Permanent magnet synchronous motor robust finite set model predictive flux linkage control method

Abstract

The application discloses a robust finite set model predictive flux linkage control method of a permanent magnet synchronous motor, and belongs to the technical field of permanent magnet synchronous motor control. The method uses a generalized proportional integral observer model to identify q-axis inductance and permanent magnet flux linkage. In view of the underdetermined problem in identification, the d-axis inductance is identified by using a d-axis current difference equation. According to the inverter voltage vector output characteristics of FCS-MPFC, the current is sampled twice in one sampling period, so as to eliminate the dead zone in the identification equation. The generalized proportional integral observer is used to improve the flux linkage prediction model, and the robustness is further improved. The application can quickly and accurately identify the dq-axis inductance and permanent magnet flux linkage, and can eliminate the influence of the inverter dead zone on parameter identification and prediction control. The application can realize high-precision control of torque, can effectively improve the parameter robustness of the system, and can improve the operation quality of the control system.

Description

Technical Field

[0001] This invention relates to the field of permanent magnet synchronous motor control, and specifically to a robust finite set model predictive flux linkage control method for permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in high-performance drive applications such as new energy vehicles and industrial servo systems due to their high efficiency, high power density, and high starting torque. Finite set model predictive control (FCS-MPC) has become a research hotspot in recent years due to its advantages such as simple concept, wide applicability, ease of considering system nonlinear constraints, and no need for controller design. Finite set model predictive flux control (FCS-MPFC), developed from finite set model predictive torque control (FCS-MPTC), equates the control of torque and flux linkage amplitude to the control of the stator flux linkage vector under the dq coordinate axis, eliminating the weighting coefficients in the cost function of FCS-MPTC. However, traditional FCS-MPFC suffers from strong parameter dependence in flux linkage calculation, affecting the actual control performance. Therefore, improving the parameter robustness of FCS-MPFC has gradually become a research hotspot in academia.

[0003] In terms of parameter robustness, the traditional FCS-MPFC mainly involves flux linkage instruction calculation, flux linkage observation, and flux linkage prediction. If the parameters can be identified online, the computational accuracy of each step can be obtained, thereby improving the parameter robustness of the system.

[0004] Currently, commonly used parameter identification methods include recursive least squares, model reference adaptive (MRAS), extended Kalman filter, and artificial intelligence algorithms. However, these methods all encounter the problem of underranked identification equations when performing multi-parameter identification, leading to non-unique identification results and getting trapped in local optima.

[0005] Reference 1, "Multi-parameter identification of permanent magnet synchronous motor based on improved antlion optimization algorithm" [Journal of Electric Power System and Automation 2023, 35(06): 66-72], proposes an antlion optimization algorithm based on an adaptive normal cloud model for parameter identification of permanent magnet synchronous motors, which can solve the underrank problem and has high accuracy. However, this algorithm requires a large number of iterations, resulting in a slow convergence speed. In addition, the accuracy of different parameter identifications is affected by the fitness function and the corresponding weight coefficients. Therefore, the identification accuracy of different parameters will restrict each other, and the tuning of the weight coefficients is difficult.

[0006] Reference 2, “A Robust DPCC for IPMSM Based on a FullParameter Identification Method” [IEEE Transactions on Industrial Electronics 2023, 70(8): 7695-7705], combines the deadbeat predictive current control method with the high-frequency current injection method, which allows for online observation of the inductance, permanent magnet flux linkage, and resistance values ​​of a permanent magnet synchronous motor. However, the injected disturbance signal will cause pulsation in the motor torque. In addition, the injection of the disturbance signal requires a modulation stage, so this type of method is not suitable for FCS-MPFC.

[0007] Furthermore, the inverter dead time causes a deviation between the stator voltage value calculated from the switching state and the actual value, which in turn affects the accuracy of parameter identification and the control effect of FCS-MPFC, reducing the robustness of the control system.

[0008] Reference 3, "Robust Current Control Strategy for Five-Phase Permanent Magnet Synchronous Motor Based on Enhanced Disturbance Observer" [Journal of Electrical Engineering Technology, 2020, 35(06): 1219-1230], uses the dead-zone voltage as a disturbance term and establishes an enhanced disturbance observer through the mathematical model of the motor to compensate for the total disturbance, including the dead-zone voltage. However, the compensation effect of this method depends on the accuracy of the parameters and is not suitable for parameter identification. Furthermore, the feedback of the compensation amount requires a modulation stage. Therefore, this type of method is not suitable for FCS-MPFC.

[0009] In summary, the existing technology has the following problems:

[0010] 1. There is no modulation stage in FCS-MPFC, while existing methods to improve system robustness all require the presence of a modulation stage in the control system, which is not suitable for FCS-MPFC.

[0011] 2. Currently, there is a lack of a method and system suitable for FCS-MPFC that can construct full-rank identification equations and achieve online identification of multiple parameters without injecting perturbation current.

