Motor current control device and motor current control method

Abstract

To compensate for distortion of a voltage with oversampling deadbeat current control applied to motor drive.SOLUTION: A motor current control device 12 comprises: current detection units 10U, 10V, and 10W for detecting a current supplied to a motor 6; and a control unit that controls turning on / off of a plurality of semiconductor switching elements. The control unit comprises: a current compensation calculation unit 9 that samples an actual current detected by the current detection units at each sampling period, which is set shorter than a switching period of the semiconductor switching elements, obtains an actual voltage for that sampling period, and calculates a current compensation value based on a difference between a temporal transition of the actual current during a carrier half-period in which a commanded voltage is applied and the actual current, and on a difference between the commanded voltage and the actual voltage; and an OSDB current control unit 2 that calculates the commanded voltage based on an ideal current obtained by adding the current compensation value to the sampled actual current, a commanded current, and an induced voltage generated in windings of an AC motor.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Embodiments of the present invention relate to an apparatus and method for controlling the current of a motor.

Background Art

[0002] In recent years, various systems have used electric power as a power source, and in Japan, approximately half of the electric power consumption is consumed by motors. Furthermore, in drive systems such as for transportation, industry, and household use, the demand for highly accurate and high-performance motor control is increasing. To achieve these, high stability and responsiveness of current control are required. To improve the responsiveness of current and torque, predictive control, such as deadbeat control, is a control method for realizing a high current control band. According to this deadbeat control, the control error can be made zero within a short control time, thereby realizing a high-speed control response.

[0003] Deadbeat control has the feature that high responsiveness can be realized at a constant switching frequency compared to other predictive controls such as model predictive control. In model predictive control, since a modulator is not used, there is a problem that the switching frequency fluctuates and the frequency components of current ripple cannot be managed. Therefore, deadbeat control is selected for many applications.

[0004] Predictive control requires a lot of arithmetic processing, so the implementation on a real machine system is complicated. Originally, it is necessary to execute sampling and control processing within one control cycle, but this is often impossible with a standard controller, and it is necessary to compensate for the delay until sampling and control processing are executed. However, if an FPGA (Field Programmable Gate Array) or the like is used to improve the arithmetic processing speed, the arithmetic time is significantly shortened, so that sampling and control processing can be executed even within one control cycle.

[0005] In addition, predictive control is highly dependent on model parameters for its performance. When the error from the actual value is large, the performance in both transient and steady states significantly degrades. Furthermore, the performance also degrades due to the non-linearity of the inverter voltage, such as the voltage drop during dead time and the non-ideal switching of power devices.

[0006] For example, in Non-Patent Document 1, in order to reduce the performance degradation due to parameter mismatch and voltage non-linearity, high-speed calculations are performed using an FPGA, enabling more current samplings than the current sampling generally performed in a carrier half-cycle in PWM control, so-called oversampling. Thereby, in motor drive control operating at a sampling frequency that is an even multiple of the switching frequency of the inverter, a configuration for realizing oversampling deadbeat (hereinafter referred to as OSDB) current control is disclosed. The OSDB current controller is obtained by discretizing the system equations over a variable sample time, increasing the sampling frequency, and enhancing the robustness against parameter variations.

Prior Art Documents

Non-Patent Documents

[0007]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0008] In the OSDB current control, the dq-axis voltage does not match the average voltage actually applied by the inverter over the sampling time. Therefore, by adding a compensation term to the dq-axis voltage obtained by the OSDB current control to obtain the final dq-axis voltage, the distortion of the voltage supplied by the PWM control is compensated. However, the mismatch between this dq-axis voltage and the actual voltage is fundamentally caused by the fact that the actual current input to the OSDB current control fluctuates due to the applied voltage, and the OSDB current control is implemented based on the fluctuated actual current.

[0009] Therefore, regarding the deadbeat current control applied to motor drive, a motor current control device and a motor current control method are provided that can eliminate the influence of current fluctuations due to the applied voltage and compensate for voltage distortion.

