Robust-enhanced sensorless permanent magnet synchronous motor predictive current control method

Through the adaptive superspiral sliding mode perturbation observer and recursive least squares online parameter identification technology, the oscillation and vibration problems of sensorless permanent magnet synchronous motor during parameter mismatch are solved, and high dynamic response and robustness are achieved Control enhances the stability and parameter identification accuracy of the FOC algorithm.

CN118984098AActive Publication Date: 2024-11-19TAIZHOU RES INST ZHEJIANG UNIV OF TECH

Patent Information

Application Number
CN202411136515.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-19
Publication Date
2024-11-19
Estimated Expiration
2044-08-19

AI Technical Summary

Technical Problem

When the parameters of sensorless permanent magnet synchronous motors are mismatched, the FOC algorithm easily leads to oscillation of electrical angle and electrical angular velocity and even algorithm divergence. Traditional sliding mode disturbance observers have jitter problems, making it difficult to achieve high dynamic response and robust control.

Method used

An adaptive superspiral sliding mode perturbation observer is designed, combined with a recursive least squares online parameter identifier with forgetting factor to achieve multi-parameter online identification and robustness enhancement of motor parameters, through the adaptive superspiral algorithm Sliding mode vibration is suppressed and identification parameter errors are fed back to the sensorless observer to enhance its robustness.

Benefits of technology

The high dynamic response and strong robustness control of sensorless permanent magnet synchronous motor are realized, which effectively suppresses the vibration of the sliding mode disturbance observer, improves the robustness of the current loop and the observer, and enhances the overall stability of the FOC algorithm. and parameter identification accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118984098A_ABST
    Figure CN118984098A_ABST
Patent Text Reader

Abstract

The invention discloses a robust enhanced sensorless permanent magnet synchronous motor predictive current control method. The method comprises the following steps: designing a self-adaptive super-spiral sliding mode disturbance observer based on a prediction current equation of the permanent magnet synchronous motor; establishing a predicted voltage equation and an online parameter identifier; and according to the motor operation parameters, the sliding mode disturbance observer and the prediction voltage equation, obtaining a prediction voltage, according to the online parameter identifier, obtaining an identification parameter error, and jointly participating in vector closed-loop control to realize prediction current control. According to the method, the limitation that predictive current control only realizes current loop robustness enhancement is solved, multi-parameter online identification of the motor is realized, the robustness of the non-inductive observer is further enhanced through feedback of identification parameters, the method has a better buffeting suppression effect, lumped disturbance accurate observation under large-range parameter mismatch can be realized, and the method is suitable for popularization and application. And current harmonics are further weakened, the parameter identification precision is improved, and the stability and robustness of motor vector closed-loop control are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a motor predictive current control method, and relates to the technical field of sensorless permanent magnet synchronous motor predictive control and parameter identification, and in particular to a robust enhanced sensorless permanent magnet synchronous motor predictive current control method. Background Art

[0002] Permanent Magnet Synchronous Motor (PMSM) is widely used in electric vehicles, wind turbines, industrial drive systems and other fields. Its high efficiency and stability make it an ideal choice to replace traditional motors. High dynamic response performance and robustness of control algorithms are important indicators of sensorless control of permanent magnet synchronous motors. High dynamic response performance determines the motor's ability to quickly adjust speed and torque, enabling it to adapt to rapidly changing working conditions and load requirements. The robustness of the control algorithm ensures control stability under various disturbances and uncertainties, ensuring that the motor can still maintain efficient and stable operation in different environments and working conditions.

[0003] As the innermost loop of the permanent magnet synchronous motor control system, the current loop largely determines the dynamic response capability and steady-state error level of the output electromagnetic torque of the permanent magnet synchronous motor. Model Predictive Control (MPC) has gradually replaced the traditional PI (proportional integral controller) controller as an advanced control strategy for the current loop due to its simple design process and fast dynamic response in dealing with multiple control variables and multiple system constraints. Factors such as temperature rise, demagnetization and magnetic saturation effects during the operation of the permanent magnet synchronous motor cause a certain degree of drift in the resistance, inductance and flux linkage parameters. In order to eliminate the parameter dependence of predictive control and improve the robustness of the current loop, extended state disturbance observers and sliding mode disturbance observers have been proposed and applied. However, the traditional extended state disturbance observer is limited by bandwidth, and the low-order sliding mode disturbance observer inevitably has the problem of jitter caused by fixed gain. In particular, for the sliding mode disturbance observer, the design of a complex reaching law may lead to difficulties in parameter tuning and worsening of the jitter problem. It is worth noting that traditional predictive control strategies are mostly used in sensor control systems. For sensorless permanent magnet synchronous motor FOC (Field-Oriented Control) control, parameter mismatch of the sensorless observer can lead to oscillation of the observer's electrical angle and electrical angular velocity and even divergence of the algorithm. Summary of the invention

[0004] In order to solve the problems existing in the background technology, the present invention provides a robust enhanced sensorless permanent magnet synchronous motor predictive current control method. When the motor parameters are mismatched, the sensorless observer will have estimated electrical angle and electrical angular velocity errors and cause the divergence of the FOC algorithm. The present invention designs a sliding mode disturbance observer combined with a jitter suppression algorithm, which cooperates with current loop predictive control and parameter identification. While ensuring the high dynamic response capability of the current loop, the robustness of the current loop and the sensorless observer can be improved, so that the sensorless permanent magnet synchronous motor FOC control has the advantages of high dynamic response and strong robustness, which has important research value and application potential.

