Torque and speed cooperative based wind turbine rotor turning drive control method

Abstract

The application relates to the technical field of wind generator driving control, and discloses a wind generator rotor turning-over driving control method based on torque and speed cooperation, which comprises the following steps: controlling a motor to execute forward and reverse quasi-static micro-motion before operation to collect average electromagnetic torque; separating pure gravity and friction components by using torque arithmetic and difference, correcting gravity model gain and configuring friction compensation; combining real-time angle calculation to feed forward torque, superimposing friction compensation and speed loop feedback to generate current instruction; and inhibiting integral component when load torque and speed are in the same direction; the application utilizes physical symmetry to decouple gravity and friction, eliminates model parameter error, avoids integral lag and speed out-of-control hidden danger when a variable polarity load passes through a dead point, and ensures the stability and safety of a hundred-ton rotor turning-over process.

Description

Technical Field

[0001] This invention relates to a wind turbine rotor turning drive control method based on torque-speed coordination, belonging to the field of wind turbine drive control technology. Background Technology

[0002] In the current manufacturing process of large direct-drive or semi-direct-drive wind turbines, the rotor turning operation of hundreds of tons requires the drive system to smoothly turn the rotor from a horizontal position to a vertical position or to precisely adjust its attitude within a specific angle range. At present, the industry generally adopts a vector control-based variable frequency drive system, which realizes closed-loop control of motor speed and torque through cascaded PI regulators of speed loop and current loop. This control architecture performs stably when handling conventional constant torque loads or slowly changing loads, and is the standard configuration in the field of wind power equipment manufacturing.

[0003] However, the rotor turning process is characterized by large inertia, nonlinearity, and variable polarity gravitational potential energy load. When the rotor's center of gravity crosses the geometric vertex (dead point), the direction of the gravitational load instantly reverses from resisting rotation to dragging rotation. To address the challenge of controlling variable polarity loads, existing technologies attempt to adjust the output torque by monitoring the load position. For example, Chinese invention patent CN109973304B discloses a rotor rotation control system and control method for wind turbine generators. The solution uses a processor to determine the switching position of the bending moment load on the shaft and outputs adjustment commands to the hydraulic drive unit to adjust the pressure value of the telescopic cylinder or switch the push-pull state to balance the load change. At this time, the integral term of the conventional speed loop PI regulator has an inherent memory effect and desaturation delay, and cannot instantly reverse with the step change of the load torque. The motor still outputs a huge positive electromagnetic torque the instant the load becomes power. The two superimposed form a strong boosting torque, causing the rotor to accelerate uncontrollably and produce uncontrollable mechanical shock or runaway accidents. Existing technologies often increase the redundancy of the mechanical braking device, introduce high-cost torque sensors for direct torque control, or add complex online parameter identification and adaptive modules to the control algorithm.

[0004] Therefore, the technical problem to be solved by this invention is how to resolve the conflict between integral lag and torque response at the zero-crossing point of a variable polarity load in a conventional PI control strategy without increasing external hardware costs and system complexity, and to achieve smooth and safe rotation of a large inertia rotor throughout the entire process, especially in the dead-point region. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A wind turbine rotor turning drive control method based on torque-speed coordination, comprising the following steps:

[0006] The drive motor is controlled to perform forward micro-rotation and reverse micro-rotation actions respectively at the initial position before the start of the turning operation. The absolute values ​​of the angular velocities of the forward micro-rotation and reverse micro-rotation actions are equal and are in the quasi-static range.

[0007] The first average electromagnetic torque of the drive motor during the forward micro-rotation motion and the second average electromagnetic torque during the reverse micro-rotation motion are collected.

[0008] The arithmetic sum of the first average electromagnetic torque and the second average electromagnetic torque is calculated to extract the pure gravitational torque component. The gain coefficient of the gravitational torque mapping model is corrected using the pure gravitational torque component. The gravitational torque mapping model characterizes the functional relationship between the rotor's theoretical gravitational load torque and the rotor's mechanical angle.

[0009] The arithmetic difference between the first average electromagnetic torque and the second average electromagnetic torque is calculated to extract the system friction torque component, and an independent friction compensation channel is configured.

[0010] During the driving process, the actual rotor mechanical angle and actual speed of the drive motor are collected in real time; the feedforward compensation torque is calculated based on the actual rotor mechanical angle and the corrected gravity torque image model; and the friction compensation torque is generated based on the direction of the actual speed and the system friction torque component.

[0011] The speed regulator is used to process the speed deviation between the actual speed and the given speed to generate feedback adjustment torque; the feedforward compensation torque, friction compensation torque and feedback adjustment torque are vector-superimposed to generate the final q-axis current setpoint.

[0012] The control drive motor outputs electromagnetic torque according to the final q-axis current setpoint, and when the direction of the feedforward compensation torque is detected to be the same as the direction of the actual speed, the suppression operation of the integral component in the speed regulator is triggered.

