A Model-Free Control Method for Permanent Magnet Synchronous Motor Based on Disturbance Observation

Abstract

The present invention provides a model-free control method for a permanent magnet synchronous motor based on disturbance observation. During the current predictive control process, only two observer control parameters need to be adjusted, and it does not depend on any motor parameters, overcoming the shortcomings of the prior art that are easily affected by parameter drift and model mismatch caused by factors such as temperature, magnetic field saturation, and operating conditions, reducing the harmonic content of current predictive control during parameter perturbation, avoiding oscillation and current static error, and improving the robustness of motor control, thereby achieving many beneficial effects that the prior art does not have.

Description

Technical Field

[0001] The present invention belongs to the technical field of permanent magnet synchronous motor control, and particularly relates to a method for realizing predictive control of current under the condition of motor parameter mismatch based on a Luenberger disturbance observer and an ultra-local model. Background Technique

[0002] In the drive of a permanent magnet synchronous motor, the current loop located at the innermost side of the control structure plays a very important role and directly affects the dynamic and steady-state performance of the motor drive system. At present, predictive current control has gradually become the mainstream method for permanent magnet synchronous motor control due to its advantages of being easy to handle multi-variable situations, having a fast dynamic response, being easy to include time variables and non-linearities, etc. In deadbeat predictive current control, the d-q axis voltage vectors are predicted from the reference current values, the feedback stator current, and the rotor position. After being modulated by the inverter SVPWM, the voltage is applied to the permanent magnet synchronous motor, so that within one control cycle, the actual current can follow the reference current. However, since the permanent magnet synchronous motor control system contains many non-linear factors, such as inevitable disturbances and parameter variations under operating conditions, and the above deadbeat predictive control also inevitably has the disadvantage that the accuracy is too dependent on the control model, and its performance depends to a large extent on the actual motor. When the motor is running, the motor parameters change, resulting in parameter drift. At the same time, these parameters are also affected by internal and external unknown disturbances, which leads to a reduction in motor control performance, poor anti-interference performance, and low robustness. The mismatch of electromagnetic parameters will further cause the reference voltage calculated by the predictive current controller to deviate from the required value. Summary of the Invention

[0003] In view of the above technical problems existing in the art, the present invention provides a model-free control method for a permanent magnet synchronous motor based on disturbance observation, which specifically includes the following steps:

[0004] Step 1: Construct a motor voltage equation containing system parameter disturbances;

[0005] Step 2: Obtain an ultra-local model for model-free current predictive control of a permanent magnet synchronous motor based on the established motor voltage equation;

[0006] Step 3: Construct a Luenberger disturbance observer and a system state equation without model parameters with current and system disturbances as state variables; analyze the stability of the system to obtain the poles of the Luenberger disturbance observer;

[0007] Step 4: Use the Luenberger disturbance observer to observe the system disturbance, substitute it into the ultra-local model to calculate the reference voltage, and use it for the model-free current predictive control.

[0008] Further, the motor voltage equation constructed in Step 1 specifically adopts the following form:

[0009]

[0010] Among them, R0, L0, and ψ f0 are the nominal parameters of the stator resistance, inductance, and rotor permanent magnet flux linkage respectively, ω e is the electrical angular velocity of the motor rotor, i d , i q are the d-axis and q-axis currents of the motor respectively, U d , U q are the d-axis and q-axis voltages of the motor respectively, f d , f q are the d-axis and q-axis system disturbances of the motor respectively, represents the differential of the parameter.

[0011] Furthermore, the super-local model described in step 2 is derived from the motor voltage equation to the following form:

[0012]

[0013]

[0014]

[0015]

[0016] Among them, F d , F q are the disturbances caused by the changes of the unknown quantities of the motor system on the d-axis and q-axis of the motor respectively, ε is the gain value of the input voltage, and it is a constant related to the nominal value of the motor stator inductance.

