Motor control method and motor control device

Abstract

To implement sensorless control capable of ensuring the estimation accuracy of a magnetic pole position and control stability of a motor even when the frequency of a high-frequency signal to be superimposed is variable.SOLUTION: A motor control method for superimposing a high-frequency signal on a basic power command value to determine a power command value, estimating a magnetic pole position θ of a motor 200 based on a current detection values id_det, iq_det, and controlling the motor 200 based on the estimated magnetic pole position θ and the power command value includes superimposing a high-frequency signal including a plurality of frequency components on a power command value, calculating a peak frequency fk at which the intensity of the response current peaks, extracting a high-frequency current component corresponding to a maximum peak frequency fmax from the current detection value, estimating the magnetic pole position θ based on the extracted high-frequency current component, calculating corrected current detection values i'd_det, i'q_det by removing the high-frequency current component from the current detection value, and calculating a basic power command value based on the corrected current detection values.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a motor control method and a motor control device.

Background Art

[0002] Conventionally, a high-frequency signal (high-frequency voltage) having a frequency higher than the frequency of the driving power is superimposed on the driving power (basic power command value) applied to drive a motor and supplied to the motor, and the high-frequency current that is the response is extracted and processed to estimate the rotor phase (magnetic pole position) of the motor. A control method (sensorless control) is known. In particular, Patent Document 1 discloses a control method for reducing electromagnetic noise from a motor by making the frequency of the superimposed high-frequency voltage variable.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, when the frequency of the superimposed high-frequency voltage is made variable, the high-frequency current component included in the response current may interfere with the basic current determined according to the output (driving force) required for the motor, resulting in a decrease in control stability.

[0005] Therefore, an object of the present invention is to realize sensorless control that can ensure the estimation accuracy of the magnetic pole position and the control stability of the motor even when the frequency of the superimposed high-frequency signal is made variable.

Means for Solving the Problems

[0006] According to an aspect of the present invention, a power command value is determined by superimposing a high-frequency signal on a predetermined basic power command value, a response current to the input of the power command value is detected, a magnetic pole position of a motor is estimated based on the obtained current detection value, and a motor control method is provided for controlling the motor based on the estimated magnetic pole position and the power command value.

[0007] In particular, in this motor control method, a high-frequency signal including a plurality of frequency components is superimposed on the basic power command value, a peak frequency representing a frequency at which the intensity of the response current peaks is calculated, a high-frequency current component corresponding to the peak frequency is extracted from the current detection value, the magnetic pole position is estimated based on the high-frequency current component, a corrected current detection value obtained by removing the high-frequency current component from the current detection value is calculated, and the power command value is calculated based on the corrected current detection value.

Advantages of the Invention

[0008] According to the present invention, even if the frequency of the superimposed high-frequency signal is variable, it is possible to ensure the estimation accuracy of the magnetic pole position and the control stability of the motor.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0010] Hereinafter, each embodiment of the present invention will be described with reference to the drawings.

[0011] [First Embodiment] FIG. 1 is a block diagram showing the configuration of the motor control device 100. As shown in FIG. 1, as the motor control device 100, instead of a configuration that measures the magnetic pole position (rotor position) of the motor 200 and the rotational speed of the motor 200 (hereinafter referred to as "rotational speed N m "), an apparatus that calculates based on an estimation algorithm described later and controls the power supplied to the motor 200 by operating the inverter 18 using the calculated estimated value is assumed. That is, the motor control system including the motor control device 100, the inverter 18, and the motor 200 of the present embodiment can be configured as a sensorless system. In particular, as the motor 200 to be controlled, an in-vehicle driving motor or a power generation motor is assumed. Further, the motor 200 is, for example, a three-phase AC IPM (Interior Permanent Magnet) motor, and the q-axis inductance L q and the d-axis inductance L d are configured as a salient pole motor showing mutually different values.

[0012] The motor control device 100 includes a current command generation unit 11, a first voltage command generation unit 12, a second voltage command generation unit 13, a final voltage command generation unit 14, a control mode signal generation unit 15, a first coordinate conversion unit 16, a PWM conversion unit 17, a phase / rotation speed estimation unit 19, a high-frequency voltage command generation unit 20, a second coordinate conversion unit 23, and a high-frequency signal removal unit 25. The motor control device 100 is configured by a computer (controller) having a program for realizing the functions of each unit. Also, the hardware constituting the computer may be composed of one or more units.

[0013] The current command generation unit 11 generates (calculates) and outputs a d-axis current command value i * , a DC voltage V dc , and a rotation speed estimated value N m ' based on the torque command value T d * and the q-axis current command value i q * .

