Control method and system for torque ripple suppression of permanent magnet synchronous motor

Abstract

The application discloses a control method and system for inhibiting torque ripple of a permanent magnet synchronous motor, and relates to the technical field of motor control. The application comprises the following steps: S1, signal acquisition and coordinate transformation; S2, constructing a discrete sliding mode torque observer to online estimate instantaneous torque ripple; S3, generating q-axis compensation voltage through a PI compensation controller; and S5, modulating a driving output. The final dq-axis modulation voltage instruction is subjected to Park inverse transformation and SVPWM modulation to generate an inverter driving signal, which drives the permanent magnet synchronous motor to operate and real-time offset the instantaneous torque ripple. The application adopts a continuous saturation function to replace an ideal sign function in view of the inherent chattering problem of a traditional sliding mode observer, and sets a reasonable boundary layer thickness, thereby ensuring robustness, effectively inhibiting high-frequency chattering, avoiding the adverse effect of high-frequency noise on the steady-state accuracy of a current loop and a torque loop, and realizing the algorithm structure with clearness and a small amount of multiplication and addition operations.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, and in particular to a method and system for suppressing torque ripple in permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in electric vehicles, industrial servos, aerospace, and other fields due to their advantages such as small size, high power density, and high efficiency. In the control process of PMSMs, the smoothness of the output torque is crucial to system performance. Excessive fluctuations in the motor's output torque will directly cause speed pulsation, system vibration, and noise, reducing control accuracy and operational stability. In severe cases, it can lead to fatigue damage to the transmission mechanism, accelerated bearing wear, and even cause motor overcurrent or runaway, seriously affecting the motor's lifespan and system reliability.

[0003] Currently, technologies for reducing torque ripple are mainly divided into motor body optimization methods and control strategy optimization methods. Although motor body optimization methods (such as magnetic pole shape optimization, skewed slots, etc.) can reduce cogging torque at the source, they increase manufacturing costs and process difficulty, and their suppression capability is fixed and cannot adapt to load and temperature changes. Among control strategy optimization methods, harmonic injection methods (such as patent CN112910336B) require extensive offline calibration of injection parameters and are extremely sensitive to changes in motor parameters; model predictive control methods (such as patent CN118677308A) have fast dynamic response but large steady-state torque ripple and are highly dependent on accurate motor models; active disturbance rejection control methods (such as patent CN116317720A) refer to all disturbances as "concentrated disturbances," making it difficult to accurately separate torque ripple components at specific frequencies, and the estimation accuracy drops significantly when the harmonic frequency exceeds the observer bandwidth; extended Kalman filtering methods (such as patent CN202210742080) have high computational complexity, are sensitive to the selection of noise covariance matrix, and have high engineering implementation thresholds; the combination of sliding mode observer and active disturbance rejection methods (such as patent CN107359837B) has strong robustness, but the parameter tuning is complex, and sliding mode chattering affects steady-state accuracy. The aforementioned methods generally suffer from drawbacks such as reliance on precise models or offline calibration, inability to perceive instantaneous torque fluctuations in real time, complex algorithms, and difficulty in simultaneously achieving steady-state and dynamic performance. These limitations restrict their application in demanding fields such as high-end manufacturing and electric vehicles. Therefore, it is necessary to develop a robust, full-condition control method that requires no precise models or offline calibration, has low computational cost, and can perceive and directly compensate for instantaneous torque fluctuations in real time. Summary of the Invention

[0004] The purpose of this invention is to provide a method and control system for suppressing torque ripple in permanent magnet synchronous motors, thereby solving at least one of a number of technical problems.

[0005] The core technical challenge lies in the inability to acquire instantaneous torque fluctuation information directly and in a timely manner for targeted compensation. Traditional sliding mode observers also suffer from inherent chattering issues.

[0006] The present invention provides the following solution: According to a first aspect of the present invention, a control method for suppressing torque ripple in a permanent magnet synchronous motor is provided, comprising: Step S1: Signal acquisition and coordinate transformation; The three-phase voltage, three-phase current, rotor electrical angle, and electrical angular velocity of the permanent magnet synchronous motor are collected. The voltage and current components under the two-phase rotating dq axis are obtained through Clark and Park transformations. The nominal electromagnetic torque is obtained based on the deviation between the given speed and the actual electrical angular velocity. Step S2: Online estimation of instantaneous torque ripple ; Step S3: Obtain the q-axis compensation voltage; The instantaneous torque ripple is input into the PI compensation controller, and the q-axis compensation voltage is calculated by combining the relationship between the electromagnetic torque of the permanent magnet synchronous motor and the q-axis current. ; Step S4: The compensation voltage is fused with the fundamental voltage; q-axis compensation voltage Superimposed on traditional field-oriented control, the output dq-axis fundamental voltage and The final dq-axis modulation voltage command is obtained. and ; Step S5: Modulate the drive output; The final dq axis modulation voltage command is converted into an inverter drive signal through Park inverse transformation and SVPWM modulation, which drives the permanent magnet synchronous motor to run and cancels instantaneous torque pulsation in real time.

