Model-free predictive control harmonic and torque ripple suppression method for permanent magnet synchronous motor

Abstract

The invention discloses a model-free predictive control harmonic and torque ripple suppression method for a permanent magnet synchronous motor. The method comprises the following steps: sampling physical information of the permanent magnet synchronous motor; performing Clarke transformation on the obtained three-phase current, and performing Park transformation by applying the sampled rotor position to obtain actual current under the dq axis; according to the obtained dq-axis actual current and the motor winding voltage, updating an ESO-based hyper-local model after forward Euler method discrete transformation; calculating a control voltage by applying a traditional dead-beat current control method through a disturbance update value obtained through calculation; and carrying out inverse Park conversion on the deadbeat control voltage obtained after calculation, then carrying out traditional SVPWM modulation to obtain a three-phase PWM waveform, and inputting the three-phase PWM waveform into a driver to control the permanent magnet synchronous motor. According to the method, current ripples and torque ripples of model-free predictive control of the permanent magnet synchronous motor can be remarkably reduced, high-frequency harmonics in the operation process of the motor can be remarkably reduced, and the calculation burden is reduced.

Description

Technical Field

[0001] The present invention relates to a model-free predictive control method for permanent magnet synchronous motors, and particularly to a method for suppressing harmonics and torque ripple in model-free predictive control of permanent magnet synchronous motors. Background Art

[0002] Existing model-free predictive control permanent magnet synchronous control schemes based on resonators require manual adjustment of gain and frequency parameters, are difficult to adapt to the dynamic operating conditions of the motor (such as speed changes and load fluctuations), and the parameter coupling problem is prominent when multiple resonators are connected in parallel. The existing problems are as follows: First, the ability to suppress high-frequency harmonics is insufficient. Due to bandwidth limitations, traditional extended state observers (ESOs) are difficult to effectively suppress high-frequency harmonics, resulting in a significant increase in the total harmonic distortion (THD) of the current and torque ripple when the motor operates at high speed.

[0003] Second, parameter tuning is complex and lacks adaptability. Existing resonators require manual adjustment of gain and frequency parameters, are difficult to adapt to the dynamic operating conditions of the motor (such as speed changes and load fluctuations), and the parameter coupling problem is prominent when multiple resonators are connected in parallel.

[0004] Third, the computational complexity is high. The traditional complex resonator structure requires complex operations, and the computational amount surges when multiple resonators are connected in parallel, making it difficult to implement in real-time control systems. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for suppressing harmonics and torque ripple in model-free predictive control of permanent magnet synchronous motors, which can at least solve one of the technical problems in the related technologies to a certain extent.

[0006] To solve the above technical problems, the present invention adopts the following technical solutions: A method for suppressing harmonics and torque ripple in model-free predictive control of permanent magnet synchronous motors, comprising the following steps: 1. Sample the physical information of the permanent magnet synchronous motor, including: three-phase current, rotor position, and speed; 2. Perform Clarke transformation on the three-phase current obtained in step 1, and then apply the rotor position sampled in step 1 to perform Park transformation to obtain the actual current in the dq axis; 3. Update the hyperlocal model based on ESO after discrete transformation by the forward Euler method according to the actual current in the dq axis obtained in step 2 and the motor winding voltage; in the hyperlocal model, its feedback quantity can be expressed as the sum of the traditional current error term and the oscillator model term with real parameters; 4. Calculate the control voltage by applying the traditional deadbeat current control method through the disturbance update value calculated in step 3; 5. The deadbeat control voltage obtained after the calculation in step 4 is subjected to inverse Park transformation, and then traditional SVPWM modulation is performed to obtain three-phase PWM waveforms, which are input into the driver to control the permanent magnet synchronous motor.

[0007] Further, the process of obtaining the oscillator model with real parameters in step 3 is as follows: First, for the traditional oscillator model with complex parameters, its conjugate transfer function is obtained; Then, the conjugate transfer function obtained is added to the traditional oscillator model with complex parameters itself to obtain an oscillator model with real parameters; Then, according to the disturbance frequency, pre-distortion correction bilinear transformation is performed on it to obtain an oscillator model with real parameters in the discrete domain; Finally, for multiple oscillator models with real parameters for different disturbance frequencies, they are summed to obtain an oscillator model with real parameters for multiple disturbance frequencies.

