A current harmonic dynamic suppression method for brushless motor vector control

Abstract

The application discloses a current harmonic dynamic suppression method for brushless motor vector control and relates to the technical field of motor control, and comprises the following steps: S1, real-time current sampling and coordinate transformation: collecting three-phase stator currents of the motor, and obtaining d-axis current i_d and q-axis current i_q in a synchronous rotating coordinate system through Clarke transformation and Park transformation; and S2, sectional harmonic spectrum scanning. The current harmonic dynamic suppression method for brushless motor vector control, in which a disturbance observer feedforward compensation is combined with adaptive resonance control. The disturbance observer eliminates harmonic disturbance generated by inverter nonlinearity (dead zone, tube voltage drop) from a "source"; the adaptive resonance controller suppresses harmonics introduced by motor body (tooth slot, back electromotive force distortion) and load characteristics from a "propagation path". The two work in a clear division and cooperation, covering a complete link from harmonic generation to harmonic response.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, specifically to a method for dynamic suppression of current harmonics in vector control of a brushless motor. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) and brushless DC motors (BLDCs) are widely used in industrial servo systems, new energy vehicles, aerospace, and other fields due to their high power density and high efficiency. Vector control (FOC) is a general control framework for achieving high-performance speed regulation of these motors. However, in practical applications, due to various factors such as inverter nonlinearity (dead zone effect, tube voltage drop), non-ideal characteristics of the motor itself (coiling effect, back EMF harmonics, magnetic circuit saturation), and load disturbances, a large number of harmonic components inevitably exist in the motor phase current. Current harmonics not only increase the iron and copper losses of the motor and reduce system efficiency, but also cause torque pulsation, electromagnetic noise, and mechanical vibration, severely restricting the application of motors in high-precision, low-noise applications.

[0003] Existing technologies for suppressing current harmonics can be mainly categorized as follows:

[0004] (1) Fixed-order resonant control method: This method uses a resonant controller (PR / R controller) with preset harmonic frequencies such as 5th, 7th, 11th, and 13th to provide high gain at the corresponding frequencies to achieve harmonic suppression. The disadvantage of this method is that the harmonic order is fixed and cannot be dynamically adjusted according to the changes in the dominant harmonic order under actual operating conditions. When the motor operating conditions change (such as changes in speed and load), the harmonic spectrum distribution drifts, and the preset fixed-order resonant controller may be "inaccurate" for the actual dominant harmonics and "over-suppress" the non-dominant harmonics, resulting in wasted control resources and reduced suppression effect.

[0005] (2) Repetitive control method: Based on the internal model principle, the error signal within the fundamental period is learned and compensated cycle by cycle, which can theoretically suppress all integer harmonics. However, repetitive control has an inherent defect: slow dynamic response, requiring at least one fundamental period to establish the compensation effect. Under the condition of frequent changes in motor speed, the fundamental period changes continuously, making it difficult for repetitive control to track effectively.

[0006] (3) Adaptive notch filter method: Harmonic frequencies are identified online, and the center frequency of the notch filter is adaptively adjusted to filter out harmonics. The core of this method is "filtering" rather than "compensation". While filtering out harmonics, the notch filter may introduce additional phase lag, affecting the stability margin of the current loop.

[0007] (4) Inverter dead time compensation method: Feedforward compensation of inverter dead time based on phase current polarity. This type of method only targets the single harmonic source of inverter nonlinearity and does not involve harmonic components introduced by the motor body (cogging, back EMF distortion) and load characteristics.

[0008] (5) Model Predictive Control Harmonic Suppression Method: Harmonic constraint terms are introduced into the cost function of model predictive control, and the optimal voltage vector is solved through online optimization. This type of method involves a large amount of computation, is highly dependent on the accuracy of the motor model, and the suppression effect decreases significantly when parameters are perturbed.

[0009] In summary, existing technologies lack a method that can identify the dominant harmonic order in the current in real time and dynamically, and adaptively adjust the suppression strategy accordingly, while simultaneously meeting the dual requirements of "eliminating inverter nonlinear harmonics at the source" and "suppressing motor body and load harmonics along the path". Summary of the Invention

[0010] To address the shortcomings of existing technologies, this invention provides a method for dynamic suppression of current harmonics in vector control of brushless motors, which solves the problems mentioned in the background section.

