An Adaptive High Power Density Permanent Magnet Motor Control Method and System

Abstract

This invention relates to the field of wind power generation control technology and discloses an adaptive high power density permanent magnet motor control method and system, including: real-time monitoring of the zero-crossing timing of the motor current waveform and using the number of occurrences of timing abnormal events as the basis for identifying the system stress state; when a stress state is identified, instantaneously injecting d-axis current to form flexible electromagnetic damping and simultaneously limiting q-axis current output to reserve a safety margin. This invention transforms the current distortion signal from control error into a system state messenger and responds in a coordinated manner combining flexible suppression and active protection, avoiding the overheating and mechanical fatigue accumulation problems caused by overcompensation when dealing with disturbances in traditional control methods. This enables the control system to obtain stress-resistant self-protection capabilities and improves the operational reliability of wind turbine generators under uncertain operating conditions.

Description

Technical Field

[0001] This invention relates to an adaptive high power density permanent magnet motor control method and system, belonging to the field of wind power generation control technology. Background Technology

[0002] High-power-density permanent magnet synchronous motors are widely used due to their high efficiency and compactness. Their control technology generally follows the field-oriented control method, which treats the motor as an equivalent linear system. This method is stable in traditional industrial applications, but its inherent design limitations gradually become apparent in the uncertain scenario of wind power generation. Wind power systems are essentially complex systems that are constantly subjected to wide-frequency, strong random wind speed disturbances and grid fluctuations. In pursuit of accurate power point tracking, existing control methods often adopt rigid control strategies with high gain and wide bandwidth. As a result, when dealing with sudden disturbances, the controller will inject instantaneous current far exceeding the steady-state requirements. This adversarial compensation behavior directly leads to the risk of performance degradation of permanent magnets due to instantaneous overheating in high-power-density motors with limited thermal margins. At the same time, the electromagnetic torque output by the motor and the inherent mechanical vibration of the transmission chain form a continuous rigid confrontation, accelerating fatigue damage to key components such as bearings.

[0003] To address these challenges, the industry has explored relevant adaptive control strategies, such as introducing complex nonlinear models or state observers to improve control accuracy. However, these improvement paths have increased the computational burden of the system and its dependence on model parameters to some extent, without fundamentally changing the control approach centered on precise compensation. When the system encounters complex disturbances not covered by the model or when parameters drift due to long-term operation, its control robustness still faces challenges.

[0004] Specifically, existing technologies suffer from the following shortcomings: 1. They lack a low-cost and reliable mechanism to perceive the stress state of the system under external disturbances in real time, resulting in a lack of self-protection capability in the control system when dealing with emergencies; 2. There is an inherent contradiction between the two goals of pursuing power generation performance and ensuring the long-term reliability of equipment, often sacrificing the latter for the former, lacking a control method that can dynamically balance and coordinate the two; 3. Existing control methods are unable to effectively distinguish and respond to system disturbances caused by different physical sources such as grid faults and mechanical degradation, and their response strategies appear simplistic and lack adaptability. Therefore, how to design a control method that can instantly identify the stress state of the system without relying on complex models and increasing additional hardware costs, and automatically switch to a self-protection mode that can effectively mitigate disturbance energy and prevent damage to core components from a mechanistic perspective, while also identifying disturbances from different sources to achieve more adaptive coordinated control, is the technical problem to be solved by this invention. Summary of the Invention

[0005] This invention provides an adaptive high power density permanent magnet motor control method. Its main purpose is to solve the problem that existing control methods, when dealing with strong uncertain disturbances in wind power systems, lack effective stress state perception and self-protection mechanisms, resulting in damage to the motor and transmission chain and insufficient adaptability.

[0006] To achieve the above objectives, this invention provides an adaptive high-power-density permanent magnet motor control method, applied to wind power generation systems. The method includes the following steps:

[0007] Step a: Monitor the zero-crossing points of the three-phase current waveform of the permanent magnet motor stator in real time, and calculate the actual time interval between each consecutive zero-crossing point;

[0008] Step b: Compare the actual time interval with the theoretical time interval determined based on the current synchronization frequency of the permanent magnet motor. When the absolute value of the difference between the actual time interval and the theoretical time interval reaches a set counting threshold within a set statistical time window, an excitatory damping injection trigger command is generated.

[0009] Step c: In response to the stress-damping injection trigger command, d-axis current is injected into the permanent magnet motor for a set duration; and during the set duration, the maximum output command value of the q-axis current used to generate torque is limited to a protection command value lower than its maximum output command value when the stress-damping injection trigger command is not triggered.

[0010] Preferably, in step b, the time threshold for the absolute value of the difference to exceed is one percent to five percent of the theoretical time interval, and in step c, the injection amplitude of the d-axis current is five percent to twenty percent of the rated current of the permanent magnet motor.