Means for Solving the Problem

[0010] The motor current control device of the embodiment includes a power converter having a plurality of semiconductor switching elements, which converts DC power into AC power and supplies it to an AC motor, a current detection unit that detects the current supplied to the AC motor, and a control unit that controls the on / off of the plurality of semiconductor switching elements. The control unit samples the actual current detected by the current detection unit at each sampling period set shorter than the switching period of the semiconductor switching element, obtains the actual voltage at each sampling period, and obtains a current compensation value from the difference between the time transition of the actual current and the actual current in the carrier half cycle when the command voltage is applied and the difference between the command voltage and the actual voltage. The control unit has a current control unit that calculates the command voltage from the ideal current obtained by adding the current compensation value to the sampled actual current, the command current, and the induced voltage generated in the windings of the AC motor.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Mode for Carrying Out the Invention

[0012] (First Embodiment) Hereinafter, the first embodiment will be described with reference to FIGS. 1 to 11. First, the principle of this embodiment will be described with reference to Expressions (1) to (17), FIGS. 9 to 11, and FIG. 1. First, the deadbeat (DB) control rule is derived from the voltage equations of the dq-axis coordinate axes of a PMSM (Permanent Magnet Synchronous Motor). The voltage equations are shown in Mathematical Expression (1).

[0013]

Number

[0014] Here, Vd is the d-axis voltage, Vq is the q-axis voltage, id is the d-axis current, iq is the q-axis current, R is the phase resistance, Ld is the d-axis inductance, Lq is the q-axis inductance, ω is the electrical angular frequency, and Φ is the magnetic flux linked by the permanent magnet.

[0015] To implement the DB current control digitally, discretization of the plant equations is required. When expressing Equation (1) in terms of state variables, it becomes Equation (2).

[0016]

Equation

[0017] The third term on the right side of Equation (2) represents the induced voltage. The discrete state equation of the plant shown in Equation (2) at time k is Equation (3).

[0018]

Equation

[0019] Here, Ts is the sampling period.

[0020] Regarding the coefficient matrices G and H of the state equation, when adopting the approximation up to the first term of the Maclaurin series, the final matrix of the PMSM state space model is Equation (4).

[0021]

Equation

[0022] The general DB control law can be derived from Equations (3) and (4). By substituting the dq-axis current commands into the terms of (idk + 1) and (iqk + 1) and solving for the voltages Vdk and Vqk, Equation (5) is obtained. (idk + 1) and (iqk + 1) terms and solving for the voltages Vdk and Vqk, Equation (5) is obtained.

[0023]

Equation

[0024] Here, the dq-axis current commands at time (k + 1) are required. However, since the sampling time is sufficiently small compared to the system time constant, it can be assumed to be equal to the dq-axis current commands at time k. The dq-axis voltages calculated by Equation (5) are voltages that can completely follow the current commands within one sampling time when saturation does not occur.

[0025] Next, the OSDB control will be described with reference to FIG. 9. Tx is the sampling period, and Tc is the carrier period. When shortening the sampling period Tx without changing the switching frequency, i.e., the carrier frequency, if the general DB control rule is simply applied, the ratio (Tx / Tc) decreases, and thus the system becomes unstable. The dq-axis voltages calculated by Equation (5) are optimal only when they are equal to the inverter output voltages averaged over the sampling period Tx. This is the case only for the carrier half period (Tc / 2) when using a PWM modulator. The oversampling coefficient nc is defined by Equation (6). nc = {Tc / (2Tx)} …(6)

[0026] Therefore, from the oversampling coefficient nc and the sampling time (k + i), the current tracking time, i.e., the variable sampling time to the peak of the triangular-wave carrier in the PWM modulator Ts(i) is defined by Equation (7).

[0027] Ts(i) = Tc / 2 - iTx = (nc - i)Tx …(7) Here, i = 0, 1, 2, …, nc - 1, and the variable i is reset to zero every carrier half period Tc / 2. To consider the variable sampling time Ts(i) of the controller, substituting Equation (7) into Equations (3) and (4) of the discrete state equations gives Equation (8).

[0028]

Number

[0029] Coefficient matrix G Ts(i) and H Ts(i) is the same as Equation (4), where the sampling time Ts is replaced by Ts(i) in Equation (7). The output voltage by OSDB control is obtained by solving Equation (8) for the dq-axis voltages Vd k+i and Vq k+i is obtained by solving for them.

[0030]

Number

[0031] The dq-axis voltages (k + nc) are calculated by Equation (9) and applied nc times for the oversampling number. The sampling time for each application decreases by the sampling period Tx. As described above, since predictive control is affected by parameter errors, it causes a significant degradation in the performance of DB control. In general DB control, when parameter variations occur, the dq-axis voltages based on the PMSM (Permanent Magnet Synchronous Motor) model are applied over the carrier half-cycle (Tc / 2). On the other hand, in OSDB control, since the dq-axis voltages based on the PMSM model can be calculated with a shorter sampling period, the influence of prediction errors due to parameter variations can be reduced.