[0011] To achieve the above objectives, the present invention provides a method for dynamic suppression of current harmonics in vector control of a brushless motor, comprising the following steps:

[0012] S1. Real-time current sampling and coordinate transformation: Collect the three-phase stator current of the motor, and obtain the d-axis current i_d and q-axis current i_q in the synchronous rotating coordinate system through Clarke transformation and Park transformation;

[0013] S2. Segmented harmonic spectrum scanning: Perform segmented spectrum analysis on the i_q or i_d components—divide the preset harmonic frequency range into low-frequency, mid-frequency and high-frequency bands, and perform spectrum scanning in each band with different spectral resolutions to obtain the amplitude and phase information of each harmonic.

[0014] S3. Dynamic identification of dominant harmonic order: Compare the amplitude of each harmonic with a dynamic threshold, mark the harmonic order whose amplitude exceeds the dynamic threshold as the dominant harmonic order, and generate the dominant harmonic order set H_dom for the current control cycle.

[0015] S4. Adaptive Resonant Controller On-Demand Configuration: Activate the corresponding adaptive resonant controller for each dominant harmonic order in H_dom, and deactivate resonant controllers that do not belong to H_dom.

[0016] S5. Disturbance Observer Feedforward Compensation: Construct a disturbance observer with i_d, i_q and the current voltage command as inputs to estimate the equivalent disturbance voltage of the inverter nonlinearity on the d / q axis, and superimpose it onto the output of the inner current loop in a feedforward manner;

[0017] S6. Harmonic suppression voltage synthesis and output: The outputs of each activated adaptive resonant controller are summed to obtain the harmonic compensation voltage, which is then superimposed on the output voltage of the current inner loop PI regulator to form the final d / q axis voltage command.

[0018] S7. Parameter Adaptive Update and Loop: Update the resonant frequency of each activated resonant controller according to the electric angular velocity of the motor, and adjust the resonant bandwidth according to the frequency band of the dominant harmonic order, and then return to S1.

[0019] Furthermore, in step S2: the low-frequency band is 1 to 10 times the base frequency, the mid-frequency band is 11 to 30 times the base frequency, and the high-frequency band is 31 times the base frequency or higher up to the Nyquist frequency; the frequency step size of the low-frequency band is less than or equal to 1 / 2 of the base frequency, the frequency step size of the mid-frequency band is less than or equal to 1 times the base frequency, and the frequency step size of the high-frequency band is less than or equal to 2 times the base frequency.

[0020] Furthermore, the feature is that the spectral scanning in step S2 is implemented using a sliding window discrete Fourier transform or the Goertzel algorithm.

[0021] Furthermore, the feature is that, in step S3, the dynamic threshold is obtained by multiplying the fundamental amplitude of the current by a preset proportional coefficient, the default value of the proportional coefficient is 2% to 5% of the fundamental amplitude, and the proportional coefficient can be adjusted in segments according to the operating conditions of the motor.

[0022] Furthermore, the feature is that, in step S4, the transfer function of each adaptive resonant controller is:

[0023] G_h(s) = (2.K_h.omega_bh.s) / (s2 + 2.omega_bh.s + omega_h2)

[0024] Where omega_h = h.omega_e is the angular frequency of the h-th harmonic, omega_e is the electric angular velocity of the motor, K_h is the resonant gain, and omega_bh is the resonant bandwidth; the resonant gain K_h is adaptively adjusted according to the ratio of the amplitude of the harmonic to the amplitude of the fundamental wave—the larger the ratio, the larger the value of K_h.

[0025] Furthermore, in step S5, the disturbance observer is implemented using a reduced-order Luneburg observer or a sliding mode observer; when a reduced-order Luneburg observer is used, its observer poles are located at 3 to 5 times the bandwidth of the current loop.

[0026] Furthermore, in step S7, the adjustment of the resonant bandwidth adopts an asymmetric allocation strategy: a narrower bandwidth is allocated to the dominant harmonics in the low-frequency band to provide high Q value for accurate suppression, and a wider bandwidth is allocated to the dominant harmonics in the high-frequency band to provide frequency robustness.

[0027] Furthermore, in step S6, the harmonic compensation voltage passes through a dynamic limiting stage before being superimposed onto the current loop output. The limiting value of the dynamic limiting stage is determined in real time based on the margin of the inverter DC bus voltage.

[0028] Furthermore, in step S2, a segmented spectral scan is performed simultaneously on the i_q and i_d components, and in step S6, harmonic compensation voltages for the d-axis and q-axis are synthesized respectively.