[0011] Preferably, in step c, the protection command value of the q-axis current is 70% to 90% of its maximum output command value when the stress damping injection trigger command is not triggered.

[0012] Preferably, after step c is completed, the cumulative count of occurrences in step b is cleared to zero.

[0013] Preferably, the method further includes: real-time monitoring of the grid voltage phase angle at the grid connection point and calculating the rate of change of the grid voltage phase angle; when the calculated rate of change of the grid voltage phase angle exceeds a phase change rate threshold characterizing a grid fault, before the calculated rate of change of the grid voltage phase angle recovers to below the phase change rate threshold, increasing the counting threshold in step b and switching the action of injecting d-axis current in step c to injecting reactive current.

[0014] Preferably, the method further includes: during each execution of step c, synchronously acquiring the q-axis current of the permanent magnet motor; extracting the harmonic energy within a predetermined mechanical fault frequency band from the acquired q-axis current data to form a harmonic fingerprint of the current damping event; comparing the harmonic fingerprint of the current damping event with a health status baseline calculated based on the harmonic fingerprints of multiple historical damping events; when the comparison result shows that the harmonic energy shows a continuous increasing trend, generating an early warning signal for early degradation of the transmission chain and increasing the counting threshold in step b.

[0015] Preferably, the method further includes: performing online resonant mode identification when the number of stress damping injection trigger commands generated in step b is zero within a set calm time window. The online resonant mode identification includes: injecting a spectrum-rich micro-perturbation excitation current signal into the d-axis current channel of the permanent magnet motor; synchronously acquiring the q-axis current of the permanent magnet motor as the output response signal; calculating the current resonant damping ratio of the wind power generation system drive train based on the micro-perturbation excitation current signal and the output response signal; and adaptively adjusting the counting threshold in step b based on the calculated resonant damping ratio.

[0016] Preferably, the method further includes: real-time monitoring of the effective value of the grid voltage at the grid connection point; and, during step c, the injection amplitude of the d-axis current. It is determined by feedforward compensation based on the real-time magnitude of the monitored effective value of the grid voltage, and its compensation relationship satisfies: ,in, Injecting amplitude to the compensated d-axis current. This is the reference amplitude of the d-axis current when the effective value of the grid voltage is at its rated value. The rated grid voltage, This refers to the effective value of the grid voltage as monitored in real time.

[0017] Preferably, in step b, the time threshold, statistical time window, and counting threshold for the absolute value of the difference exceeding the threshold are all adaptively adjusted according to the operating conditions of the permanent magnet motor.

[0018] An adaptive high-power-density permanent magnet motor control system, applied to a wind power generation system, includes:

[0019] The timing monitoring module is configured to monitor the zero-crossing points of the three-phase current waveform of the permanent magnet motor stator in real time and calculate the actual time interval between each consecutive zero-crossing point.

[0020] The trigger decision module is connected to the timing monitoring module and is configured to compare the actual time interval calculated by the timing monitoring module with the theoretical time interval determined based on the current synchronization frequency of the permanent magnet motor. When the absolute value of the difference between the actual time interval and the theoretical time interval reaches a set counting threshold within a set statistical time window, an excitatory damping injection trigger command is generated.

[0021] The collaborative control module, connected to the trigger decision module, is configured to inject d-axis current into the permanent magnet motor for a set duration in response to an excitable damping injection trigger command; and, during the set duration, to limit the maximum output command value of the q-axis current used to generate torque to a protection command value lower than its maximum output command value when the excitable damping injection trigger command is not triggered.

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

[0023] 1. By real-time monitoring of the timing characteristics of the current waveform at the zero-crossing point and using the abnormal accumulation of these timing characteristics as a trigger condition, a system stress state perception method independent of complex models is established. Then, after identifying the stress state, the d-axis current is injected synchronously to form flexible electromagnetic damping, and the q-axis current output is actively limited to reserve a safety margin. This collaborative control method, which integrates disturbance perception, flexible suppression, and active protection, avoids the overcurrent impact and mechanical fatigue accumulation problems caused by the pursuit of rapid response in traditional control methods when dealing with external disturbances. This enables the control system to effectively mitigate external shocks without sacrificing safety, and improves the operational reliability of wind turbine generators under uncertain operating conditions.

[0024] 2. By further introducing the monitoring of the voltage phase angle change rate of the grid connection point, the control system gains the ability to identify the source of disturbance. When the source of disturbance is identified as a grid fault, the system can immediately adjust its internal abnormal event counting threshold and switch the control target to prioritize providing reactive power support. This dynamic reconfiguration of the control mode based on disturbance source identification makes motor control no longer a single internal stabilization strategy, but can intelligently balance the two objectives of ensuring its own safety and maintaining grid friendliness, thus demonstrating strong adaptability and system-level resilience in complex grid environments.

