Permanent magnet motor predictive current control method and system based on iterative sliding mode observer

Abstract

The invention discloses a permanent magnet motor predictive current control method and system based on an iterative sliding-mode observer, and is used for solving the problem that the performance of motor control is reduced due to the fact that periodic large current fluctuation related to the rotating speed is generated due to inevitable dead time and flux linkage distortion during operation of a motor. According to the technical scheme, a brand-new iterative sliding-mode observer ISMO is provided, the sliding-mode observer (SMO) and the ILC technology are fused, and the sliding-mode observer ISMO and the ILC technology are applied to a current loop. Wherein the ILC part is responsible for suppressing periodic disturbances such as nonlinearity and magnetic flux harmonics of an inverter, and the SMO part is used for observing and processing residual non-periodic disturbances. The DPCC equation is corrected and compensated through the current and disturbance information observed by the ISMO, so that the control error and the current harmonic during the operation of the permanent magnet synchronous motor are effectively reduced, and the anti-interference capability and the operation stability of the control system are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of permanent magnet motor control, and in particular relates to a permanent magnet motor predictive current control method and system based on an iterative sliding mode observer. Background Art

[0002] Permanent magnet synchronous motor (PMSM) and its digital control system have been widely used in industrial applications due to its advantages such as high power density, high control accuracy and high efficiency. As the inner loop of the entire control system, the current controller plays an important role in improving the steady-state accuracy and dynamic response of the system. In order to improve the tracking accuracy and dynamic response speed of current control, a variety of control strategies have been developed, including:

[0003] Proportional-integral (PI) control: has small steady-state error and simple structure, but parameter adjustment is difficult in the face of parameter changes and interference;

[0004] Robust control: Compared with PI control, it has faster dynamic response and good robustness, but it may produce larger ripples in high-precision systems;

[0005] Predictive control: Based on discrete models, it includes finite control set model predictive control (FCS-MPC) and dead zone predictive current control (DPCC). FCS-MPC predicts the current and minimizes the cost function by selecting a finite voltage vector, while DPCC uses a continuous voltage vector and space vector PWM (SVPWM) to simplify calculations and improve steady-state tracking performance.

[0006] In order to solve the performance problems of DPCC under parameter mismatch and external interference, as well as the delay problem in digital control systems, there are several improvement methods:

[0007] Robust DPCC: constructed by adding discrete time integral terms to the DPCC prediction structure. Although it is easy to implement and insensitive to PMSM system parameter changes, it may affect the dynamic performance;

[0008] DPCC based on PI compensation: uses a fuzzy algorithm to determine the weight coefficient of the PI compensation link, reduces the impact of the integral chain on dynamic performance, and eliminates oscillations and errors caused by parameter mismatch; DPCC based on online parameter identification: uses a fully discretized model to reconstruct the characteristic vector through disturbance observer and current injection, solving the parameter mismatch problem.

[0009] The above methods can reduce the impact of parameter mismatch on DPCC to varying degrees. However, periodic disturbances, such as the dead time of switching devices, gear torque, stator current measurement error, flux harmonics and harmonic effects, which account for a large part of the cumulative disturbances in current control, are still not well suppressed, especially the inevitable dead time and flux distortion during motor operation, which will lead to periodic large current fluctuations related to the speed, seriously affecting the motor control performance. If these periodic disturbances are suppressed, the controller performance can be improved and the observer gain can be further reduced. Summary of the invention

[0010] The purpose of the present invention is to solve the technical problem that the prior art cannot effectively suppress periodic disturbances, resulting in the control performance to be improved, and then provide a permanent magnet motor predictive current control method and system based on an iterative sliding mode observer. The present invention has developed and found that the periodic error caused by dead time and flux harmonics has a significant impact, and then proposed a new iterative sliding mode observer ISMO, which combines the sliding mode observer (SMO) and ILC technology and applies it to the current loop. Among them, the ILC part is responsible for suppressing periodic disturbances such as inverter nonlinearity and flux harmonics, while the SMO part observes and processes the remaining non-periodic disturbances. The DPCC equation is corrected and compensated by the current and disturbance information observed by ISMO, thereby ensuring the stability of the system and improving the dynamic performance. Among them, the iterative learning control (ILC) introduced in the present invention uses the error between the output and the target to correct the input signal, improve the system tracking performance, and is particularly suitable for control systems with repeated operations. ILC does not rely on an accurate system model, has a simple learning rate and low computational complexity, and can effectively suppress periodic disturbances in PMSM operation.