[0007] Furthermore, the expressions for the Clark transformation matrix and the Park transformation matrix in step S1 are as follows: ; ; This is the Clark transformation matrix from the three-phase stationary coordinate system to the two-phase stationary αβ coordinate system. The Park transformation matrix is ​​given by the formula for the two-phase stationary αβ coordinate system to the two-phase rotating dq coordinate system. The rotor electrical angle includes the electrical angle between the rotor d-axis axis and the α-axis axis of the two-phase stationary coordinate system, which changes in real time as the rotor rotates. The cosine value of the rotor electrical angle is used as the in-phase projection coefficient between the two coordinate systems, corresponding to the amplitude coupling calculation of the d-axis and α-axis components and the q-axis and β-axis components, respectively. The sine value of the rotor electrical angle is used as the orthogonal projection coefficient between the two coordinate systems, corresponding to the amplitude coupling calculation of the d-axis and β-axis components; The negative value of the sine of the rotor electrical angle is used as the orthogonal projection coefficient between the q-axis and α-axis components. Since the electrical angle of the q-axis leads the d-axis by 90°, its α-axis projection corresponds to the negative value of the sine, ensuring that the d and q axes always maintain an orthogonal relationship. The formulas for converting three-phase voltage and three-phase current are as follows: , ; ; In the formula, It is a three-phase voltage. It is a three-phase current. For two-phase stationary coordinate system voltages, For two-phase stationary coordinate system currents, For the dq axis voltage components, For the dq axis current components.

[0008] Furthermore, online estimation of instantaneous torque ripple. include: Based on the discrete mechanical motion equations of the motor, a continuous saturated function is used to replace the ideal sign function. Speed ​​and load torque observations are obtained through sliding mode convergence. The total electromagnetic torque estimate is then calculated by back-calculating the sum of the nominal and total electromagnetic torque estimates. The difference between the nominal and total electromagnetic torque estimates yields the instantaneous torque pulsation, including cogging torque, back EMF harmonics, and inverter dead-zone effects. .

[0009] Furthermore, it also includes: the discrete mechanical motion equations of the motor are derived from the continuous dynamic equations of the motor via the forward Euler method and the sampling period. Discretization yields the result; The dynamic continuity equation of the electric motor is: ; In the formula, The moment of inertia of the motor. The mechanical angular velocity of the motor. For electromagnetic torque, The total load torque including friction. The damping coefficient; Discretized equations of motion for the motor: ; In the formula, The sampling period of the control system; k is the kth sampling time. Let k be the mechanical angular velocity at time k. Let K be the electromagnetic torque at time k. Let be the load torque at time k.

[0010] Furthermore, with , Let be the state variable, and the observation equation be: Speed ​​observation equation: ; Load torque observation equation: ; in, The rotational speed observation value at time k is... Let k be the load torque observation value at time k. To account for the deviation in rotational speed observation, , For observer gain greater than 0, It is a continuously saturated function.

[0011] Furthermore, in step S2, the sliding surface is defined. Select gain The sliding surface converges to the boundary layer in a finite number of steps. ,in Boundary layer thickness; After sliding mode stabilization, the total electromagnetic torque is estimated by back-calculating based on the discrete mechanical motion of the motor: ; Instantaneous torque ripple calculation formula: ; In the formula, Let K be the nominal electromagnetic torque at time k. This is the estimated value of the total electromagnetic torque at time k. This refers to instantaneous torque pulsation.

[0012] Furthermore, the continuous-domain transfer function of the PI compensation controller in step S3 is: ; In the formula, This is the proportional compensation coefficient, with a value ranging from 0.01 to 0.5. This is the integral compensation coefficient, with a value ranging from 0.05 to 0.2 times. ; The discrete domain transfer function is obtained through bilinear transformation. Formula for calculating the q-axis compensation voltage of a surface-mounted permanent magnet synchronous motor: ; in, This represents the number of pole pairs of the motor. It is a permanent magnet flux linkage. Let be the q-axis compensation voltage at time k.

[0013] Furthermore, when the controlled motor is a built-in permanent magnet synchronous motor, step S3 synchronously generates the d-axis compensation voltage. It is used to compensate for torque pulsation caused by reluctance torque.

[0014] Furthermore, the final fusion formula for the dq-axis modulation voltage command in step S4 is as follows: ， ; In the formula, The dq-axis fundamental voltage output by traditional magnetic field-oriented control. This is the dq axis compensation voltage; The d-axis compensation voltage of the surface-mounted permanent magnet synchronous motor is set to 0.

[0015] According to a second aspect of the present invention, a control system for suppressing torque ripple in a permanent magnet synchronous motor is provided, comprising: The signal acquisition and coordinate transformation module is used to acquire the three-phase voltage, three-phase current, rotor electrical angle and electrical angular velocity of the permanent magnet synchronous motor, and obtain the voltage and current components under the two-phase rotating dq axis through Clark transformation and Park transformation; the nominal electromagnetic torque is obtained based on the deviation between the given speed and the actual electrical angular velocity. The online estimation module is used for online estimation of instantaneous torque ripple. ; The q-axis compensation voltage acquisition module is used to input instantaneous torque pulsation into the PI compensation controller, and calculate the q-axis compensation voltage by combining the correspondence between the electromagnetic torque of the permanent magnet synchronous motor and the q-axis current. ; The compensation voltage and fundamental voltage fusion module is used to integrate the q-axis compensation voltage. Superimposed on traditional field-oriented control, the output dq-axis fundamental voltage and The final dq-axis modulation voltage command is obtained. and ; The modulation drive output module is used to generate an inverter drive signal by converting the final dq axis modulation voltage command into an inverter drive signal through Park inverse transformation and SVPWM modulation, thereby driving the permanent magnet synchronous motor to run and canceling instantaneous torque pulsation in real time.