[0008] Further, in the process of obtaining the oscillator model with real parameters, for multiple different disturbance frequencies, the number is a natural number greater than or equal to 1, and the disturbance frequencies are arbitrarily selected.

[0009] Further, the disturbance frequency is determined using the collected motor speed.

[0010] Further, the parameters of the oscillator model with real parameters are real-time online tuned through the gain online tuning method. The gain online tuning method includes: Calculating the complex gain of the denominator of the transfer function of the traditional super-local model observer at a given disturbance; Taking the gain phase of the oscillator model with real parameters as the phase of the above complex gain, and taking the gain of the oscillator model with real parameters as the product of the square of the bandwidth of the traditional observer and a given constant; the above calculation methods of the phase and gain uniquely determine the gain of the oscillator model with real parameters.

[0011] Further, in step 3, the update of the ESO-based super-local model after discrete transformation by the forward Euler method is performed according to formula (1): (1) Where, is the proposed oscillator model with real parameters; is the frequency of the disturbance; Where, ESO is the extended state observer; Where, ( k ) represents the value of the current control cycle, and ( k + 1) represents the updated value; Among them, is the predicted value of the current on the dq axes of the permanent magnet synchronous motor, is the voltage of the motor windings on the dq axes, is the control gain, is the predicted value of the disturbance, is the actual current obtained according to Step 1 and the predicted current is the current estimation error calculated therefrom, β 1 ，β 2 is the gain of the ESO, is the control period.

[0012] The beneficial effects of the present invention are as follows: 1. The present invention can significantly reduce the current ripple and torque ripple of the model-free predictive control (MFPCC) of the permanent magnet synchronous motor.

[0013] 2. The present invention can significantly reduce the high-frequency harmonics during the operation of the motor.

[0014] 3. In the model-free predictive control process, the present invention introduces a single or multiple parallel oscillator models with real parameters, simplifies the traditional oscillator model with complex parameters into an oscillator model with real parameters, and reduces the computational burden.

[0015] 4. The present invention performs real-time online tuning of the parameters of the oscillator model with real parameters through a gain online tuning method (automatically tunes the oscillator parameters online through the time-varying disturbance frequency under different motor speeds), avoiding the manual parameter tuning required by the traditional scheme (the traditional manual tuning method requires offline and manual debugging of parameters), and improving the deployment efficiency and convenience of the present invention. Description of the Drawings

[0016] Figure 1 is the control block diagram of the present invention; Figure 2 is the three-phase current waveform diagram of the permanent magnet synchronous motor in the prior art; Figure 3 is the three-phase current waveform diagram of the permanent magnet synchronous motor of the present invention; Figure 4 is the comparison diagram of the current harmonic content between the present invention and the traditional method.

[0017] The drawings are only for illustrative purposes and should not be construed as a limitation of this patent; for better illustration of this embodiment, some components in the drawings are omitted, enlarged or reduced, and do not represent the dimensions of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted. Detailed implementation manners

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention.

[0019] As Figure 1 shown, this embodiment discloses a method for suppressing harmonics and torque ripple in a permanent magnet synchronous motor without model prediction control, including the following steps: 1. Sample the physical information of the permanent magnet synchronous motor, including: three-phase current, rotor position, and speed.

[0020] 2. Perform Park transformation and Clarke transformation on the three-phase current obtained in step 1 to obtain the actual current under the dq axis .

[0021] 3. Update the Ultra-local model based on ESO after discrete transformation by the forward Euler method: (1) Among them, is the proposed oscillator model with real parameters; is the frequency of the disturbance; Among them, the Ultra-local model is a traditional ultra-local model, and the ESO is an extended state observer; Among them, ( k ) represents the value of the current control cycle, and ( k + 1) represents the updated value; Among them, is the predicted value of the dq-axis current of the permanent magnet synchronous motor, is the dq-axis motor winding voltage, is the control gain, is the predicted value of the disturbance, is the actual current obtained according to step 1 and the predicted current The current estimation error calculated; β 1 ，β 2 is the gain of the ESO, is the control cycle.