[0011] To this end, the present invention provides the following technical solutions:

[0012] On the one hand, the present invention provides a permanent magnet motor predictive current control method based on an iterative sliding mode observer, comprising the following steps:

[0013] Data acquisition: Collect the electrical angular velocity and motor position angle of the permanent magnet synchronous motor, as well as the three-phase current of the permanent magnet synchronous motor and perform coordinate conversion to obtain the d-axis and q-axis currents in the rotating coordinate system;

[0014] Iterative sliding mode observation: the d-axis and q-axis currents and the currently obtained d-axis and q-axis desired voltages and electrical angular velocities are input into an iterative sliding mode observer to obtain observation quantities, wherein the observation quantities include d-axis and q-axis current observation values, voltage compensation quantities, harmonic disturbance compensation quantities, and parameter mismatch compensation quantities;

[0015] The iterative sliding mode observer is an ILC module based on iterative learning control added to the sliding mode observer SMO, and the ILC module is provided with flux linkage harmonic disturbance compensation and voltage error compensation for dead time;

[0016] Deadbeat current prediction: The deadbeat current prediction module constructed by combining the observed quantity obtained by the sliding mode observer with the reference current input of the d and q axes obtains the new expected voltage of the d and q axes;

[0017] Then, the new desired voltages of the d and q axes are subjected to inverse Park transformation and input into the control inner loop of the permanent magnet motor to generate control information to control the operation of the PMSM module of the permanent magnet synchronous motor.

[0018] Optionally, the mathematical model of the flux harmonic disturbance compensation is expressed as:

[0019]

[0020] In the formula, and is the flux harmonic disturbance compensation of the d and q axes in the current sampling period, is the compensation of magnetic flux harmonic disturbance of d and q axes in the previous sampling period, T is the sampling period, β 1 , k 1d is the positive coefficient set for the d-axis on the iterative sliding mode observer, β 3 , k 1q It is the positive coefficient set for the q-axis on the iterative sliding mode observer.

[0021] Optionally, the mathematical model of the voltage error compensation is expressed as:

[0022]

[0023]

[0024] In the formula, and They are the voltage error compensation of the d and q axes of the current sampling period, and are the voltage error compensation of the d and q axes of the previous sampling period, T is the sampling period, β 2 , k 1d are all positive coefficients set for the d-axis on the iterative sliding mode observer, β 4 , k 1q They are all positive coefficients set for the q-axis on the iterative sliding mode observer.

[0025] Optionally, the observation formula of the iterative sliding mode observer is expressed as:

[0026]

[0027]

[0028] in:

[0029]

[0030]

[0031] In the formula, are the current observation values ​​of the d and q axes respectively The differential, R s is the stator resistance, L is the stator inductance, ω e is the electrical angular velocity, and ω e >ε, a positive constant set by ε, i d 、i q are the actual currents of the d and q axes, u d 、u q are the expected voltages of the d and q axes respectively, are the parameter mismatch compensation of d and q axes respectively, They are parameter mismatch compensation The differential of They are the voltage error compensation of d and q axes respectively. are the compensation of magnetic flux harmonic disturbances on the d and q axes, ν d 、v q are the sliding mode control functions corresponding to the d and q axes on the iterative sliding mode observer; k 1d , k 2d , k 3d are all positive coefficients set for the d-axis on the iterative sliding mode observer; k 1q , k 2q , k 3q They are all positive coefficients set for the q-axis on the iterative sliding mode observer;

[0032] They represent the difference between the estimated current and the actual current after d-axis and q-axis observation respectively.

[0033] Optionally, the deadbeat current prediction module obtains the expected voltage of the d and q axes using a formula expressed as:

[0034]

[0035]

[0036] In the formula, They represent the expected voltages of the d and q axes, respectively, L is the stator inductance, T is the sampling period, and R s is the stator resistance, k is the kth sampling moment, They represent the reference currents of the d and q axes at the k+2 sampling time, respectively. They represent the current observation values ​​of the d and q axes at the k+1 sampling time, respectively. are the parameter mismatch compensation amounts of the d and q axes at k sampling moments, respectively. are the voltage error compensation of d and q axes at k sampling moments, are the flux harmonic disturbance compensation of d and q axes at k sampling moments, ω e (k) represents the electrical angular velocity at time k, ψ f is the rotor flux.

[0037] It can be seen that the future reference current value is needed, which can be assumed to be equal to the actual value, because T s is short enough relative to the dynamic behavior of the system. Therefore, the reference current value can be considered to be at T s The output desired voltage makes the actual current close to the predicted value.

[0038] Optionally, the process of performing an inverse Park transformation on the desired voltages of the d and q axes and inputting them into the control inner loop of the permanent magnet motor to generate control information to control the operation of the PMSM module of the permanent magnet synchronous motor is as follows:

[0039] The desired voltages of the d and q axes are input into the inverse Park transformation module in coordination with the motor position angle to obtain the voltages in the two-phase stationary coordinate system;

[0040] Then it is input into the SVPWM generation module, and after modulation, it generates six inverter switch pulse signals to drive the permanent magnet synchronous motor PMSM module to operate.

[0041] Optionally, the voltage error derived from the dead time is expressed in the d and q axes as:

[0042]

[0043]

[0044] In the formula, Δu d , Δu q Respectively represent the voltage error derived from the dead time on the d and q axes, u dead is the dead zone voltage, γ is the angle difference between the q axis and the current vector, k is a positive integer, k represents the sampling time or sampling point, ω e represents electrical angular velocity, and t represents time.

[0045] Optionally, the flux harmonic disturbance compensation is set for the periodic disturbance caused by flux harmonics, and the flux harmonics are expressed in the d and q axes as:

[0046]

[0047]

[0048] In the formula, r d 、r q is the flux harmonic on the d and q axes, k represents the sampling time / sampling point, ψ d，6k and ψ q,6k is the amplitude of the 6kth harmonic, θ e is the motor position angle.