[0016] The above solution achieves the following beneficial technical effects: This invention focuses on permanent magnet synchronous motors and addresses the core challenge of torque ripple suppression technology: the inability to directly and in real-time acquire instantaneous torque ripple information and perform targeted compensation. It proposes a strategy based on a discrete sliding mode torque observer and direct ripple compensation. First, a discrete sliding mode torque observer is constructed, which uses only the easily measurable speed and current commands in the system to estimate the instantaneous torque fluctuation component online in real time. Then, the fluctuation component is directly sent to the pulsation compensation controller to generate a compensation voltage and superimpose it on the voltage command of the traditional field-oriented control, forming a closed-loop negative feedback system of "sensing fluctuations → calculating compensation amount → real-time cancellation".

[0017] Its core advantages lie in the following: the discrete sliding mode observer exhibits strong robustness to motor parameter deviations and load disturbances, requiring no precise moment of inertia or damping coefficient, nor offline harmonic calibration, thus significantly reducing the difficulty of engineering implementation; the direct compensation mechanism has a fast response speed and can effectively suppress composite torque ripples from various sources, including cogging torque, back EMF harmonics, and inverter dead-zone effects. This invention effectively solves the contradiction between suppression accuracy and operating condition adaptability in traditional methods, significantly improving the steady-state stability and dynamic response performance of the system.

[0018] This invention also addresses the inherent chattering problem of traditional sliding mode observers by replacing the ideal sign function with a continuous saturation function and setting a reasonable boundary layer thickness. While ensuring robustness, it effectively suppresses high-frequency chattering and avoids the adverse effects of high-frequency noise on the steady-state accuracy of the current loop and torque loop. The algorithm structure is clear and can be implemented with only a few multiply-accumulate operations. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the control method flow provided in a specific embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of an overall control principle block provided in a specific embodiment of the present invention.

[0021] Figure 3 This is a flowchart of a control method for suppressing torque ripple in a permanent magnet synchronous motor, provided by one or more embodiments of the present invention.

[0022] Figure 4 This is a structural diagram of a control system for suppressing torque ripple in a permanent magnet synchronous motor, provided by one or more embodiments of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0024] Figure 4 This is a structural diagram of a control system for suppressing torque ripple in a permanent magnet synchronous motor, provided by one or more embodiments of the present invention.

[0025] Example 1, as Figure 4 The control system for suppressing torque ripple in the permanent magnet synchronous motor shown includes: Speed ​​loop PI controller unit, current loop PI controller unit, SVPWM+ inverter unit, PMSM unit, discrete sliding mode torque observer unit, and PI compensation controller unit; The speed loop PI controller unit is used to generate a nominal electromagnetic torque command based on the deviation between the given speed and the actual speed, so as to realize the speed closed-loop zero steady-state error control. Its input terminal is connected to the external speed command signal and the actual electrical angular velocity signal fed back by the PMSM unit, and its output terminal is connected to the torque command input terminal of the current loop PI controller unit and the reference torque input terminal of the discrete sliding mode torque observer unit, respectively. The current loop PI controller unit is used to output the dq-axis fundamental voltage command for field-oriented control based on the deviation between the nominal electromagnetic torque corresponding to the dq-axis current setpoint and the actual dq-axis current; its input terminal is connected to the nominal electromagnetic torque command output by the speed loop PI controller unit and the actual dq-axis current signal fed back by the PMSM unit, and its output terminal is connected to the first input terminal of the voltage fusion unit. The discrete sliding mode torque observer unit is used to construct a sliding mode observation architecture based on the discrete mechanical motion equations of the motor. It uses a continuous saturation function to replace the sign function to suppress sliding mode chattering, estimates the total electromagnetic torque of the motor in real time, and extracts the instantaneous torque pulsation by subtracting the nominal electromagnetic torque from the estimated total electromagnetic torque. Its input terminal is connected to the nominal electromagnetic torque command output by the speed loop PI controller unit and the actual electric angular velocity signal fed back by the PMSM unit, and its output terminal is connected to the pulsation signal input terminal of the PI compensation controller unit. The PI compensation controller unit is used to generate dq-axis compensation voltage commands based on the amplitude and polarity of instantaneous torque ripple, and to cancel the periodic torque ripple caused by cogging torque, back EMF harmonics, and inverter dead zone in real time; its input terminal is connected to the instantaneous torque ripple signal output by the discrete sliding mode torque observer unit, and its output terminal is connected to the second input terminal of the voltage fusion unit. The voltage fusion unit is used to directly superimpose the dq-axis fundamental voltage command and the dq-axis compensation voltage command to generate the final dq-axis modulation voltage command. Its input terminals are respectively connected to the voltage output terminal of the current loop PI controller unit and the compensation voltage output terminal of the PI compensation controller unit, and its output terminal is connected to the modulation input terminal of the SVPWM+ inverter unit. The SVPWM+ inverter unit integrates a Park inverse converter module, a space vector pulse width modulation module, and a three-phase power inverter circuit. It is used to convert the final modulation voltage in the rotating coordinate system into a three-phase AC drive voltage output. Its input terminal is connected to the final dq axis modulation voltage command output by the voltage fusion unit, and its output terminal is connected to the stator winding input terminal of the PMSM unit. The PMSM unit is the controlled permanent magnet synchronous motor body of the system, used to output electromagnetic torque to drive the load. Its stator winding input terminal is connected to the three-phase drive voltage output by the SVPWM+ inverter unit, and the signal output terminal is respectively fed back to the corresponding feedback input terminals of the speed loop PI controller unit, the current loop PI controller unit, and the discrete sliding mode torque observer unit.

[0026] Figure 3 This is a flowchart of a control method for suppressing torque ripple in a permanent magnet synchronous motor, provided by one or more embodiments of the present invention.