[0029] This invention provides a dynamic current harmonic suppression method for brushless motor vector control, which has the following advantages: This dynamic current harmonic suppression method for brushless motor vector control dynamically identifies and suppresses harmonics on demand, resulting in high suppression efficiency. Unlike fixed-order resonant control methods that preset specific orders such as 5th, 7th, 11th, and 13th, this invention identifies the dominant harmonic order under the current operating condition in real time through segmented spectrum scanning, activating the resonant controller only for dominant harmonics that truly exceed the threshold. When changes in operating conditions cause the dominant harmonic order to drift, the resonant controller configuration automatically adjusts accordingly, ensuring that suppression resources are always concentrated on the harmonics that most need to be addressed.

[0030] This invention combines disturbance observer feedforward compensation with adaptive resonant control. The disturbance observer eliminates harmonic disturbances generated by inverter nonlinearity (dead zone, tube voltage drop) at the "source"; the adaptive resonant controller suppresses harmonics introduced by the motor body (cogging, back EMF distortion) and load characteristics along the "propagation path." The two components have clear division of labor and work synergistically, covering the complete link from harmonic generation to harmonic response.

[0031] This invention allocates different resonant controller bandwidths according to the frequency band of the dominant harmonic: low-frequency harmonics (such as the 5th and 7th harmonics) change slowly, so a narrow bandwidth (high Q value) is allocated to achieve precise suppression; high-frequency harmonics (such as the 31st harmonic and above) may sweep the frequency rapidly due to speed fluctuations, so a wide bandwidth (low Q value) is allocated to ensure robustness in the event of frequency mismatch.

[0032] The resonant frequency of the resonant controller, omega_h = h*omega_e, is continuously updated with the electric angular velocity omega_e, ensuring that the resonant controller is always aligned with the target harmonic frequency across the entire speed range of the motor, thus avoiding the "detuning" phenomenon caused by speed changes.

[0033] Before being superimposed, the harmonic compensation voltage is dynamically limited based on the DC bus voltage margin to prevent voltage saturation due to excessive harmonic compensation voltage in high-speed weak magnetic field regions or heavy load conditions, thus ensuring the overall stability of the vector control system. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the overall control block diagram of the dynamic suppression method for current harmonics in the vector control of a brushless motor according to the present invention. Detailed Implementation

[0035] Example 1

[0036] A dynamic current harmonic suppression method for vector control of a brushless motor is disclosed in this embodiment. The controlled object is a surface-mounted permanent magnet synchronous motor (SPMSM) with a rated power of 1.5kW, a rated speed of 3000rpm, and 4 pole pairs (8 poles). The dynamic current harmonic suppression method of this invention is implemented on a vector control platform based on a DSP (TMS320F28379D). The vector control framework is a standard id=0 control, the PWM frequency is 10kHz, and the current loop control frequency is synchronized with the PWM frequency.

[0037] The overall control structure of this method embeds a harmonic dynamic suppression module at the output of the current inner loop of the standard vector control framework. This module receives the dq-axis currents i_d and i_q and the PI regulator output voltages u_d_pi and u_q_pi as inputs, and outputs harmonic compensation voltages u_d_comp and u_q_comp and disturbance compensation voltage, which are superimposed to form the final voltage commands u_d* and u_q*.

[0038] The method in this embodiment is performed according to the following steps:

[0039] S1. Real-time current sampling and coordinate transformation: The two-phase stator currents are synchronously sampled at a frequency of 10kHz using an ADC (the third phase is calculated by ia+ib+ic=0). The currents i_alpha and i_beta in the alpha-beta coordinate system are obtained by Clarke transformation, and then the dq-axis currents i_d and i_q are obtained by Park transformation. The rotor electrical angle required for Park transformation is provided by an incremental encoder, and the electrical angular velocity omega_e is obtained by M / T method speed measurement.

[0040] S2. Segmented Harmonic Spectrum Scan: Segmented spectrum scanning is performed on the i_q components. In this embodiment, the low-frequency band is set to 1 to 10 times the fundamental frequency (the fundamental frequency is 200Hz at a rated speed of 3000rpm, corresponding to a maximum of 2kHz), the mid-frequency band is set to 11 to 30 times the fundamental frequency (maximum 6kHz), and the high-frequency band is set to 31 times the fundamental frequency and above up to the Nyquist frequency (5kHz). The low-frequency band performs sliding window DFT with a frequency step size of 0.5 times the fundamental frequency, the mid-frequency band with a step size of 1 times the fundamental frequency, and the high-frequency band with a step size of 2 times the fundamental frequency. The scanning of each frequency band is performed in parallel, and the total calculation time consumed is approximately 15 microseconds (at a DSP main frequency of 200MHz), which does not affect the real-time performance of the 10kHz control cycle (100 microseconds).