[0032] In general DB control, if there are no parameter variations, the current command can be completely followed with one sampling period of the carrier period or the carrier half-cycle. At this time, the dq-axis voltages calculated by DB control are equal to the average voltages actually applied by the inverter over the sampling time. However, in OSDB control, since these do not match, it is necessary to compensate the dq-axis voltages. According to the PMSM model, from Equation (8), the dq-axis currents at time (k + i) can be predicted from the previous (k + i - 1) by Equation (10).

[0033]

Number

[0034] Here, the PMSM matrices G Tx and H Tx correspond to Equation (4) and are discretized at time Tx. Note that the dq-axis voltages in Equations (10C) and (10R) are respectively applied with the control voltage and the actual voltage of the OSDB control. Substituting Equation (10) into Equation (9), the dq-axis voltages at time (k + i) depending on the control voltage and the actual voltage can be obtained.

[0035]

Number

[0036] Here, the voltage compensation term in Non-Patent Document 1 is calculated by subtracting the actual voltage formula of Equation (11R) from the control voltage formula of Equation (11C).

[0037]

Number

[0038] The final control law of the OSDB control in Non-Patent Document 1 becomes as shown in Equation (13) by adding the voltage compensation term of Equation (12) to Equation (9).

[0039]

Number

[0040] From Equation (12), for the calculation of the compensation term of the OSDB control, both the previous control voltage and the actual voltage are required. The previous control voltage can be stored in the memory of the controller, but the actual voltage needs to be calculated. Figure 10 shows a single-phase PWM signal generated by comparing with a triangular carrier and the corresponding inverter voltage. The switching times t+ and t- can be obtained from the triangular carrier and the single-phase voltage command Vref. Using these switching times, the actual voltage applied by the inverter can be calculated as follows.

[0041]

Number

[0042] Equation (14I) corresponds to the period during which the level of the triangular wave carrier increases, and equation (14D) corresponds to the period during which the above level decreases.

[0043] Here, the control block diagram of Non-Patent Document 1 is shown in FIG. 11. In Non-Patent Document 1, as shown in equation (13), the compensation term is added to the dq-axis voltage obtained by the OSDB control rule, and the resulting value is used as the final dq-axis voltage. This makes the dq-axis voltage match the average voltage actually applied by the inverter over the sampling time.

[0044] As described above, the discrepancy between this dq-axis voltage and the actual voltage is fundamentally caused by the fact that the actual current input to the OSDB current control fluctuates due to the applied voltage, and the OSDB current control is implemented based on the fluctuated actual current. Therefore, in this embodiment, the detected actual current is corrected, and the ideal current is obtained by eliminating the influence of current fluctuations due to the applied voltage, and this is input to the OSDB current controller.

[0045] Regarding the dq-axis current at time (k + i) shown in equation (10), when the control value and the actual value are applied to the dq-axis current and the dq-axis voltage respectively, equation (15) is obtained.

[0046]

Number

[0047] Equation (15C) shows the time transition of the actual dq-axis current in the half cycle of the carrier when the dq-axis command voltage is applied. By subtracting the actual dq-axis current of equation (15R) from equation (15C), the current compensation term can be derived.

[0048]

Number

[0049] Also, at the apex of the triangular carrier wave, since the current ripple due to PWM control increases and decreases, and as a result, the time transition of the actual dq-axis current and the actual dq-axis current match, the current compensation term at time (i = 0) is set to zero. Further, the time transition of the actual dq-axis current in Equation (16) is obtained by Equation (15C), and at time (k + 1, that is, i = 1), the dq-axis current at time (k) is the detected actual current. The final control law of the OSDB control in this embodiment is represented by Equation (17) by adding the current compensation term of Equation (16) to Equation (9).

[0050] [Number]

[0051] The control block diagram of this embodiment corresponding to the above control law is shown in FIG. 1. In addition, the corresponding equations are described in the blocks using the above equations. In the speed control unit 1, the voltage Vdc, Vdq, and the speed command ωref are input from the outside, and the speed ω is also input. The voltage Vdc is the voltage of the driving DC power supply supplied to the inverter unit, and the voltage VCdq is the square root of the sum of the squares of the dq-axis command voltages VCd and VCq shown in the figure. The speed ω is the speed detected by the θ·ω detection unit 7 for the motor 6 which is an AC motor. Also, in FIG. 1, the signal lines of the rotational position θ and the speed ω detected by the θ·ω detection unit 7 are shown bundled together as one.

