MTPA method based on virtual signal injection considering inverter nonlinearity

Abstract

The present invention discloses a maximum torque current ratio control method based on virtual signal injection considering nonlinear factors of inverter. The method uses a voltage model, a torque model, an electromagnetic power model, a high-frequency signal injection model and a nonlinear voltage drop model of a PMSM to obtain an accurate VSIC-MTPA torque model; then, the nonlinear voltage of the inverter is extracted according to the nonlinear voltage model of the inverter, the d-axis inductance parameter value is obtained according to the d-axis inductance parameter observation model, and the nonlinear voltage of the inverter and the d-axis inductance parameter value are brought into the derived torque model to improve the accuracy of the torque model; finally, the extracted nonlinear voltage of the inverter is compensated to the current loop control output, the high-order harmonics caused by the nonlinear voltage drop are reduced, and the steady-state control performance of the motor is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of motor control, and in particular relates to a maximum torque current ratio control method based on virtual signal injection taking into account nonlinear factors of an inverter. Background Art

[0002] Permanent Magnet Synchronous Motor (PMSM) is widely used in electric vehicle main drive systems due to its advantages such as high efficiency and high power density. However, its parameters are easily affected by operating conditions and temperature changes and have strong time-varying properties, making it difficult to achieve precise control.

[0003] As a commonly used high-performance vector control strategy, the MTPA (maximum torque to current ratio control) strategy can make full use of the permanent magnet torque and reluctance torque of the salient-pole PMSM to minimize the current output and achieve the desired torque output. As a commonly used control method, the table lookup method determines the optimal MTPA operating point based on prior test data. It does not require a large amount of computing resources and has good stability and applicability, but requires a large number of experiments to obtain accurate prior data, and the calibrated data is not universal for different motors. The traditional formula method MTPA (EQ-MTPA) control strategy has the advantages of simple calculation and fast response, but the electromagnetic parameters of automotive motors are highly time-varying, and the influence of inverter nonlinear errors causes the actual operating point of the motor to deviate from the maximum torque to current ratio target. At present, the online MTPA optimization control strategy based on high-frequency signal injection is mainly divided into the MTPA control method of real signal injection (RSIC-MTPA) and the MTPA control method of virtual signal injection (VSIC-MTPA). RSIC-MTPA obtains the relevant information of the optimal current phase angle by extracting the high-frequency harmonics of the torque or speed after the high-frequency signal is injected, while VSIC-MTPA obtains the MTPA partial derivative information of the torque about the optimal phase angle by establishing a mathematical model of the motor and approximately solving the torque after the signal is injected. Compared with the above two methods, VSIC-MTPA does not require real high-frequency signal injection and does not generate additional torque / speed fluctuations and losses, but how to reduce the dependence on electromagnetic parameters in the calculation process of torque partial derivatives is the main problem to be solved.

[0004] The traditional VSIC-MTPA (TVSIC-MTPA) proposed in the literature [Liang Kangyou, Yuan Ling, Tan Yuhang. (2016). Virtual signal injection MTPA control method for permanent magnet synchronous motor. Power Electronics Technology, 50(6), 4.] calculates the electromagnetic torque by dividing the electromagnetic power by the angular velocity, where the electromagnetic power is approximated by subtracting the copper loss from the input power. It should be noted that due to the measurement bandwidth limitation, the voltage required in the input power calculation is replaced by the voltage reference value output by the controller, which leads to the problem that the torque still deviates from the maximum torque current ratio target at the working point where the partial derivative of the current vector angle is zero. In order to solve this problem, the literature [Qiu Jianqi, Song Pan, Chen Zhuoyi. (2022). Improved virtual signal injection MTPA control of permanent magnet synchronous motor. Journal of Electric Machines and Control (009), 026] establishes a more accurate torque model based on the torque model used by TVSIC-MTPA, but this method is based on the premise that the d-axis inductance value is constant. Summary of the invention

[0005] In view of the above, the present invention provides a maximum torque-to-current ratio control method based on virtual signal injection taking into account inverter nonlinear factors, which can effectively improve the maximum torque-to-current ratio steady-state control accuracy of the permanent magnet synchronous motor.

[0006] A maximum torque current ratio control method based on virtual signal injection considering inverter nonlinear factors comprises the following steps:

[0007] (1) Collect the three-phase stator current of PMSM and convert it into the stator current vector in the dq coordinate system i dq , and then calculate and extract the nonlinear voltage drop Δ of the inverter u dq ;

[0008] (2) For the nonlinear voltage drop Δ u dq Perform low-pass filtering to obtain its DC voltage drop ;

[0009] (3) According to the stator current vector i dq And DC voltage drop The observed value of the d-axis stator inductance is obtained by identification ;

[0010] (4) Based on the observed value of the d-axis stator inductance and DC voltage drop Calculate the torque error caused by the voltage drop and the torque error required to compensate ;

[0011] (5) Torque error and It is introduced into the torque model in VSIC-MTPA to calculate a more accurate torque for controlling the motor.