[0025] 3. This invention treats each d-axis current injection as an active perturbation detection of the drivetrain and achieves online self-sensing of the drivetrain's health status by synchronously analyzing the harmonic response fingerprint of the q-axis current. This approach, combining control behavior with state diagnosis, enables the system to distinguish between transient external shocks and chronic internal degradation, and adaptively adjust the sensitivity of its control strategy accordingly. This not only avoids control misjudgments caused by mechanical degradation but also provides a health status monitoring method for wind turbine generators without increasing hardware costs. Furthermore, by performing feedforward compensation of the d-axis current injection amplitude based on the real-time monitored effective grid voltage during damping injection, the core suppression capability of the control method remains effective under conditions such as grid voltage fluctuations. Moreover, by actively injecting perturbation excitation during stable system operation to identify the drivetrain's resonant modes online and adjusting the control threshold in a closed loop accordingly, the entire control method possesses the ability to resist system characteristic drift caused by changes in equipment aging conditions, thus improving its long-term operational stability and adaptability. Attached Figure Description

[0026] Figure 1 This is a flowchart of the adaptive flexible damping control process under the three-phase current drive of the permanent magnet motor stator of the present invention.

[0027] Figure 2 This is a comparison curve of motor temperature rise over time under different control strategies of the present invention;

[0028] Figure 3 This is a flowchart of the system functional modules for adaptive flexible damping control and health status monitoring of the present invention.

[0029] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail 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.

[0031] The adaptive high-power-density permanent magnet motor control method and system disclosed in this invention mainly includes a timing monitoring module, a trigger decision module, and a cooperative control module. The timing monitoring module is configured to continuously and uninterruptedly sample the current signal inside the permanent magnet motor in real time and transmit the quantized timing feature data stream to the trigger decision module. The trigger decision module analyzes and identifies the input data stream based on preset logic criteria to determine whether the system has entered a stress state, and generates a corresponding trigger command when the conditions are met. After receiving the trigger command, the cooperative control module is responsible for executing a series of motor electromagnetic state adjustment operations involving multiple cooperative actions.

[0032] In a specific application scenario, when a wind turbine generator is operating under highly uncertain conditions such as grid fluctuations or strong gusts, mechanical resonance in the drive train or external grid disturbances will cause microsecond-level distortions in the stator current waveform. This distortion provides an immediate endogenous signal for the system to enter a stress state. To capture this signal, the timing monitoring module of this invention is configured to use the hardware comparator and high-precision timer built into a standard microcontroller to monitor the zero-crossing points of the three-phase stator current waveform in real time and calculate the actual time interval between each consecutive zero-crossing point. This hardware-based zero-crossing detection method consumes less computational resources than performing a complete waveform Fourier analysis, thereby enabling the generation of raw data streams reflecting the dynamic stability of the system at a high update frequency.

[0033] Given that single waveform jitter may originate from occasional electromagnetic noise, to ensure high confidence in triggering subsequent protection actions, the triggering decision module employs a statistical decision logic based on event accumulation. This module first determines a theoretical time interval based on the current synchronization frequency of the permanent magnet motor measured by the phase-locked loop (PLL). This theoretical time interval is the synchronization electrical cycle divided by 6. Then, the actual time interval calculated by the timing monitoring module is compared with this theoretical time interval. When the absolute value of the difference between the actual time interval and the theoretical time interval exceeds a preset time threshold, it is identified as a zero-crossing anomaly and accumulated. The specific value of the time threshold is not a fixed empirical value but is determined through a set of deterministic procedures. The calibration procedure includes collecting zero-crossing interval data for at least 10 minutes under stable and low wind speed turbulence conditions during the wind turbine commissioning phase. The statistical mean and standard deviation are calculated, and the time threshold is set as the mean plus three times the standard deviation. A typical setting is one to five percent of the theoretical time interval. Then, when the cumulative count reaches a preset count threshold within a preset statistical time window, the trigger decision module generates an excitatory damping injection trigger command. This event counting rather than amplitude detection mechanism makes the judgment logic insensitive to the instantaneous intensity of the disturbance, but only focuses on its persistence, thereby effectively separating real system-level disturbances from occasional measurement noise.

[0034] Upon receiving the stress-induced damping injection trigger command, the coordinated control module then executes a set of coordinated electromagnetic control actions over a predetermined duration. One of these actions is to inject a predetermined amplitude into the permanent magnet motor. shaft current, this The amplitude of the shaft current is typically set to 5% to 20% of the motor's rated current. Its function is to slightly weaken the rigid magnetic coupling between the stator and rotor, forming a flexible electromagnetic damping. This allows the electromagnetic system to temporarily transform from a rigid connection into a flexible damping body, effectively absorbing and dissipating disturbance energy from the mechanical or electrical grid side. Secondly, it is used in the injection... During the same time period of the shaft current, the torque will be generated. The maximum output command value of the shaft current is limited to a protection command value, which is typically set to 70% to 90% of the maximum output command value when the stress-induced damping injection trigger command is not triggered. This synchronous active limiting measure, in mechanism, allows the system to actively relinquish its extreme torque output capability when it determines that it has entered a stress state, thereby reserving a safety margin for the permanent magnet and power semiconductor devices and avoiding the risks of overcurrent and overheating caused by overcompensation. After the set duration ends, the cooperative control module automatically terminates. Injection of shaft current and release of the control The shaft current is limited, and at the same time, the trigger decision module resets the cumulative count of occurrences to zero, so that the system returns to the normal maximum power point tracking (MPPT) control mode, ready to deal with the next disturbance event.