[0013] Preferably, in the steps of calculating the arithmetic sum of the first average electromagnetic torque and the second average electromagnetic torque to extract the pure gravitational torque component and calculating the arithmetic difference between the first average electromagnetic torque and the second average electromagnetic torque to extract the system frictional torque component, the pure gravitational torque component... and system friction torque component The following relationship must be satisfied: ,in, The first average electromagnetic torque, The second average electromagnetic torque; the step of correcting the gain coefficient of the gravitational torque mapping model using the pure gravitational torque component includes using the pure gravitational torque component. The preset rotor mass and lever arm product parameters are calibrated and updated.

[0014] Preferably, the method further includes a backlash crossing soft handover control step, which calculates and predicts the rotor angle position corresponding to the zero-crossing point of the gravity load torque in real time based on the gravity torque mapping model, and sets the backlash crossing interval with the angle position as the center; when the actual rotor mechanical angle is detected to enter the backlash crossing interval, the torque change rate limiting logic is activated; the torque change rate limiting logic forcibly constrains the differential amplitude of the final q-axis current setpoint with respect to time to not exceed the preset soft handover slope threshold, until the actual rotor mechanical angle leaves the backlash crossing interval.

[0015] Preferably, the method further includes a dynamic damping injection step, wherein when the direction of the feedforward compensation torque is detected to be the same as the direction of the actual rotational speed, a reverse damping torque component is superimposed on the final q-axis current setpoint; the amplitude of the reverse damping torque component is positively correlated with the amplitude of the actual rotational speed, and is used to suppress speed overshoot.

[0016] Preferably, the triggering conditions for the suppression operation also include: the real-time load torque amplitude estimated by the disturbance observer is less than a preset safety threshold or the actual rotor mechanical angle is within a preset potential energy release angle range.

[0017] Preferably, the step of configuring an independent friction compensation channel specifically includes: establishing a symbolic function logic block that is only related to the rotational speed direction; using the extracted system frictional torque component as a fixed-gain input symbolic function logic block; during the driving process, the symbolic function logic block outputs a frictional compensation torque in the same direction as the current motion direction according to the real-time symbol of the actual rotational speed, so as to counteract the Coulomb friction of the mechanical transmission chain.

[0018] Preferably, the step of correcting the gain coefficient of the gravity torque mapping model using the pure gravity torque component specifically includes: obtaining the initial mechanical angle of the drive motor at the initial position; calculating the sine value of the initial mechanical angle, dividing the pure gravity torque component by the sine value to obtain the corrected maximum gravity torque amplitude; updating the amplitude parameter in the gravity torque mapping model to the corrected maximum gravity torque amplitude to ensure that the feedforward compensation torque matches the actual physical load across the entire angle range.

[0019] Preferably, the absolute values ​​of the angular velocities of the forward and reverse micro-rotation actions are set to be less than one percent of the rated speed of the drive motor. This ensures that the viscous friction component and inertial torque component of the drive motor can be ignored when collecting the first and second average electromagnetic torques, so that the collected torque data mainly reflects the gravitational torque and Coulomb friction torque.

[0020] Preferably, in the backlash crossing soft transition control step, the soft transition slope threshold is set based on the gear module of the wind turbine gearbox and the maximum allowable instantaneous impact load. The torque change rate limiting logic is used to convert the gear tooth surface contact process from rigid impact to flexible transition.

[0021] Preferably, the step of the speed regulator processing the speed deviation to generate feedback regulating torque specifically includes: inputting the speed deviation into the proportional-integral controller; when the suppression operation is not triggered, the proportional-integral controller calculates the output according to the preset proportional coefficient and integral coefficient; when the suppression operation is triggered, the integral term output of the proportional-integral controller is forcibly set to zero or kept at the current value and no longer accumulated, and the control of the drive motor is maintained only through the proportional term and the feedforward compensation torque.

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] 1. In wind turbines based on torque-speed coordination, a collaborative mechanism of model feedforward and integral state switching is used to address the risk of runaway control due to high inertia and variable polarity loads at dead points. A gravity torque image model is constructed as the dominant component of feedforward control. Combined with the zero-crossing region integral term suppression logic, the feedforward path directly outputs a reverse balancing torque that matches the amplitude of the gravity component based on the angle position, undertaking most of the load demand. When the feedback path detects the moment of load polarity reversal, it resets or freezes the integral component, and the damping injection logic takes over the system stability control. The dual-channel collaborative approach cuts off the lag effect of historical error accumulation in the integral link of the conventional speed regulator on the current control output, so that the motor electromagnetic torque follows the sign change of the gravity load with zero delay, eliminating the resultant force boost phenomenon caused by integral desaturation delay, and ensuring the controllability and stability of the rotor speed during the potential energy release stage of the 100-ton-level rotor.

[0024] 2. Based on the bidirectional micro-motion scanning parameter decoupling mechanism, adaptive calibration of the model without external sensors is achieved. Utilizing the path independence of conservative force work and the path correlation of dissipative force work, the drive motor performs a symmetrical quasi-static micro-motion scanning action before operation, algebraically separating the mixed load torque into pure gravity components and pure friction components. Relying on the motor's own current and position feedback signals, the gravity model correction coefficient, which is not affected by the friction of the mechanical transmission chain, and the friction benchmark reflecting the lubrication and wear state of the equipment are extracted from the total load. The corrected parameters are fed back to the feedforward model and an independent friction compensation channel is established to eliminate the deviation between theoretical CAD data and actual physical objects, ensure the physical fidelity of the feedforward control command across the entire angle range, solve the problem of low-speed crawling or reverse dead zone caused by model mismatch, and improve the adaptability of the control system to equipment with different service years.