[0022] 4. Calculate the deadbeat control voltage through the updated value of the disturbance calculated in step 3 (2) Among them is the current control command.

[0023] 5. By performing an inverse Park transformation on the deadbeat control voltage obtained after the calculation in step 4, and then performing traditional SVPWM modulation, a three-phase PWM waveform is obtained and input into the driver to control the permanent magnet synchronous motor.

[0024] The oscillator model with real parameters in step 3 is obtained as follows: First, for the traditional oscillator model with complex parameters, find its conjugate transfer function; then add the obtained conjugate transfer function to the traditional oscillator model with complex parameters itself to get an oscillator model with real parameters; then perform pre-distortion correction bilinear transformation on it according to the perturbation frequency to obtain an oscillator model with real parameters in the discrete domain; finally, sum up multiple oscillator models with real parameters for different perturbation frequencies to obtain an oscillator model with real parameters for multiple perturbation frequencies.

[0025] The specific steps are as follows: (1) The traditional oscillator model with complex parameters is as follows (3) where is a complex parameter, is the perturbation frequency, is the imaginary unit, is the Laplace transform operator.

[0026] (2) Simplify the traditional oscillator model with complex parameters (3) to the real domain, including the following steps: First, the conjugate transfer function of the complex transfer function in formula (3) is as follows (4) Then, the oscillator model with real parameters is expressed in the form of the sum of the above (3) and (4) as follows: (5) where and respectively represent the real part and the imaginary part of the complex parameter ; thus all parameters in the proposed oscillator model with real parameters are real numbers.

[0027] (3) Perform pre-distortion correction bilinear transformation on the oscillator model in the above formula (5) to obtain: (6) where is the complex variable symbol; where and ​Two parameters of the transfer function of the oscillator model after discretization:

[0028] (4) By converting the above discretized model formula (6) into the time domain, the proposed oscillator model with real parameters can be obtained. .

[0029] (5) By summing up multiple oscillator models with real parameters at different perturbation frequencies, the proposed oscillator model with real parameters can be obtained ; where ; where represents the summation operation.

[0030] In step (5), for multiple different perturbation frequencies, the number can be a natural number greater than or equal to 1, and the perturbation frequencies can be arbitrarily selected.

[0031] For example, in this embodiment, it is selected as the 6nth order harmonic generated under the dq axis during motor operation, where n represents the harmonic order. The perturbation frequency is determined using the collected motor speed, and the specific calculation formula is .

[0032] Where is the frequency corresponding to the motor speed, and its calculation formula is , where is the collected motor speed, is the number of pole pairs of the motor.

[0033] In other embodiments, the perturbation frequencies can be arbitrarily selected and are not limited to the above frequency calculation formula.

[0034] This embodiment also proposes to perform real-time online tuning of the parameters of the oscillator model with real parameters and through a gain online tuning method, without the need for offline debugging based on traditional manual tuning methods.

[0035] The gain online tuning method includes: calculating the complex gain of the denominator of the transfer function based on the traditional super-local model observer at a given perturbation; taking the gain phase of the oscillator model with real parameters as the phase of the above complex gain, and taking the gain of the oscillator model with real parameters as the product of the square of the bandwidth of the traditional observer and a given constant; the above calculation methods of phase and gain uniquely determine the gain of the oscillator model with real parameters.

[0036] Specifically, it includes the following steps: 1) The classical ESO differential equations of the model-free predictive control of the Ultra-local model: (7) Where, is the differential operator, is the predicted value of the current on the dq axes of the permanent magnet synchronous motor, is the voltage of the dq-axis motor winding, is the control gain, is the predicted value of the disturbance, is the actual current obtained according to Step 1 and the predicted current The current estimation error calculated, β 1 ，β 2 is the gain of the ESO; Where, Ultra-local model is the traditional super-local model, and ESO is the extended state observer.