[0035] Furthermore, to enhance the adaptability of the control method in complex power grid environments, this invention also integrates a disturbance source identification mechanism based on the rate of change of the grid voltage phase angle. The physical causes of disturbances encountered by wind power systems can be categorized into internal mechanical disturbances and external grid-side disturbances. Different effective response strategies exist for different disturbance sources. To address this issue, this method also includes real-time monitoring of the grid voltage phase angle at the grid connection point and calculation of its rate of change. This process can reuse the output data of the phase-locked loop (PLL) module already present in the wind turbine controller for grid connection, obtained by differential calculation of the phase angles of two consecutive sampling cycles. When the calculated rate of change of the grid voltage phase angle exceeds a phase change rate threshold characterizing a grid fault, the system immediately identifies the primary problem as originating from an external grid fault. In this case, the control logic is immediately reconfigured, specifically by temporarily and significantly increasing the counting threshold in the trigger decision module to avoid erroneous damping injection actions caused by severe current distortion during grid faults. Simultaneously, the original flexible damping injection mechanism in the coordination control module is reconfigured. The axis current control channel temporarily switches to a strategy of injecting reactive current to meet the requirements of the power grid for reactive power support. Until the rate of change of the voltage phase angle of the power grid is detected to return to below the threshold of the phase change rate, all temporarily reconfigured parameters and control logic will automatically return to the original settings, so that the entire control system can achieve a dynamic intelligent trade-off between ensuring its own safety and maintaining grid friendliness.

[0036] Furthermore, to enable the system to combat adaptive degradation caused by the evolution of the equipment's own health status, this invention also includes an online self-sensing mechanism for the health status of the drivetrain; this mechanism will coordinate the operation of the control module each time it executes... Each shaft current injection action is considered an active perturbation detection excitation of the transmission chain system; in each injection execution... During the shaft current period, the system will synchronously collect the permanent magnet motor's... The shaft current data is collected and extracted using the efficient Goertzel algorithm. In the shaft current data, the harmonic energy within the predetermined mechanical fault frequency band constitutes a harmonic fingerprint of the current damping event. The system continuously records and maintains a health status baseline calculated based on the harmonic fingerprints of multiple historical damping events. When a newly captured harmonic fingerprint is compared with this health status baseline, and the result shows that the harmonic energy exhibits a continuous increasing trend, the system determines that the drivetrain has entered an early deterioration state and generates an early deterioration warning signal for the drivetrain. At the same time, the system automatically increases the counting threshold to acknowledge the existence of a chronic vibration source in the system, thereby appropriately reducing the sensitivity of the control scheme to the response of this chronic vibration, avoiding unnecessary energy loss, and allowing it to better focus on dealing with shock disturbances.

[0037] To ensure that the flexible damping control maintains stable performance under various operating conditions, this invention also designs a feedforward compensation mechanism to cope with changes in damping performance caused by grid voltage fluctuations; this mechanism monitors the effective value of the grid voltage at the grid connection point in real time. And during the injection When the shaft current operates, for Amplitude of shaft current injection Feedforward compensation is performed, and the compensation relationship satisfies: ,in, The effective value of the grid voltage is at the rated value. time The shaft current reference amplitude is ensured by this compensation logic, which guarantees the injected current even under conditions such as a voltage dip in the mains that weakens the air gap flux of the motor. The shaft current can be increased accordingly to produce an equivalent electromagnetic damping effect, thereby broadening the effectiveness boundary of the control method to more extreme operating conditions.

[0038] Finally, to enable the control strategy to adapt to the slow drift of the transmission chain resonant modes due to factors such as equipment aging and temperature changes, this invention also includes a closed-loop mechanism for online resonant mode identification and adaptive adjustment of control parameters. The trigger condition for this mechanism is that the number of stress-damped injection trigger commands is zero within a set calm time window. Under this condition, the system performs an online resonant mode identification, which includes: sending signals to the permanent magnet motor... A micro-perturbation excitation current signal with a rich spectrum is injected into the shaft current channel; the permanent magnet motor is simultaneously acquired. The shaft current is used as the output response signal. Based on the micro-perturbation excitation current signal and the output response signal, the current resonance damping ratio of the wind power generation system's transmission chain is calculated. The system then adaptively adjusts the counting threshold in the trigger decision module based on the latest calculated resonance damping ratio. For example, when a significant decrease in the damping ratio of a certain resonance point is identified, the system will automatically lower the counting threshold to improve its response sensitivity to the weak link. This enables the entire control system to resist the drift of system characteristics caused by equipment aging and changes in operating conditions.