[0017] Furthermore, the construction process of the Luenberger disturbance observer without model parameters described in step 3 is as follows:

[0018] Based on the assumption that the change rates of the d-axis and q-axis system disturbances f d , f q are zero, that is , select the d-axis and q-axis currents and the system disturbance as the system state variables, construct the state equation of this Luenberger disturbance observer and perform discretization to obtain:

[0019]

[0020] Among them, k1 and k2 are the observer gains, T s is the sampling time, the superscript "∧" represents the estimated predicted value of the corresponding parameter, k is a certain moment,

[0021]

[0022] Furthermore, in step 3, the poles are configured to analyze the system stability. The specific form of the system state characteristic equation is constructed as follows:

[0023] |λI - G| = -[λ 2 +(k1 - 2)λ + 1 - k1 - T s εk2] 2 = 0

[0024] where λ is the characteristic root of the characteristic equation, i.e., the pole of the system, and I is the identity matrix.

[0025] Solving this characteristic equation, the poles of the disturbance observer can be obtained as:

[0026]

[0027] Make the poles be distributed within the unit circle in the z-domain to maintain stability, and thereby determine the respective value ranges of k1 and k2.

[0028] Furthermore, in step 4, specifically using the system disturbance and estimated prediction values observed by the Romberg disturbance observer, substituting them into the super-local model and discretizing, the following reference voltage is calculated:

[0029]

[0030] where, are the reference voltages respectively, are the reference currents respectively, and α is a constant related to the nominal value of the motor stator resistance.

[0031] For the method provided by the present invention above, during the current predictive control process, it only needs to adjust two observer control parameters and does not depend on any motor parameters, overcoming the disadvantages of the prior art being susceptible to parameter drift and model mismatch caused by factors such as temperature, magnetic field saturation, and operating state, reducing the harmonic content during parameter perturbation in current predictive control, avoiding oscillation and current static error, and improving the robustness of motor control, thereby being able to achieve many beneficial effects that the prior art does not have. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is the framework of the permanent magnet synchronous motor current predictive control system constructed based on the present invention;

[0033] Figure 2 is the overall flow of the method provided by the present invention;

[0034] Figure 3 is the result of deadbeat current predictive control executed under the condition of no motor parameter perturbation at a sampling frequency of 20 kHz;

[0035] Figure 4 The current prediction control result of implementing the present invention without motor parameter disturbance at a sampling frequency of 20 kHz;

[0036] Figure 5 The deadbeat current prediction control result executed under the condition of motor parameter disturbance at a sampling frequency of 20 kHz;

[0037] Figure 6 The current prediction control result of implementing the present invention under the condition of motor parameter disturbance at a sampling frequency of 20 kHz. Detailed implementation manners

[0038] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0039] Figure 1 A block diagram of a current prediction control system that can be used to implement the method provided by the present invention is shown. The control system mainly includes a PI speed loop, an improved current loop, a Clarke / Park transformation module, a permanent magnet synchronous motor module, an SVPWM model, and a three-phase voltage inverter. The abc three-phase currents of the motor are detected by current sensors and then transformed into dq-axis currents through coordinate transformation. The speed sensor provides real-time speed signals and rotor position signals. Given a reference speed, the q-axis reference current can be obtained by the PI control of the speed loop. Since the vector control strategy of i d =0 is adopted, the d-axis reference current is 0. According to the improved current loop, current prediction control is performed to obtain the dq-axis reference voltages acting in the next period. After inverse coordinate transformation, the αβ-axis voltages are obtained and used as the input of the SVPWM (space vector modulation) module, which is output to the inverter power module to finally drive the motor to operate. Among them, the improved current loop consists of a Luenberger disturbance observer and a hyper-local model-free current prediction controller.

[0040] The method provided by the present invention that can be applied to the above system, as Figure 2 shown, specifically includes the following steps:

[0041] Step 1: Construct a motor voltage equation containing system parameter disturbances;

[0042] Step 2: Obtain a hyper-local model for model-free current prediction control of a permanent magnet synchronous motor based on the established motor voltage equation;

[0043] Step 3: Using current and system disturbance as state variables, construct a Luenberger disturbance observer and a system state equation without model parameters; analyze the stability of the system to obtain the poles of the Luenberger disturbance observer.