[0049] In another aspect, the present invention provides a system based on the above control method, comprising at least:

[0050] Position sensor, used to collect the electrical angular velocity and motor position angle of the permanent magnet synchronous motor;

[0051] Current sensor, used to collect the three-phase current of the permanent magnet synchronous motor;

[0052] Clark transformation module, used to transform the three-phase current of permanent magnet synchronous motor into a two-phase stationary coordinate system;

[0053] Park transformation module, used to transform the current of the two-phase stationary coordinate system into the two-phase rotating coordinate system;

[0054] Iterative sliding mode observer: used to calculate and obtain observation quantities using the d-axis and q-axis currents and the currently obtained d-axis and q-axis desired voltages and electrical angular velocities, wherein the observation quantities include d-axis and q-axis current observation values, voltage compensation quantities, harmonic disturbance compensation quantities, and parameter mismatch compensation quantities;

[0055] The iterative sliding mode observer is an ILC module based on iterative learning control added on the basis of the sliding mode observer SMO, and the ILC module is provided with flux linkage harmonic disturbance compensation and voltage error compensation for dead time;

[0056] A deadbeat current prediction module, used to calculate and obtain new expected voltages of the d and q axes by using the observed quantity obtained by the sliding mode observer in coordination with the reference current inputs of the d and q axes;

[0057] An inverse Park transformation module, used for performing an inverse Park transformation on the new desired voltages of the d and q axes;

[0058] The control module is used to utilize the desired voltage after the inverse Park transformation to input the control inner loop of the permanent magnet motor to generate control information to control the operation of the PMSM module of the permanent magnet synchronous motor.

[0059] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program is called by a processor to implement:

[0060] A predictive current control method for permanent magnet motor based on iterative sliding mode observer.

[0061] Beneficial Effects

[0062] Compared with the existing method, the advantages of the present invention are:

[0063] The technical solution of the present invention aims at the problem that the inevitable dead time and flux distortion during motor operation will cause periodic large current fluctuations related to the speed, thereby reducing the performance of motor control. By proposing a new sliding mode observer, it is effectively solved. The iterative sliding mode observer fully considers the characteristics of disturbances, wherein the disturbance caused by parameter decoupling is bounded, while the voltage error caused by the dead time and the harmonic component caused by the flux distortion are periodic disturbances. To this end, SMO can use the switching function to approximate the bounded disturbance, and the iterative learning strategy can suppress the periodic disturbance according to the previously recorded error. Therefore, the iterative sliding mode observer proposed in the present invention integrates the sliding mode observer (SMO) and ILC technology, and applies it to the current loop, making full use of the ILC part to suppress periodic disturbances such as inverter nonlinearity and flux harmonics. In the ILC module, flux harmonic disturbance compensation and voltage error compensation for dead time are provided, which effectively improves the control performance and further ensures the stable operation of the motor. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is the block diagram of the deadbeat prediction current module;

[0065] Figure 2 This is the ISMO principle block diagram, (a) is the d-axis, (b) is the q-axis;

[0066] Figure 3 It is the structural block diagram of DPCC_ISMO control system;

[0067] Figure 4 It is a block diagram of PWM structure, that is, inserting Transport delay into the PWM submodule in SVPWM to simulate dead time;

[0068] Figure 5 Schematic diagram of the comparison of steady-state waveforms of d-axis current of three models under periodic disturbance, (a) corresponds to DPCC model, (b) corresponds to SMO-DPCC model, and (c) corresponds to ISMO model;

[0069] Figure 6 Schematic diagram of the comparison of the steady-state waveforms of the q-axis current of the three models, (a) corresponds to the DPCC model, (b) corresponds to the SMO-DPCC model, and (c) corresponds to the ISMO model;

[0070] Figure 7Schematic diagram of FFT analysis of phase A current, (a) corresponds to DPCC model, (b) corresponds to SMO-DPCC model, and (c) corresponds to ISMO model;

[0071] Figure 8 Schematic diagram of the comparison of d-axis current waveforms of three models under inductance mismatch, (a) corresponds to DPCC model, (b) corresponds to SMO-DPCC model, and (c) corresponds to ISMO model;

[0072] Fig. 9 This is a schematic diagram comparing the q-axis current waveforms of the three models under flux mismatch, (a) corresponds to the DPCC model, (b) corresponds to the SMO-DPCC model, and (c) corresponds to the ISMO model. DETAILED DESCRIPTION

[0073] The technical solution of the present invention proposes a permanent magnet motor predictive current control method and system based on an iterative sliding mode observer, and the ILC part of the proposed iterative sliding mode observer effectively suppresses periodic disturbances such as inverter nonlinearity and flux harmonics, while the SMO part of the iterative sliding mode observer observes and processes the remaining non-periodic disturbances, further ensuring the stable operation of the motor. The present invention will be further described below in conjunction with an embodiment.

