

7. The low wind speed environment wind power efficient generation capture control method according to claim 1, characterized in that, The specific rule for triggering the iterative correction of control parameters in step S4 is as follows: when the deviation values ​​calculated for three consecutive preset verification cycles all exceed the preset allowable deviation range, the iterative correction process of control parameters is triggered.

8. The low wind speed environment wind power efficient generation capturing control method according to claim 7, characterized in that, In step S4, when iteratively correcting the control parameters, a particle swarm optimization algorithm is used to optimize the control parameters of pitch angle, electromagnetic torque, and yaw angle. The iteration terminates when the deviation between the actual output power of the generator and the theoretical optimal output power is within a preset allowable deviation range.

9. The low wind speed environment wind power high efficiency generation capture control method of claim 1, wherein, The method further includes: when it is determined that the target wind turbine has entered a preset low wind speed condition, the active pitch redundancy adjustment function of the turbine is turned off, and only the single-channel pitch adjustment path is retained.

10. The low wind speed environment wind power efficient generation capture control method of claim 1, wherein, The method further includes: when the incoming wind speed is detected to be lower than the wind speed threshold of the wind turbine, controlling the impeller to maintain an idling state so that the impeller speed is within a preset standby speed range.

Images