[0037] 2) Performing Laplace transform on it and finding its transfer function, we can obtain (8) For the denominator of the above transfer function at a certain specific disturbance frequency, find the magnitude of its gain value (9) 3) For the required gain , use formula (9) to calculate its phase (set the phase angle of the oscillator gain to the phase of the gain of the denominator of the transfer function of the extended state observer of the model-free predictive control at a specific disturbance frequency): (10) Where, refers to the phase angle of the complex number.

[0038] 4) Calculate the magnitude of the required gain : (11) Where, δ is a positive real number much smaller than the unit 1, and its value range can be between 0.001 - 0.1. For example, in this embodiment, δ takes 0.01. is the bandwidth of the classical ESO described in (8).

[0039] 5) Then, decompose the obtained in the above steps 3) - 4) into the real part and the non-real part: (12) Those calculated in the above formula (12) and are the parameters of the oscillator model with real parameters .

[0040] In the model-free predictive control process of the present invention, a single or multiple parallel oscillator models with real parameters are introduced, and the traditional oscillator with complex parameters is simplified to an oscillator model with real parameters, reducing the computational burden. And the present invention performs real-time online tuning of the parameters of the oscillator model with real parameters through a gain online tuning method, avoiding the manual parameter tuning required by the traditional scheme, and improving the deployment efficiency and convenience of the present invention.

[0041] The permanent magnet synchronous motor control platform applied to the harmonic and torque ripple suppression method of model-free predictive control of permanent magnet synchronous motor proposed in this embodiment includes a permanent magnet synchronous motor controller, a permanent magnet synchronous motor, and other accompanying servo systems.

[0042] The present invention can significantly reduce the current ripple and torque ripple of the MFPCC control of the permanent magnet synchronous motor.

[0043] As Figure 2 shown, it is a schematic diagram of the three-phase current waveform / torque ripple waveform of the permanent magnet synchronous motor controlled by using the oscillator principle, but without using a real-gain oscillator and using the method of manually tuning the oscillator parameters, at 1000 revolutions and a load of 0.127 Nm. Channels 1-3 are the three-phase currents, and channel 4 is the torque fluctuation.

[0044] Figure 2 Among them: the vertical coordinate ratio of the three-phase current waveform is 200 mA / div, the vertical coordinate ratio of the torque ripple waveform is 5 ms / div, and the horizontal coordinate ratio of both is 5 ms / div. Figure 2 The effective value of the three-phase current among them is 233.6 mA (RMS), and the torque ripple value is 1.74%.

[0045] As Figure 3 shown, it is a schematic diagram of the three-phase current waveform / torque ripple waveform of the permanent magnet synchronous motor controlled by the method proposed by the present invention. Channels 1-3 are the three-phase currents, and channel 4 is the torque fluctuation.

[0046] Figure 3 Among them: the vertical coordinate ratio of the three-phase current waveform is 200 mA / div, the vertical coordinate ratio of the torque ripple waveform is 5 ms / div, and the horizontal coordinate ratio of both is 5 ms / div. Figure 3 The effective value of the three-phase current among them is 232.6 mA (RMS), and the torque ripple value is 0.63%.

[0047] By Figure 3and Figure 2 By comparison, it can be seen that the method proposed in the present invention significantly reduces the current ripple and torque ripple.

[0048] The present invention can also significantly reduce the 6k±1 harmonics during the operation of the motor, where k = 1, 2, 3...... The comparison of the current harmonic content of the method proposed in the present invention under different operating conditions of the motor with the traditional method is shown as Figure 4 shown.

[0049] Figure 4 In the figure, Method 1 is the traditional model-free predictive control method without using an oscillator, Method 2 is the method that uses the oscillator principle but does not use a real-number gain oscillator and uses manually tuned oscillator parameters, and Method 3 is the method of the present invention. Compared with Method 1 and Method 2, the method of the present invention significantly reduces each current harmonic component.

[0050] The above embodiments are only used to illustrate rather than limit the technical solutions of the present invention. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent substitutions can still be made to the present invention, and any modification or partial substitution without departing from the spirit and scope of the present invention should be covered by the scope of the claims of the present invention.