[0012] 3. Currently, there is no proposed dead zone compensation method that is applicable to both multi-parameter online identification and FCS-MPFC. Summary of the Invention

[0013] The technical problem to be solved by this invention is to realize a multi-parameter identification method applicable to FCS-MPFC, solve the underrank problem of identification equation in multi-parameter identification, and eliminate the influence of inverter dead-zone voltage on multi-parameter identification and FCS-MPFC, thereby realizing a robust FCS-MPFC.

[0014] The objective of this invention is achieved by providing a robust finite set model predictive flux linkage control method for permanent magnet synchronous motors, comprising the following steps:

[0015] Step 1, set the control cycle duration to T. s Let the current control period be k; at the beginning of the current control period k, the rotor electric angular velocity of the permanent magnet synchronous motor is acquired. and rotor electrical angle At time τ after the start of the current control period k, at time (T) s At time -τ), the stator current is sampled twice, and the sampled values ​​are recorded as the first stator current sample value i for the current control cycle. a1 i b1 i c1 and the current control cycle secondary stator current sampling value i a2 i b2 i c2 Then, after transforming from the stationary coordinate system to the rotating dq coordinate system, the dq-axis component of the stator current sample value for the current control cycle is obtained. and the dq-axis component of the secondary stator current sample value in the current control cycle

[0016] Step 2, record the three-phase switch state of the inverter as switch state S. a S b S c Record the inverter three-phase switch state in the current control cycle k as the current control cycle switch state. The voltage on the DC side of the inverter is denoted as voltage U. dc Based on the current control cycle switching state and voltage U dc The dq-axis component of the stator voltage of the permanent magnet synchronous motor during the current control cycle is calculated. The AC side of the inverter is connected to the stator of the permanent magnet synchronous motor;

[0017] Step 3: Construct a generalized proportional-integral observer model and calculate the dq-axis components of the stator current observation value for the current control cycle. and the dq-axis components of the stator current disturbance observation during the current control cycle

[0018] Step 4: Given the boundary conditions for q-axis inductance identification and permanent magnet flux linkage identification, obtain the filter coefficient h for q-axis inductance identification. Lq The filter coefficient h for permanent magnet flux identification ψf Given the boundary conditions for d-axis inductance identification, obtain the filter coefficient h for d-axis inductance identification. Ld ;

[0019] Step 5: Based on the d-axis component of the stator current disturbance observation value during the current control cycle. The filter coefficient h identified by the q-axis inductance Lq The q-axis inductance is identified using the first identification equation, and the q-axis component of the inductance deviation identification value for the current control cycle is obtained. and the q-axis component of the inductance identification value in the current control cycle Based on the current control cycle inductance identification value q-axis component The filter coefficient h identified by the d-axis inductance Ld The d-axis inductance is identified using the d-axis current difference equation, and the d-axis component of the inductance identification value for the current control cycle is obtained. Based on the q-axis component of the stator current disturbance observation during the current control cycle The filter coefficient h for permanent magnet flux identification ψf The permanent magnet flux linkage is identified using the second identification equation to obtain the permanent magnet flux linkage deviation identification value for the current control cycle. and the permanent magnet flux identification value of the current control cycle

[0020] Step 6: Identify the dq-axis component of the inductance value in the current control cycle. and the permanent magnet flux identification value of the current control cycle The current control cycle flux reference value dq axis component is calculated. and the dq-axis components of the flux linkage observation during the current control period

[0021] Step 7, based on the current control cycle inductance identification value dq axis component and the dq-axis components of the flux linkage observation during the current control period The predicted flux linkage component dq-axis of the next control cycle is calculated. And the dq-axis components of the flux linkage prediction disturbance observations during the current control period

[0022] Step 8, based on the inverter three-phase switch status S a S b S c This yields seven voltage vectors, denoted as voltage vector u. j j is the index of the voltage vector, j = 0, 1, ... 6;

[0023] The calculation yields the voltage vector u j The dq-axis component of the flux linkage prediction value in the (k+2)th control cycle under the action is denoted as the dq-axis component of the target flux linkage prediction value.

[0024] Step 9, convert the dq axis components of the target flux linkage prediction value. Substituting the value function, the voltage vector u is calculated. j The corresponding value function value g j The expression for the value function is as follows:

[0025]

[0026] Among the seven calculated value function values, the value function value with the smallest value is selected and recorded as the minimum value function value. The voltage vector corresponding to the minimum value function value is selected as the inverter output voltage vector for the next control cycle to control the switching state of the inverter switching transistor.