[0039] Example 1: In a grid-connected operation scenario of a high-power-density permanent magnet wind turbine generator set deployed with the control method of the present invention, the unit encountered a grid voltage dip event lasting for hundreds of milliseconds caused by a fault in a distant power grid line. After the event was resolved, the unit entered the power recovery process. At this time, the grid voltage had recovered to near the rated value, but the entire mechanical system of the unit, including the tower and blades, was still in a low-frequency transient oscillation caused by the previous grid disturbance.

[0040] At the initial moment of voltage recovery, according to the conventional maximum power point tracking strategy, the control system instructs the generator to quickly restore power output. However, under the traditional control paradigm, this restoration command would be directly translated into a large amplitude... Shaft current command; however, just at that Just before the shaft current command was to be executed, the unit's internal control system executed different response sequences. Due to the residual oscillation of the mechanical system, there were slight fluctuations in the motor speed. These fluctuations were directly reflected in the stator three-phase current waveform, causing the actual time interval between consecutive zero crossings calculated by the timing monitoring module to deviate continuously and beyond the preset time threshold. The abnormal event counter in the trigger decision module accumulated to the count value threshold within tens of milliseconds, thereby generating a stress-induced damping injection trigger command.

[0041] In response to the stress-induced damping injection trigger command, the coordinated control module did not perform a significant boost. Instead of acting on the shaft current, it performed a series of coordinated operations: first, it injected a current into the motor for 0.5 seconds with an amplitude of 15% of the rated current. The shaft current, through this action, creates flexible electromagnetic damping, increasing the damping ratio of the transmission chain. This allows the original oscillating energy of the mechanical system to be rapidly dissipated into copper and iron losses within the motor, rather than being countered by rigid electromagnetic torque against the mechanical structure. Simultaneously, the coordinated control module will... The maximum output command value of the shaft current is temporarily limited to 80% of the normal maximum value, which is a protection command value. This active output limitation avoids the instantaneous current surge exceeding the rated value that might have occurred in order to quickly restore torque. The simultaneous occurrence of the injection of flexible damping and the limitation of torque command resolves the technical contradiction between rapid power recovery and the suppression of mechanical oscillation and the protection of permanent magnet thermal safety.

[0042] After the 0.5-second flexible damping control was completed, the mechanical oscillation amplitude of the unit had significantly converged, the current waveform had returned to stability, and the abnormal event counter had been cleared. Subsequently, the control system smoothly increased [the unit's] speed under stable conditions without mechanical resonance interference. The shaft current allows the generator set to stably recover to the expected power output level within seconds in a manner without overshoot or oscillation. Throughout the process, the temperature fluctuation of the motor windings is suppressed within the normal range, and the alternating stress on the transmission chain gearbox and bearings does not increase. This control method transforms the distortion characteristics of the current waveform from a control error that needs to be quickly eliminated into an endogenous messenger characterizing the system entering a stress state. Furthermore, by logically binding and synchronously executing the two originally separate strategies of flexible suppression and active protection, the system can automatically and temporarily adjust the priority of control targets when it senses a reduction in its own load margin, prioritizing the operational reliability of core components.

[0043] Example 2: The control method of the present invention was verified in a hardware-in-the-loop (HIL) simulation test platform. This platform integrates mathematical models of a high-power-density permanent magnet motor, a drive train, and a power grid, and is connected to a physical controller running a control algorithm to reproduce the dynamic process of a wind turbine generator encountering external disturbances in a laboratory environment. In the experiment, the system responses of the test group using the control method of the present invention and the control group using the traditional field-oriented high-bandwidth current loop control method were compared when the same drive train torque disturbance was applied. The frequency of the torque disturbance signal applied in the experiment was set to effectively stimulate the weak links of the system while taking into account the representativeness of the engineering scenario. Therefore, the disturbance frequency was set to 90% of the first-order torsional resonance frequency of the simulated drive train model, and the amplitude was set to 15% of the rated torque. The initial conditions of the experiment were that the motor was running stably at 80% of the rated speed and 80% of the rated load. The disturbance signal was injected after the system entered steady state and lasted for 30 seconds.

[0044] During the test in the control group, after the torque disturbance was injected, the motor's The shaft current exhibited a large fluctuation at the same frequency as the disturbance. The controller attempted to counteract the external torque fluctuation by rapidly adjusting the electromagnetic torque, resulting in a momentary current surge. However, during the experimental group's testing, when the same torque disturbance was injected, the system triggered an excitatory damping injection within a short time, and this was observed... The fluctuation amplitude of shaft current was significantly suppressed, and the output of electromagnetic torque was smoother; Table 1 shows the comparison data of key performance indicators recorded by the two sets of tests during the 30-second disturbance period.