[0025] 3. Based on the zero torque interval prediction gradient constraint logic, the mechanical transmission chain tooth backlash crossing impact load is suppressed. The gravity torque image model is used for look-ahead prediction to locate the critical angle range of the load torque crossing zero. Within the interval, the current setpoint change rate constraint logic is activated. When the load direction changes and the gear meshing surface disengages, the electromagnetic torque change gradient is constrained, so that the motor completes the torque crossing zero process at a controlled and gradual rate. The active softening process avoids the instantaneous acceleration of the motor rotor's unloaded backlash stroke. The tooth surface contact process after the gear crosses the gap is changed from rigid impact to flexible transition, reducing the mechanical fatigue damage of alternating load to the reducer tooth root and motor bearing, and extending the service life of large flipping tooling. Attached Figure Description

[0026] Figure 1 This is a flowchart of the drive control method integrating parameter calibration and collaborative logic of the present invention;

[0027] Figure 2 This is a comparison of the overshoot suppression and damping torque response of the rotor over dead zone in this invention;

[0028] Figure 3 This is a schematic diagram of the system hardware and software logic architecture based on physical and digital dual-domain interaction of the present invention. Detailed Implementation

[0029] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0030] This invention provides a wind turbine rotor reversal drive control method based on torque-speed coordination. This method constructs a multi-dimensional collaborative control architecture including a gravitational potential energy feedforward channel, a friction compensation channel, a speed feedback channel, and a state monitoring logic unit. The feedforward channel calculates the theoretical gravitational torque based on real-time angle to bear the main load; the friction channel injects a constant compensation amount based on the direction signal; the feedback channel processes residual errors; and the state monitoring unit is responsible for taking over the integral state of the speed regulator and the torque change rate limit in the over-dead-point region when the load polarity reverses. All functional modules interact in real time through a data bus to jointly generate the final q-axis current setpoint to drive the motor. In the reversal operation of large wind turbine rotors, this method addresses the rotor eccentricity mass... To address the issues of polarity reversal due to the massive gravitational load and interference from the mixed frictional torque of the mechanical transmission chain, this invention performs a friction-gravity decoupling calibration procedure before the formal turning operation begins. This controls the drive motor to perform forward and reverse micro-rotation movements from its initial position. During this process, the speed loop is set to a high-stiffness mode, and the drive motor performs uniform lifting and retraction at angular velocities with equal absolute values ​​and within the quasi-static range (e.g., below 1% of the rated speed). This quasi-static condition is used to ignore viscous friction and inertial torque components, ensuring that the electromagnetic torque output by the motor mainly consists of gravitational torque and Coulomb friction torque. The system collects the first average electromagnetic torque of the drive motor during the forward micro-rotation movement in real time. and the second average electromagnetic torque during the reverse micro-rotation action. Utilizing the physical property that the work done by conservative forces is path-independent while the work done by dissipative forces is path-dependent, the system's internal computing unit performs algebraic decoupling operations: calculating the arithmetic sum of the first average electromagnetic torque and the second average electromagnetic torque and dividing by 2, i.e. Extract the pure gravitational torque component without friction interference. Obtain the initial mechanical angle of the drive motor in its initial position, calculate the sine value of this angle, divide the pure gravitational torque component by this sine value, and inversely calculate the corrected maximum gravitational torque amplitude, which is the product of the actual mass and the radius of the center of mass. This correction value is updated to the system's preset gravity torque mapping model, ensuring that the physical model upon which subsequent feedforward control is based matches the actual workpiece's mass distribution characteristics. Simultaneously, the system calculates the arithmetic difference between the first average electromagnetic torque and the second average electromagnetic torque and divides it by 2. Extract the system friction torque component that reflects the current lubrication and wear state of the mechanical transmission chain. This component is stored as a reference value and used to configure an independent friction compensation channel.