[0005] The technical solution adopted by the present invention is:

[0006] The robust enhanced sensorless permanent magnet synchronous motor predictive current control method of the present invention comprises:

[0007] S1: Establish the predictive current equation of sensorless permanent magnet synchronous motor, and design an adaptive super-helical sliding mode disturbance observer based on the predictive current equation; design an adaptive super-helical sliding mode disturbance observer to realize online observation of lumped disturbances.

[0008] S2: Establish a predictive voltage equation based on zero-beat current control and an online parameter identifier using the recursive least squares method with a forgetting factor. The predictive voltage equation realizes online compensation of disturbances while considering the one-beat delay problem of the actual digital system. The online parameter identifier realizes online identification of multiple motor parameters such as resistance, inductance and flux linkage errors.

[0009] S3: The d-axis and q-axis stator currents and stator voltages and the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the current moment are input into the adaptive super-helical sliding mode disturbance observer for processing. After processing, the adaptive super-helical sliding mode disturbance observer outputs the d-axis and q-axis stator currents and the estimated values ​​of the lumped disturbances at the next moment. The d-axis and q-axis stator currents and the estimated values ​​of the lumped disturbances at the next moment, the d-axis and q-axis reference currents and the rotor electrical angular velocity at the current moment are input into the predicted voltage equation based on the zero-beat current control for processing. After processing, the predicted voltage equation of the zero-beat current control outputs the d-axis and q-axis stator voltages at the next moment, thereby performing vector control FOC closed-loop control of the sensorless permanent magnet synchronous motor.

[0010] S4: The d-axis and q-axis stator currents and stator voltages, rotor electrical angular velocity, and estimated values ​​of the d-axis and q-axis lumped disturbances output by the adaptive super-helical sliding mode disturbance observer of the sensorless permanent magnet synchronous motor at the current moment are input into the recursive least squares online parameter identifier with forgetting factor for processing. After processing, the recursive least squares online parameter identifier with forgetting factor outputs the parameter error vector to the sensorless observer in the vector control FOC closed-loop control, and finally realizes the predictive current control of the sensorless permanent magnet synchronous motor. The identified parameter error is fed back to the sensorless observer online to realize the robust enhancement of the observer.

[0011] In step S1, the prediction current equation of the permanent magnet synchronous motor is as follows:

[0012]

[0013] in, and are the predicted stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment; T s is the sampling interval, i.e. the execution cycle of the vector control FOC closed-loop control; R s and L s are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; u d (k) and u q (k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment; f d (k) and f q (k) are the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

[0014] The premise for establishing the ideal voltage equation of the sensorless permanent magnet synchronous motor in the rotating coordinate system is to ignore the cross-saturation effect of the motor inductance, the deviation of the motor rated parameters, and other factors. In order to characterize the voltage deviation caused by the motor parameter mismatch, the ideal voltage equation is reconstructed into the actual voltage equation, as follows:

[0015]

[0016] Among them, u d 、u q 、i d and iq are the stator voltage and stator current of d-axis and q-axis respectively; f d and f q are the lumped disturbances of the d-axis and q-axis caused by the deviation of the actual parameters of the motor from the rated parameters; L d and L q They are the stator inductances of the d-axis and q-axis respectively. Taking the surface-mounted permanent magnet synchronous motor as an example, the inductances of the d-axis and q-axis are equal and unified as the stator rated phase inductance L s ; After further discretization through the forward Euler method, the actual voltage equation under discrete conditions can be obtained, that is, the predicted current equation.

[0017] In step S1, the adaptive super-helical sliding mode disturbance observer is specifically as follows:

[0018]

[0019]

[0020]

[0021]

[0022] in, and are the estimated values ​​of the stator current of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling time. and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling time; T s is the sampling interval; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; and are the estimated values ​​of the stator current of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; u d (k) and u q (k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; U dSTA and U qSTAare the sliding mode functions based on the adaptive superhelical algorithm for the d-axis and q-axis of the sensorless permanent magnet synchronous motor respectively; δ is the disturbance term gain, δ>0; λ1 and λ2 are the first and second positive gains; s d and q are the sliding surfaces formed by the d-axis and q-axis current errors, respectively, and the input error s d 、s q Adaptation is achieved through the hyperbolic tangent function inside the sliding mode function to suppress the sliding mode chattering problem; tanh() is a hyperbolic tangent function with adjustable boundary layer. The traditional superhelical algorithm itself is a second-order sliding mode algorithm and has a certain chattering suppression effect, but its built-in sign function is a step switching function, which cannot achieve adaptation to the input error under fixed gain. When the error decreases, it will be affected by the fixed gain and cause sliding mode chattering; τ is the time variable; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment respectively; x is the input of the hyperbolic tangent function tanh(), and m is the positive gain of the hyperbolic tangent function tanh(). When the input error is within the gain constraint range, the hyperbolic tangent function can realize error adaptation and further improve the jitter problem caused by fixed gain.