[0027] Preferably, the inverter three-phase switch state S in step 2 is... a S b S c The specific actions are as follows:

[0028] S a =1 indicates that the inverter's A-phase bridge arm switch S a1 On, switch S a2 Turn off;

[0029] S a =0 indicates that the inverter A-phase bridge arm switch S a1 Turn off, switch S a2 Conduction;

[0030] S b =1 indicates that the inverter's B-phase bridge arm switch S b1 On, switch S b2 Turn off;

[0031] S b =0 indicates that the inverter's B-phase bridge arm switch S b1 Turn off, switch S b2 Conduction;

[0032] S c =1 indicates that the inverter's C-phase bridge arm switch S c1 On, switch S c2 Turn off;

[0033] S c =0 indicates that the inverter's C-phase bridge arm switch S c1 Turn off, switch S c2 Conduction;

[0034] Step 2 describes the current control cycle's permanent magnet synchronous motor stator voltage periodic average value dq-axis component. The calculation formula is as follows:

[0035]

[0036] Preferably, the expression for the generalized proportional-integral observer model described in step 3 is:

[0037]

[0038] in, The d-axis component of the stator current observation value from the previous control cycle. This represents the q-axis component of the stator current observation from the previous control cycle. This represents the d-axis component of the secondary stator current sample value from the previous control cycle. T represents the q-axis component of the secondary stator current sample value from the previous control cycle. c T is the time interval between two stator current samplings. c =T s -2τ,R s For stator resistance, The d-axis component of the stator inductance identification value from the previous control cycle. The q-axis component of the stator inductance identification value from the previous control cycle. This is the nominal value of the q-axis stator inductance. γ is the nominal value of the permanent magnet flux linkage, γ is the observer proportional parameter, and κ is the observer integral parameter. The d-axis component of the stator current disturbance observation value from the previous control cycle. The q-axis component of the stator current disturbance observation value from the previous control cycle.

[0039] Preferably, the boundary conditions for q-axis inductance identification and permanent magnet flux identification in step 4 are as follows:

[0040]

[0041] Among them, I min ω is the boundary value of the stator current amplitude. min ω is the boundary value of electric angular velocity.c For the bandwidth of the low-pass filter in q-axis inductance identification and permanent magnet flux identification, T c T is the time interval between two stator current samplings. c =T s -2τ;

[0042] The boundary conditions for d-axis inductance identification in step 4 are as follows:

[0043]

[0044] In the formula, η is the scaling factor for d-axis inductance identification, η∈(0,1), ω Ld Identify the bandwidth of the low-pass filter for the d-axis inductor. The d-axis component of the stator inductance identification value from the previous control cycle.

[0045] Preferably, the first identification equation in step 5 is as follows:

[0046]

[0047] in, The q-axis component of the inductance deviation identification value from the previous control cycle. This is the nominal value of the q-axis stator inductance;

[0048] The d-axis current difference equation described in step 5 is as follows:

[0049]

[0050] in, R is the d-axis component of the inductance identification value from the previous control cycle. s T is the stator resistance. c T is the time interval between two stator current samplings. c =T s -2τ;

[0051] Step 5: The second identification equation is as follows:

[0052]

[0053] in, This is the permanent magnet flux deviation identification value from the previous control cycle. This is the nominal value of the permanent magnet flux linkage.

[0054] Preferably, the current control cycle flux reference value dq-axis component in step 6 The calculation formula is as follows:

[0055]

[0056] in, The q-axis component of the flux linkage reference value from the previous control cycle. The given value for motor torque in the current control cycle, where p is the number of motor pole pairs;

[0057] Step 6 describes the dq-axis component of the current control cycle flux linkage observation. The calculation formula is as follows:

[0058]

[0059] Preferably, the dq-axis component of the flux linkage prediction value in the next control cycle described in step 7... The calculation formula is as follows:

[0060]

[0061] in, The d-axis component of the flux linkage prediction value for the current control cycle. For the current control cycle, the predicted flux linkage value along the q-axis is γ. ψ R is the scaling parameter for the prediction model. s Stator resistance;

[0062] Step 7 describes the dq-axis component of the current control cycle flux prediction disturbance observation. The calculation formula is as follows:

[0063]

[0064] in, The d-axis component of the flux linkage prediction disturbance observation from the previous control cycle. For the q-axis component of the flux linkage prediction disturbance observation from the previous control cycle, κ ψ These are the integral parameters for the prediction model.

[0065] Preferably, the voltage vector u in step 8 j The corresponding inverter three-phase switch state S a S b S c The specific status is as follows:

[0066] The switching state corresponding to voltage vector u0 is S a =0, S b =0, S c =0;

[0067] The switching state corresponding to voltage vector u1 is S a =1,S b =0, S c =0;

[0068] The switching state corresponding to voltage vector u2 is Sa =1,S b =1,S c =0;

[0069] The switching state corresponding to voltage vector u3 is S. a =0, S b =1,S c =0;

[0070] The switching state corresponding to voltage vector u4 is S a =0, S b =1,S c =1;

[0071] The switching state corresponding to voltage vector u5 is S. a =0, S b =0, S c =1;

[0072] The switching state corresponding to voltage vector u6 is S. a =1,S b =0, S c =1;

[0073] Step 8 describes the dq-axis component of the target flux linkage prediction value. The calculation formula is as follows:

[0074]

[0075] Among them, u dj Voltage vector u j The d-axis component, u qj Voltage vector u j The q-axis component, R s This is the stator resistance.