[0074] Based on the above core idea, the present invention provides a permanent magnet motor predictive current control method based on iterative sliding mode observer, comprising the following steps:

[0075] Data acquisition: Collect the electrical angular velocity and motor position angle of the permanent magnet synchronous motor, as well as the three-phase current of the permanent magnet synchronous motor and perform coordinate conversion to obtain the d-axis and q-axis currents in the rotating coordinate system;

[0076] Iterative sliding mode observation: The d-axis and q-axis currents and the currently obtained d-axis and q-axis desired voltages and electrical angular velocity are input into the iterative sliding mode observer to obtain the observed quantities, which include the d-axis and q-axis current observation values, voltage compensation, harmonic disturbance compensation, and parameter mismatch compensation.

[0077] Deadbeat current prediction: The deadbeat current prediction module constructed by combining the observation quantity obtained by the sliding mode observer with the reference current input of the d and q axes obtains the new expected voltage of the d and q axes;

[0078] The new desired voltage of the d and q axes is input into the inverse Park transformation module in coordination with the motor position angle to obtain the voltage in the two-phase stationary coordinate system; then it is input into the SVPWM generation module, and after modulation, a six-way inverter switch pulse signal is generated to drive the permanent magnet synchronous motor PMSM module to operate.

[0079] It should be noted that after obtaining the voltage in the two-phase stationary coordinate system, how to generate control instructions to drive the permanent magnet synchronous motor PMSM module to operate is achieved by the existing technology. The above is an example, and other feasible existing methods also fall within the protection scope of the present invention.

[0080] In order to fully demonstrate the necessity of the iterative sliding mode observer proposed in the present invention, the present invention conducts the following analysis:

[0081] (1) Parameter sensitivity analysis of traditional DPCC

[0082] When the iron loss and magnetic saturation can be ignored, the current and motion equations of the surface-mounted permanent magnet synchronous motor (PMSM) in the synchronous rotating coordinate system can be expressed as:

[0083]

[0084] Among them, u d and u q is the stator voltage of the d-axis and q-axis, i d and i q is the current of d-axis and q-axis, ω e and ω m are the electrical angular velocity and the mechanical angular velocity, respectively. Represents the mechanical angular velocity ω m The differential of e is the electromagnetic torque, B is the damping coefficient, J is the moment of inertia, T L is the load torque, R s is the stator resistance, t is the time, ψ d and ψ q is the stator flux of the d-axis and q-axis, which can be expressed as:

[0085] ψ d =Li d +ψ f (4)

[0086] ψ q =Li q (5)

[0087] Te=1.5pψ f i q (6)

[0088] Where L is the stator inductance, ψ f is the rotor flux, and p is the number of pole pairs.

[0089] Current prediction control can predict the control instructions of the next cycle in the current control cycle, thereby essentially solving the hysteresis of PI regulation and improving the bandwidth of the system. The given voltage vector is calculated by using the mathematical model of the system and the sampling value at the current moment, and the inverter outputs the voltage vector, so that the actual current can track the given current in the next sampling cycle.

[0090] DPCC (deadbeat predictive current control) uses the current equation to obtain the expected value of the voltage vector. Construct a mathematical model of deadbeat predictive current, such as Figure 1 As shown. Using the first-order Euler discretization, the current at the next moment (i dq (k+1), i qp (k+1), which can be expressed as:

[0091]

[0092] Where T is the sampling period, and k is the kth sampling moment / sampling point. Considering the one-beat delay compensation of the applied voltage, and assuming that ω e (k) and ω e (k+1) is the same. Therefore, in order to obtain the reference current i q (k+2) and i d (k+2), the expected voltage vector output by the deadbeat predictive control ( and ) can be expressed as:

[0093]

[0094]

[0095] In the DPCC control process, if the motor parameter values ​​are accurate, the actual current value at the k+1th moment is equal to the given current value at the kth moment. However, in the actual control process, there will be a phenomenon of motor parameter mismatch, so there will be a deviation between the actual current value at the k+1th moment and the given current value at the kth moment.

[0096] By analyzing the mismatch of the three parameters of the motor, resistance, inductance and flux linkage, we can get

[0097]

[0098] In the formula, e d 、e q They represent the deviation between the actual current value of the d and q axes and the current reference value under the disturbance of the internal parameters of the motor, ΔL s , ΔR s , Δψ fThey represent stator inductance error, stator resistance error, and rotor flux error respectively.

[0099] From the above formula, we can know that the rotor flux error Δψ f It will only affect the q-axis current output by the DPCC controller, but not the d-axis current, while the stator resistance value error and the stator inductance value error will affect both the d-axis and q-axis currents. Through the above analysis, it can be found that any parameter mismatch in the control system will cause a deviation in the current response, and the resulting current deviation and its rate of change will also be affected by variables such as angular velocity and current size.

[0100] In summary, the stator inductance error ΔL s The current deviation is significantly affected, and the stator resistance value R s The current deviation during mismatch can be ignored, and the rotor flux error Δψ f It does not affect the output d-axis current, but will cause a large q-axis current error. The traditional DPCC is significantly affected by the parameter mismatch of the inductance and flux parameters, so it is necessary to consider adding an observer to improve the applicability of the control model.