[0031] During real-time drive after model calibration, the system operates in torque-speed coordinated control mode. The controller acquires the actual rotor mechanical angle and actual speed of the drive motor in real time through the encoder. On the feedforward path, the system calculates the feedforward compensation torque required to overcome gravity at the current angle based on the actual rotor mechanical angle and the corrected gravity torque image model. This torque component varies sinusoidally with the angle, responding to changes in the load gravity torque. Simultaneously, the system establishes a sign function logic block that is only related to the speed direction. The extracted system friction torque component is used as a fixed gain input to this logic block. The sign function logic block directly outputs the friction compensation torque in the same direction as the current motion direction based on the real-time sign of the actual speed, including positive or negative, to counteract the Coulomb friction of the mechanical transmission chain. On the feedback path, the system uses a speed regulator to process the speed deviation between the actual speed and the given speed, generating feedback adjustment torque. Finally, the system vector-superimposes the feedforward compensation torque, friction compensation torque, and feedback adjustment torque to generate the final q-axis current setpoint and controls the drive motor to output electromagnetic torque according to this setpoint. To address potential speed fluctuations that may occur at the critical moment when the rotor center of gravity crosses the dead point, the system... The system employs state-monitoring-based integral suppression logic. During the drive process, the controller monitors the relationship between the feedforward command and the motion state in real time. When the direction of the feedforward compensation torque is detected to be the same as the direction of the actual speed, it indicates that the rotor has entered the potential energy release stage. The system triggers the suppression operation of the integral component in the speed regulator. The logic switch inside the speed regulator is activated, forcing the output of the integral term of the proportional-integral controller to zero or maintaining it at the current value without accumulation. The control of the drive motor is maintained through the proportional term and the feedforward compensation torque. The triggering conditions for the suppression operation also include: the amplitude of the real-time load torque estimated by the disturbance observer is less than the preset safety threshold or the actual rotor mechanical angle is within the preset potential energy release angle range. This cuts off the influence of historical errors on the current control, causing the motor torque to reverse with the change of gravity direction. To enhance the damping characteristics of the system, when the direction of the feedforward compensation torque is detected to be the same as the direction of the actual speed, the system executes a dynamic damping injection step, superimposing a reverse damping torque component into the final q-axis current setpoint. The amplitude of the reverse damping torque component is positively correlated with the amplitude of the actual speed, which is used to suppress speed overshoot and maintain the operating stability of the rotor in the potential energy release region.To address the backlash issue in wind turbine gearboxes, this invention employs a soft-junction control step for backlash crossing. The system utilizes a calibrated gravity torque mapping model to calculate and predict the rotor angle position corresponding to the zero-crossing point of the gravity load torque in real time. Using this angle position as the center, the system sets a backlash crossing interval. When the actual rotor mechanical angle enters the backlash crossing interval, the system activates torque change rate limiting logic, constraining the differential amplitude of the final q-axis current setpoint with respect to time to not exceed a preset soft-junction slope threshold. The soft-junction slope threshold is set based on the gear module of the wind turbine gearbox and the maximum allowable instantaneous impact load. By limiting the rate of change of current, the electromagnetic torque output by the motor crosses the zero point with a controlled slope, allowing the driving and driven gears in the gear train to re-mesh with the opposite tooth surface at a lower relative speed after disengaging and crossing the backlash. This transforms the gear tooth surface contact process from rigid impact to a flexible transition.

[0032] Example 1: This example combines the industrial application scenario of assembling the stator and rotor of a megawatt-level direct-drive wind turbine. In the overturning operation of a 100-ton-level permanent magnet direct-drive wind turbine rotor, the drive system faces challenges such as the uncertainty of the rotor's center of mass position due to manufacturing tolerances and the nonlinear fluctuation of the transmission chain friction torque caused by changes in ambient temperature. When the rotor crosses the geometric vertex (top dead center) during the overturning process, the risk of gearbox backlash impact and speed runaway caused by the polarity reversal of the gravity load becomes a control difficulty. To address the physical contradiction between the deterministic model and the time-varying load, the system executes a friction-gravity decoupling calibration procedure before the formal lifting action is initiated. It controls the drive motor to perform bidirectional micro-scanning at an angular velocity with equal absolute values ​​and within the quasi-static range, and collects the first average electromagnetic torque during the forward micro-rotation action in real time. and the second average electromagnetic torque during the reverse micro-rotation action The system utilizes the physical symmetry that the work done by conservative forces is path-independent while the work done by dissipative forces is path-dependent to separate the pure gravitational torque component through arithmetic operations. Friction torque component of the system Gain coefficient of the reverse correction gravitational moment mapping model It also features an independent friction compensation channel.

[0033] During the drive process, the system uses a modified gravity torque image model to generate feedforward compensation torque to bear the main load, eliminating the influence of model parameter deviation on control accuracy. When the rotor runs to the potential energy release zone, that is, when the direction of the feedforward compensation torque is detected to be the same as the direction of the actual speed, the system uses the state monitoring unit to trigger the suppression operation of the integral component in the speed regulator and superimposes the reverse damping torque component that is positively correlated with the speed amplitude. At the same time, the system activates the torque change rate limiting logic in the load zero-crossing region based on model prediction, constraining the differential amplitude of the final q-axis current setpoint with respect to time. This control strategy, through the synergy of high-precision model feedforward and state-related damping injection, cuts off the reverse lag path caused by the integral effect in conventional PI control without relying on external torque sensors. It transforms the rigid impact of the reducer tooth surface in the zero-torque crossing zone into a controlled flexible transition, ensuring the smooth operation of the hundred-ton-class large inertia variable polarity load throughout the entire overturning process and avoiding mechanical shock and slippage accidents.