[0044] Step 4: Use the Luenberger disturbance observer to observe the system disturbance, substitute it into the hyper-local model to calculate the reference voltage, and use it for model-free current predictive control.

[0045] First, establish the mathematical model of the permanent magnet synchronous motor in the synchronous rotating coordinate system:

[0046]

[0047] Since the permanent magnet synchronous motor is a non-linear system, during the operation of the motor, the motor parameters will inevitably change and bring disturbances. To improve the control robustness of the motor under various working conditions, the parameter disturbance term needs to be considered:

[0048]

[0049]

[0050] In the above equations, ΔR, ΔL, Δψ f are system parameter disturbances, R0, L0, ψ f0 are the parameters of the stator resistance, inductance and rotor permanent magnet flux as specified on the motor nameplate, ω e is the electrical angular velocity of the motor rotor, i d , i q are the d-axis and q-axis currents of the motor respectively, U d , U q are the d-axis and q-axis voltages of the motor respectively, f d , f q are the d-axis and q-axis system disturbances of the motor respectively, represents the differential of the parameter, T d , T q are the d-axis and q-axis noises and other unknown disturbance terms of the motor respectively.

[0051] Therefore, the motor voltage equation constructed in Step 1 specifically adopts the following form:

[0052]

[0053] The hyper-local model in Step 2 is derived from the motor voltage equation to the following form:

[0054]

[0055]

[0056]

[0057]

[0058] Among them, F d and F q are respectively the disturbances caused by the changes of the unknown quantities of the motor system on the d-axis and q-axis of the motor. ε is the gain value of the input voltage, which is a constant related to the nominal value of the stator inductance of the motor.

[0059] The present invention introduces a Romberg disturbance observer with excellent performance in parameter disturbance observation to observe the system disturbance. Based on the assumption that the change rates of the d-axis and q-axis system disturbances f d and f q are zero, that is , the d-axis and q-axis currents and the system disturbance are selected as the system state variables to construct the state equation:

[0060]

[0061]

[0062]

[0063] Correspondingly, the Romberg disturbance observer is constructed as:

[0064]

[0065]

[0066] After discretizing the Romberg disturbance observer, considering that the sampling time T s is small enough, it can be considered that T s R0 / L0 = 0, and T s ω e = 0. Therefore, a Romberg disturbance observer without model parameters can be constructed:

[0067]

[0068] Among them, k1 and k2 are the observer gains, T s is the sampling time, the superscript "∧" represents the estimated predicted value of the corresponding parameter, k is a certain moment,

[0069]

[0070] In step 3, the poles are configured to analyze the system stability. The system state characteristic equation is specifically constructed in the following form:

[0071] |λI - G| = -[λ 2+(k1 - 2)λ + 1 - k1 - T s εk2] 2 = 0

[0072] Where λ is the eigenvalue of the characteristic equation, i.e., the pole of the system, and I is the identity matrix.

[0073] Solving this characteristic equation, the poles of the disturbance observer can be obtained as:

[0074]

[0075] Make the poles distribute inside the unit circle in the z-domain to maintain stability, and thereby determine the respective value ranges of k1 and k2.

[0076] In step 4, specifically, the system disturbance and estimated prediction values observed by the Romberg disturbance observer are substituted into the super-local model and discretized, and the following reference voltages are calculated:

[0077]

[0078] Where, are the reference voltages respectively, are the reference currents respectively, and α is a constant related to the nominal value of the motor stator resistance.

[0079] By comparing the specific examples based on the present invention with the prior art, the beneficial effects of the present invention can be more intuitively reflected. Figure 3 and Figure 4 respectively give the simulation results of deadbeat current predictive control and model-free current predictive control based on the Romberg disturbance observer without motor parameter perturbation at a sampling frequency of 20 kHz. The test conditions are that a speed reference value of 1000 rpm is given, and then the load torque is suddenly changed from 5 Nm to 10 Nm at t = 0.02 s. The first channel shows the reference and actual currents on the q-axis, and the second channel shows the actual current on the d-axis. The results show that although the traditional deadbeat predictive current has good dynamic performance without parameter perturbation, there is a current static error, and the followability of the q-axis current to the reference current is not good; while the improved model-free current predictive control based on the Romberg disturbance observer has no current static error on the q-axis without parameter perturbation, has good followability to the reference current, and the dynamic and static performance of motor control are improved.