[0101] (2) Dead time and flux harmonic analysis

[0102] A typical SVPWM is an inverter based on a two-level topology, which contains six insulated gate bipolar transistors (IGBTs). Ideally, each phase has only two states for the upper and lower arms: one is that the upper arm is ON and the lower arm is OFF; the other is that the upper arm is OFF and the lower arm is ON. However, in order to avoid the on / off delay of the switching device causing the bridge arms to be turned on at the same time, a dead time is introduced to make the bridge arms turn on at the same time. A three-phase PWM voltage source inverter (VSI) based on IGBT (insulated gate bipolar transistor, Insulate-Gate Bipolar Transistor—IGBT) with PMSM load is defined as a positive current when the current flows from the inverter to the motor. The voltage applied to the motor is determined by the current direction during the dead time. The voltage error (Δu a , Δu b , Δu c ) can be expressed as:

[0103]

[0104] In the formula, sign is the switch function, i a 、i b 、i c is the phase current, dead zone voltage u dead Defined as:

[0105]

[0106] Among them, Tdead 、T on 、T off 、T pwm 、u dc They are dead time, on time, off time, PWM waveform period, and DC bus voltage. After Fourier series expansion, they can be transformed into dq axis voltage error:

[0107]

[0108]

[0109] Among them, γ is the angle difference between the q axis and the current vector, and k is a positive integer, which represents the sampling time / sampling point. They are composed of DC components, sixth harmonics and multiple harmonics. When the angular velocity changes, the amplitude of the harmonic component remains unchanged, but the frequency is different.

[0110] The manufacturing error and topological defects of the motor will cause the distortion of the flux, which will lead to the distortion of the back electromotive force. The harmonic of the A-phase flux (A) generated by the permanent magnet can be expressed as:

[0111]

[0112] Among them, ψ a, 2i -1 is the amplitude of the (2i-1)th harmonic.

[0113] Therefore, the entire flux harmonic can be converted into the flux harmonic of the dq axis using Park transformation:

[0114]

[0115]

[0116] Among them, ψ d,6k and ψ q,6k is the amplitude of the 6kth harmonic.

[0117] The voltage equation can be rewritten as:

[0118]

[0119]

[0120] in,

[0121]

[0122]

[0123] These components are also composed of the sixth harmonic and its multiple harmonics. However, their amplitude and frequency will change with the angular velocity, and they are related to ω e Directly proportional.

[0124] Through the above analysis, it can be found that the periodic errors caused by dead time and flux harmonics have a significant impact. Therefore, it is necessary to consider introducing ILC in the observer to compensate for these periodic disturbances in order to improve the accuracy and applicability of the control model.

[0125] To this end, the technical solution of the present invention provides the following iterative sliding mode observer:

[0126] The disturbance caused by parameter mismatch can only be guaranteed to be bounded, while the voltage error caused by dead time and the harmonic component caused by flux distortion are periodic disturbances. SMO can use the switching function to approximate bounded disturbances, and the iterative learning strategy can suppress periodic disturbances based on previously recorded errors. Therefore, the iterative learning strategy and SMO are combined to handle periodic and bounded disturbances. Figure 1 and Figure 2 As shown in Figure 1, ILC is added to the traditional SMO to estimate and compensate for periodic disturbances. d , q-axis current i q 、DPCC output d-axis voltage u d , q-axis voltage u q , after calculation, we get parameter mismatch compensation and Voltage compensation and Harmonic disturbance compensation and Then, the differentials of the estimated currents on the d-axis and q-axis are obtained by using equations (23) and (24), and the observed current value on the d-axis is obtained after integration: q-axis current observation value Output to DPCC module.

[0127] The iterative sliding mode observer proposed in the present invention can be constructed as:

[0128]

[0129]

[0130] in:

[0131]

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139] Where, the superscript ^ is the estimated value of the actual value, the superscript · is the differential of the parameter, and k 1d , k 2d , k 3d , k 1q , k 2q , k 3q , β 1 , β 2 , β 3 , β 4 is the positive coefficient of the observer; are the differences between the estimated current and the actual current after d-axis and q-axis observations, respectively;

[0140] and To compensate for parameter mismatch, and Compensation for flux harmonic disturbances and is the voltage error compensation caused by the dead time. In order to ensure the stability of the proposed ISMO, it is necessary to satisfy the unidirectional operation of the motor and the angular velocity is greater than a small positive constant, which can be expressed as ω e >ε. The designed ISMO can be used to compensate for the lumped disturbance and predict the next instantaneous current.

[0141] Based on the above analysis, the structure of enhanced DPCC can be expressed as:

[0142]

[0143]

[0144] In the formula, They represent the expected voltages of the d and q axes respectively.

[0145] Simulation Example

[0146] 1. Periodic disturbance test:

[0147] Dead zone effect and flux harmonics will lead to periodic disturbances. In order to verify the effectiveness of the proposed method, three control strategies, namely traditional DPCC, SMO-DPCC and ISMO proposed in the optimal embodiment of the present invention, are used to compare the output waveforms of dq-axis current feedback.

[0148] The parameters of the permanent magnet synchronous motor are shown in Table 1.