[0023] In step S2, the predicted voltage equation based on deadbeat current control is as follows:

[0024]

[0025] Among them, u d (k+1) and u q (k+1) are the predicted stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment predicted in advance, so as to compensate for the one-beat delay of the actual digital system; R s and L s are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; T s is the sampling interval; and are the reference currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment, based on i d =0 The d-axis reference current controlled by is always zero, that is, The q-axis reference current is provided by the speed loop output and the speed loop execution frequency is approximately 1 / 20 to 1 / 10 of the current loop execution frequency. It is worth noting that the mechanical frequency of the permanent magnet synchronous motor is much smaller than the electrical frequency, so the electrical angular velocity in adjacent cycles can be considered equal; and are the estimated values ​​of the stator current of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment; ωe (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling time, The disturbance observer is used to observe and compensate online; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

[0026] The command voltage u calculated at the kth sampling time of the actual digital control system d (k) and u q (k) is actually applied to the inverter module only at the k+1th moment, and the inverter module takes about half of the sampling period to apply the voltage to the permanent magnet synchronous motor winding, and a one-step advance prediction voltage equation is further established based on the prediction current equation.

[0027] In step S2, the recursive least squares online parameter identifier with forgetting factor is specifically as follows:

[0028]

[0029] Φ(k)=[ΔR s ,ΔL s ,Δψ f ] T

[0030]

[0031] Among them, Φ(k) and Φ(k-1) are the parameter error vectors to be identified of the sensorless permanent magnet synchronous motor at the kth and k-1th sampling moments, respectively, ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux error of the sensorless permanent magnet synchronous motor respectively; P(k) and P(k-1) are the covariance matrices at the kth and k-1th sampling moments respectively, and the order is the number of objects to be identified; y(k) is the disturbance input of the online parameter identifier at the kth sampling moment, including the estimated value of the lumped disturbance of the d-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment and the estimated value of the lumped disturbance on the q axis is the measurement input of the online parameter identifier at the kth sampling moment; κ is the forgetting factor of the least squares parameter identification method.

[0032] Measurement input The details are as follows:

[0033]

[0034] The recursive least squares parameter identification equation with forgetting factor is designed based on the lumped disturbance formula to realize the online identification of motor multi-parameters including resistance, inductance and flux linkage errors.

[0035] In the step S4, the recursive least squares method with a forgetting factor outputs a parameter error vector to the sensorless observer in the vector control FOC closed-loop control after processing, and the motor parameter error correction is performed on the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor in the synchronous rotating coordinate system, or the motor parameter error correction is performed on the stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor in the stationary coordinate system. After the motor parameter error correction, the sensorless observer outputs the electrical angular velocity and electrical angle to realize the FOC control of the sensorless permanent magnet synchronous motor.

[0036] The identification parameter error is fed back online to the sensorless observer to achieve robust enhancement of the observer. Generally, the sensorless observer based on back electromotive force or permanent magnet flux signal is established under the voltage equation of the stationary coordinate system of the permanent magnet synchronous motor or the voltage equation of the synchronous rotating coordinate system, that is, the identification parameter error is fed back and updated to the voltage equation of the sensorless observer to achieve robust enhancement.

[0037] The stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor in the synchronous rotating coordinate system after the motor parameter error correction are specifically as follows:

[0038]

[0039] Among them, u d ′(k) and u q ′(k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment after the motor parameter error correction; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux linkage error of the sensorless permanent magnet synchronous motor respectively; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

[0040] The stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor in the stationary coordinate system after the motor parameter error correction are specifically as follows:

[0041]

[0042] Among them, u α ′(k) and u′ β (k) are the stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment after the motor parameter error correction; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux linkage error of the sensorless permanent magnet synchronous motor respectively; i α (k) and i β (k) are the stator currents of the α-axis and β-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; ψ f is the permanent magnet flux of the sensorless permanent magnet synchronous motor; θ e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment.

[0043] The electronic device of the present invention comprises: a memory and a processor coupled to each other, wherein the memory stores program data, and the processor calls the program data to execute the method as described above.

[0044] The computer-readable storage medium of the present invention stores program data thereon, and when the program data is executed by a processor, the method described above is implemented.

[0045] The method of the present invention can realize the input error adaptation of the sliding mode disturbance observer and effectively suppress the inherent chattering phenomenon of the traditional sliding mode, realize the accurate observation of the lumped disturbance and the suppression of the total harmonic distortion of the phase current. At the same time, the recursive least squares parameter identification algorithm with forgetting factor based on the lumped disturbance equation can realize the multi-parameter online identification of the motor resistance, inductance and flux error parameters, and the identification parameter error feedback is updated to the sensorless observer equation to realize the robust enhancement of the sensorless observer.

[0046] The beneficial effects of the present invention are:

[0047] The present invention solves the limitation that most current predictive current control strategies can only realize the robustness enhancement of the current loop and realizes the online identification of multiple parameters of the motor, and further enhances the robustness of the sensorless observer through the feedback of the identification parameters. The present invention expands the application of the traditional predictive current control algorithm in the FOC control of sensorless permanent magnet synchronous motors, realizes the adaptation of the input error and effectively suppresses the inherent jitter of the sliding mode through the super spiral algorithm based on the hyperbolic tangent function, realizes the accurate online observation and compensation of the disturbance, and suppresses the harmonic distortion of the phase current at the same time. The adaptive super spiral sliding mode disturbance observer has the advantages of strong robustness of the sliding mode algorithm, and has better jitter suppression effect than the traditional low-order sliding mode observer. It can realize the accurate observation of the lumped disturbance under a large range of parameter mismatch, and further weaken the current harmonics and improve the accuracy of parameter identification, thereby improving the overall stability and robustness of the sensorless permanent magnet synchronous motor FOC algorithm. The recursive least squares method with forgetting factor online identifies the motor parameter error and feeds back to the sensorless observer to realize the robust enhancement of the sensorless permanent magnet synchronous motor control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of applying the robust enhanced sensorless permanent magnet synchronous motor predictive current control method of the present invention to participate in vector FOC control;