[0076] Compared with the prior art, the beneficial effects of the present invention are reflected in:

[0077] 1. Compared with traditional FCS-MPFC control, this invention can solve the shortcomings of FCS-MPFC in terms of strong parameter dependence, eliminate the influence of motor parameter deviation and inverter dead zone on control, thereby eliminating control error and achieving high-precision motor torque control, and can improve the dynamic performance of torque control and reduce motor stator current harmonics.

[0078] 2. Compared with the parameter identification method of the antlion optimization algorithm based on the adaptive normal cloud model in Reference 1, the identification accuracy of each parameter in this invention will not affect each other, and the convergence speed is faster.

[0079] 3. Compared with the high-frequency current signal injection for multi-parameter identification in Reference 2, the present invention does not require the injection of disturbance signal during identification, and will not cause pulsation of motor torque.

[0080] 4. Compared with existing inverter dead-time compensation schemes, the dead-time compensation method proposed in this paper can simultaneously solve the inverter dead-time problem existing in multi-parameter identification and FCS-MPFC. Furthermore, compared with the dead-time compensation scheme based on an enhanced disturbance observer in Reference 3, the dead-time compensation method in this invention does not require the construction of a disturbance observer, making it simpler to implement, and can eliminate the effects of inverter dead-time without requiring motor parameters and modulation stages. Attached Figure Description

[0081] Figure 1 This is a control block diagram of the robust finite set model predictive flux control method for permanent magnet synchronous motors in this invention.

[0082] Figure 2 This refers to the inverter topology involved in this invention.

[0083] Figure 3 This is a schematic diagram of the inverter voltage vector of the present invention.

[0084] Figure 4 The waveform diagram of d-axis inductance identification proposed in this invention is shown when the motor inductance and permanent magnet flux have a deviation of -30% or +30% at the same time.

[0085] Figure 5 The waveform diagram for q-axis inductance identification proposed in this invention is shown when both the motor inductance and permanent magnet flux have a deviation of -30% or +30%.

[0086] Figure 6 The permanent magnet flux linkage identification waveform of the method proposed in this invention is shown when both the motor inductance and the permanent magnet flux linkage have a deviation of -30% or +30%.

[0087] Figure 7 The motor torque waveform diagram of the method proposed in this invention is shown when the motor inductance and permanent magnet flux have a deviation of -30% or +30% at the same time.

[0088] Figure 8 The torque waveform of a traditional FCS-MPFC motor when both the motor inductance and permanent magnet flux have a deviation of -30% or +30%. Detailed Implementation

[0089] The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0090] Figure 1This is a control block diagram of the robust finite set model predictive flux control method for permanent magnet synchronous motors in this invention. Figure 1 As can be seen, the present invention includes the following steps:

[0091] Step 1, set the control cycle duration to T. s Let the current control period be k; at the beginning of the current control period k, the rotor electric angular velocity of the permanent magnet synchronous motor is acquired. and rotor electrical angle At time τ after the start of the current control period k, at time (T) s At time -τ), the stator current is sampled twice, and the sampled values ​​are recorded as the first stator current sample value i for the current control cycle. a1 i b1 i c1 and the current control cycle secondary stator current sampling value i a2 i b2 i c2 Then, after transforming from the stationary coordinate system to the rotating dq coordinate system, the dq-axis component of the stator current sample value for the current control cycle is obtained. and the dq-axis component of the secondary stator current sample value in the current control cycle

[0092] In this embodiment, T s =50μs, τ=8μs.

[0093] In this embodiment, the dq-axis component of the stator current sample value in the current control cycle and the dq-axis component of the secondary stator current sample value in the current control cycle The coordinate transformation formula is as follows:

[0094]

[0095]

[0096] in, This represents the electrical angle of the motor rotor during the first current sampling of the current cycle. The motor rotor electrical angle during the second current sampling of the current cycle is calculated as follows:

[0097]

[0098]

[0099] Step 2, record the three-phase switch state of the inverter as switch state S. a S b S cRecord the inverter three-phase switch state in the current control cycle k as the current control cycle switch state. The voltage on the DC side of the inverter is denoted as voltage U. dc Based on the current control cycle switching state and voltage U dc The dq-axis component of the stator voltage of the permanent magnet synchronous motor during the current control cycle is calculated. The AC side of the inverter is connected to the stator of the permanent magnet synchronous motor.

[0100] The current control cycle permanent magnet synchronous motor stator voltage periodic average value dq axis component The calculation formula is as follows:

[0101]

[0102] Depend on Figure 1 As can be seen, an inverter is connected to the stator of the permanent magnet synchronous motor, and the DC side of the inverter is connected in parallel with a DC voltage source.

[0103] Figure 2 The inverter topology involved in this invention is composed of... Figure 2 As can be seen, the inverter has three phase arms, and each phase arm contains two switching transistors.