[0045] Table 1: Comparison of Experimental Data

[0046]

[0047] See Table 1 for the peak values ​​of the experimental group. Both shaft current and peak winding temperature rise decreased compared to the control group. The mechanism lies in the fact that the collaborative control module in this invention actively controls the load after recognizing the stress state. The maximum output command value of the shaft current is limited to a protection command value, thus avoiding overcurrent. Meanwhile, the standard deviation of the electromagnetic torque in the test group is much smaller than that in the control group. This phenomenon is due to the synchronous injection... The flexible electromagnetic damping formed by the shaft current effectively absorbs the disturbance energy from the transmission chain, thus alleviating the rigid resistance between the electromagnetic system and the mechanical system.

[0048] Example 3: This example combines Figures 1 to 3 This document describes an adaptive high-power-density permanent magnet motor control method and system, such as... Figure 1 As shown, firstly, the system samples the three-phase stator current of the permanent magnet motor and sends it to the timing monitoring module. This module is responsible for real-time monitoring of the current waveform's zero-crossing point, calculating the actual time interval, and forming zero-crossing timing data. This timing data is transmitted to the trigger decision module, which compares the actual time interval with the theoretical time interval. When the accumulated anomaly reaches a threshold, a trigger command is generated, namely, a stress-induced damping injection trigger command. After the trigger command is issued, the control flow enters the collaborative control module. This module injects d-axis current to form flexible electromagnetic damping and simultaneously limits the q-axis current output to adjust the motor's electromagnetic state. Simultaneously, to enhance the system's intelligent disturbance response, this process introduces the measurement data of the grid voltage phase angle at the grid connection point for disturbance source identification, to determine whether the disturbance originates from a grid fault or the mechanical side. When a grid fault is identified, the counting threshold can be increased and the control target switched. Furthermore, to achieve proactive health monitoring, the collaborative control module collects q-axis current data and transmits it to the online self-sensing module for the drivetrain health status. This module analyzes the q-axis current harmonic fingerprint to assess the drivetrain's deterioration trend.

[0049] like Figure 2As shown, the horizontal axis represents time (s), and the vertical axis represents temperature rise (K). The dashed curve represents the temperature rise trend under the traditional FOC control method, and the solid curve represents the temperature rise response curve under the method of this invention. Figure 2 As can be seen, under the same disturbance conditions, traditional FOC control, due to its rigid current response characteristics, causes the motor to generate a large current surge in a short time, which in turn leads to heat accumulation in the windings. This causes the motor temperature to rise rapidly and reach a steady-state temperature rise level of nearly 23K in about 25 seconds. However, when using the method of this invention, by injecting a flexible d-axis current to form electromagnetic damping after detecting the disturbance and simultaneously limiting the q-axis current output, the instantaneous current overshoot is effectively suppressed, energy loss is reduced, and thus the overall temperature rise of the motor is significantly suppressed, keeping the temperature rise below 6K throughout the entire 30 seconds.

[0050] like Figure 3 As shown, the process begins with the three-phase stator current output from the permanent magnet motor being input to the 1.0 current monitoring timing module for real-time current zero-crossing detection and calculation of the actual time interval. Subsequently, the data is sent to the 2.0 stress state determination module. If the detection result meets the abnormal accumulation condition, a stress trigger command is generated based on the control threshold read from the D1 control parameter library. Simultaneously, under grid-connected operation conditions, the system obtains the grid connection point voltage phase angle from the grid connection point and sends it to the 4.0 disturbance source identification module to determine if a grid fault indicator exists. If so, the control strategy is adjusted. After successful stress state determination, the control process enters the 3.0 coordination phase. The damping control module generates corresponding d / q-axis control commands and outputs them. Simultaneously, it collects q-axis current data, which is transmitted to the 5.0 drivetrain health analysis module and compared with the reference model in the D2 health status baseline. Harmonic fingerprint indices are extracted to determine if a drivetrain degradation warning exists. If a warning is triggered, a signal is sent to the warning and diagnostic system. This module also feeds back the current evaluation results to the 6.0 online adaptive tuning module, which performs parameter updates under the condition of quiescent operation, thereby updating the contents of the D1 control parameter library and the D2 health status baseline.