[0027] Example 2, as Figure 3 The control method for suppressing torque ripple in a permanent magnet synchronous motor, as shown, includes: Step S1: Signal acquisition and coordinate transformation; The three-phase voltage, three-phase current, rotor electrical angle and electrical angular velocity of the permanent magnet synchronous motor are collected. The voltage and current components in the two-phase rotating dq coordinate system are obtained through Clark transformation and Park transformation. The deviation between the given speed and the actual electrical angular velocity is input into the speed loop PI controller, which outputs the nominal electromagnetic torque. Step S2: Construct a discrete sliding mode torque observer to estimate instantaneous torque ripples online. ; Step S3: Generate q-axis compensation voltage using a PI compensation controller; The instantaneous torque pulsation is input into the PI compensation controller, and the q-axis compensation voltage is calculated by combining the relationship between the electromagnetic torque of the permanent magnet synchronous motor and the q-axis current. ; Step S4: Fusion of compensation voltage and fundamental voltage: Step q-axis compensation voltage The dq-axis fundamental voltage superimposed on the output of a conventional field-oriented control (FOC) and The final dq-axis modulation voltage command is obtained. and ; Step S5, Modulation drive output: The final dq axis modulation voltage command is converted into an inverter drive signal by Park inverse transformation and SVPWM modulation, which drives the permanent magnet synchronous motor to run and cancels instantaneous torque pulsation in real time.

[0028] In this embodiment, the expressions for the Clark transformation matrix and the Park transformation matrix in step S1 are as follows: ; ; The formulas for voltage and current transformation from a three-phase stationary coordinate system to a two-phase stationary αβ coordinate system, and from a two-phase stationary αβ coordinate system to a two-phase rotating dq coordinate system are as follows: , ; ; In the formula, It is a three-phase voltage. It is a three-phase current. For two-phase stationary coordinate system voltages, For two-phase stationary coordinate system currents, For the dq axis voltage components, For the dq axis current components.

[0029] In this embodiment, a discrete sliding mode torque observer is constructed to estimate instantaneous torque ripples online. include: A discrete sliding mode torque observer is built based on the discrete mechanical motion equations of the motor. A continuous saturated function is used to replace the ideal sign function to suppress sliding mode chattering. The observed speed and load torque values ​​are obtained through sliding mode convergence. The total electromagnetic torque estimate is then calculated by back-calculating the total electromagnetic torque. The difference between the nominal electromagnetic torque and the total electromagnetic torque estimate yields the instantaneous torque pulsation, including cogging torque, back EMF harmonics, and inverter dead-zone effects. .

[0030] In this embodiment, it further includes: the discrete mechanical motion equation of the motor is derived from the continuous dynamic equation of the motor through the forward Euler method and the sampling period. Discretization yields the result; The dynamic continuity equation of the electric motor is: ; In the formula, The moment of inertia of the motor. The mechanical angular velocity of the motor. For electromagnetic torque, The total load torque including friction. The damping coefficient; Discretized equations of motion for the motor: ; In the formula, k is the kth sampling time. Let k be the mechanical angular velocity at time k. Let K be the electromagnetic torque at time k. Let be the load torque at time k.

[0031] In this embodiment, the discrete sliding mode torque observer in step S2 uses... , Let be the state variable, and the observation equation be: Speed ​​observation equation: ; Load torque observation equation: ; in, The rotational speed observation value at time k is... Let k be the load torque observation value at time k. To account for the deviation in rotational speed observation, , For observer gain greater than 0, It is a continuously saturated function.

[0032] In this embodiment, the sliding surface is defined in step S2. Select the observer gain The sliding surface converges to the boundary layer in a finite number of steps. ,in Boundary layer thickness; After sliding mode stabilization, the total electromagnetic torque is estimated by back-calculating based on the discrete mechanical motion of the motor: ; Instantaneous torque ripple calculation formula: ; In the formula, Let K be the nominal electromagnetic torque at time k. This is the estimated value of the total electromagnetic torque at time k. This refers to instantaneous torque pulsation.

[0033] In this embodiment, the continuous domain transfer function of the PI compensation controller in step S3 is: ; In the formula, This is the proportional compensation coefficient, with a value ranging from 0.01 to 0.5. This is the integral compensation coefficient, with a value ranging from 0.05 to 0.2 times. ; The discrete domain transfer function is obtained through bilinear transformation. Formula for calculating the q-axis compensation voltage of a surface-mounted permanent magnet synchronous motor: ; in, This represents the number of pole pairs of the motor. It is a permanent magnet flux linkage. Let be the q-axis compensation voltage at time k.

[0034] In this embodiment, when the controlled motor is a built-in permanent magnet synchronous motor, step S3 synchronously generates the d-axis compensation voltage. It is used to compensate for torque pulsation caused by reluctance torque.

[0035] In this embodiment, the fusion formula for the final dq-axis modulation voltage command in step S4 is: ， ; In the formula, The dq-axis fundamental voltage output by traditional magnetic field-oriented control. This is the dq axis compensation voltage; The d-axis compensation voltage of the surface-mounted permanent magnet synchronous motor is set to 0.

[0036] In this embodiment, the continuous saturation function A piecewise linear function is used; the boundary layer thickness is adaptively adjusted based on the motor sampling frequency and the noise amplitude of the current sensor. In low-noise conditions, δ is reduced to improve the accuracy of pulsation observation; in high-noise conditions, it is increased. To suppress observer noise amplification.

[0037] In this embodiment, in step S2, the acquired electrical angular velocity is pre-processed with a low-pass filter to reduce the high-frequency noise input from the sensor to the discrete sliding mode torque observer, thereby avoiding high-frequency noise generated by instantaneous torque pulsation.