[0104] Step 3: Construct a generalized proportional-integral observer model and calculate the dq-axis components of the stator current observation value for the current control cycle. and the dq-axis components of the stator current disturbance observation during the current control cycle

[0105] The expression for the generalized proportional-integral observer model is:

[0106]

[0107] in, The d-axis component of the stator current observation value from the previous control cycle. This represents the q-axis component of the stator current observation from the previous control cycle. This represents the d-axis component of the secondary stator current sample value from the previous control cycle. T represents the q-axis component of the secondary stator current sample value from the previous control cycle. c T is the time interval between two stator current samplings. c =T s -2τ,R s For stator resistance, The d-axis component of the stator inductance identification value from the previous control cycle. The q-axis component of the stator inductance identification value from the previous control cycle. This is the nominal value of the q-axis stator inductance. γ is the nominal value of the permanent magnet flux linkage, γ is the observer proportional parameter, and κ is the observer integral parameter. The d-axis component of the stator current disturbance observation value from the previous control cycle. The q-axis component of the stator current disturbance observation value from the previous control cycle.

[0108] Step 4: Given the boundary conditions for q-axis inductance identification and permanent magnet flux linkage identification, obtain the filter coefficient h for q-axis inductance identification. Lq The filter coefficient h for permanent magnet flux identification ψf Given the boundary conditions for d-axis inductance identification, obtain the filter coefficient h for d-axis inductance identification. Ld .

[0109] The boundary conditions for q-axis inductance identification and permanent magnet flux linkage identification are as follows:

[0110]

[0111] Among them, I min ω is the boundary value of the stator current amplitude. min ω is the boundary value of electric angular velocity. c For the bandwidth of the low-pass filter in q-axis inductance identification and permanent magnet flux identification, T c T is the time interval between two stator current samplings. c =T s -2τ;

[0112] The boundary conditions for d-axis inductance identification are as follows:

[0113]

[0114] In the formula, η is the scaling factor for d-axis inductance identification, η∈(0,1), ω Ld Identify the bandwidth of the low-pass filter for the d-axis inductor. The d-axis component of the stator inductance identification value from the previous control cycle.

[0115] Step 5: Based on the d-axis component of the stator current disturbance observation value during the current control cycle. The filter coefficient h identified by the q-axis inductance Lq The q-axis inductance is identified using the first identification equation, and the q-axis component of the inductance deviation identification value for the current control cycle is obtained. and the q-axis component of the inductance identification value in the current control cycle Based on the current control cycle inductance identification value q-axis component The filter coefficient h identified by the d-axis inductance Ld The d-axis inductance is identified using the d-axis current difference equation, and the d-axis component of the inductance identification value for the current control cycle is obtained. Based on the q-axis component of the stator current disturbance observation during the current control cycle The filter coefficient h for permanent magnet flux identification ψf The permanent magnet flux linkage is identified using the second identification equation to obtain the permanent magnet flux linkage deviation identification value for the current control cycle. and the permanent magnet flux identification value of the current control cycle

[0116] The first identification equation is as follows:

[0117]

[0118] in, The q-axis component of the inductance deviation identification value from the previous control cycle. This is the nominal value of the q-axis stator inductance;

[0119] The d-axis current difference equation is as follows:

[0120]

[0121] in, T is the d-axis component of the inductance identification value from the previous control cycle. c T is the time interval between two stator current samplings. c =T s -2τ;

[0122] The second identification equation is as follows:

[0123]

[0124] in, This is the permanent magnet flux deviation identification value from the previous control cycle. This is the nominal value of the permanent magnet flux linkage.

[0125] Step 6: Identify the dq-axis component of the inductance value in the current control cycle. and the permanent magnet flux identification value of the current control cycle The current control cycle flux reference value dq axis component is calculated. and the dq-axis components of the flux linkage observation during the current control period

[0126] The current control cycle flux reference value dq axis component The calculation formula is as follows:

[0127]

[0128] in, The q-axis component of the flux linkage reference value from the previous control cycle. The given value for motor torque in the current control cycle, where p is the number of motor pole pairs;

[0129] The dq-axis component of the current control cycle flux observation value The calculation formula is as follows:

[0130]

[0131] Step 7, based on the current control cycle inductance identification value dq axis component and the dq-axis components of the flux linkage observation during the current control period The predicted flux linkage component dq-axis of the next control cycle is calculated. And the dq-axis components of the flux linkage prediction disturbance observations during the current control period

[0132] The dq-axis component of the flux linkage prediction value in the next control cycle The calculation formula is as follows:

[0133]

[0134] in, The d-axis component of the flux linkage prediction value for the current control cycle. For the current control cycle, the predicted flux linkage value along the q-axis is γ. ψ R is the scaling parameter for the prediction model. s Stator resistance;

[0135] The dq-axis component of the current control cycle flux prediction disturbance observation value The calculation formula is as follows:

[0136]

[0137] in, The d-axis component of the flux linkage prediction disturbance observation from the previous control cycle. For the q-axis component of the flux linkage prediction disturbance observation from the previous control cycle, κ ψ These are the integral parameters for the prediction model.

[0138] Step 8, based on the inverter three-phase switch status S a S b S c This yields seven voltage vectors, denoted as voltage vector u. j j is the index of the voltage vector, j = 0, 1, ... 6;

[0139] The calculation yields the voltage vector u j The dq-axis component of the flux linkage prediction value in the (k+2)th control cycle under the action is denoted as the dq-axis component of the target flux linkage prediction value.