[0014] Here, the torque command value T * is the target value of the output torque of the motor 200 defined by the upper control device. In particular, when the motor 200 is used as the driving source for vehicle running, it is determined to an appropriate value according to the required driving force (such as the operation amount for the accelerator pedal). Also, the DC voltage V dc is the output voltage of the battery 21 for supplying the driving power of the motor 200. The DC voltage V dc is detected by the voltage sensor 22. Note that the DC voltage V dc may be acquired as an estimated value obtained by a battery controller (not shown) or the like instead of the detected value by the voltage sensor 22. Further, the rotation speed estimated value N m ' is an estimated value of the rotation speed N m calculated according to the control state of the motor 200. The rotation speed estimated value N m ' is calculated (estimated) by the phase / rotation speed estimation unit 19 according to an estimation algorithm (high-frequency voltage injection method) described later.

[0015] The d-axis current command value i d* is the command value for the d-axis current i of the motor 200. The q-axis current command value i d is the command value for the q-axis current i of the motor 200. Hereinafter, for the sake of notation simplification, the d-axis current i q * and the q-axis current i q are collectively referred to as the dq-axis current i d and i q and the d-axis current command value i d , i q and the q-axis current command value i d * are collectively referred to as the dq-axis current command value i q * d * q , i * * .

[0016] The first voltage command generation unit 12 calculates and outputs the first d-axis voltage command value V * , the DC voltage V dc , the estimated rotational speed N m ′, the dq-axis current command value i d * , i q * , and the detected values of the dq-axis current i d , i q . That is, the first voltage command generation unit 12 calculates the first d-axis voltage command value V d1 * and the first q-axis voltage command value V q1 * (hereinafter also referred to as the "first voltage command value V d1 * , V q1 * ").

[0017] The first voltage command value V d1 * , V q1 * is a voltage command value for controlling the motor 200 by so-called current vector control. That is, the first voltage command generation unit 12 calculates, for example, according to the following formula (1), the dq-axis current i d , i q and the dq-axis current command value i d * , iq * The deviation (i d -i d * , i q -i q * ) Based on PI control and the decoupling of the dq axes, the first voltage command values V d1 * , V q1 * are calculated.

[0018]

Number

[0019] Note that "s" in Equation (1) is a differential operator. Also, "K p1 " is the proportional gain, and "K i1 " is the integral gain. Furthermore, "V d-dcpl * " is the d-axis interference voltage, and "V q-dcpl * " is the q-axis interference voltage. In this embodiment, the first voltage command generation unit 12 has an interference voltage table (not shown) that associates the torque command value T * , the DC voltage V dc , and the estimated rotational speed N m ' with the d-axis interference voltage V d-dcpl * and the q-axis interference voltage V q-dcpl * . Therefore, the first voltage command generation unit 12 calculates the d-axis interference voltage V * , the DC voltage V dc , and the q-axis interference voltage V m ' corresponding to the torque command value T d-dcpl * and the q-axis interference voltage V q-dcpl * by referring to this interference voltage table. The interference voltage table is preset by experiments or simulations, etc.

[0020] Note that when the first voltage command generation unit 12 generates the first voltage command values V d1 * , V q1* The dq-axis currents i d , i q used in the operation are the detected values of the currents flowing through the motor 200 and are obtained from the second coordinate conversion unit 23. The second coordinate conversion unit 23 performs coordinate conversion using, for example, the phase estimation value θ′ output from the phase / rotation speed estimation unit 19 on the currents i u , i v , i w of each phase of the motor 200 detected by the current sensor 24 to calculate the dq-axis currents i d , i q . Specifically, the second coordinate conversion unit 23 calculates the dq-axis currents i d , i q according to the following formula (2).

[0021]

Equation

[0022] Also, in this embodiment, the current sensor 24 detects the U-phase current i u and the V-phase current i v of the motor 200, and the second coordinate conversion unit 23 obtains the W-phase current i w by calculation according to the following formula (3).

[0023]

Equation

[0024] Hereinafter, the values of the dq-axis currents i d , i q calculated based on the above formulas (2) and (3) are regarded as the detected values of the dq-axis currents i d , i q . When specifically indicating that it is a detected value, it is denoted as "current detection value i d_det , i q_det ", "d-axis current detection value i d_det ", or "q-axis current detection value i q_det ".