[0049] Figure 2 : is a graph of mechanical speed, q-axis current and phase current of the conventional super spiral sliding mode and the adaptive super spiral sliding mode disturbance observer of the present invention before and after the motor parameter mismatch in the embodiment of the present invention, wherein: Figure 2 (a) is the mechanical speed, q-axis current and phase current curves of the traditional super-helical sliding mode disturbance observation. Figure 2 (b) is a graph showing the mechanical speed, q-axis current and phase current of the adaptive super-helical sliding mode disturbance observer of the present invention;

[0050] Figure 3 It is the analysis statistics of the phase current total harmonic distortion THD (Total Harmonic Distortion) corresponding to the non-sliding mode disturbance observer, the traditional super-spiral sliding mode disturbance observer and the adaptive super-spiral sliding mode disturbance observer after the motor parameter mismatch at different speeds in the embodiment of the present invention;

[0051] Figure 4 is a curve diagram of resistance identification results of least square method parameter identification using lumped perturbation equation in an embodiment of the present invention compared with traditional voltage equation;

[0052] Figure 5 is a curve diagram of the inductance identification result of least square method parameter identification using lumped disturbance equation in the embodiment of the present invention compared with the traditional voltage equation;

[0053] Figure 6 It is a curve diagram of flux linkage identification results of least squares parameter identification using the lumped disturbance equation in the embodiment of the present invention compared with the traditional voltage equation. DETAILED DESCRIPTION

[0054] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] like Figure 1 As shown, the hardware sampling circuit of the permanent magnet synchronous motor controller performs current sampling according to the FOC algorithm execution frequency and converts the sampled value into the two-phase current i at the current kth moment of the controller algorithm. a (k), i b (k). The phase current is converted into the stationary coordinate system current i via Clark transformation and Park transformation. α (k), i β (k) and the synchronous rotating coordinate system current i d (k), i q (k). The synchronous rotating coordinate system current i at the current kth moment d (k), i q (k), command voltage u d (k),u q (k) and the sensorless observer (taking the saturation function sliding mode observer SMO (Sliding Mode Observer) as an example) to estimate the electrical angular velocity ω e (k) As the input of the adaptive super-helical sliding mode disturbance observer, it performs one-step-ahead current prediction and disturbance estimation and outputs the predicted current at the k+1th moment and the perturbation observable The predicted current and disturbance observations are input into the deadbeat predicted current control module to perform one-step ahead voltage prediction and output the command voltage u at the k+1th moment. d (k+1) and u q (k+1). At the same time, the permanent magnet synchronous motor i d =0 control, d-axis reference current Always 0. Reference speed And the sensorless observer feedback speed ω e (k) constitutes an error and is input into the speed loop PI controller. The output of the PI controller is the q-axis reference current The synchronous rotating coordinate system reference current is input into the deadbeat prediction current control module and used as the expected current at the k+2th moment. In addition, the disturbance observation quantity at the kth moment, the synchronous rotating coordinate system current and the sensorless observer estimated electrical angular velocity are used as the input of the recursive least squares online parameter identification module with forgetting factor to identify the resistance error ΔR online. s , inductance error ΔL s and flux error Δψ f The identification error is fed back to the sensorless observer module to correct the observer parameter mismatch. The observer outputs the electrical angular velocity as feedback for the speed loop, the adaptive super-helical sliding mode disturbance observer, the recursive least squares method with forgetting factor online parameter identification module, etc. The output electrical angle is used for Park and Anti-Park transformation. The command voltage u at the k+1th moment d (k+1) and u q (k+1) is converted into the command voltage u in the stationary coordinate system through Anti-Park transformation α (k+1) and u β (k+1), the command voltage in the stationary coordinate system is not only fed back to the sensorless observer module, but also serves as the input of the space vector modulation module SVPWM (Space Vector Pulse Width Modulation). The modulation wave is used for the switching of the two-level inverter and realizes the FOC closed-loop control of the sensorless permanent magnet synchronous motor.

[0056] The specific embodiments of the present invention are as follows:

[0057] First, the premise for establishing the ideal voltage equation of the permanent magnet synchronous motor in the rotating coordinate system is to ignore the cross-saturation effect of the motor inductance, the deviation of the motor rated parameters and other factors. However, factors such as temperature rise, demagnetization and magnetic saturation effects during the actual operation of the permanent magnet synchronous motor cause the motor resistance, inductance and flux parameters to drift to varying degrees. The mismatch of resistance and flux parameters leads to the steady-state error of the current in the synchronous coordinate system, while the mismatch of inductance parameters leads to the oscillation of the current in the synchronous coordinate system. In order to characterize the voltage deviation caused by the motor parameter mismatch, the concept of equivalent voltage disturbance caused by parameter mismatch is introduced, and the ideal voltage equation is reconstructed into the actual voltage equation, as follows:

[0058]

[0059] Among them, u d 、u q 、i d and i q are the stator voltage and stator current of d-axis and q-axis respectively; f d and f q are the lumped disturbances of the d-axis and q-axis caused by the deviation of the actual parameters of the motor from the rated parameters; L d and Lq They are the stator inductances of the d-axis and q-axis respectively. Taking the surface-mounted permanent magnet synchronous motor as an example, the inductances of the d-axis and q-axis are equal and unified as the stator rated phase inductance L s .