[0149] Tab.1Parameters of the PMSM

[0150]

[0151]

[0152] The simulation conditions are set as: PWM wave switching frequency f pwm =10 4 kHz, dead time is 2μs, DC bus voltage is U dc =50V, sampling period T s =10μs. The proposed ISMO parameters are: k in SMO 1d =k 1q =0.006,k 2d =k 2q =9000, k 3d =k 3q =2000, iterative learning rate β 1d =β 1q =0.00001,β 2d =β 2q = 0.01. A variable step solver is used, the relative tolerance is set to 2e-6s, and the running time is set to 0.4s. The PMSM is running at 600rpm (i.e. ω e =251rad / s) speed steady state operation. Comparison test results are as follows Figure 5-6 shown.

[0153] Since no disturbance suppression measures are taken, the feedback current of the conventional DPCC strategy has obvious current pulsation. Figure 5 As shown, the DPCC current i d The fluctuation amplitude is 2.3A. After adding SMO, the current i d The fluctuation amplitude is 1.8A. When the ISMO strategy is adopted, the iterative controller effectively suppresses the periodic disturbance, and the current fluctuation is reduced. d The fluctuation amplitude is 1.2A, which is 92% and 50% lower than the other two fluctuations respectively; similarly, Figure 6 As shown, the current i under the three control strategies qThe fluctuation amplitudes are 2.2A, 1.7A and 1.2A respectively. Compared with the other two, the fluctuation of ISMO is reduced by 83% and 42% respectively. It is easy to conclude that ISMO achieves more stable current tracking.

[0154] The A phase steady-state current is analyzed by Fast Fourier Transform (FFT). The results are as follows: Figure 7 As shown. The fundamental frequency is 50Hz (600r / min). The total harmonic distortion (THD) of phase A current of DPCC, SMO-DPCC and ISMO are 6.73%, 3.76% and 2.24% respectively. The FFT results show that harmonics are generally concentrated in the 5th and 7th orders. The harmonic content of the DPCC-ISMO method proposed in the present invention is significantly reduced, which is much smaller than the harmonics of the other three methods. This shows the effectiveness of the method proposed in the present invention in compensating for VSI nonlinearity. This method can suppress harmonics by 4.49% to 1.52%. In addition, this method can also significantly reduce the clamping phenomenon of phase current.

[0155] 2. Inductor mismatch:

[0156] First, the PMSM is driven at 600 rpm (ω e =251rad / s) when the speed is stable, set the inductance L s At the start, 0.4s and 0.7s, follow 0.5L s →L s →2L s Step change. Inductance L s The d-axis current under mismatch is as follows Figure 8 As shown in the figure, when the motor speed is stable, the current under DPCC control fluctuates greatly, especially when the set value is less than the reference value, the fluctuation amplitude reaches 4.8A. The current fluctuation amplitude under SMO-DPCC control is 3.1A, which is an improvement compared with DPCC. The current fluctuation amplitude under the proposed ISMO control is reduced to less than 2.2A, which is 118% and 41% lower than the current fluctuations of the previous two methods respectively. Similarly, the inductor L s The q-axis current results under mismatch also show that ISMO exhibits good robustness under inductance mismatch.

[0157] 3. Magnetic flux mismatch:

[0158] The PMSM is operated at 600 rpm (ω e =251rad / s) when the speed is stable, set the flux linkage ψ f At the beginning, 0.4s and 0.7s, thef → f →2ψ f Step change.

[0159] When the motor speed is stable, the current fluctuation under DPCC control is large, while the current fluctuation amplitude under the proposed ISMO control is the smallest. The d-axis currents of the three methods can follow the reference value and are not affected by the flux mismatch; the flux ψ f The q-axis current under mismatch is as follows Fig. 9 As shown in the figure, when the motor speed is stable, the current fluctuation under DPCC control is large, with an amplitude of 2.5A. The current fluctuation amplitude under SMO-DPCC control is 2.8A, and when the flux value increases, obvious jitter occurs. The current fluctuation amplitude under the proposed ISMO control is within 1.3A, which is still the best, and is reduced by 92% and 115% respectively compared with the previous two current fluctuations.

[0160] From the above simulation results, we can see that parameter mismatch will lead to non-periodic error of current, dead zone effect and flux distortion will lead to periodic error of current, and flux mismatch only affects i q Inductor mismatch will also affect i d and i q The current error will cause similar changes in the motor speed error, and the method proposed in the present invention suppresses this error through the learning process under all the above mismatch conditions. The simulation results mean that when the motor is running stably, the permanent magnet synchronous motor zero-beat predictive current control method based on iterative sliding mode disturbance observer can better control the stator current i d i q And motor speed are detected and controlled.

[0161] Embodiment 2:

[0162] An embodiment of the present invention provides a system based on the above control method, which at least includes: a position sensor, a current sensor, a Clark transformation module, a Park transformation module, an iterative sliding mode observer, a deadbeat current prediction module, an inverse Park transformation module, and a control module.