[0140] Figure 3 This is a schematic diagram of the inverter voltage vector of the present invention. Figure 3 As can be seen, the seven voltage vectors are evenly distributed, forming six identical sectors.

[0141] The target flux linkage prediction value dq axis component The calculation formula is as follows:

[0142]

[0143] Among them, u dj Voltage vector u j The d-axis component, u qj Voltage vector u j The q-axis component, R s This is the stator resistance.

[0144] Step 9, convert the dq axis components of the target flux linkage prediction value. Substituting the value function, the voltage vector u is calculated. j The corresponding value function value g j The expression for the value function is as follows:

[0145]

[0146] Among the seven calculated value function values, the value function value with the smallest value is selected and recorded as the minimum value function value. The voltage vector corresponding to the minimum value function value is selected as the inverter output voltage vector for the next control cycle to control the switching state of the inverter switching transistor.

[0147] To verify the effectiveness of this invention, experimental verification was conducted. Control system experimental parameters: motor rated power P N =1.5kW, rated voltage U N =220V, rated current i N =6.2A, rated torque T N =9Nm, rated speed n N =1500r / min, number of pole pairs p=5, stator resistance R s = 0.937Ω. Controller sampling frequency f s =20kHz.

[0148] Figure 4The waveform diagram shows the d-axis inductance identification of the method proposed in this invention when both the motor inductance and permanent magnet flux have a deviation of -30% or +30%. (The motor is running at 1000 r / min, with an initial torque of 0 Nm, and a torque step input of 5 Nm at a running time t of 2 s. The solid line waveform represents the d-axis inductance identification waveform when both the motor inductance and permanent magnet flux have a deviation of +30%, and the dashed line waveform represents the d-axis inductance identification waveform when both the motor inductance and permanent magnet flux have a deviation of -30%.)

[0149] Figure 5 The waveform diagram shows the q-axis inductance identification using the method proposed in this invention when both the motor inductance and permanent magnet flux linkage have a deviation of -30% or +30%. (The motor is running at 1000 r / min, with an initial torque of 0 Nm, and a torque step input of 5 Nm at a running time t of 2 s. The solid line waveform represents the q-axis inductance identification waveform when both the motor inductance and permanent magnet flux linkage have a deviation of +30%, and the dashed line waveform represents the q-axis inductance identification waveform when both the motor inductance and permanent magnet flux linkage have a deviation of -30%.)

[0150] Figure 6 The waveform diagrams for permanent magnet flux identification using the method proposed in this invention are shown when both motor inductance and permanent magnet flux have a deviation of -30% or +30%. (The motor is running at 1000 r / min, with an initial torque of 0 Nm, and a torque step input of 5 Nm at a running time t of 2 s. The solid line waveform represents the permanent magnet flux identification waveform when both motor inductance and permanent magnet flux have a deviation of +30%, and the dashed line waveform represents the permanent magnet flux identification waveform when both motor inductance and permanent magnet flux have a deviation of -30%.)

[0151] exist Figure 4 , Figure 5 and Figure 6 In the middle, the current is zero before 2s, which does not meet the parameter identification condition. At this time, L d The initial values ​​are 4.585 mH and 8.515 mH in the two cases, respectively. q The values ​​are 7.455 mH and 13.845 mH, respectively, ψ fThe values ​​are 0.161 Wb and 0.299 Wb, respectively. With a torque step input of 5 Nm at a running time t of 2 seconds, parameter identification begins. The steady-state average values ​​of the identification results from 5 to 10 seconds for both cases are: d-axis inductance 6.557 mH and 6.565 mH, respectively; q-axis inductance 10.648 mH and 10.662 mH, respectively; and permanent magnet flux linkage 0.2312 Wb and 0.2305 Wb, respectively. The deviations in the identification results for the two cases are: 0.12% for d-axis inductance, 0.13% for q-axis inductance, and 0.30% for permanent magnet flux linkage. It can be seen that the identification results are consistent and show no significant deviation between the two cases. With a 30% deviation in dynamic performance, the rise times of the d-axis inductance are 1.10s and 1.20s, the q-axis inductance are 0.40s and 0.58s, and the permanent magnet flux linkage is 0.27s and 0.25s, which can meet the actual requirements.

[0152] Figure 7 The diagram shows the motor torque waveform of the method proposed in this invention when both the motor inductance and permanent magnet flux have a deviation of -30% or +30%. (The motor is running at 1000 r / min, with an initial torque of 0 Nm, and a torque step of 5 Nm given at a running time t of 2 s. The solid line waveform represents the motor torque waveform when both the motor inductance and permanent magnet flux have a deviation of +30%, and the dashed line waveform represents the motor torque waveform when both the motor inductance and permanent magnet flux have a deviation of -30%.)