[0060] After further discretization by the forward Euler method, the actual voltage equation under discrete conditions, that is, the predicted current equation, can be obtained as follows:

[0061]

[0062] in, and are the predicted stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment; T s is the sampling interval, i.e. the execution cycle of the vector control FOC closed-loop control; R s and L s are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; u d (k) and u q (k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment; f d (k) and f q (k) are the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ψ f is the permanent magnet flux of the sensorless permanent magnet synchronous motor. In this embodiment, T s =0.00005, that is, the FOC execution frequency is 20kHz.

[0063] Then, an adaptive super-helical sliding mode disturbance observer is designed based on the predicted current equation; an adaptive super-helical sliding mode disturbance observer is designed to realize online observation of the lumped disturbance. The adaptive super-helical sliding mode disturbance observer is as follows:

[0064]

[0065]

[0066]

[0067]

[0068] in, and are the estimated values ​​of the stator current of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling time. and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling time; T s is the sampling interval; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; and are the estimated values ​​of the stator current of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; u d (k) and u q (k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; U dSTA and U qSTA are the sliding mode functions based on the adaptive superhelical algorithm for the d-axis and q-axis of the sensorless permanent magnet synchronous motor respectively; δ is the disturbance term gain, δ>0; λ1 and λ2 are the first and second positive gains; s d and q are the sliding surfaces formed by the d-axis and q-axis current errors, respectively, and the input error s d 、s q Adaptation is achieved through the hyperbolic tangent function inside the sliding mode function to suppress the sliding mode chattering problem; tanh() is a hyperbolic tangent function with adjustable boundary layer. The traditional superhelical algorithm itself is a second-order sliding mode algorithm and has a certain chattering suppression effect, but its built-in sign function is a step switching function, which cannot achieve adaptation to the input error under fixed gain. When the error decreases, it will be affected by the fixed gain and cause sliding mode chattering; τ is the time variable; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; x is the input of the hyperbolic tangent function tanh( ), and m is the positive gain of the hyperbolic tangent function tanh( ). When the input error is within the gain constraint range, the hyperbolic tangent function can realize error adaptation and further improve the jitter problem caused by fixed gain.

[0069] By using adaptive super-helical sliding mode disturbance observation, the inherent chattering problem of the traditional low-order sliding mode can be effectively reduced and accurate observation of the predicted current and disturbance can be achieved. In this embodiment, δ1000, λ1=900, λ2=150, and m=5 are taken.

[0070] The command voltage u calculated at the kth sampling time in the actual digital control system d (k),u q (k) is actually applied to the inverter module at the k+1th moment, and the inverter module takes about half of the sampling period to apply the voltage to the permanent magnet synchronous motor winding. Therefore, by predicting the current and disturbance observation compensation in advance and further establishing the prediction voltage equation based on the deadbeat current control based on the prediction current equation, the details are as follows:

[0071]

[0072] Among them, u d (k+1) and u q (k+1) are the predicted stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment predicted in advance, so as to compensate for the one-beat delay of the actual digital system; R s and L s are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; T s is the sampling interval; and are the reference currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment, based on i d =0 The d-axis reference current controlled by is always zero, that is, The q-axis reference current is provided by the speed loop output and the speed loop execution frequency is approximately 1 / 20 to 1 / 10 of the current loop execution frequency. It is worth noting that the mechanical frequency of the permanent magnet synchronous motor is much smaller than the electrical frequency, so the electrical angular velocity in adjacent cycles can be considered equal; and are the estimated values ​​of the stator current of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling time, The disturbance observer is used to observe and compensate online; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

[0073] The predicted voltage equation realizes online compensation of disturbances by considering the one-beat delay problem of actual digital systems.

[0074] Then, a recursive least squares online parameter identifier with forgetting factor is established; the online parameter identifier realizes the online identification of multiple parameters of the motor including resistance, inductance and flux linkage error. The recursive least squares online parameter identifier with forgetting factor is as follows:

[0075]

[0076] Φ(k)=[ΔR s ,ΔL s ,Δψ f ] T

[0077]

[0078] Among them, Φ(k) and Φ(k-1) are the parameter error vectors to be identified of the sensorless permanent magnet synchronous motor at the kth and k-1th sampling moments, respectively, ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux error of the sensorless permanent magnet synchronous motor respectively; P(k) and P(k-1) are the covariance matrices at the kth and k-1th sampling moments respectively, and the order is the number of objects to be identified; y(k) is the disturbance input of the online parameter identifier at the kth sampling moment, including the estimated value of the lumped disturbance of the d-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment and the estimated value of the lumped disturbance on the q axis is the measurement input of the online parameter identifier at the kth sampling moment; κ is the forgetting factor of the least squares parameter identification method. The value of the forgetting factor should comprehensively consider the convergence speed and identification stability of the recursive least squares parameter identification algorithm. In this embodiment, k=1e3 and κ=0.99995.

[0079] Measurement input The details are as follows:

[0080]

[0081] The recursive least squares parameter identification equation with forgetting factor is designed based on the lumped disturbance formula to realize the online identification of motor multi-parameters including resistance, inductance and flux linkage errors.

[0082] The d-axis and q-axis stator currents and stator voltages and rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the current moment are input into an adaptive super-helical sliding mode disturbance observer for processing. After processing, the adaptive super-helical sliding mode disturbance observer outputs the d-axis and q-axis stator currents and estimated values ​​of the lumped disturbances at the next moment. The d-axis and q-axis stator currents and estimated values ​​of the lumped disturbances at the next moment, the d-axis and q-axis reference currents and rotor electrical angular velocity at the current moment are input into a predicted voltage equation based on zero-beat current control for processing. After processing, the predicted voltage equation of zero-beat current control outputs the d-axis and q-axis stator voltages at the next moment, thereby performing vector control FOC closed-loop control of the sensorless permanent magnet synchronous motor.