[0041] S3. Dynamic Identification of Dominant Harmonic Orders: The amplitude values ​​I_h of each harmonic obtained in S2 are compared with the dynamic threshold I_th = I1 × 3% (I1 is the fundamental frequency amplitude). In this embodiment, when running at a rated speed of 3000 rpm and a rated load, the amplitude values ​​of the 5th (1000 Hz), 7th (1400 Hz), 11th (2200 Hz), and 13th (2600 Hz) harmonics are detected to exceed the threshold, generating the dominant harmonic order set H_dom = {5, 7, 11, 13}.

[0042] S4. Adaptive Resonant Controller On-Demand Configuration: For H_dom = {5, 7, 11, 13}, activate four resonant controllers: ARC_5, ARC_7, ARC_11, and ARC_13. The transfer function of each ARC is discretized by bilinear transform and implemented in the DSP. In this embodiment, the 5th and 7th harmonics are in the low-frequency range, and their K_h is 1.2 times the base value; the 11th and 13th harmonics are in the mid-frequency range, and their K_h is 1.0 times the base value.

[0043] S5. Disturbance Observer Feedforward Compensation: A reduced-order Luenberger observer is used to estimate the equivalent dq-axis disturbance voltage. The observer takes i_d, i_q and u_d*, u_q* from the previous cycle as inputs. The observer poles are positioned at 2000Hz, which is four times the current loop bandwidth (500Hz). The estimated equivalent disturbance voltage is fed forward and added to the current loop output.

[0044] S6. Harmonic Suppression Voltage Synthesis and Output: The outputs of the four resonant controllers ARC_5 to ARC_13 are summed to obtain u_q_comp (mainly focusing on the q-axis under id=0 control). This is then added to the disturbance observer output and superimposed with the current loop PI output u_q_pi. After dynamic limiting (limiting value = DC bus voltage x 0.9 - |u_q_pi|), the final u_q* is formed. This is then converted to 6 PWM signals to drive the three-phase inverter bridge after Park inverse transformation and SVPWM modulation.

[0045] S7. Parameter Adaptive Update and Loop: At the end of each control cycle (100 microseconds), based on the latest measured electrical angular velocity omega_e from the encoder, update the resonant frequency h*omega_e and resonant bandwidth omega_bh of ARC_5 to ARC_13. After the update is complete, return to S1.

[0046] Comparison before and after implementation: Before implementation, the 5th and 7th harmonics accounted for 6.2% and 4.8% of the fundamental frequency, respectively, and the THD was 9.1%; after implementing the method of this invention, the 5th and 7th harmonics decreased to 0.8% and 0.6%, respectively, and the THD decreased to 2.3%, verifying the effectiveness of this method.

[0047] Example 2

[0048] The difference between this embodiment and Embodiment 1 is that the motor operates under low-speed, heavy-load conditions (500 rpm, rated torque), where the inverter dead-time effect dominates as the source of harmonics. The dynamic threshold scaling factor in S3 is reduced from 3% to 2% to more sensitively capture harmonics. After identification by S2-S3, H_dom = {5, 7}, only ARC_5 and ARC_7 are activated. Simultaneously, the observer gain in S5 is appropriately increased to accommodate the nonlinear enhancement of the dead-time effect at low speeds. The results show that under low-speed, heavy-load conditions, THD decreases from 15.3% before implementation to 3.1% after implementation, verifying the adaptive capability of this method under different operating conditions.

[0049] Example 3

[0050] The difference between this embodiment and Embodiment 1 is that during the transient process of the motor speed changing from 1500 rpm to 3000 rpm, the spectrum scan result in S2 reflects the change of harmonic distribution from {3, 5, 9} to {5, 7, 11, 13} in real time. S3 dynamically updates the H_dom set, and S4 automatically switches the resonant controller configuration. In S7, the resonant frequency is continuously updated with omega_e, ensuring the frequency tracking accuracy of the resonant controller during the transient speed change and avoiding the "detuning" problem of traditional fixed-frequency resonators when the speed changes.

[0051] 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 present invention.