[0051] Example 4: A wind turbine generator set that has been installed and entered the commissioning stage has its control system executing a set of parameter self-tuning and baseline calibration procedures to match its control parameters with the specific mechanical structure and on-site operating conditions of the generator set. This procedure aims to find a set of localized parameters for the core threshold in the trigger decision module and the baseline model in the drivetrain health status monitoring module. The procedure first enters an offline calibration stage to determine the trigger sensitivity of the stress-damped injection. In this stage, a series of low-intensity mechanical torque disturbance signals with known frequency and amplitude are injected into the drive system through the motor controller. At the same time, the timing monitoring module and the trigger decision module record the frequency and time distribution of zero-crossing abnormal events generated under this excitation at a high sampling rate. Through this calibration process, while being able to stably detect disturbances equivalent to 0.1 times the rated torque, the ability to suppress background noise below 0.02 times the rated torque is maximized, thereby establishing a balance between response sensitivity and false trigger robustness, and finally determining the count value threshold and statistical time window in the trigger decision module.

[0052] Then, the program enters the damping effectiveness calibration stage to determine the damping performance in the coordinated control module. The injected amplitude of the shaft current; the system executes a series of short-duration, step-incremental currents. Shaft current injection test, with injection amplitude ranging from 2% to 25% of rated current; synchronous monitoring during each injection. The ripple energy attenuation rate of the shaft current near the first-order resonant frequency of the transmission chain, and the minimum value that can attenuate this ripple energy by 60% within 0.2 seconds. The shaft current injection amplitude is determined as the protection command value for the subsequent operation of the unit. This procedure ensures that the damping injection action not only meets the performance requirements for suppressing vibration, but also reduces the additional copper loss and thermal effect to a low level.

[0053] After the initial calibration of the above parameters was completed, the unit was put into normal operation and started a health status baseline online learning process that lasted for 1000 operating hours; during this period, each stress-induced damping injection event triggered by a minor disturbance was recorded. Harmonic fingerprints of the shaft current are recorded and stored in a sample database. When the number of valid samples accumulated in the database reaches 500, the system uses principal component analysis (PCA) to reduce the dimensionality of these high-dimensional harmonic fingerprint data and establishes a multidimensional Gaussian mixture model (GMM) that can encompass all healthy sample points with a confidence interval of over 99.7%. This model constitutes the quantitative health status baseline of the unit's drivetrain. For newly generated harmonic fingerprints during subsequent operation, the system calculates their Mahalanobis distance under the GMM model. Once this distance exceeds the preset three-times-standard-deviation control limit five times consecutively, it is determined to be a statistically significant deterioration trend, triggering corresponding early warning signals and adaptive adjustment of control parameters. After executing this calibration and learning procedure, the core decision parameters and state diagnosis model of the wind turbine generator control system are set to specific values ​​that match the actual operating characteristics of the unit, thereby improving the unit's operational reliability and efficiency throughout its life cycle.

[0054] Example 5: On a certified hardware-in-the-loop simulation system, the robustness of the control logic of a controller to be deployed in the field is verified. The verification program first simulates a mechanical shock to the transmission chain caused by emergency braking and records the peak value of the grid voltage phase angle change rate caused by this purely mechanical disturbance. Subsequently, the program simulates multiple grid voltage sag events that conform to the grid connection guidelines and records the minimum value of the grid voltage phase angle change rate under these grid faults. The phase change rate threshold used to distinguish disturbance sources is set to be no less than 150% of the peak value measured under the mechanical shock scenario and no more than 80% of the minimum value measured under the grid fault scenario, thereby establishing a decision boundary based on measured data for the identification of disturbance sources.

[0055] In the grid voltage sag test phase of this verification procedure, the feedforward compensation reference parameters of the flexible damping are simultaneously tested. Calibration is performed when the system simulates a primary grid voltage drop to 50% of the rated voltage. During operation, the controller performs a damping performance test by injecting different amplitudes. Shaft current: Determine an actual injected current value that can achieve the preset oscillation suppression effect. Reference parameters Based on the relation The calculation shows that, among which This represents the actual voltage value during the voltage drop; this calibration method directly correlates the core parameters of feedforward compensation with a damping performance target that can be reproduced through physical testing.

[0056] Example 6: When performing online resonant mode identification on a normally operating wind turbine generator set, the controller follows a set of execution and verification procedures. This procedure first calibrates the parameters of the micro-disturbance excitation current signal. The signal amplitude is set to an upper limit that will not cause a change in the unit's average output power exceeding 0.1% over ten minutes after injection. The signal sequence length is determined by a minimum value that provides the required frequency resolution based on the required transmission chain resonant frequency interval. Under the condition of satisfying the calm time window, the controller performs a pre-state verification before injecting the micro-disturbance excitation current signal. This verification confirms that the standard deviation of the generator speed is below a preset stable operating threshold within the past minute. After this verification passes, the excitation signal is injected. During signal injection, an independent monitoring logic monitors in real time. If the instantaneous value of the shaft current exceeds 120% of its average value before injection, the identification process will be immediately stopped, and the data obtained in this instance will be discarded.