[0083] The d-axis and q-axis stator currents and stator voltages, rotor electrical angular velocity, and the estimated values ​​of the d-axis and q-axis lumped disturbances output by the adaptive super-helical sliding mode disturbance observer of the sensorless permanent magnet synchronous motor at the current moment are input into the recursive least squares online parameter identifier with forgetting factor for processing. After processing, the recursive least squares online parameter identifier with forgetting factor outputs the parameter error vector to the sensorless observer in the vector control FOC closed-loop control, and finally realizes the predictive current control of the sensorless permanent magnet synchronous motor. The identified parameter error is fed back to the sensorless observer online to realize the robust enhancement of the observer.

[0084] The recursive least squares method with forgetting factor is used to process the online parameter identifier and output the parameter error vector to the sensorless observer in the vector control FOC closed-loop control. The motor parameter error is corrected for the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor in the synchronous rotating coordinate system, or the motor parameter error is corrected for the stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor in the stationary coordinate system. After the motor parameter error is corrected, the sensorless observer outputs the electrical angular velocity and electrical angle to realize the FOC control of the sensorless permanent magnet synchronous motor.

[0085] Finally, the identification parameter error is fed back online to the sensorless observer to achieve robust enhancement of the observer. Generally, the sensorless observer based on back electromotive force or permanent magnet flux signal is established under the voltage equation of the stationary coordinate system of the permanent magnet synchronous motor or the voltage equation of the synchronous rotating coordinate system, that is, the identification parameter error is fed back and updated to the voltage equation of the sensorless observer to achieve robust enhancement. In this embodiment, the sensorless observer is a sliding mode observer based on a saturation function. The observer is established on the voltage equation of the stationary coordinate system. The stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor in the stationary coordinate system after the motor parameter error correction are as follows:

[0086]

[0087] Among them, u α ′(k) and u β′(k) are the stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment after the motor parameter error correction; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux linkage error of the sensorless permanent magnet synchronous motor respectively; i α (k) and i β (k) are the stator currents of the α-axis and β-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; ψ f is the permanent magnet flux of the sensorless permanent magnet synchronous motor; θ e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment.

[0088] The stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor in the synchronous rotating coordinate system after the motor parameter error correction are as follows:

[0089]

[0090] Among them, u d ′(k) and u q ′(k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment after the motor parameter error correction; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux linkage error of the sensorless permanent magnet synchronous motor respectively; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

[0091] The identification parameter error is fed back online to the sensorless observer to achieve robust enhancement of the observer. Generally, the sensorless observer based on back electromotive force or permanent magnet flux signal is established under the voltage equation of the stationary coordinate system of the permanent magnet synchronous motor or the voltage equation of the synchronous rotating coordinate system, that is, the identification parameter error is fed back and updated to the voltage equation of the sensorless observer to achieve robust enhancement.

[0092] The sliding mode observer formula after parameter error feedback update is as follows:

[0093]

[0094] in, represents the estimated value of the variable x; σ is the gain of the saturation function sliding mode; sat( ) is the traditional saturation function; They are the estimated back-EMF components of the α-axis and β-axis in the stationary coordinate system respectively. The back-EMF components output the electrical angular velocity and electrical angle through the phase-locked loop, and the output is fed back to the disturbance observer, parameter identification module and coordinate transformation module to realize the closed-loop FOC control of the sensorless permanent magnet synchronous motor.

[0095] The method in this embodiment expands the application of the traditional predictive current control algorithm in the FOC control of sensorless permanent magnet synchronous motors. The hyperhelical algorithm based on the hyperbolic tangent function realizes the self-adaptation of the input error and effectively suppresses the inherent jitter of the sliding mode, realizes the accurate online observation and compensation of the disturbance, and suppresses the harmonic distortion of the phase current. The recursive least squares method with forgetting factor based on the lumped disturbance equation is used to identify the motor parameter error online and feedback update to the sensorless observer to realize the robust enhancement of the sensorless permanent magnet synchronous motor control system.

[0096] like Figure 2 (a) and Figure 2 As shown in (b), the mechanical speed, q-axis current and phase current curves of the traditional super-helical sliding mode and the adaptive super-helical sliding mode algorithms before and after the motor parameter mismatch at 1200r / m and 50% rated load are shown. The main parameters of the three-phase permanent magnet synchronous motor and the control method in this embodiment are: phase resistance 0.5Ω, phase inductance 1mH, permanent magnet flux 0.012Wb, rated speed 3000rpm, maximum speed 3500rpm, carrier frequency 20kHz (carrier period 50us). In order to simulate the real effect of motor parameter mismatch, the main parameters in the control algorithm are changed to: phase resistance 1.0Ω, phase inductance 2mH, permanent magnet flux 0.006Wb after 0.1s. When the motor parameters are mismatched, due to the online observation and compensation of the disturbance observer, the speeds of the traditional super-helical sliding mode disturbance observation and the adaptive super-helical sliding mode disturbance observation both have obvious steady-state errors, and the phase current still maintains a smooth sine-like curve. It is worth noting that the q-axis current ripple of the adaptive super-helical algorithm is smaller, and the q-axis current represents the torque, which shows that the adaptive algorithm can effectively reduce the torque ripple to achieve high-performance torque control. In addition, the improvement of the q-axis current also directly affects the accurate observation of the disturbance and further improves the accuracy of the recursive least squares parameter identification with the forgetting factor.