[0034] Example 2: This example aims to quantitatively and objectively verify the actual engineering performance of the aforementioned wind turbine rotor turning drive control method based on torque-speed coordination by constructing a high-fidelity industrial field verification environment. The experiment relies on a wind turbine rotor-stator assembly test platform with a rated power of 2.5MW. This platform is equipped with a permanent magnet synchronous drive motor with a rated torque of 18000Nm and a position feedback unit with a 23-bit high-resolution absolute encoder. To simulate the complex electromagnetic environment and mechanical disturbances of a real industrial field, Gaussian white noise with a signal-to-noise ratio of 25dB is actively superimposed on the encoder feedback signal during the experiment. Random mechanical vibration interference with an amplitude of 5% of the rated load and a frequency of 15Hz is applied at the load end to construct stringent engineering boundary conditions. The core purpose of the experiment is to quantify... This study verifies the ability of this invention to suppress speed overshoot and buffer backlash impact on mechanical transmission chains when handling the over-dead-point condition of a 100-ton rotor. The experiment employed a friction-gravity decoupling calibration procedure. With the drive motor in its initial horizontal position (0°), the control system commanded the motor to perform forward and reverse micro-scans at a quasi-static angular velocity of 0.05 rpm, with the scanning angle range set at 0.5°. The data acquisition system recorded the q-axis current during this process at a sampling frequency of 2 kHz and converted it into electromagnetic torque. In the raw data not processed by the method of this invention, due to the randomness of static friction and signal noise, the instantaneous torque fluctuation reached ±450 Nm. The system used the aforementioned algebraic decoupling algorithm to average the 1024 data points within the acquisition window, measuring the first average electromagnetic torque during the forward micro-scan rotation. The second average electromagnetic torque during the reverse micro-rotation motion is 14520.6 Nm. The torque is 11280.4 Nm. Based on the principle of physical symmetry, the pure gravitational torque component is calculated by the system. The system friction torque component is 12900.5 Nm. The measured value is 1620.1 Nm. Value, the system reverse correction of the gravitational moment mapping model The parameters show that the deviation between the corrected model output and the standard force gauge calibration value is only 0.8%, which is better than the 5.4% deviation calculated based on theoretical CAD parameters before correction.

[0035] A dynamic drive test was conducted on the entire rotor flipping process. Two comparison groups were set up: the control group used a traditional speed-current loop cascaded PI control strategy without feedforward and integral suppression logic; the experimental group of this invention fully utilized a collaborative control strategy including gravity feedforward, friction compensation, integral suppression, and dynamic damping injection. The test condition was set to flip the rotor uniformly from 85° to 95° (i.e., crossing the 90° dead point), with a given speed of 1.5 rpm. In the control group, when the rotor crossed the 90° dead point, due to the desaturation delay of the integral term, the motor output torque failed to reverse in time, causing the rotor to accelerate instantaneously under gravity drag, with a measured peak speed of 2.85 rpm and an overshoot of up to 90%, triggering low-frequency mechanical oscillations lasting approximately 3.5 seconds. In the experimental group of this invention, when the direction of the feedforward compensation torque was detected to be consistent with the actual speed direction, the integral suppression logic intervened, and a reverse damping torque component was injected. Data shows that the motor's electromagnetic torque smoothly switches from forward drive to reverse braking within an extremely narrow range of 89.8° to 90.2°. The peak speed when the rotor passes over the top is strictly limited to 1.58 rpm, the overshoot is only 5.3%, and no obvious mechanical oscillation is observed. To further verify the effectiveness of the backlash crossing soft handover control step, the torque change rate data near the zero point is analyzed. In the control group, due to the lack of gradient limitation, the change rate of electromagnetic torque when crossing the zero point is as high as 12000 Nm / s, which causes the gearbox gear to experience a high-intensity rigid impact after crossing the gap. The instantaneous impact peak value monitored by the acceleration sensor reaches 4.2g. In contrast, the sample of this invention activates the torque change rate limiting logic after detecting that the rotor has entered the preset backlash crossing range, and forcibly constrains the differential amplitude of the q-axis current setpoint to a level where the corresponding torque change rate does not exceed 2000 Nm / s (see Table 1).

[0036] Table 1: Comparison of dynamic response data of the two control strategies at critical moments of dead ends.

[0037]

[0038] The data in Table 1 clearly demonstrates the inherent logical necessity of the present invention. By establishing a negative electromagnetic torque that balances the gravitational component in the intermediate state (at 90°), the overshoot of the final output speed is significantly suppressed. By limiting the rate of change of the torque at zero crossing, the impact of tooth surface contact is reduced. The present invention solves the control problems caused by integral lag and mechanical backlash when dealing with large inertia variable polarity loads.