[0052] The speed control unit 1 generates the dq-axis current commands Idref and Iqref based on the input signals and inputs them to the next-stage OSDB current control unit 2. In addition to these, in the OSDB current control unit 2, the dq-axis currents Id and Iq from the three-phase → dq coordinate conversion unit 11, the speed ω, and the current compensation terms idC and iqC from the current compensation term calculation unit 9 are input. The OSDB current control unit 2 generates the dq-axis command voltages VCd and VCq according to the control law shown in Equation (17) and inputs them to the dq → three-phase coordinate conversion unit 3 and the current compensation term calculation unit 9.

[0053] The dq-axis to three-phase coordinate conversion unit 3 converts the dq-axis command voltages VCd and VCq into three-phase command voltages V U , V V and V W according to the rotational position θ, and inputs them to the PWM generation unit and the actual voltage calculation unit 4. The voltage Vdc is input to the PWM generation unit and the actual voltage calculation unit 4, which generates three-phase PWM signals U±, V±, and W± and inputs them to the inverter unit 5. Also, according to Equation (14), three-phase actual voltages VR U , VR V and VR W are generated and input to the three-phase to dq-axis coordinate conversion unit 8. The three-phase to dq-axis coordinate conversion unit 8 converts the three-phase actual voltages VR U , VR V and VR W into dq-axis actual voltages VRd and VRq and inputs them to the current compensation term calculation unit 9.

[0054] The inverter unit 5, which is a power converter, is specifically composed of, for example, semiconductor switching elements such as power MOSFETs or IGBTs connected in a three-phase bridge configuration, although not shown in the figure. Current detection units 10U, 10V, and 10W are arranged at the output terminals of each phase of the inverter unit 5, and the phase currents IU, IV, and IW detected by them are input to the three-phase to dq-axis coordinate conversion unit 11. The three-phase to dq-axis coordinate conversion unit 11 converts the sampled three-phase currents IU, IV, and IW into dq-axis currents Id and Iq according to the rotational position θ and inputs them to the current compensation term calculation unit 9. When the current compensation term calculation unit 9 samples the dq-axis command voltages VCd, VCq and the dq-axis actual voltages VRd, VRq, it calculates current compensation terms idC and iqC according to Equation (16) and inputs them to the OSDB current control unit 2.

[0055] In the above, excluding the motor 6 constitutes the motor current control device 12. Also, excluding the inverter unit 5 and the current detection unit 10 from the motor current control device 12 constitutes the control unit 13.

[0056] For the motor current control device 12 of the present embodiment, in the control block diagram of Non-Patent Document 1 shown in FIG. 11, in the OSDB current control unit 2 and the block corresponding to the current compensation term calculation unit 9, the mathematical formulas used are obtained by substituting Formula (17) with Formula (13) and Formula (16) with Formula (12), respectively.

[0057] The effects of the present embodiment were verified by simulation using MATLAB (registered trademark) / Simulink. FIG. 2 shows the simulation conditions and motor constants. Sampling is performed 5 times per carrier half cycle and current control is executed. The carrier half cycle is 25 μs and the sampling period Tx is 5 μs.

[0058] With the d-axis current Id set to zero, the transient response was verified when only the q-axis current Iq was changed stepwise. FIG. 3 shows the three-phase voltage command of the OSDB control without compensation when there is no error between the parameters of the control model and the parameters of the motor, and FIG. 4 shows the step response waveform of the q-axis current. It can be seen that distortion has occurred in the voltage command due to the mismatch between the voltage command and the actual voltage, and the step response has deteriorated.

[0059] In contrast, FIGS. 5 and 6 are diagrams corresponding to FIGS. 3 and 4 in the case of current compensation according to the present embodiment. The distortion of the voltage command has been significantly improved. Also, the step response of the q-axis current Iq has almost no time delay, the overshoot has been reduced from 9.6% to 0.1%, and the steady-state deviation has also changed from 3.2% to 0%, and ideal results have been obtained in both the transient and steady states.

[0060] Next, we verified the case where there are errors in the parameters of the control model and the motor parameters. Figures 7 and 8 compare the performance between general DB control and OSDB control with compensation when there is a ±30% error in the dq-axis inductance. It can be seen that for the transient response, that is, the overshoot and the delay of the current response, when there are the parameter errors shown in Figures 7 and 8, the OSDB control with compensation prevents performance degradation compared to general DB control. Thus, the effects of the oversampling and current compensation method in this embodiment were confirmed.