[0097] like Figure 3 As shown, the phase current THD analysis statistics of the motor under the condition of 50% rated load and different speeds corresponding to the no-disturbance observer, the traditional super-helical sliding mode disturbance observation and the adaptive super-helical sliding mode disturbance observation after the motor parameter mismatch. The minimum phase current THD under the condition of no-disturbance observer is 25.86%, while the maximum phase current THD under the condition of super-helical sliding mode disturbance observer does not exceed 3.02%. In contrast, the adaptive super-helical sliding mode disturbance observer has a lower THD than the traditional super-helical sliding mode disturbance observer at different speeds, achieving better harmonic suppression effect, which is conducive to accurate disturbance observation and parameter identification.

[0098] like Figure 4 As shown, the resistance identification result curve of the least square method parameter identification using the lumped perturbation equation of the present invention is compared with the traditional voltage equation. The resistance error identification result of the recursive least square method with forgetting factor based on the lumped perturbation equation is 0.5209Ω, and the corrected resistance identification result is 0.4791Ω. Compared with the traditional recursive least square method resistance identification result of 0.3714Ω based on the voltage equation with forgetting factor, the identification error is reduced from 0.1286Ω to 0.0209Ω, and the identification accuracy is significantly improved.

[0099] like Figure 5 As shown, the inductance identification result curve of the least square method parameter identification using the lumped perturbation equation of the present invention is compared with the traditional application of the voltage equation. The inductance error identification result of the recursive least square method with forgetting factor based on the lumped perturbation equation is 0.7232mH, and the corrected inductance identification result is 1.2768mH. Compared with the traditional recursive least square method inductance identification result of 1.264mH based on the voltage equation with forgetting factor, the identification accuracy is similar.

[0100] like Figure 6 As shown, the flux linkage identification result curve of least squares parameter identification using the lumped perturbation equation of the present invention is compared with the traditional application of voltage equation. The flux linkage error identification result of the recursive least squares method with forgetting factor based on the lumped perturbation equation is -0.005676Wb, and the corrected flux linkage identification result is 0.011676Wb. Compared with the traditional recursive least squares flux linkage identification result of 0.01256Wb based on the voltage equation, the identification error is reduced from 0.00056Wb to 0.000327Wb, and the identification accuracy is slightly improved.

[0101] By adopting the above technical solutions, the input error of the sliding mode disturbance observer is self-adaptive and the inherent jitter phenomenon of the traditional sliding mode is effectively suppressed, and the accurate observation of the lumped disturbance and the suppression of the total harmonic distortion of the phase current are realized. At the same time, the recursive least squares parameter identification algorithm with forgetting factor based on the lumped disturbance equation can realize the multi-parameter online identification of the motor resistance, inductance and flux error parameters. The identification accuracy is improved compared with the traditional recursive least squares method with forgetting factor based on the voltage equation. The identification parameter error feedback is updated to the sensorless observer equation to realize the robust enhancement of the sensorless observer.

[0102] The above description is only a preferred embodiment of the present invention, and does not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. A robust enhanced sensorless permanent magnet synchronous motor predictive current control method, characterized in that: include: S1: Establish the predictive current equation of sensorless permanent magnet synchronous motor and design the adaptive super-helical sliding mode disturbance observer based on the predictive current equation; S2: Establish a predictive voltage equation based on deadbeat current control and an online parameter identifier using the recursive least squares method with a forgetting factor; S3: input the d-axis and q-axis stator currents and stator voltages and rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the current moment into the adaptive super-helical sliding mode disturbance observer for processing, and the adaptive super-helical sliding mode disturbance observer outputs the d-axis and q-axis stator currents and estimated values ​​of the lumped disturbance at the next moment after processing, and inputs the d-axis and q-axis stator currents and estimated values ​​of the lumped disturbance at the next moment, the d-axis and q-axis reference currents and rotor electrical angular velocity at the current moment into the predicted voltage equation based on the deadbeat current control for processing, and the predicted voltage equation of the deadbeat current control outputs the d-axis and q-axis stator voltages at the next moment after processing, so as to perform vector control FOC closed-loop control of the sensorless permanent magnet synchronous motor; S4: The d-axis and q-axis stator currents and stator voltages, rotor electrical angular velocity and estimated values ​​of the d-axis and q-axis lumped disturbances output by the adaptive super-helical sliding mode disturbance observer of the sensorless permanent magnet synchronous motor at the current moment are input into a recursive least squares online parameter identifier with a forgetting factor for processing. After processing, the recursive least squares online parameter identifier with a forgetting factor outputs a parameter error vector to the sensorless observer in the vector control FOC closed-loop control, and finally realizes predictive current control of the sensorless permanent magnet synchronous motor.