[0038] In this embodiment, the stator temperature and load torque of the motor are collected in real time, and the moment of inertia is corrected according to the temperature change. With damping coefficient The equivalent value, and synchronously adaptively adjust the observer gain. This improves the robustness of observations across the entire temperature range and wide load range.

[0039] In this embodiment, upper and lower threshold values ​​are set for the instantaneous torque ripple obtained in step S2. When the instantaneous torque ripple exceeds the threshold, a limiting process is performed to avoid excessive compensation voltage leading to inverter overcurrent and current loop saturation.

[0040] In this embodiment, the instantaneous torque pulsation is decomposed in real time using Fourier transform to distinguish between the low-frequency harmonic components of the cogging torque and the high-frequency harmonic components of the inverter dead zone. The proportional compensation coefficient and integral compensation coefficient of the PI compensation controller are adjusted in segments for harmonic components of different frequencies to improve the suppression effect of multi-frequency harmonics.

[0041] It is worth noting that although this system / device only discloses the above-mentioned modules / units, it does not mean that this system / device is limited to the above-mentioned basic functional modules. On the contrary, what this invention intends to express is that, based on the above-mentioned basic functional modules, those skilled in the art can add one or more functional modules in combination with the prior art to form an infinite number of embodiments or technical solutions. That is to say, this system is open rather than closed. It cannot be assumed that the scope of protection of the claims of this invention is limited to the above-disclosed basic functional modules just because this embodiment only discloses a few basic functional modules.

[0042] Example 3, in a specific embodiment, discloses a control method for suppressing torque ripple in a permanent magnet synchronous motor, the steps of which are as follows: Figure 1 As shown: Step 1, signal acquisition and coordinate transformation; Step 2, online estimation of discrete sliding mode torque observer; Step 3, generation of compensation voltage based on PI compensation controller; Step 4, fusion of compensation voltage and fundamental voltage; Step 5, output of final voltage command.

[0043] Step 1, Signal Acquisition and Coordinate Transformation, details are as follows: 1. Signal acquisition refers to obtaining the actual values ​​of three-phase voltage, three-phase current, motor rotor electrical angle, and electrical angular velocity under the current state of the PMSM; The actual values ​​of the three-phase voltage and three-phase current are obtained by sampling circuits acquiring the real-time values ​​from the voltage sensor and the current sensor fixed on the motor stator, respectively, and are denoted as . , , as well as , , The electric angle and electric angular velocity of the motor rotor are obtained through a position sensor and are denoted as the electric angle of the motor rotor. Electric angular velocity ; 2. Using coordinate transformation, based on the actual values ​​of the three-phase voltage and three-phase current, obtain... , Off-axis voltage and current components , , , ; The calculation process is as follows: , ; ; In the formula, , The coordinate transformation matrix has the following specific form: ; ; 3. Speed ​​loop calculation: Compare the given speed with the actual feedback speed. The deviation value is input to the speed loop PI controller, and the output reference torque (nominal electromagnetic torque) is given to the torque control loop; 4. Output: The output of this step is the voltage and current signals in a two-phase rotating coordinate system. , , , Motor rotor electrical angle As input to subsequent observers; II. The specific details of the online estimation of the discrete sliding mode torque observer in step 2 are as follows: like Figure 2 As shown, a discrete sliding mode observer is constructed to estimate the instantaneous torque ripple of a motor. Specifically, based on the motor's mechanical motion equations, the observable speed signal and the unknown load torque are used as state variables to construct a discrete sliding mode observer. By using sliding mode, the observed speed converges to the actual speed, thereby estimating the total electromagnetic torque acting on the motor shaft and separating the instantaneous torque ripple component.

[0044] Because traditional sliding mode observers use ideal sign functions as switching functions, they inevitably lead to high-frequency chattering, introduce noise and affect the steady-state accuracy of the current loop and torque loop, while also weakening the system's lifespan and operational stability. Furthermore, traditional methods treat various disturbances in a general way, making it difficult to accurately separate torque ripple components at specific frequencies.

[0045] Therefore, this embodiment uses a continuous saturation function instead of a sign function and sets a reasonable boundary layer thickness ( While ensuring the robustness of sliding mode, it effectively suppresses chattering. At the same time, by constructing the difference between the nominal electromagnetic torque and the observed total torque, the instantaneous pulsation is directly extracted, realizing accurate and real-time estimation of multi-source composite torque pulsation such as cogging torque, back EMF harmonics, and inverter dead zone effect, providing accurate torque pulsation information for subsequent compensation.

[0046] 1. Establish the mechanical motion equations of the electric motor Neglecting elastic deformation, the mechanical dynamic equation of the motor is: ; Where J is the moment of inertia. This is the actual instantaneous electromagnetic torque. This refers to the load torque (including friction, etc.). The damping coefficient is used. The equation is discretized (sampling period). (using the forward Euler method) ; 2. Define state variables and construct the observer structure. Define measurable state quantities as well as The discrete sliding mode observer is constructed as follows: Speed ​​observation equation: ; Load torque observation equation: ; The parameters are defined as follows: ), These are the observed values ​​for rotational speed and load torque, respectively. This is due to the deviation in the observed rotational speed. , For observer gain, the sliding mode condition must be guaranteed; It is a continuous saturation function used to replace the ideal sign function to suppress chattering.

[0047] 3. Sliding mode convergence analysis Define the sliding surface Subtracting the speed dynamics from the observer's speed dynamics from the actual speed dynamics yields the deviation dynamic equation.