[0012] Furthermore, the step (1) calculates and extracts the nonlinear voltage drop Δ of the inverter by the following expression: u dq :

[0013] ;

[0014]

[0015] in: is the three-phase stator current of PMSM, θ e is the rotor electrical angle of PMSM, is the nonlinear voltage drop produced by the switch tube in the inverter, sgn( ) represents the sign function, is the coordinate transformation matrix.

[0016] Furthermore, the nonlinear voltage drop generated by the switch tube in the inverter The expression is as follows:

[0017]

[0018] in: T dead is the dead time of the switch tube, T on is the on-time of the switch tube, T off is the turn-off time of the switch tube, u t is the conduction voltage drop of the switch tube, u diode is the conduction voltage drop of the freewheeling diode in the inverter, U dc is the DC bus voltage of the inverter, T s is the switching cycle.

[0019] Furthermore, in step (3), the d-axis stator inductance observation value is obtained by identifying the following expression: :

[0020]

[0021] in: i d and i q They are the stator current vectorsi dq The d-axis component and q-axis component of R s is the stator phase resistance of PMSM, ψ m is the permanent magnet flux linkage of PMSM, ω e is the rotor electrical angular velocity of the PMSM, is the reference value of the q-axis component of the stator voltage, DC voltage drop The q-axis component of .

[0022] Furthermore, in step (4), the torque error is calculated by the following expression: :

[0023]

[0024] in: and The DC voltage drop is The d-axis component and q-axis component of i q is the stator current vector i dq The q-axis component of and They are the d-axis component and q-axis component of the stator current containing high-frequency components, ω m is the mechanical angular velocity of the PMSM.

[0025] Furthermore, in step (4), the torque error is calculated by the following expression: :

[0026]

[0027] in: p n is the number of pole pairs of PMSM, Δ θ is the injected high frequency changing angle signal, i q is the stator current vector i dq The q-axis component of It is the q-axis component of the stator current containing high-frequency components.

[0028] Furthermore, in step (5), the torque error and The torque model introduced into VSIC-MTPA is expressed as follows:

[0029]

[0030] in: Represents the torque containing high-frequency components, and are the d-axis component reference value and q-axis component reference value of the stator voltage respectively, R s is the stator phase resistance of PMSM, ω m is the mechanical angular velocity of the PMSM, i d and i q They are the stator current vectors i dq The d-axis component and q-axis component of and They are the d-axis component and q-axis component of the stator current containing high-frequency components respectively.

[0031] Furthermore, the stator current d-axis component and the q-axis component The expression is as follows:

[0032]

[0033] in: ω h and A They are respectively the high frequency changing angle signal Δ θ The frequency and amplitude of I s and θ They are the stator current vectors i dq Amplitude and angle.

[0034] In order to further improve the control accuracy of the VSIC-MTPA strategy, the present invention takes into account the nonlinear factors of the inverter, and derives the error terms of the torque model through theoretical derivation and analysis of the torque model in the VSIC-MTPA control strategy. The motor parameters contained in the error terms are observed in real time to achieve accurate error compensation. At the same time, by extracting the nonlinear voltage drop of the inverter, more accurate dq-axis voltage values ​​are used for torque calculation and parameter identification, which improves the parameter identification accuracy and torque model accuracy, thereby improving the steady-state control performance of PMSM under the VSIC-MTPA control strategy.

[0035] The present invention derives a new permanent magnet synchronous motor VSIC-MTPA torque model and proposes an improved VSIC-MTPA control strategy. The method fully considers the influence of the inverter nonlinear voltage drop and the change of inductance parameters on the MTPA optimization result, and the obtained current operating point is closer to the optimal MTPA operating point. In addition, the present invention further improves the current control accuracy by simultaneously feeding back the inverter nonlinear voltage to the control voltage. The experimental results verify the feasibility and stability of the method of the present invention under different working conditions, and can achieve more accurate MTPA control and better steady-state control performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Schematic diagram of the MTPA partial derivative information acquisition method based on virtual signal injection, where Eq. (8) represents formula (8).

[0037] Figure 2 is a control flow chart of the maximum torque current ratio control method of the present invention, wherein T p represents Park transform, T C represents the Clarke transform, Represents the inverse Park transform.