2. The robust enhanced sensorless permanent magnet synchronous motor predictive current control method according to claim 1, characterized in that: In step S1, the prediction current equation of the permanent magnet synchronous motor is as follows: in, and are the predicted stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment; T s is the sampling interval, i.e. the execution cycle of the vector control FOC closed-loop control; R s and L s are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; u d (k) and u q (k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment; f d (k) and f q (k) are the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

3. The robust enhanced sensorless permanent magnet synchronous motor predictive current control method according to claim 1, characterized in that: In step S1, the adaptive super-helical sliding mode disturbance observer is specifically as follows: in, and are the estimated values ​​of the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment respectively; and are the estimated values ​​of the lumped disturbances of the sensorless permanent magnet synchronous motor on the d-axis and q-axis at the k+1th sampling time; T s is the sampling interval; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; and are the estimated values ​​of the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; u d (k) and u q (k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; U dSTA and U qSTA are the sliding mode functions of the d-axis and q-axis of the sensorless permanent magnet synchronous motor respectively; δ is the disturbance term gain, δ>0; λ1 and λ2 are the first and second positive gains; s d and q are the sliding surfaces formed by the d-axis and q-axis current errors respectively; tanh() is the hyperbolic tangent function; τ is the time variable; i d (k) and i q (k) are the d-axis and q-axis stator currents of the sensorless permanent magnet synchronous motor at the kth sampling moment, respectively.

4. The robust enhanced sensorless permanent magnet synchronous motor predictive current control method according to claim 1, characterized in that: In step S2, the predicted voltage equation based on deadbeat current control is as follows: Among them, u d (k+1) and u q (k+1) are the predicted stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment; R s and L s are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; T s is the sampling interval; and are the reference currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment respectively; and are the estimated values ​​of the stator current of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; and are the estimated values ​​of the lumped disturbances of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k+1th sampling time; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

5. The robust enhanced sensorless permanent magnet synchronous motor predictive current control method according to claim 1, characterized in that: In step S2, the recursive least squares online parameter identifier with forgetting factor is specifically as follows: Φ(k)=[ΔR s ,ΔL s ,Dp f ] T Among them, Φ(k) and Φ(k-1) are the parameter error vectors to be identified of the sensorless permanent magnet synchronous motor at the kth and k-1th sampling moments, respectively, ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux error of the sensorless permanent magnet synchronous motor respectively; P(k) and P(k-1) are the covariance matrices at the kth and k-1th sampling moments respectively, and the order is the number of objects to be identified; y(k) is the disturbance input of the online parameter identifier at the kth sampling moment, including the estimated value of the lumped disturbance of the d-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment and the estimated value of the lumped disturbance on the q axis is the measurement input of the online parameter identifier at the kth sampling moment; κ is the forgetting factor of the least squares parameter identification method.

6. The robust enhanced sensorless permanent magnet synchronous motor predictive current control method according to claim 1, characterized in that: In the step S4, after the recursive least squares method with forgetting factor is processed by the online parameter identifier, the parameter error vector is output to the sensorless observer in the vector control FOC closed-loop control, and the motor parameter error correction is performed on the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor in the synchronous rotating coordinate system, or the motor parameter error correction is performed on the stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor in the stationary coordinate system.

7. The robust enhanced sensorless permanent magnet synchronous motor predictive current control method according to claim 6, characterized in that: The stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor in the synchronous rotating coordinate system after the motor parameter error correction are specifically as follows: Among them, u d ′(k) and u q ′(k) are the stator voltages of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment after the motor parameter error correction; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux linkage error of the sensorless permanent magnet synchronous motor respectively; i d (k) and i q (k) are the stator currents of the d-axis and q-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; ψ f is the permanent magnet flux linkage of the sensorless permanent magnet synchronous motor.

8. The robust enhanced sensorless permanent magnet synchronous motor predictive current control method according to claim 6, characterized in that: The stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor in the stationary coordinate system after the motor parameter error correction are specifically as follows: Among them, u α ′(k) and u β ′(k) are the stator voltages of the α-axis and β-axis of the sensorless permanent magnet synchronous motor at the kth sampling moment after the motor parameter error correction; R s and L s They are the stator rated phase resistance and stator rated phase inductance of the sensorless permanent magnet synchronous motor respectively; ΔR s , ΔL s and Δψ f are the resistance error, inductance error and flux linkage error of the sensorless permanent magnet synchronous motor respectively; i α (k) and i β (k) are the stator currents of the α-axis and β-axis of the sensorless permanent magnet synchronous motor at the k-th sampling moment; ω e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment; ψ f is the permanent magnet flux of the sensorless permanent magnet synchronous motor; θ e (k) is the rotor electrical angular velocity of the sensorless permanent magnet synchronous motor at the kth sampling moment.

9. An electronic device, characterized in that: include: A memory and a processor coupled to each other, wherein the memory stores program data, and the processor calls the program data to execute the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having program data stored thereon, characterized in that: When the program data is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Citation Information

Patent Citations

  • Permanent magnet synchronous motor predicted current control method and system based on FOSMC

    CN110022105A

  • Double-ring dead-beat prediction control method for permanent magnet synchronous motor based on disturbance estimation compensation

    CN110165951A

  • Predictive control method and system for super-spiral sliding-mode observer of permanent magnet synchronous motor

    CN114362626A

  • Dead-beat predictive current loop control method for permanent magnet synchronous motor

    CN115347841A

  • Permanent magnet motor passive control method based on adaptive gain superspiral sliding mode control

    CN117792184A

Cited By

  • Real-time position identification method and system for permanent-magnet semi-direct-drive wind generating set and medium

    CN119787908A

  • Self-adaptive current decoupling control method for sliding mode of permanent magnet synchronous motor

    CN119966307A

  • Permanent magnet synchronous motor sensorless control system for enhancing system robustness

    CN120074314A

  • Permanent magnet synchronous motor inductance parameter identification method based on data pre-screening

    CN120566974A

  • Reluctance motor control method and system based on three-dimensional model parameter observation

    CN121923546A