[0153] Figure 8 This is a diagram showing the motor torque waveform of a traditional FCS-MPFC motor when both the motor inductance and permanent magnet flux linkage have a deviation of -30% or +30%. (The motor is running at 1000 r / min, with an initial torque of 0 Nm, and a torque step input of 5 Nm at a running time t of 2 s. The solid line waveform represents the motor torque waveform when both the motor inductance and permanent magnet flux linkage have a deviation of +30%, and the dashed line waveform represents the motor torque waveform when both the motor inductance and permanent magnet flux linkage have a deviation of -30%.)

[0154] pass Figure 7 and Figure 8 The comparison shows that the proposed solution can achieve precise control of motor torque regardless of the deviation in motor parameters. The steady-state average torque values ​​after 2 seconds are 5.058 Nm and 5.056 Nm in the two cases, respectively. In contrast, the traditional FCS-MPFC has a steady-state torque value of 7.140 Nm when the motor parameters have a -30% deviation, with the smaller parameter value causing a significant error in motor torque; and a steady-state torque value of 3.874 Nm when the motor parameters have a +30% deviation, with the larger parameter value causing a smaller motor torque, both of which fail to accurately track the given value.

Claims

1. A robust finite set model predictive flux linkage control method for permanent magnet synchronous motors, characterized in that, Includes the following steps: Step 1, set the control cycle duration to T. s Let the current control period be k; at the beginning of the current control period k, the rotor electric angular velocity of the permanent magnet synchronous motor is acquired. and rotor electrical angle At time τ after the start of the current control period k, at time (T) s At time -τ), the stator current is sampled twice, and the sampled values ​​are recorded as the first stator current sample value i for the current control cycle. a1 i b1 i c1 and the current control cycle secondary stator current sampling value i a2 i b2 i c2 Then, after transforming from the stationary coordinate system to the rotating dq coordinate system, the dq-axis component of the stator current sample value for the current control cycle is obtained. and the dq-axis component of the secondary stator current sample value in the current control cycle Step 2, record the three-phase switch state of the inverter as switch state S. a S b S c Record the inverter three-phase switch state in the current control cycle k as the current control cycle switch state. The voltage on the DC side of the inverter is denoted as voltage U. dc Based on the current control cycle switching state and voltage U dc The dq-axis component of the stator voltage of the permanent magnet synchronous motor during the current control cycle is calculated. The AC side of the inverter is connected to the stator of the permanent magnet synchronous motor; Step 3: Construct a generalized proportional-integral observer model and calculate the dq-axis components of the stator current observation value for the current control cycle. and the dq-axis components of the stator current disturbance observation during the current control cycle Step 4: Given the boundary conditions for q-axis inductance identification and permanent magnet flux linkage identification, obtain the filter coefficient h for q-axis inductance identification. Lq The filter coefficient h for permanent magnet flux identification ψf Given the boundary conditions for d-axis inductance identification, obtain the filter coefficient h for d-axis inductance identification. Ld ; Step 5: Based on the d-axis component of the stator current disturbance observation value during the current control cycle. The filter coefficient h identified by the q-axis inductance Lq The q-axis inductance is identified using the first identification equation, and the q-axis component of the inductance deviation identification value for the current control cycle is obtained. and the q-axis component of the inductance identification value in the current control cycle Based on the current control cycle inductance identification value q-axis component The filter coefficient h identified by the d-axis inductance Ld The d-axis inductance is identified using the d-axis current difference equation, and the d-axis component of the inductance identification value for the current control cycle is obtained. Based on the q-axis component of the stator current disturbance observation during the current control cycle The filter coefficient h for permanent magnet flux identification ψf The permanent magnet flux linkage is identified using the second identification equation to obtain the permanent magnet flux linkage deviation identification value for the current control cycle. and the permanent magnet flux identification value of the current control cycle Step 6: Identify the dq-axis component of the inductance value in the current control cycle. and the permanent magnet flux identification value of the current control cycle The current control cycle flux reference value dq axis component is calculated. and the dq-axis components of the flux linkage observation during the current control period Step 7, based on the current control cycle inductance identification value dq axis component and the dq-axis components of the flux linkage observation during the current control period The predicted flux linkage component dq-axis of the next control cycle is calculated. And the dq-axis components of the flux linkage prediction disturbance observations during the current control period Step 8, based on the inverter three-phase switch status S a S b S c This yields seven voltage vectors, denoted as voltage vector u. j j is the index of the voltage vector, j = 0, 1, ... 6; The calculation yields the voltage vector u j The dq-axis component of the flux linkage prediction value in the (k+2)th control cycle under the action is denoted as the dq-axis component of the target flux linkage prediction value. Step 9, convert the dq axis components of the target flux linkage prediction value. Substituting the value function, the voltage vector u is calculated. j The corresponding value function value g j The expression for the value function is as follows: Among the seven calculated value function values, the value function value with the smallest value is selected and recorded as the minimum value function value. The voltage vector corresponding to the minimum value function value is selected as the inverter output voltage vector for the next control cycle to control the switching state of the inverter switching transistor.