[0038] Figure 3 The current amplitude under different control strategies when the parameters are mismatched I s , d-axis current i d , q-axis current i q , motor torque T e Schematic diagram of control effect comparison, where the superscript eq , t , nd Corresponding to EQ-MTPA, TVSIC-MTPA, and the control strategy of the present invention, the horizontal axis t Indicates time, and Sampling Interval indicates the sampling interval.

[0039] Figure 4 Figure 2 shows the MTPA partial derivative information waveform and d-axis current deviation waveform under different control strategies when the parameters are mismatched, where (a) corresponds to TVSIC-MTPA, (b) corresponds to the control strategy of the present invention, and the ordinate Δ i d is the d-axis current deviation, is the DC component, the horizontal axis t Indicates time, and Sampling Interval indicates the sampling interval.

[0040] Figure 5 is the current amplitude under different control strategies when the parameters are mismatched I s Comparison diagram, where the vertical axis I s is the current amplitude, the horizontal axis T e is the motor torque, with the superscript eq , t , nd , rl They correspond to EQ-MTPA, TVSIC-MTPA, the control strategy of the present invention, and the real MTPA trajectory respectively.

[0041] Figure 6 The schematic diagram of MTPA operation trajectory curve under different control strategies when the parameters are mismatched, where the horizontal axis i d is the d-axis current, the ordinate i q is the q-axis current, , , , They correspond to the EQ-MTPA trajectory, the TVSIC-MTPA trajectory, the MTPA trajectory under the control strategy of the present invention, and the real MTPA trajectory respectively.

[0042] Figure 7 The control effect waveform diagram of the control strategy of the present invention under the working condition of speed 750rpm and torque 50Nm, wherein (a) corresponds to no inverter nonlinear voltage compensation, and (b) corresponds to inverter nonlinear voltage compensation.

[0043] Figure 8 The control effect waveform diagram of the control strategy of the present invention under the working condition of speed 750rpm and torque 250Nm, wherein (a) corresponds to no inverter nonlinear voltage compensation, and (b) corresponds to inverter nonlinear voltage compensation.

[0044] Figure 7 and Figure 8 Middle vertical coordinate i d is the d-axis current, i q is the q-axis current, i a for a Phase current, Mag (% of fund) represents the percentage of each frequency harmonic amplitude relative to the fundamental frequency amplitude, THD represents total harmonic distortion, and the horizontal axis Freq represents frequency. t Indicates time, and Sampling Interval indicates the sampling interval. DETAILED DESCRIPTION

[0045] In order to describe the present invention more specifically, the technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0046] The mathematical model of the permanent magnet synchronous motor in the dq axis coordinate system can be expressed as follows:

[0047] (1)

[0048] (2)

[0049] in: u d 、u q are the d-axis and q-axis voltage values ​​respectively; i d 、i q are the d-axis and q-axis current values ​​respectively; R s is the stator phase resistance; ω e is the rotor electrical angular velocity; L d 、L q are the d-axis and q-axis inductance values ​​respectively; ψ m is the permanent magnet flux; T e is the electromagnetic torque; p n is the number of motor pole pairs.

[0050] The dq axis current value is converted from the current vector amplitude I s and angle θ To express:

[0051] (3)

[0052] If the frequency injected into the current phase angle is ω h A high frequency changing angle signal Δ with amplitude A θ , then the dq axis current can be expressed as:

[0053] (4)

[0054] Substituting the above formula into the torque model, we can get the torque expression containing high-frequency components: , and then expand it according to the Taylor series, that is:

[0055] (5)

[0056] It is known that Using the center frequency ω h The band-pass filter (BPF) can be used to filter the A sinusoidal signal proportional to ,in k is the bandpass filter gain, multiplying the sinusoidal signal by sin( ω h t ), that is:

[0057] (6)

[0058] The DC component can be obtained by using a low-pass filter (LPF) ,like Figure 1 As shown, by integrating the DC component and controlling it to 0, the d-axis current deviation Δ i d , the MTPA operating point can be derived i d - i q expression:

[0059] (7)

[0060] in: , k i is Δ i d The integral coefficient of .