[0039] Example 3: This example combines Figures 1 to 3 This section describes the rotor overturning drive control method for wind turbine generators based on torque-speed coordination, as follows: Figure 1 As shown, the system executes a parameter decoupling calibration procedure. By controlling the motor to perform forward and reverse quasi-static micro-motion, it separates the pure gravity and friction components and corrects the model gain. Then, it enters the real-time state acquisition stage, using the actual rotor mechanical angle and actual rotational speed as input signals. These signals are split into three parallel processing channels: one is the friction compensation channel, which uses a sign function logic output to compensate based on the system's friction torque component; the second is the gravity torque mapping model, which calculates the feedforward compensation torque based on the corrected gravity model; and the third is the speed regulator PI, which processes speed deviations to generate feedback regulation torque. The outputs of these three channels converge to the state monitoring and collaborative control logic module. This module, when detecting a load in the same direction, performs integral suppression to force zero or freeze the integral term, superimposes the reverse damping torque component, and limits the torque change rate in the backlash crossing interval. Finally, the processed components are processed in a vector superposition unit. The summation operation, including feedforward compensation, friction compensation, feedback adjustment, and damping components, generates the final q-axis current setpoint to output electromagnetic torque to drive the motor, and maintains cyclic control through closed-loop feedback.

[0040] like Figure 2 As shown in the figure, the graph uses the rotor angle as the horizontal axis, covering the range from 85° to 95°. The left vertical axis represents the rotational speed in rpm, and the right vertical axis represents the damping torque in Nm. The dashed line in the figure shows the undamped injection speed curve, with a significant peak in speed near the 90° position. The solid line shows the damped injection speed curve. After using the method of this invention, the peak speed is reduced and the change is gradual. At the same time, the dotted line, the reverse damping torque curve, shows the trend of the damping torque changing dynamically with the speed fluctuation. It intervenes rapidly during the speed rise phase and decays accordingly after the speed falls, showing a regulation characteristic that is positively correlated with the speed deviation.

[0041] like Figure 3As shown, the architecture is divided into a physical entity domain on the left for execution and load, and a digital computing domain on the right for algorithms and strategies. The physical entity domain includes a hundred-ton wind turbine rotor as a variable polarity gravity load object, a geared transmission chain with backlash nonlinearity, and a permanent magnet drive motor as an electromagnetic torque output source. The digital computing domain mainly consists of a parameter self-calibration engine, a torque-speed coordination core, and a state observer. The self-calibration engine is responsible for executing micro-motion scanning logic, gravity / friction separation algorithm, and model gain correction, and passes the correction parameters to the coordination core. The coordination core integrates gravity image model feedforward, integral suppression PI adjustment feedback, and backlash soft junction gradient limitation protection functions. The state observer is responsible for determining the load polarity in real time and feeding back the load status. In terms of inter-domain interaction, the information flow transmits angle and speed data from the physical domain to the digital domain, while the energy flow acts in the opposite direction in the form of current commands to drive the physical entity to run.

[0042] Example 4: This example constructs an online dynamic compensation mechanism based on a disturbance observer, running in the computing core of the drive controller. Its primary task is to isolate and quantify dynamic load disturbances not covered by the static gravity torque model in real time on a microsecond-level timescale. The system establishes a full-order state observer, which uses the motor stator voltage equation and mechanical motion equation as its mathematical foundation. It takes the measured stator current and rotor mechanical angular velocity as input vectors. Within each control cycle, the observer calculates the theoretical state estimate using preset nominal motor parameters and compares it with the actual feedback value from the sensors. The deviation between the two is no longer simply considered noise, but is amplified by the feedback matrix and used to correct the extended state variable inside the observer. This variable represents the total lumped disturbance torque, including friction torque drift, model parameter errors, and external unknown disturbances. Based on the real-time estimated lumped disturbance torque, the system executes an adaptive feedforward correction strategy. The controller sets up frequency domain splitting logic based on a low-pass filter to... The system is decomposed into steady-state and transient components. The slowly changing steady-state component mainly corresponds to the viscous friction torque that drifts with temperature. As a slowly changing bias value, it is directly superimposed on the reference value of the friction compensation channel. This allows for automatic compensation of frictional characteristic deviations caused by oil temperature changes without interrupting production for recalibration. For the rapidly changing transient component, the system introduces a feedforward input of the current loop to improve the dynamic stiffness of the system against sudden mechanical shocks. This dual-frequency domain current shunting compensation mechanism ensures that even under non-ideal operating conditions where there are time-varying deviations between the model parameters and the actual physical object, the electromagnetic torque command can still be anchored to the actual load demand.

[0043] To further eliminate control singularities introduced by backlash nonlinearity at the moment of dead point, the system is configured with a torque gradient limiting algorithm in the discrete time domain. Unlike conventional analog filters, this algorithm applies torque gradient limiting in every interrupt cycle of the digital signal processor. Execution: The system calculates the target torque command in real time at the current moment. Compared with the final output torque command of the previous cycle The difference between them is considered when the system detects that the rotor is in the gravitational potential energy release region and the absolute value of the difference exceeds a preset gradient threshold. When the algorithm forces a truncation of the current output, the electromagnetic torque corresponding to the final output q-axis current setpoint is then determined. Strictly constrained as ,in The values ​​are not chosen empirically, but are based on the allowable contact stress limit of the gearbox gears. Gear pitch circle radius and system sampling period Through physical formulas Rigorous calculations show that, among them This represents the effective contact area of ​​the tooth surface. To calculate the equivalent rotational inertia of the motor shaft, the algorithm ensures that the rate of change of electromagnetic torque is always kept within the safety boundary of the mechanical structure, thus completely eliminating the destructive impact on the tooth surface caused by command step from the underlying logic of the control algorithm.