[0163] Among them, the position sensor is used to collect the electrical angular velocity and motor position angle of the permanent magnet synchronous motor; the current sensor is used to collect the three-phase current of the permanent magnet synchronous motor; the Clark transformation module is used to perform coordinate transformation of the three-phase current of the permanent magnet synchronous motor into a two-phase stationary coordinate system; the Park transformation module is used to perform coordinate transformation of the current of the two-phase stationary coordinate system into a two-phase rotating coordinate system; the iterative sliding mode observer is used to use the d and q axis currents and the currently obtained d and q axis expected voltages and electrical angular velocity to calculate and obtain the observed quantities, and the observed quantities include the d and q axis current observation values, voltage compensation quantities, harmonic disturbance compensation quantities and parameter mismatch compensation quantities; the deadbeat current prediction module is used to use the observed quantities obtained by the sliding mode observer to coordinate the d and q axis reference current inputs to calculate and obtain the new d and q axis expected voltages; the inverse Park transformation module is used to perform inverse Park transformation on the new d and q axis expected voltages; the control module is used to use the expected voltage after the inverse Park transformation to input the control inner loop of the permanent magnet motor to generate control information to control the operation of the PMSM module of the permanent magnet synchronous motor.

[0164] In some embodiments, the control module includes an SVPWM generation module and an inverter, wherein the SVPWM generation module uses the desired voltage after inverse Park transformation to generate six inverter switch pulse signals through modulation to drive the permanent magnet synchronous motor PMSM module to operate.

[0165] It should be understood that the specific implementation process of the above-mentioned unit modules refers to the method content, and the present invention will not be specifically repeated here, and the division of the above-mentioned functional module units is only a division of logical functions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. At the same time, the above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units.

[0166] Embodiment 3:

[0167] The present invention provides a computer-readable storage medium storing a computer program, which is called by a processor to implement: a permanent magnet motor predictive current control method based on an iterative sliding mode observer.

[0168] In some embodiments, the following process is specifically implemented:

[0169] Data acquisition: Collect the electrical angular velocity and motor position angle of the permanent magnet synchronous motor, as well as the three-phase current of the permanent magnet synchronous motor and perform coordinate conversion to obtain the d-axis and q-axis currents in the rotating coordinate system;

[0170] Iterative sliding mode observation: The d-axis and q-axis currents and the currently obtained d-axis and q-axis desired voltages and electrical angular velocity are input into the iterative sliding mode observer to obtain the observed quantities, which include the d-axis and q-axis current observation values, voltage compensation, harmonic disturbance compensation, and parameter mismatch compensation.

[0171] Deadbeat current prediction: The deadbeat current prediction module constructed by combining the observation quantity obtained by the sliding mode observer with the reference current input of the d and q axes obtains the new expected voltage of the d and q axes;

[0172] The new desired voltage of the d and q axes is input into the inverse Park transformation module in coordination with the motor position angle to obtain the voltage in the two-phase stationary coordinate system; then it is input into the SVPWM generation module, and after modulation, a six-way inverter switch pulse signal is generated to drive the permanent magnet synchronous motor PMSM module to operate.

[0173] For the specific implementation process of each step, please refer to the description of the above method.

[0174] The readable storage medium is a computer-readable storage medium, which may be an internal storage unit of the controller described in any of the foregoing embodiments, such as a hard disk or memory of the controller. The readable storage medium may also be an external storage device of the controller, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the controller. Further, the readable storage medium may also include both an internal storage unit of the controller and an external storage device. The readable storage medium is used to store the computer program and other programs and data required by the controller. The readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0175] Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned readable storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0176] It should be emphasized that the examples described in the present invention are illustrative rather than restrictive, and therefore the present invention is not limited to the examples described in the specific embodiments. Any other embodiments derived by those skilled in the art based on the technical solution of the present invention that do not depart from the purpose and scope of the present invention, whether modified or replaced, also fall within the scope of protection of the present invention.

Claims

1. A method for predictive current control of a permanent magnet motor based on an iterative sliding mode observer, characterized in that: The following steps are involved: Data acquisition: Collect the electrical angular velocity and motor position angle of the permanent magnet synchronous motor, as well as the three-phase current of the permanent magnet synchronous motor and perform coordinate conversion to obtain the d-axis and q-axis currents in the rotating coordinate system; Iterative sliding mode observation: the d-axis and q-axis currents and the currently obtained d-axis and q-axis desired voltages and electrical angular velocities are input into an iterative sliding mode observer to obtain observation quantities, wherein the observation quantities include d-axis and q-axis current observation values, voltage compensation quantities, harmonic disturbance compensation quantities, and parameter mismatch compensation quantities; The iterative sliding mode observer adds an ILC module based on iterative learning control to the sliding mode observer SMO, and the ILC module is provided with flux linkage harmonic disturbance compensation and voltage error compensation for dead time; Deadbeat current prediction: The deadbeat current prediction module constructed by combining the observed quantity obtained by the sliding mode observer with the reference current input of the d and q axes obtains the new expected voltage of the d and q axes; Then, the new desired voltages of the d and q axes are subjected to inverse Park transformation and input into the control inner loop of the permanent magnet motor to generate control information to control the operation of the PMSM module of the permanent magnet synchronous motor.

2. The method according to claim 1, characterized in that: The mathematical model of the flux harmonic disturbance compensation is expressed as: In the formula, and is the flux harmonic disturbance compensation of the d and q axes in the current sampling period, are the compensation of magnetic flux harmonic disturbance of d and q axes in the previous sampling period, T is the sampling period, β1, k 1d is the positive coefficient set for the d-axis on the iterative sliding mode observer, β3, k 1q It is the positive coefficient set for the q-axis on the iterative sliding mode observer.