[0061] Based on the above Figure 2 As shown, the torque model derivation process of the maximum torque current ratio control method based on virtual signal injection of the present invention is as follows:

[0062] First, according to the electromagnetic power P m Divided by the mechanical angular velocity ω m The torque can be calculated by eliminating the influence of motor parameters:

[0063] (8)

[0064] Considering the steady state of the motor, substituting equation (1) into equation (8) yields:

[0065] (9)

[0066] According to formula (4), the high-frequency injected angle signal Δ θ It changes slightly around 0, so cosΔ θ ≈1, and then we can get the following formula:

[0067] ;

[0068] (10)

[0069]

[0070] In formula (10), i d Substituting the expression into formula (8) yields:

[0071] (11)

[0072] It can be seen that compared with directly using the torque model (2) for high-frequency signal injection, the torque model (8) used has a certain deviation. Therefore, the accuracy of the torque model in the virtual signal injection method can be improved by error compensation. The torque model after error compensation is shown in formula (12):

[0073] (12)

[0074] The new torque model includes the dq axis voltage value. Since the dq axis voltage is difficult to obtain by measurement in practice, the reference value is used. The calculation is not accurate, so the nonlinear voltage drop of the inverter Δ u dq The impact on the calculation of MTPA related information is , then the torque model after considering the nonlinear voltage drop of the inverter is:

[0075] (13)

[0076] For the inverter nonlinear voltage drop and inductance parameters included in the torque model, the present invention extracts and compensates them in an observation manner to improve the accuracy of the torque model.

[0077] The nonlinear voltage drop of the inverter switch is mainly caused by the dead time. T dead , Opening time T on , off time T off , Switching tube conduction voltage drop u t and the freewheeling diode conduction voltage drop u diode The nonlinear voltage drop model generated by each switch tube can be expressed as:

[0078] (14)

[0079] The nonlinear voltage drop of the inverter converted to the dq axis coordinate system can be expressed as:

[0080] (15)

[0081] in: , Δ u dq =[Δ u d Δ u q ] T , θ e is the rotor electrical angle, i abc =[ i a i b i c ] T , sgn() represents the sign function, which determines the positive or negative sign of the independent variable and takes values ​​of 1, 0, and -1.

[0082] The nonlinear voltage drop of the inverter under the extracted dq axis can be compensated to the control voltage at the same time u dq In order to reduce the influence of dead zone on control accuracy and reduce the high-order harmonics in the three-phase current, the DC component of the inverter voltage drop in the torque model and inductance parameter observation model is used. Calculation is performed to reduce the impact of high-order harmonics on the calculation accuracy of the model. The DC component can be obtained through a low-pass filter:

[0083] (16)

[0084] in: ω c is the low-pass filter cutoff frequency.

[0085] Then the torque model used in the maximum torque current ratio control strategy based on virtual signal injection of the present invention can be expressed as:

[0086] (17)

[0087] After compensating the error term of the torque model, in order to further reduce the influence of motor parameters on model accuracy, it is necessary to identify the d-axis inductance. The identification model is shown in the following formula:

[0088] (18)

[0089] in: ω e = p n ω m .

[0090] In order to verify the feasibility and effectiveness of the present invention, we built a PMSM experimental platform for verification. The PMSM controller uses DSP (TMS320F28379D) + FPGA (5CEFA7F27I7N) as the main control chip architecture. The main parameters of the motor and the experimental platform are shown in Table 1:

[0091] Table 1: Main parameters of motor and experimental platform

[0092]

[0093] At a motor speed of 750 rpm, the MTPA curve obtained by manual calibration is as follows: Figure 6 To simulate the parameter mismatch phenomenon, the dq axis inductance parameters in the DSP control program are set to L d =2 L dn , L q =1.2 L qn . Figure 3 and Figure 5 The comparative experimental results of the TVSIC-MTPA control method and the control method of the present invention are given respectively, the sampling interval (SamplingInterval) is 5.0000us, the motor speed is 750rpm during the experiment, and the load torque is dynamically loaded; Figure 5 It can be seen that the current amplitude using the control strategy of the present invention Compared with TVSIC-MTPA current amplitude and the current amplitude of EQ-MTPA Smaller, closer to the current amplitude under the real MTPA trajectory .Depend on Figure 4 Figure 6 It can be seen that compared with the TVSIC-MTPA control strategy, the control strategy of the present invention can accurately extract the torque partial derivative information related to MTPA due to the use of a more accurate torque model, so that the calculated d-axis error current is more accurate, making the MTPA trajectory closer to the real MTPA trajectory.

[0094] Will Figure 3 The dq axis current shown in is plotted into the MTPA operation trajectory curve, as shown in Figure 6As shown, the MTPA trajectory under the control method of the present invention Compared with TVSIC-MTPA trajectory and EQ-MTPA trajectory Closer to the real MTPA trajectory , it can be seen that the method of the present invention can achieve more accurate MTPA control.