[0061] As described above, according to this embodiment, the inverter unit 5 converts DC power into AC power and supplies it to the motor 6. The current detection unit 10 detects the current supplied to the motor 6. The control unit 13 controls the on / off of the semiconductor switching elements constituting the inverter unit 5. At this time, the current compensation calculation unit 9 samples the actual currents iRd and iRq every sampling period Tx set shorter than the switching period of the semiconductor switching elements, obtains the actual voltages VRd and VRq every sampling period Tx, and calculates the current compensation values idC and iqC from the differences between the time transitions iCd and iCq of the actual current and the actual currents iRd and iRq in the carrier half-cycle when the command voltages VCd and VCq are applied, and the differences between the command voltages VCd and VCq and the actual voltages VRd and VRq.

[0062] The OSDB current control unit 2 calculates the command voltages VCd and VCq from the ideal current obtained by adding the current compensation values idC and iqC to the dq-axis currents Iq and Id obtained by converting the three-phase currents sampled by the three-phase → dq coordinate conversion unit 11, the command currents Idref and Iqref, and the induced voltage generated in the windings of the motor 6. With this configuration, when performing OSDB current control, the influence of current fluctuations due to the applied voltage can be eliminated and voltage distortion can be compensated, so that more stable drive control can be performed than before.

[0063] (Second Embodiment) Hereinafter, the same parts as those in the first embodiment will be denoted by the same reference numerals and the description thereof will be omitted, and different parts will be described. Regarding the first term on the right side in Equation (16), as shown in Equation (18), only the previous value of the current compensation term may be used.

[0064]

Number

[0065] In this case, the control block diagram becomes as shown in FIG. 12, and the path for inputting the actual currents Iq and Id to the current compensation term calculation unit 9A becomes unnecessary. Therefore, the calculation amount in the current compensation term calculation unit 9A can be reduced.

[0066] (Other Embodiments) Regarding the carrier frequency, voltage value, motor constant, etc., they may be appropriately changed according to individual designs. Regarding the sampling frequency, it may be appropriately set within a range where the number of samplings within one carrier period is "4" or more.

[0067] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are included in the invention described in the claims and its equivalent scope.

Description of Reference Numerals

[0068] In the drawings, 1 is a speed control unit, 2 is an OSDB current control unit, 3 is a dq→3-phase coordinate conversion unit, 4 is a PWM generation unit and an actual voltage calculation unit, 5 is an inverter unit, 6 is a motor, 8 and 11 are 3-phase→dq coordinate conversion units, 9 is a current compensation term calculation unit, and 13 is a control unit.

Claims

1. A power converter having a plurality of semiconductor switching elements, converting DC power into AC power and supplying the AC power to an AC motor, a current detection unit that detects a current supplied to the AC motor, and a control unit that controls on / off of the plurality of semiconductor switching elements, and the control unit samples the actual current detected by the current detection unit at a sampling period set shorter than the switching period of the semiconductor switching element, obtains an actual voltage for each sampling period, and obtains a current compensation value from the difference between the time transition of the actual current and the actual current in a carrier half-cycle when a command voltage is applied and the difference between the command voltage and the actual voltage; a current compensation calculation unit, a motor current control device having a current control unit that calculates the command voltage from an ideal current obtained by adding the current compensation value to the sampled actual current, a command current, and an induced voltage generated in the winding of the AC motor.

2. The motor current control device according to claim 1, wherein the current compensation calculation unit obtains the current compensation value by adding the difference between the command voltage and the actual voltage to the current compensation value in the previous control cycle.

3. When converting DC power into AC power and supplying the AC power to an AC motor by using a plurality of semiconductor switching elements, detecting a current supplied to the AC motor, sampling the detected actual current at a sampling period set shorter than the switching period of the semiconductor switching element, obtaining an actual voltage for each sampling period, obtaining a current compensation value from the difference between the time transition of the actual current and the actual current in a carrier half-cycle when a command voltage is applied and the difference between the command voltage and the actual voltage, a motor current control method for calculating the command voltage from an ideal current obtained by adding the current compensation value to the sampled actual current, a command current, and an induced voltage generated in the winding of the AC motor.

4. The motor current control method according to claim 3, wherein the current compensation value is obtained by adding the difference between the command voltage and the actual voltage to the current compensation value in the previous control cycle.

Images