3. The method according to claim 1, characterized in that: The mathematical model of the voltage error compensation is expressed as: In the formula, and They are the voltage error compensation of the d and q axes of the current sampling period, and are the voltage error compensation of the d and q axes of the previous sampling period, T is the sampling period, β2, k 1d are all positive coefficients set for the d-axis on the iterative sliding mode observer, β4, k 1q They are all positive coefficients set for the q-axis on the iterative sliding mode observer.

4. The method according to claim 1, characterized in that: The observation formula of the iterative sliding mode observer is expressed as: in: In the formula, are the current observation values ​​of the d and q axes respectively The differential, R s is the stator resistance, L is the stator inductance, ω e is the electrical angular velocity, and ω e >ε, a positive constant set by ε, i d 、i q are the actual currents of the d and q axes, u d 、u q are the stator voltages of the d and q axes respectively, and the actual input is the expected voltage at the previous moment. are the parameter mismatch compensation of d and q axes respectively, They are parameter mismatch compensation The differential of They are the voltage error compensation of d and q axes respectively. are the compensation of magnetic flux harmonic disturbances on the d and q axes, ν d 、v q are the sliding mode control functions corresponding to the d and q axes on the iterative sliding mode observer; k 1d , k 2d , k 3d are all positive coefficients set for the d-axis on the iterative sliding mode observer; k 1q , k 2q , k 3q They are all positive coefficients set for the q-axis on the iterative sliding mode observer; They represent the difference between the estimated current and the actual current after d-axis and q-axis observation respectively.

5. The method according to claim 1, characterized in that: The deadbeat current prediction module obtains the expected voltage of the d and q axes as follows: In the formula, They represent the expected voltages of the d and q axes, respectively, L is the stator inductance, T is the sampling period, and R s is the stator resistance, k is the kth sampling moment, They represent the reference currents of the d and q axes at the k+2 sampling time, respectively. They represent the current observation values ​​of the d and q axes at the k+1 sampling time, respectively. are the parameter mismatch compensation amounts of the d and q axes at k sampling moments, respectively. are the voltage error compensation of d and q axes at k sampling moments, are the flux harmonic disturbance compensation of d and q axes at k sampling moments, ω e (k) represents the electrical angular velocity at time k, ψ f is the rotor flux.

6. The method according to claim 1, characterized in that: The process of performing inverse Park transformation on the desired voltages of the d and q axes and inputting them into the control inner loop of the permanent magnet motor to generate control information to control the operation of the PMSM module of the permanent magnet synchronous motor is as follows: The desired voltages of the d and q axes are input into the inverse Park transformation module in coordination with the motor position angle to obtain the voltages in the two-phase stationary coordinate system; Then it is input into the SVPWM generation module, and after modulation, it generates six inverter switch pulse signals to drive the permanent magnet synchronous motor PMSM module to operate.

7. The method according to claim 1, characterized in that: The voltage error derived from the dead time is expressed on the d and q axes as: In the formula, Δu d , Δu q Respectively represent the voltage error derived from the dead time on the d and q axes, u dead is the dead zone voltage, γ is the angle difference between the q axis and the current vector, k is a positive integer, k represents the sampling time or sampling point, ω e represents electrical angular velocity, and t represents time.

8. The method according to claim 1, characterized in that: The flux harmonic disturbance compensation is set for the periodic disturbance caused by flux harmonics. The flux harmonics are expressed in the d and q axes as: In the formula, r d 、r q is the flux harmonic on the d and q axes, k represents the sampling time / sampling point, ψ d，6k and ψ q,6k is the amplitude of the 6kth harmonic, θ e is the motor position angle.

9. A system based on the method according to any one of claims 1 to 8, characterized in that: At least: Position sensor, used to collect the electrical angular velocity and motor position angle of the permanent magnet synchronous motor; Current sensor, used to collect the three-phase current of the permanent magnet synchronous motor; Clark transformation module, used to transform the three-phase current of permanent magnet synchronous motor into a two-phase stationary coordinate system; Park transformation module, used to transform the current of the two-phase stationary coordinate system into the two-phase rotating coordinate system; Iterative sliding mode observer: used to calculate and obtain observation quantities using the d-axis and q-axis currents and the currently obtained d-axis and q-axis desired voltages and electrical angular velocities, wherein the observation quantities include d-axis and q-axis current observation values, voltage compensation quantities, harmonic disturbance compensation quantities, and parameter mismatch compensation quantities; The iterative sliding mode observer is an ILC module based on iterative learning control added on the basis of the sliding mode observer SMO, and the ILC module is provided with flux linkage harmonic disturbance compensation and voltage error compensation for dead time; A deadbeat current prediction module, used to calculate and obtain new expected voltages of the d and q axes by using the observed quantity obtained by the sliding mode observer in coordination with the reference current inputs of the d and q axes; An inverse Park transformation module, used for performing an inverse Park transformation on the new desired voltages of the d and q axes; The control module is used to utilize the desired voltage after the inverse Park transformation to input the control inner loop of the permanent magnet motor to generate control information to control the operation of the PMSM module of the permanent magnet synchronous motor.

10. A computer-readable storage medium, characterized in that: A computer program is stored, which is called by a processor to implement: The method according to any one of claims 1 to 8.

Images