[0025] The second voltage command generation unit 13 is the torque command value T *, DC voltage V dc , estimated rotation speed N m ′, and detected current value i d_det , i q_det , based on these, the second d-axis voltage command value V d2 * and the second q-axis voltage command value V q2 * are calculated. In the following, the second d-axis voltage command value V d2 * and the second q-axis voltage command value V q2 * are appropriately referred to as "second voltage command value V d2 * , V q2 * ".

[0026] The second voltage command value V d2 * , V q2 * is a voltage command value for controlling the motor 200 by so-called voltage phase control. That is, the second voltage command generation unit 13 calculates the voltage norm command value V a which is a command value for the voltage norm V a * and the voltage phase command value α * which is a command value for the voltage phase α, and uses these to calculate the second voltage command value V d2 * , V q2 * .

[0027] Specifically, the second voltage command generation unit 13 calculates the voltage norm command value V dc based on, for example, the DC voltage V * and the modulation ratio command value MF a * which is a command value for the modulation ratio, according to the following formula (4).

[0028]

Equation

[0029] Next, the second voltage command generation unit 13 is the torque command value T* , DC voltage V dc , and the estimated rotational speed N m ′, based on this, the target voltage phase value α ff * is calculated. The target voltage phase value α ff * is the target value of the voltage phase α by feedforward control. In this embodiment, the second voltage command generation unit 13 has a torque command value T * , DC voltage V dc , and the estimated rotational speed N m ′, and a voltage phase target value table (not shown) that associates the voltage phase target value α ff * with them. Therefore, the second voltage command generation unit 13 refers to this voltage phase target value table to calculate the voltage phase target value α * corresponding to the torque command value T dc , DC voltage V m ′, and the estimated rotational speed N ff * . Note that the voltage phase target value table is preset by experiments or simulations, etc.

[0030] Furthermore, the second voltage command generation unit 13 calculates an estimated torque value T d_det , i q_det which is an estimated value of the torque (hereinafter also simply referred to as "torque T") output by the motor 200, based on the detected current value i m ′ and the estimated rotational speed N est . In particular, in this embodiment, the second voltage command generation unit 13 refers to a torque estimated value table (not shown) that associates the dq-axis current i d , i q and the estimated rotational speed N m ′ with the torque estimated value T est to obtain the torque estimated value T d_det , i q_det corresponding to the detected current value i m ′ and the estimated rotational speed N est . Note that the torque estimated value table is preset by experiments or simulations, etc.

[0031] Also, the second voltage command generation unit 13 calculates a torque command value T * and a torque estimated value T est to calculate a voltage phase correction value α fb * . The voltage phase correction value α fb * is a correction value for the voltage phase target value α ff * and is calculated by feedback control. For example, the second voltage command generation unit 13 calculates the voltage phase correction value α * using PI control based on the deviation (T est fin * est - T fb ) between the torque command value T * p2 and the torque estimated value T

[0032]

Equation

[0033] Note that "K p2 " in Equation (5) is the proportional gain, and "K i2 " is the integral gain.

[0034] Then, the second voltage command generation unit 13 adds the voltage phase correction value α ff * calculated as described above to the voltage phase target value α fb * to calculate the voltage phase command value α * .

[0035] Next, the second voltage command generation unit 13 calculates the second voltage command value V a * and V * by vector conversion based on the voltage norm command value V d2 * and the voltage phase command value α q2 * as shown in the following Equation (6).

[0036] [Number]

[0037] The final voltage command generation unit 14 selects one of the first voltage command values V d1 * , V q1 * and the second voltage command value V d2 * , V q2 * based on the control mode signal Msw input from the control mode signal generation unit 15, and outputs it as the final voltage command value V d * , V q * . That is, the final voltage command generation unit 14 selectively switches the control mode between the current vector control mode using the first voltage command value V d1 * , V q1 * and the voltage phase control mode using the second voltage command value V d2 * , V q2 * according to the control mode signal Msw.

[0038] And in this embodiment, the final voltage command generation unit 14 adds (superimposes) the high-frequency voltage V d1 * , V q1 * or the second voltage command value V d2 * , V q2 * input from the high-frequency voltage command generation unit 20 to the selected first voltage command value V dh * , V qh * and outputs it.

[0039] The control mode signal generation unit 15 uses the DC voltage V dc and the final voltage command value V d * , V q *Based on this, a control mode signal Msw is generated.

[0040] Specifically, the control mode signal generation unit 15 calculates a modulation ratio MF according to the following formula (7), based on the DC voltage V dc and the final voltage command value V d * , V q * Based on this, a modulation ratio MF is calculated.