[0057] After a successful online resonant mode identification, the system determines the adjustment value of the counting threshold in the trigger decision module by consulting an internal nonlinear mapping table based on the calculated resonant damping ratio. This mapping table maps the resonant damping ratio in different intervals to a specific counting threshold. The lower the damping ratio, the lower the corresponding counting threshold. This correspondence allows the control system to adjust its sensitivity to vibration according to the actual dynamic characteristics of the transmission chain.

[0058] 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.

[0059] 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. An adaptive high power density permanent magnet motor control method, applied to a wind power generation system, comprising the following steps: Step a: Monitor the continuous zero-crossing points of the three-phase current waveform of the permanent magnet motor stator in real time, and calculate the actual time interval between each continuous zero-crossing point; Step b: Compare the actual time interval with the theoretical time interval determined based on the current synchronization frequency of the permanent magnet motor. When the absolute value of the difference between the actual time interval and the theoretical time interval reaches a set counting threshold within a set statistical time window, an excitatory damping injection trigger command is generated. Step c: In response to the stress-damping injection trigger command, d-axis current is injected into the permanent magnet motor for a set duration; and during the set duration, the maximum output command value of the q-axis current used to generate torque is limited to a protection command value lower than its maximum output command value when the stress-damping injection trigger command is not triggered.

2. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, In step b, the time threshold for the absolute value of the difference to exceed is one percent to five percent of the theoretical time interval. In step c, the injection amplitude of the d-axis current is five percent to twenty percent of the rated current of the permanent magnet motor.

3. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, In step c, the protection command value of the q-axis current is 70% to 90% of its maximum output command value when the stress damping injection trigger command is not triggered.

4. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, After step c is completed, the cumulative count of occurrences in step b is reset to zero.

5. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, The method further includes: real-time monitoring of the grid voltage phase angle at the grid connection point and calculation of the rate of change of the grid voltage phase angle; when the calculated rate of change of the grid voltage phase angle exceeds a phase change rate threshold characterizing a grid fault, before the calculated rate of change of the grid voltage phase angle recovers to below the phase change rate threshold, the counting threshold in step b is increased, and the action of injecting d-axis current in step c is switched to injecting reactive current.

6. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, The method further includes: synchronously acquiring the q-axis current of the permanent magnet motor during each execution of step c; extracting the harmonic energy within a predetermined mechanical fault frequency band from the acquired q-axis current data to form a harmonic fingerprint of the current damping event; comparing the harmonic fingerprint of the current damping event with a health status baseline calculated based on the harmonic fingerprints of multiple historical damping events; and generating an early warning signal for early degradation of the transmission chain when the comparison result shows that the harmonic energy exhibits a continuous increasing trend, and increasing the counting threshold in step b.

7. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, The method further includes: performing online resonant mode identification when the number of stress damping injection trigger commands generated in step b is zero within a set calm time window. The online resonant mode identification includes: injecting a spectrum-rich micro-perturbation excitation current signal into the d-axis current channel of the permanent magnet motor; synchronously acquiring the q-axis current of the permanent magnet motor as the output response signal; calculating the current resonant damping ratio of the wind power generation system drive train based on the micro-perturbation excitation current signal and the output response signal; and adaptively adjusting the counting threshold in step b based on the calculated resonant damping ratio.

8. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, The method also includes: real-time monitoring of the effective value of the grid voltage at the grid connection point; and, during step c, the injection amplitude of the d-axis current. It is determined by feedforward compensation based on the real-time magnitude of the monitored effective value of the grid voltage, and its compensation relationship satisfies: ,in, Injecting amplitude to the compensated d-axis current. This is the reference amplitude of the d-axis current when the effective value of the grid voltage is at its rated value. The rated grid voltage, This refers to the effective value of the grid voltage as monitored in real time.

9. The adaptive high power density permanent magnet motor control method according to claim 1, characterized in that, In step b, the time threshold, statistical time window, and counting threshold for the absolute value of the difference exceeding the threshold are all adaptively adjusted according to the operating conditions of the permanent magnet motor.

10. An adaptive high-power-density permanent magnet motor control system, applied to a wind power generation system, characterized in that, The system includes: The timing monitoring module is configured to monitor the continuous zero-crossing points of the three-phase current waveform of the permanent magnet motor stator in real time and calculate the actual time interval between each continuous zero-crossing point. The trigger decision module is connected to the timing monitoring module and is configured to compare the actual time interval calculated by the timing monitoring module with the theoretical time interval determined based on the current synchronization frequency of the permanent magnet motor. When the absolute value of the difference between the actual time interval and the theoretical time interval reaches a set counting threshold within a set statistical time window, an excitatory damping injection trigger command is generated. The collaborative control module, connected to the trigger decision module, is configured to inject d-axis current into the permanent magnet motor for a set duration in response to an excitable damping injection trigger command; and, during the set duration, to limit the maximum output command value of the q-axis current used to generate torque to a protection command value lower than its maximum output command value when the excitable damping injection trigger command is not triggered.

Images