[0048] When observer gain When large enough, the sliding surface is attracted to the boundary layer within a finite number of steps. .

[0049] The equivalent value of the saturation function under sliding mode motion. This precisely compensates for the unknowns in the model, thus forcing the observations to... , Approaching the true value.

[0050] 4. Estimate the total electromagnetic torque and extract instantaneous torque pulsations. After the sliding mode is established, the observer provides the following: and The total electromagnetic torque estimate is derived by inversely using the discrete mechanical equations: ; because To ensure smooth observations, this calculation does not amplify noise. Instantaneous torque fluctuation is defined as the difference between the nominal electromagnetic torque and the actual electromagnetic torque; therefore: ; The Tr value reflects the periodic pulsations caused by factors such as cogging torque, back EMF harmonics, and inverter dead zone. It is a key output value of this step and is sent to step 3 for compensation.

[0051] Third, steps 3 to 5 constitute a closed loop for the direct suppression and execution of torque fluctuations.

[0052] Step 3, the generation of compensation voltage based on the PI compensation controller, is detailed below: By constructing a PI-type compensation controller, the instantaneous torque ripple Tr estimated by the discrete sliding mode observer is used as the input to generate the q-axis compensation voltage. Simultaneously, a direct fusion method of "measure and compensate" is adopted to integrate the compensation voltage with the fundamental voltage output by the traditional FOC. The final voltage command is obtained by direct superposition. Finally, through Park inverse transformation and SVPWM modulation, an inverter drive signal is generated to cancel out instantaneous torque pulsation in real time.

[0053] This direct compensation mechanism responds quickly, effectively suppresses multi-source pulsations such as cogging torque and back EMF harmonics, and does not change the original control framework, making it easy to implement in engineering.

[0054] Step 3 converts the instantaneous torque ripple Tr estimated in Step 2 into a compensation voltage that can be used to offset it. This is achieved by designing a compensation controller, specifically a proportional-integral (PI) controller, with Tr as the input and the q-axis compensation current command as the output. The transfer function is as follows: ; in This is the proportional compensation coefficient. This is the integral compensation coefficient. The function of this controller is: when Tr>0 (i.e., the actual torque is greater than the commanded value), it outputs a negative compensation current; otherwise, it outputs a positive compensation current.

[0055] Step 4, the fusion of compensation voltage and fundamental voltage, is detailed below: Converting the compensation current to the compensation voltage, and considering that the electromagnetic torque is approximately proportional to the q-axis current (ignoring the reluctance torque for surface-mounted motors), the final compensation voltage can be obtained as follows: ; in for The discretization result of the bilinear transform. For an embedded permanent magnet synchronous motor, if reluctance torque is taken into account, it can be discretized along the d-axis. Compensation items are also introduced. However, in engineering practice, significant results can usually be obtained by compensating only the q-axis. The compensation strength is determined, and the value is usually taken in the range of 0.01 to 0.5. To eliminate steady-state pulsating residuals, one can take In practical applications, fine-tuning is performed based on motor inertia and sampling frequency.

[0056] Based on the aforementioned compensation voltage, it is further integrated with the fundamental voltage obtained from vector control using a direct fusion method of "measuring and compensating on the spot". , Direct fusion is performed to obtain the final modulation voltage command: ; ; in It is usually set to zero (for surface-mounted motors) or slightly compensated based on the reluctance torque effect.

[0057] Step 5, output the final voltage command as follows: Combine the above-mentioned fused d-axis and q-axis voltage commands. , The signal is converted into an inverter drive signal through Park inverse transformation and SVPWM modulation, and then output to the motor control system to form a closed-loop control.

[0058] Example 4: The strategy in the above examples is generated based on the following more specific research: Currently, permanent magnet synchronous motors (PMSMs) are widely used due to their high power density and high efficiency. However, factors such as cogging torque, back EMF harmonics, and inverter nonlinearity cause periodic torque fluctuations in the motor, severely affecting system stability and lifespan. Existing torque fluctuation suppression technologies are mainly divided into two categories: one is motor body optimization methods (such as magnetic pole shape optimization, skewed slots, etc.), which can reduce harmonics at the source, but increase manufacturing costs and process difficulty, and the suppression capability is fixed and cannot adapt to changes in operating conditions. The other is control strategy optimization methods, which mainly include harmonic injection, sliding mode observer, model predictive control (MPC), active disturbance rejection control (ADRC), and extended Kalman filtering. Harmonic injection methods (such as CN112910336B) require extensive offline experiments to calibrate injection parameters, are extremely sensitive to changes in motor parameters, and have poor adaptability to operating conditions. Sliding mode observer methods (such as CN107359837B) are robust, but struggle to accurately separate instantaneous torque fluctuations and suffer from inherent chattering issues. MPC methods (such as CN202410771521) have fast dynamic responses, but suffer from large steady-state torque ripples and are highly dependent on accurate motor models, with their suppression effect deteriorating when parameters drift. ADRC methods (such as CN116317720A) treat all disturbances as "concentrated disturbances," resulting in a significant drop in estimation accuracy when harmonic frequencies exceed the observer bandwidth, and complex parameter tuning. Extended Kalman filtering methods (such as CN202210742080) involve matrix operations, have a large computational load, demand high-performance processors, and experience reduced accuracy when the model is mismatched. Therefore, it is necessary to develop a relatively simple, robust control method for all operating conditions that does not require precise models or offline calibration, has low computational requirements, and can sense and directly compensate for instantaneous torque fluctuations in real time, in order to resolve the contradiction between stability and suppression accuracy.