Claims

1. A method for dynamic suppression of current harmonics in vector control of a brushless motor, characterized in that, Includes the following steps: S1. Real-time current sampling and coordinate transformation: Collect the three-phase stator current of the motor, and obtain the d-axis current i_d and q-axis current i_q in the synchronous rotating coordinate system through Clarke transformation and Park transformation; S2. Segmented harmonic spectrum scanning: Perform segmented spectrum analysis on the i_q or i_d components—divide the preset harmonic frequency range into low-frequency, mid-frequency and high-frequency bands, and perform spectrum scanning in each band with different spectral resolutions to obtain the amplitude and phase information of each harmonic. S3. Dynamic identification of dominant harmonic order: Compare the amplitude of each harmonic with a dynamic threshold, mark the harmonic order whose amplitude exceeds the dynamic threshold as the dominant harmonic order, and generate the dominant harmonic order set H_dom for the current control cycle. S4. Adaptive Resonant Controller On-Demand Configuration: Activate the corresponding adaptive resonant controller for each dominant harmonic order in H_dom, and deactivate resonant controllers that do not belong to H_dom. S5. Disturbance Observer Feedforward Compensation: Construct a disturbance observer with i_d, i_q and the current voltage command as inputs to estimate the equivalent disturbance voltage of the inverter nonlinearity on the d / q axis, and superimpose it onto the output of the inner current loop in a feedforward manner; S6. Harmonic suppression voltage synthesis and output: The outputs of each activated adaptive resonant controller are summed to obtain the harmonic compensation voltage, which is then superimposed on the output voltage of the current inner loop PI regulator to form the final d / q axis voltage command. S7. Parameter Adaptive Update and Loop: Update the resonant frequency of each activated resonant controller according to the electric angular velocity of the motor, and adjust the resonant bandwidth according to the frequency band of the dominant harmonic order, and then return to S1.

2. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 1, characterized in that, In step S2: the low frequency band is 1 to 10 times the base frequency, the mid frequency band is 11 to 30 times the base frequency, and the high frequency band is 31 times the base frequency or higher up to the Nyquist frequency; the frequency step size of the low frequency band is less than or equal to 1 / 2 of the base frequency, the frequency step size of the mid frequency band is less than or equal to 1 times the base frequency, and the frequency step size of the high frequency band is less than or equal to 2 times the base frequency.

3. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 1, characterized in that, The spectral scanning in step S2 is implemented using a sliding window discrete Fourier transform or the Goertzel algorithm.

4. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 1, characterized in that, In step S3, the dynamic threshold is obtained by multiplying the fundamental amplitude of the current by a preset proportional coefficient. The default value of the proportional coefficient is 2% to 5% of the fundamental amplitude, and the proportional coefficient can be adjusted in segments according to the operating conditions of the motor.

5. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 1, characterized in that, In step S4, the transfer function of each adaptive resonant controller is: G_h(s) = (2.K_h.omega_bh.s) / (s2 + 2.omega_bh.s + omega_h2) Where omega_h = h.omega_e is the angular frequency of the h-th harmonic, omega_e is the electric angular velocity of the motor, K_h is the resonant gain, and omega_bh is the resonant bandwidth; the resonant gain K_h is adaptively adjusted according to the ratio of the amplitude of the harmonic to the amplitude of the fundamental wave—the larger the ratio, the larger the value of K_h.

6. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 1, characterized in that, In step S5, the disturbance observer is implemented using a reduced-order Luneburg observer or a sliding mode observer; when a reduced-order Luneburg observer is used, its observer poles are located at 3 to 5 times the bandwidth of the current loop.

7. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 5, characterized in that, In step S7, the adjustment of the resonant bandwidth adopts an asymmetric allocation strategy: a narrower bandwidth is allocated to the dominant harmonics in the low-frequency band to provide high Q value for accurate suppression, and a wider bandwidth is allocated to the dominant harmonics in the high-frequency band to provide frequency robustness.

8. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 1, characterized in that, In step S6, the harmonic compensation voltage passes through a dynamic limiting stage before being superimposed on the current loop output. The limiting value of the dynamic limiting is determined in real time based on the margin of the inverter DC bus voltage.

9. The method for dynamic suppression of current harmonics in vector control of a brushless motor according to claim 1, characterized in that, In step S2, segmented spectrum scanning is performed on both the i_q and i_d components. In step S6, harmonic compensation voltages for the d-axis and q-axis are synthesized respectively.

Images