[0044] Example 5: To ensure the robustness of the control system under different electromagnetic interference environments and mechanical assembly tolerances, and to address the engineering uncertainties in the calibration of key control parameters, this example constructs a standardized adaptive tuning procedure for system parameters. To address the high-frequency jitter that may be caused by the friction compensation channel in the zero-speed crossing region of the motor, the system performs a zero-speed noise statistical scan. With the drive motor enabled and maintaining a zero-speed hover, the system continuously acquires speed feedback signals for 1000 sampling cycles and calculates the standard deviation of this signal sequence. Based on this, the width of the linear transition interval of the friction compensation function is set. The friction compensation torque is required to be less than the absolute value of the rotational speed. Within the range, it linearly decays to zero with the rotational speed, thereby eliminating the root cause of oscillation caused by the frequent flipping of the sign function triggered by measurement noise at the physical level. At the same time, regarding the setting of the integral suppression logic trigger threshold, the system monitors the output residual of the disturbance observer in the no-load constant speed operation mode, extracts the peak value of the residual signal, and strictly sets the safety threshold to 1.2 times the peak value to prevent the observer's background noise during normal operation from being misjudged as a load polarity reversal signal.

[0045] Damping coefficient in the reverse damping torque component during the potential energy release phase To address the challenge of quantitative tuning, the system is configured with critical damping optimization logic based on step response. Under steady-state conditions where the rotor is horizontal and the gravitational torque is completely balanced by the feedforward channel, the controller injects a step disturbance signal with an amplitude of 2% of the rated torque into the current loop and records the speed response curve in real time. The system calculates the overshoot and settling time of this response curve and iteratively adjusts the damping coefficient using a gradient descent algorithm. The value, up to the second-order damping ratio of the system. The process converges to the critical damping state of 0.707. This process transforms the selection of the damping coefficient from empirical estimation to physical calculation based on the actual rotational inertia of the system and the electromechanical time constant. This ensures that the motor can provide the optimal reverse electromagnetic damping that neither causes overdamped hysteresis nor underdamped oscillations when suppressing the gravitational acceleration effect.

[0046] Example 6: Calibration Procedure for Backlash Crossing Soft Intersection Parameters This example details the backlash crossing interval width in the backlash crossing soft intersection control steps. With torque gradient limit threshold Standardized procedures are established based on fundamental principles of contact mechanics and rigid body dynamics, eliminating control uncertainties arising from empirical settings through computational processes. The backlash crossing width is calibrated to ensure that the torque change rate limiting logic covers the entire process from tooth disengagement to reverse meshing. The system executes a process based on geometric parameter-defined intervals: Step 1, obtain the design value of the normal backlash of the final stage gear of the reducer. and pitch circle radius All units are millimeters; Step 2: Convert linear displacement clearance to angular displacement clearance using geometric relationships; Theoretical clearance angle The calculation is as follows: The unit is degrees; Step 3, considering the consistency deviation caused by mechanical wear and assembly errors, an interval expansion coefficient is introduced. For example, take The width of the tooth gap crossing section Determined as: The system uses the angle position corresponding to the zero-crossing point of the gravity load torque. Centered on, Set as the effective range for activating the torque change rate limit logic.

[0047] The physical quantification of the torque gradient limiting threshold is to prevent destructive contact stress from sudden changes in electromagnetic torque at the moment of tooth re-meshing. The procedure is based on the allowable contact stress and the preset transition time, and derives the maximum allowable torque increment within the discrete control cycle in reverse. According to Hertzian contact stress theory, the gear tooth surface can withstand the maximum steady-state torque. With respect to allowable contact stress of the material Proportional relationship: The variables are defined as follows: The maximum steady-state safety torque, measured in Newton-meters. The allowable contact stress limit for gear materials, expressed in Pascals. The effective contact area of ​​the tooth surface, in square meters. The unit for gear pitch circle radius is meters. To achieve a smooth transition, the system mandates that the stator current setpoint establishment process must be within a preset soft transition time. The internal linear operation is completed, from which the unit control period is derived. The maximum allowable torque change, i.e., the gradient limit threshold. The calculation formula is as follows: The variables are defined as follows: The single-cycle torque gradient limit threshold, in Newton-meters. The preset soft handover transition time is in seconds, and the parameter is selected as the system's lowest mechanical resonance period. 1.5 to 2.0 times, to suppress impact-induced mechanical oscillations, The discrete sampling period of the controller is in seconds; the discrete domain execution logic drives the controller interrupt cycle. In the process, when the actual rotor mechanical angle is within the aforementioned backlash crossing range, the system executes the gradient clamping algorithm: calculating the unconstrained target torque command. Compared with the actual output torque of the previous cycle Difference : Based on the calculated threshold right Amplitude constraints are used to generate corrected torque increments. : Update the final output torque command for the current cycle. : This ensures that the rate of change of electromagnetic torque is constant when it crosses the zero-point region, constrained by physical boundaries, and strictly limits the impact energy of tooth surface contact within the allowable range of the material.