[0080] Figure 5 and Figure 6 respectively give the cases where motor parameters are perturbed (L s = 2L0, ψ f = 2ψ f0) Simulation results of deadbeat current predictive control and model-free current predictive control based on the Luenberger disturbance observer. The test conditions are as follows: a speed reference value of 1000 rpm is given, and then the load torque is suddenly changed from 5 Nm to 10 Nm at t = 0.02 s. The first channel shows the reference and actual currents of the q-axis, and the second channel shows the actual current of the d-axis. The results show that the traditional deadbeat predictive current is quite sensitive to parameter perturbations. The dq-axis currents cannot accurately track the reference values, obvious oscillations occur, the harmonic content is large, and the control performance of the motor is greatly reduced. The improved model-free current predictive control based on the Luenberger disturbance observer eliminates the steady-state tracking error of the q-axis current in both cases, has good tracking performance, low harmonic content, and shows strong robustness to changes in machine parameters.

[0081] It should be understood that the magnitudes of the sequence numbers of the steps in the embodiments of the present invention do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0082] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A model-free control method for permanent magnet synchronous motors based on disturbance observation, characterized in that: Specifically, it includes the following steps: Step 1: Construct a motor voltage equation containing system parameter perturbations: where, R0, L0, ψ f0 are the nominal parameters of the stator resistance, inductance and rotor permanent magnet flux linkage respectively, ω e is the electrical angular velocity of the motor rotor, i d , i q are the d-axis and q-axis currents of the motor respectively, U d , U q are the d-axis and q-axis voltages of the motor respectively, f d , f q are the d-axis and q-axis system disturbances respectively, denotes the differential of the parameter; Step 2: Obtain a hyper-local model for model-free current predictive control of a permanent magnet synchronous motor based on the established motor voltage equation: where, F d and F q are respectively the disturbances caused by the changes of the unknown variables of the motor system on the d-axis and q-axis of the motor, ε is the gain value of the input voltage, which is a constant related to the nominal value of the stator inductance of the motor; Step 3: Using current and system perturbations as state variables, construct a Luenberger disturbance observer and a system state equation without model parameters through the following process: Perturbation f based on the d, q axis system d , f q The rate of change is zero, that is Based on the assumption that, the d and q axis currents and the system perturbation are selected as the system state variables, the system state equation of this Romberg perturbation observer is constructed and discretized to obtain: where k1 and k2 are observer gains, T s is the sampling time, the superscript "∧" represents the estimated predicted value of the corresponding parameter, and k is a certain moment. Analyze the stability of the system to obtain the poles of the Luenberger disturbance observer; Step 4: Use the Luenberger disturbance observer to observe the system perturbation, substitute it into the hyper-local model to calculate the reference voltage, and use it for the model-free current predictive control.

2. The method according to claim 1, wherein: In Step 3, the poles are configured to analyze the system stability. The system state characteristic equation is specifically constructed in the following form: |λI - G| = -[λ 2 +(k1 - 2)λ + 1 - k1 - T s εk2] 2 = 0 where λ is the characteristic root of the characteristic equation, i.e., the pole of the system, and I is the identity matrix; Solving this characteristic equation, the poles of the disturbance observer can be obtained as: Make the poles distributed within the unit circle in the z-domain to maintain stability, and thereby determine the respective value ranges of k1 and k2.

3. The method according to claim 2, wherein: In step 4, specifically, the system disturbance observed by using the Romberg disturbance observer and the estimated predicted value are substituted into the hyper-local model and discretized, and the following reference voltage is calculated: Among them, are the reference voltages respectively, are the reference currents respectively, and α is a constant related to the nominal value of the motor stator resistance.

Images