[0051] If terms such as "first" and "second" are used in this article to limit components, those skilled in the art should be aware that the use of "first" and "second" is only for the convenience of describing the present invention and simplifying the description. Without additional statements, the above terms have no special meanings.

[0052] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by terms such as "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0053] In the description of the present invention, it should be noted that, unless otherwise clearly specified and defined, the terms "installed", "connected", and "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

Claims

1. Method for suppressing harmonics and torque ripple of permanent magnet synchronous motor based on model-free predictive control, characterized in that Including the following steps: (1) Sampling the physical information of the permanent magnet synchronous motor, including three-phase current, rotor position, and rotational speed; (2) Performing Clarke transformation on the three-phase current obtained in step (1), and then applying the rotor position sampled in step (1) for Park transformation to obtain the actual current in the dq axes; (3) Updating the ESO-based super-local model after discrete transformation by the forward Euler method according to the actual current in the dq axes obtained in step (2) and the motor winding voltage; in the super-local model, its feedback quantity can be expressed as the sum of the traditional current error term and the oscillator model term with real parameters; (4) Calculating the control voltage by applying the traditional deadbeat current control method through the disturbance update value calculated in step (3); (5) Performing inverse Park transformation on the deadbeat control voltage obtained after calculation in step (4), and then performing traditional SVPWM modulation to obtain the three-phase PWM waveform, which is input into the driver to control the permanent magnet synchronous motor.

2. The method for suppressing harmonics and torque ripple in model-free predictive control of permanent magnet synchronous motors according to claim 1, characterized in that: The obtaining process of the oscillator model with real parameters in step (3) is as follows: First, find the conjugate transfer function of the traditional oscillator model with complex parameters; Then add the obtained conjugate transfer function to the traditional oscillator model with complex parameters itself to obtain the oscillator model with real parameters; Then perform pre-distortion correction bilinear transformation on it according to the disturbance frequency to obtain the oscillator model with real parameters in the discrete domain; Finally, sum up multiple oscillator models with real parameters for different disturbance frequencies to obtain the oscillator model with real parameters for multiple disturbance frequencies.

3. The method for suppressing harmonics and torque ripple of model-free predictive control of permanent magnet synchronous motor according to claim 2, wherein In the process of obtaining the oscillator model with real parameters, for multiple different disturbance frequencies, the number is a natural number greater than or equal to 1, and the disturbance frequencies are arbitrarily selected.

4. The method for suppressing harmonics and torque ripple of model-free predictive control of a permanent magnet synchronous motor according to claim 3, wherein The disturbance frequency is determined using the collected motor rotational speed.

5. The method for suppressing harmonics and torque ripple of model-free predictive control of permanent magnet synchronous motor according to claim 2, characterized in that: The parameters of the oscillator model with real parameters are real-time online tuned through the gain online tuning method, and the gain online tuning method includes: Calculating the complex gain of the denominator of the transfer function of the traditional super-local model observer at a given disturbance; Taking the gain phase of the oscillator model with real parameters as the phase of the above complex gain, and taking the gain of the oscillator model with real parameters as the product of the square of the bandwidth of the traditional observer and a given constant; the above calculation methods of the phase and gain uniquely determine the gain of the oscillator model with real parameters.

6. The method for suppressing harmonics and torque ripple of a permanent magnet synchronous motor based on model-free predictive control according to any one of claims 1-5, characterized in that: In step (3), the update of the ESO-based super-local model after discrete transformation by the forward Euler method is based on formula (1): （1） Among them, is the proposed oscillator model with real parameters; is the frequency of the perturbation; where ESO is the extended state observer; Among them, ( k ) represents the value of the current control cycle, and ( k + 1) represents the updated value; Among them, is the predicted value of the current in the dq axes of the permanent magnet synchronous motor, is the motor winding voltage in the dq axes, is the control gain, is the predicted value of the disturbance, is the actual current obtained according to step 1 in claim 1 and the predicted current to calculate the current estimation error, β 1 ，β 2 is the gain of the ESO, is the control period.

Images