[0059] Patent document 1 (CN119315887A) discloses a method for suppressing torque ripple based on harmonic injection. This method first constructs a joint simulation model of the motor control system and the electromagnetic model to obtain the electromagnetic torque output by the motor. Then, it performs a fast Fourier transform on the output torque to obtain the harmonic torque components with higher content from the spectrum as the target object, and determines the order of the injected harmonic current. Next, it performs parameter scanning analysis with the amplitude and phase of the harmonic current as variables to find the optimal solution of the variables when the amplitude of the target harmonic torque component is minimized, serving as the reference value for harmonic current injection. Finally, it converts the injected current signal in the three-phase stationary coordinate system into a current signal in the two-phase rotating coordinate system through Clark and Park transforms, and then superimposes the output of the current controller onto a given voltage signal, thereby suppressing motor torque ripple. This scheme essentially adopts an open-loop compensation approach, relying heavily on offline simulations and parameter scanning to calibrate the amplitude and phase of the injected harmonics, lacking online adaptive adjustment capability for changes in motor parameters and operating conditions. Patent document 2 (CN107359837B) proposes a torque control system and method for a permanent magnet synchronous motor based on a sliding mode observer and active disturbance rejection control (ADRC). First, a minimum-order extended flux sliding mode observer is designed to observe the motor's output torque in real time. Then, an ADRC current controller is used to suppress overshoot and enhance robustness. Finally, the difference between the electromagnetic torque output by the observer and the given torque signal is input to a PI controller. The q-axis current signal output by the PI controller and the dq-axis flux component output by the observer are also input to the ADRC, thus forming a closed-loop torque control and improving the reliability of torque control. While this scheme can solve the problems of integral saturation and insufficient dynamic performance of traditional PI controllers, the ADRC requires tuning multiple high-order parameters (extended state observer gain, feedback control law gain, etc.), making parameter adjustment complex and reliant on engineering experience. Even with improved genetic algorithm optimization, the computational load is large and the convergence speed is slow, resulting in a high barrier to engineering implementation. Patent document 3 (CN118677308A) discloses a model predictive control method for permanent magnet synchronous motors (PMSMs) that applies anti-phase torque harmonics. Starting with torque harmonics, it adds a compensation amount with equal amplitude but opposite phase to the torque harmonics. This reduces electromagnetic torque, cogging torque, and friction torque harmonics without extracting torque harmonics. A load torque observer is used to observe the actual output torque of the motor, and the difference between the observed value and the reference torque value output by the speed loop is used to obtain the torque deviation for compensation. While this method eliminates the need to extract torque harmonics and can reduce torque ripple to some extent, model predictive control itself is highly dependent on an accurate motor model. The control effect deteriorates as motor parameters drift with temperature and frequency. Furthermore, traditional MPC selects only one optimal vector from seven voltage vectors per cycle, which is one of the fundamental reasons for its large steady-state torque ripple.Patent document 4 (CN116317720A) discloses an improved method for suppressing torque ripple in permanent magnet synchronous motors using active disturbance rejection control (ADRC) to address the problems of numerous ADRC parameters and difficult adjustment. This method employs an improved genetic algorithm to optimize the parameters of the extended state observer of the ADRC controller. Improved crossover and mutation probability formulas enhance the reliability and applicability of crossover and mutation. The improved extended state observer estimates the system state and total disturbance, feeding the estimated values ​​back to the controller to suppress torque ripple during disturbances such as load abrupt changes. However, this method has the following drawbacks: the improved genetic algorithm has a large computational load and slow convergence speed; the iterative calculations in the global optimization process rely heavily on extensive offline optimization and calibration in the early stages, resulting in a high engineering implementation threshold; simultaneously, ADRC treats all unmodeled dynamics as "concentrated disturbances" for observation and compensation, making it difficult to accurately separate and suppress torque ripple components of specific frequencies. Furthermore, the estimation accuracy of harmonic disturbances by the traditional extended state observer is limited by the observer bandwidth, and noise amplification becomes prominent as the bandwidth increases.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control method for suppressing torque ripple in a permanent magnet synchronous motor, characterized in that, include: Step S1: Signal acquisition and coordinate transformation; The three-phase voltage, three-phase current, rotor electrical angle and electrical angular velocity of the permanent magnet synchronous motor are collected, and the voltage and current components under the two-phase rotating dq axis are obtained by Clark transformation and Park transformation. The nominal electromagnetic torque is obtained based on the deviation between the given rotational speed and the actual electrical angular velocity; Step S2: Online estimation of instantaneous torque ripple ; Step S3: Obtain the q-axis compensation voltage; The instantaneous torque pulsation is input into the PI compensation controller, and the q-axis compensation voltage is calculated by combining the correspondence between the electromagnetic torque of the permanent magnet synchronous motor and the q-axis current. ; Step S4: The compensation voltage is fused with the fundamental voltage; q-axis compensation voltage Superimposed on traditional field-oriented control, the output dq-axis fundamental voltage and The final dq-axis modulation voltage command is obtained. and ; Step S5: Modulate the drive output; The final dq axis modulation voltage command is converted into an inverter drive signal through Park inverse transformation and SVPWM modulation, which drives the permanent magnet synchronous motor to run and cancels instantaneous torque pulsation in real time.

2. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 1, characterized in that, The expressions for the Clark transformation matrix and the Park transformation matrix in step S1 are as follows: ； ； This is the Clark transformation matrix from the three-phase stationary coordinate system to the two-phase stationary αβ coordinate system. The Park transformation matrix is ​​given by the formula for the two-phase stationary αβ coordinate system to the two-phase rotating dq coordinate system. The rotor electrical angle includes the electrical angle between the rotor d-axis axis and the α-axis axis of the two-phase stationary coordinate system, which changes in real time as the rotor rotates. The cosine value of the rotor electrical angle is used as the in-phase projection coefficient between the two coordinate systems, corresponding to the amplitude coupling calculation of the d-axis and α-axis components and the q-axis and β-axis components, respectively. The sine value of the rotor electrical angle is used as the orthogonal projection coefficient between the two coordinate systems, corresponding to the amplitude coupling calculation of the d-axis and β-axis components; The negative value of the sine of the rotor electrical angle is used as the orthogonal projection coefficient between the q-axis and α-axis components. Since the electrical angle of the q-axis leads the d-axis by 90°, its α-axis projection corresponds to the negative value of the sine, ensuring that the d and q axes always maintain an orthogonal relationship. The formulas for converting three-phase voltage and three-phase current are: ， ； ； In the formula, It is a three-phase voltage. For three-phase current, For two-phase stationary coordinate system voltages, For two-phase stationary coordinate system currents, For the dq axis voltage components, For the dq axis current components.

3. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 1, characterized in that, The online estimation of instantaneous torque pulsation include: Based on the discrete mechanical motion equations of the motor, a continuous saturated function is used to replace the ideal sign function. Speed ​​and load torque observations are obtained through sliding mode convergence. The total electromagnetic torque estimate is then calculated by back-calculating the nominal electromagnetic torque. The difference between the nominal electromagnetic torque and the total electromagnetic torque estimate yields the instantaneous torque pulsation, including cogging torque, back EMF harmonics, and inverter dead-zone effects. .

4. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 3, characterized in that, Also includes: The discrete mechanical motion equations of the motor are derived from the continuous dynamic equations of the motor using the forward Euler method and a sampling period. Discretization yields the result; The motor's mechanical dynamic continuity equation is: ; In the formula, For the moment of inertia of the motor, The mechanical angular velocity of the motor. For electromagnetic torque, The total load torque including friction. The damping coefficient; Discretized equations of motion for the motor: ； In the formula, The sampling period of the control system; k is the kth sampling time. Let k be the mechanical angular velocity at time k. Let K be the electromagnetic torque at time k. Let be the load torque at time k.

5. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 4, characterized in that, by , Let be the state variable, and the observation equation be: Speed ​​observation equation: ; Load torque observation equation: ; in, Let k be the observed rotational speed. Let k be the load torque observation value at time k. To account for the deviation in rotational speed observation, , For observer gain greater than 0, It is a continuously saturated function.

6. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 5, characterized in that, In step S2, the sliding surface is defined. Select gain The sliding surface converges to the boundary layer in a finite number of steps. ,in Boundary layer thickness; After sliding mode stabilization, the total electromagnetic torque is estimated by back-calculating based on the discrete mechanical motion of the motor: ； Instantaneous torque ripple calculation formula: ; In the formula, Let K be the nominal electromagnetic torque at time k. This is the estimated value of the total electromagnetic torque at time k. This refers to instantaneous torque pulsation.

7. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 1, characterized in that, The continuous-domain transfer function of the PI compensation controller in step S3 is: ； In the formula, This is the proportional compensation coefficient, with a value ranging from 0.01 to 0.

5. This is the integral compensation coefficient, with a value ranging from 0.05 to 0.2 times. ; The discrete domain transfer function is obtained through bilinear transformation. Formula for calculating the q-axis compensation voltage of a surface-mounted permanent magnet synchronous motor: ； in, This represents the number of pole pairs of the motor. It is a permanent magnet flux linkage. Let be the q-axis compensation voltage at time k.

8. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 7, characterized in that, When the controlled motor is a built-in permanent magnet synchronous motor, step S3 synchronously generates the d-axis compensation voltage. It is used to compensate for torque pulsation caused by reluctance torque.

9. The control method for suppressing torque ripple in a permanent magnet synchronous motor according to claim 1, characterized in that, The final fusion formula for the dq-axis modulation voltage command in step S4 is: ， ； In the formula, The dq-axis fundamental voltage output by traditional magnetic field-oriented control. This is the dq axis compensation voltage; The d-axis compensation voltage of the surface-mounted permanent magnet synchronous motor is set to 0.

10. A control system for suppressing torque ripple in a permanent magnet synchronous motor, characterized in that, include: The signal acquisition and coordinate transformation module is used to acquire the three-phase voltage, three-phase current, rotor electrical angle and electrical angular velocity of the permanent magnet synchronous motor, and obtain the voltage and current components under the two-phase rotating dq axis through Clark transformation and Park transformation. The nominal electromagnetic torque is obtained based on the deviation between the given rotational speed and the actual electrical angular velocity; The online estimation module is used for online estimation of instantaneous torque ripple. ; The q-axis compensation voltage acquisition module is used to input the instantaneous torque pulsation into the PI compensation controller, and calculate the q-axis compensation voltage by combining the correspondence between the electromagnetic torque of the permanent magnet synchronous motor and the q-axis current. ; The compensation voltage and fundamental voltage fusion module is used to integrate the q-axis compensation voltage. Superimposed on traditional field-oriented control, the output dq-axis fundamental voltage and The final dq-axis modulation voltage command is obtained. and ; The modulation drive output module is used to generate an inverter drive signal by converting the final dq axis modulation voltage command into an inverter drive signal through Park inverse transformation and SVPWM modulation, thereby driving the permanent magnet synchronous motor to run and canceling instantaneous torque pulsation in real time.

Images