[0048] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A wind turbine rotor standing up driving control method based on torque speed cooperation, characterized in that, The method comprises the following steps: controlling the driving motor to perform forward and reverse micro-motion rotation actions at the initial position before the turning operation is started, the absolute values of the angular velocities of the forward and reverse micro-motion rotation actions being equal and being in a quasi-static interval; collecting a first average electromagnetic torque of the driving motor during the forward micro-motion rotation action and a second average electromagnetic torque of the driving motor during the reverse micro-motion rotation action; calculating the arithmetic sum of the first average electromagnetic torque and the second average electromagnetic torque to extract a pure gravity torque component, correcting a gain coefficient of a gravity torque mapping model by using the pure gravity torque component, and the gravity torque mapping model representing a functional relationship between a theoretical gravity load torque of a rotor and a mechanical angle of the rotor; calculating the arithmetic difference between the first average electromagnetic torque and the second average electromagnetic torque to extract a system friction torque component, and configuring an independent friction compensation channel; during the driving process, collecting an actual rotor mechanical angle and an actual speed of the driving motor in real time; calculating a feedforward compensation torque according to the actual rotor mechanical angle and the corrected gravity torque mapping model; and generating a friction compensation torque according to the direction of the actual speed and the system friction torque component; processing a speed deviation between the actual speed and a given speed by using a speed regulator to generate a feedback adjustment torque; and performing vector superposition on the feedforward compensation torque, the friction compensation torque and the feedback adjustment torque to generate a final q-axis current given value; controlling the driving motor to output an electromagnetic torque according to the final q-axis current given value, and triggering an inhibition operation on an integral component in the speed regulator when it is detected that the direction of the feedforward compensation torque is the same as the direction of the actual speed.

2. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 1, characterized in that, the step of calculating an arithmetic sum of the first average electromagnetic torque and the second average electromagnetic torque to extract a pure gravity torque component and calculating an arithmetic difference of the first average electromagnetic torque and the second average electromagnetic torque to extract a system friction torque component, the pure gravity torque component and the system friction torque component satisfies the following relationship: wherein, is the first average electromagnetic torque, is the second average electromagnetic torque; the step of correcting a gain coefficient of a gravity torque map model with the pure gravity torque component includes correcting the gain coefficient of the gravity torque map model with the pure gravity torque component a preset rotor mass and force arm product parameter is updated.

3. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 1, characterized in that, The method further comprises a cogging crossing soft handover control step, in which a gravity load torque zero-crossing point corresponding to a rotor angle position is calculated and predicted in real time based on the gravity torque mapping model, and a cogging crossing interval is set with the angle position as the center; when it is detected that the actual rotor mechanical angle enters the cogging crossing interval, torque change rate limiting logic is activated; the torque change rate limiting logic forcibly restricts the differential amplitude of the final q-axis current given value with respect to time to be less than a preset soft handover slope threshold value until the actual rotor mechanical angle exits the cogging crossing interval.

4. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 1, characterized in that, The method further comprises a dynamic damping injection step, in which a reverse damping torque component is superimposed in the final q-axis current given value when it is detected that the direction of the feedforward compensation torque is the same as the direction of the actual speed; the amplitude of the reverse damping torque component is positively correlated with the amplitude of the actual speed.

5. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 1, characterized in that, The triggering condition of the inhibition operation further comprises that the real-time load torque amplitude estimated by a disturbance observer is less than a preset safety threshold value or the actual rotor mechanical angle is in a preset potential energy release angle interval.

6. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 1, characterized in that, The step of configuring the independent friction compensation channel specifically comprises: establishing a sign function logic block related only to the direction of the speed; inputting the extracted system friction torque component into the sign function logic block as a fixed gain; and during the driving process, the sign function logic block outputs a friction compensation torque in the same direction as the current motion direction according to the real-time sign of the actual speed.

7. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 1, characterized in that, The step of correcting the gain coefficient of the gravity torque mapping model by the pure gravity torque component specifically comprises: obtaining an initial mechanical angle of the driving motor at an initial position; calculating a sine value of the initial mechanical angle, dividing the pure gravity torque component by the sine value to obtain a corrected maximum gravity torque amplitude; and updating an amplitude parameter in the gravity torque mapping model to the corrected maximum gravity torque amplitude.

8. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 3, characterized in that, In the soft handover control step of the gear gap crossing, the setting of the soft handover slope threshold value is based on the gear modulus of the gear box of the wind driven generator and the maximum instantaneous impact load allowed, and the torque change rate limiting logic is used to convert the gear tooth surface contact process from rigid impact to flexible transition.

9. The wind turbine rotor roll-over drive control method based on torque and speed coordination according to claim 1, characterized in that, The step of processing the speed deviation by the speed regulator to generate the feedback adjustment torque specifically comprises: inputting the speed deviation into a proportional integral controller; when the suppression operation is not triggered, the proportional integral controller calculates an output according to preset proportional and integral coefficients; when the suppression operation is triggered, the integral term output of the proportional integral controller is forced to be zero or kept at the current value without further accumulation, and the control on the driving motor is maintained only by the proportional term and the feedforward compensation torque.

Images