[0095] Figure 7 The comparative experimental results of the control strategy of the present invention with and without inverter nonlinear voltage compensation control are given. Mag (% of fund) represents the percentage of each frequency harmonic amplitude relative to the fundamental frequency amplitude. During the experiment, the motor speed is 750rpm and the load torque is 50Nm. It can be seen that the THD (total harmonic distortion) of the A-phase current is 3.57% with compensation control, while the THD of the A-phase current is 4.04% without compensation. Increasing the load torque to 250Nm, and keeping other conditions unchanged, the experimental waveform is as follows Figure 8 As shown, it can be seen that the values ​​of current THD with and without compensation control are 1.90% and 2.16% respectively. In addition, under different working conditions, after adding the inverter nonlinear voltage drop compensation control, the dq axis current ripple is reduced.

[0096] The above description of the embodiments is to facilitate the understanding and application of the present invention by those skilled in the art. It is obvious that those skilled in the art can easily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without creative work. Therefore, the present invention is not limited to the above embodiments. Improvements and modifications made by those skilled in the art to the present invention based on the disclosure of the present invention should be within the protection scope of the present invention.

Claims

1. A maximum torque current ratio control method based on virtual signal injection considering inverter nonlinear factors, comprising the following steps: (1) Collect the three-phase stator current of PMSM and convert it into the stator current vector i in the dq coordinate system dq , and then calculate and extract the nonlinear voltage drop Δu of the inverter dq ; (2) For the nonlinear voltage drop Δu dq Perform low-pass filtering to obtain its DC voltage drop (3) According to the stator current vector i dq And DC voltage drop The observed value of the d-axis stator inductance is obtained by identification (4) Based on the observed value of d-axis stator inductance Calculate the torque error required to compensate According to the DC voltage drop Calculate the torque error caused by the voltage drop (5) The torque error and It is introduced into the torque model in VSIC-MTPA to calculate a more accurate torque for controlling the motor.

2. The maximum torque current ratio control method according to claim 1, characterized in that: The step (1) is to calculate and extract the nonlinear voltage drop Δu of the inverter by the following expression: dq : Δu dq =2T PC ·sgn(i abc )Δu Where: i abc is the three-phase stator current of PMSM, θ e is the rotor electrical angle of PMSM, Δu is the nonlinear voltage drop generated by the switch tube in the inverter, sgn() represents the sign function, T PC is the coordinate transformation matrix.

3. The maximum torque current ratio control method according to claim 2, characterized in that: The expression of the nonlinear voltage drop Δu generated by the switch tube in the inverter is as follows: Where: T dead is the dead time of the switch tube, T on is the on-time of the switch tube, T off is the turn-off time of the switch tube, u t is the conduction voltage drop of the switch tube, u diode is the conduction voltage drop of the freewheeling diode in the inverter, U dc is the DC bus voltage of the inverter, T s is the switching cycle.

4. The maximum torque current ratio control method according to claim 1, characterized in that: In step (3), the d-axis stator inductance observation value is obtained by identifying the following expression: Where: i d and i q They are the stator current vector i dq The d-axis component and q-axis component of R s is the stator phase resistance of PMSM, ψ m is the permanent magnet flux linkage of PMSM, ω e is the rotor electrical angular velocity of the PMSM, is the reference value of the q-axis component of the stator voltage, DC voltage drop The q-axis component of .

5. The maximum torque current ratio control method according to claim 1, characterized in that: In step (4), the torque error is calculated by the following expression: in: and The DC voltage drop is The d-axis component and q-axis component of i q is the stator current vector i dq The q-axis component of and are the d-axis component and q-axis component of the stator current containing high-frequency components, ω m is the mechanical angular velocity of the PMSM.

6. The maximum torque current ratio control method according to claim 1, characterized in that: In step (4), the torque error is calculated by the following expression: Where: p n is the number of pole pairs of PMSM, Δθ is the injected high-frequency angle signal, i q is the stator current vector i dq The q-axis component of It is the q-axis component of the stator current containing high-frequency components.

7. The maximum torque current ratio control method according to claim 1, characterized in that: In step (5), the torque error and The torque model introduced into VSIC-MTPA is expressed as follows: in: Represents the torque containing high-frequency components, and are the d-axis component reference value and q-axis component reference value of the stator voltage, R s is the stator phase resistance of PMSM, ω m is the mechanical angular velocity of PMSM, i d and i q They are the stator current vector i dq The d-axis component and q-axis component of and They are the d-axis component and q-axis component of the stator current containing high-frequency components respectively.

8. The maximum torque current ratio control method according to claim 5 or 7, characterized in that: The stator current d-axis component and the q-axis component The expression is as follows: Where: h and A are the frequency and amplitude of the high-frequency angle signal Δθ, respectively. s and θ are the stator current vector i dq Amplitude and angle.

Images