2. The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors according to claim 1, characterized in that, Step 2 describes the three-phase switch state S of the inverter. a S b S c The specific actions are as follows: S a =1 indicates that the inverter's A-phase bridge arm switch S a1 On, switch S a2 Turn off; S a =0 indicates that the inverter A-phase bridge arm switch S a1 Turn off, switch S a2 Conduction; S b =1 indicates that the inverter's B-phase bridge arm switch S b1 On, switch S b2 Turn off; S b =0 indicates that the inverter's B-phase bridge arm switch S b1 Turn off, switch S b2 Conduction; S c =1 indicates that the inverter's C-phase bridge arm switch S c1 On, switch S c2 Turn off; S c =0 indicates that the inverter's C-phase bridge arm switch S c1 Turn off, switch S c2 Conduction; Step 2 describes the current control cycle's permanent magnet synchronous motor stator voltage periodic average value dq-axis component. The calculation formula is as follows:

3. The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors according to claim 1, characterized in that, The expression for the generalized proportional-integral observer model described in step 3 is: in, The d-axis component of the stator current observation from the previous control cycle. This represents the q-axis component of the stator current observation from the previous control cycle. This represents the d-axis component of the secondary stator current sample value from the previous control cycle. T represents the q-axis component of the secondary stator current sample value from the previous control cycle. c T is the time interval between two stator current samplings. c =T s -2τ,R s For stator resistance, The d-axis component of the stator inductance identification value from the previous control cycle. The q-axis component of the stator inductance identification value from the previous control cycle. This is the nominal value of the q-axis stator inductance. γ is the nominal value of the permanent magnet flux linkage, γ is the observer proportional parameter, and κ is the observer integral parameter. The d-axis component of the stator current disturbance observation value from the previous control cycle. The q-axis component of the stator current disturbance observation value from the previous control cycle.

4. The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors according to claim 1, characterized in that, The boundary conditions for q-axis inductance identification and permanent magnet flux identification in step 4 are as follows: in, I min ω is the boundary value of the stator current amplitude. min ω is the boundary value of electric angular velocity. c For the bandwidth of the low-pass filter in q-axis inductance identification and permanent magnet flux identification, T c T is the time interval between two stator current samplings. c =T s -2τ; The boundary conditions for d-axis inductance identification in step 4 are as follows: Where η is the scaling factor for d-axis inductance identification, η∈(0,1), ω Ld Identify the bandwidth of the low-pass filter for the d-axis inductor. The d-axis component of the stator inductance identification value from the previous control cycle.

5. The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors according to claim 1, characterized in that, Step 5: The first identification equation is as follows: in, The q-axis component of the inductance deviation identification value from the previous control cycle. This is the nominal value of the q-axis stator inductance; The d-axis current difference equation described in step 5 is as follows: in, R is the d-axis component of the inductance identification value from the previous control cycle. s T is the stator resistance. c T is the time interval between two stator current samplings. c =T s -2τ; Step 5: The second identification equation is as follows: in, This is the permanent magnet flux deviation identification value from the previous control cycle. This is the nominal value of the permanent magnet flux linkage.

6. The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors according to claim 1, characterized in that, Step 6 describes the current control cycle flux reference value dq-axis component. The calculation formula is as follows: in, The q-axis component of the flux linkage reference value from the previous control cycle. The given value for motor torque in the current control cycle, where p is the number of motor pole pairs; Step 6 describes the dq-axis component of the current control cycle flux linkage observation. The calculation formula is as follows:

7. The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors according to claim 1, characterized in that, Step 7 describes the predicted flux linkage dq-axis component for the next control cycle. The calculation formula is as follows: in, The d-axis component of the flux linkage prediction value for the current control cycle. For the current control cycle, the predicted flux linkage value along the q-axis is γ. ψ R is the scaling parameter for the prediction model. s Stator resistance; Step 7 describes the dq-axis component of the current control cycle flux prediction disturbance observation. The calculation formula is as follows: in, The d-axis component of the flux linkage prediction disturbance observation from the previous control cycle. For the q-axis component of the flux linkage prediction disturbance observation from the previous control cycle, κ ψ These are the integral parameters for the prediction model.

8. The robust finite set model predictive flux linkage control method for permanent magnet synchronous motors according to claim 1, characterized in that, The voltage vector u in step 8 j The corresponding inverter three-phase switch state S a S b S c The specific status is as follows: The switching state corresponding to voltage vector u0 is S a =0, S b =0, S c =0; The switching state corresponding to voltage vector u1 is S a =1,S b =0, S c =0; The switching state corresponding to voltage vector u2 is S a =1,S b =1,S c =0; The switching state corresponding to voltage vector u3 is S. a =0, S b =1,S c =0; The switching state corresponding to voltage vector u4 is S a =0, S b =1,S c =1; The switching state corresponding to voltage vector u5 is S. a =0, S b =0, S c =1; The switching state corresponding to voltage vector u6 is S. a =1,S b =0, S c =1; Step 8 describes the dq-axis component of the target flux linkage prediction value. The calculation formula is as follows: Among them, u dj Voltage vector u j The d-axis component, u qj Voltage vector u j The q-axis component, R s This is the stator resistance.

Images