[0041]

Equation

[0042] Also, the control mode signal generation unit 15 compares the calculated modulation ratio MF with a predetermined threshold value (modulation ratio threshold TH MF ). Then, when the modulation ratio MF is less than the modulation ratio threshold TH MF , for example, the control mode signal generation unit 15 generates and outputs a control mode signal M d1 * , V q1 * (current vector control mode) to be selected. On the other hand, when the modulation ratio MF is greater than or equal to the modulation ratio threshold TH sw , for example, the control mode signal generation unit 15 generates and outputs a control mode signal Msw that selects the second voltage command value V MF d2 * , V q2 * (voltage phase control mode). (voltage phase control mode) is generated and output.

[0043] Note that the modulation ratio threshold TH MF used for the switching determination from the current vector control mode to the voltage phase control mode and the modulation ratio threshold TH MF used for the switching determination from the voltage phase control mode to the current vector control mode may be set to different values. In this case, since hysteresis can be provided for the switching between the current vector control mode and the voltage phase control mode, frequent switching (so-called chattering) between each mode is suppressed.

[0044] The first coordinate conversion unit 16 calculates the final voltage command value V through coordinate conversion using, for example, the phase estimation value θ′ output by the phase / rotation speed estimation unit 19. d * ,V q * to three-phase voltage command values V u * ,V v * ,V w * . Specifically, the first coordinate conversion unit 16 calculates the three-phase voltage command values V u * ,V v * ,V w * according to the following formula (8).

[0045]

Equation

[0046] The PWM conversion unit 17 generates a PWM (Pulse Width Modulation) signal for driving the power elements of the inverter 18 based on the DC voltage V dc and the three-phase voltage command values V u * ,V v * ,V w * . Specifically, the PWM conversion unit 17 generates power element drive signals D u * ,D v * ,D w * corresponding to the three-phase voltage command values V uu * ,D ul * ,D vu * ,D vl * ,D wu * ,D wl *and inputs this to the inverter 18. When generating the PWM signal, the PWM conversion unit 17 can perform so-called dead time compensation processing and voltage utilization rate improvement processing.

[0047] The inverter 18 switches the power elements in accordance with the PWM signal to generate a DC voltage V dc is the pseudo AC voltage V u ,V v ,V w and input to each of the UVW phases of the motor 200. As a result, the motor 200 receives the torque command value T * The torque T is controlled to be output in accordance with the torque.

[0048] The phase / rotation speed estimation unit 19 estimates the current detection value i d_det ,i q_det and the second voltage command value V d2 * ,V q2 * Based on this, the estimated rotation speed N m The phase estimate θ′ is an estimate of the phase θ that defines the rotor position of the motor 200. In particular, the phase / rotation speed estimator 19 calculates and outputs the q-axis inductance L q and d-axis inductance L d The high-frequency voltage application method, which is an estimation algorithm for the magnetic pole position that utilizes the difference in m ' and phase estimate θ'.

[0049] Specifically, first, the phase / rotation speed estimation unit 19 calculates the current detection value i d_det ,i q_det From the phase error estimate θ γ Here, the phase error estimate θ γ ' is the phase error θ between the dq-axis coordinate system and the predetermined two-axis control coordinate system (γδ-axis coordinate system) γ is an estimate.

[0050] First, as a premise, the first voltage command value V d1 * ,Vq1 * or a second voltage command value V d2 * , V q2 * the high-frequency voltage V superimposed on dh * , V qh * draws an elliptical orbit having a major axis and a minor axis in directions respectively coinciding with those of the γ-axis and the δ-axis. Then, the high-frequency voltage V dh * , V qh * the high-frequency current component i responding to dh , i qh also draws an elliptical orbit in the γδ-axis coordinate system. Here, when both the phase error θ dh , i qh between the dq-axis coordinate system and the γδ-axis coordinate system described above and the major-axis phase θ re are minute, they can be regarded as coinciding with each other. Therefore, if the major-axis phase θ γ and the major-axis phase θ γe are obtained, this can be used as an estimated value of the phase error θ γe (hereinafter, also referred to as "estimated phase error θ γ ′"). γ

[0051] Therefore, the phase / rotation speed estimation unit 19 calculates the estimated phase error θ d_det , i q_det ′ by executing the following arithmetic algorithm for determining the major-axis phase θ γe from the current detection value i γ .

[0052] Specifically, the phase / rotation speed estimation unit 19 performs filter processing (hereinafter, "first-half filter processing") on the current detection value i d_det , i q_det using a band-pass filter or the like, so as to obtain the high-frequency component (hereinafter, "high-frequency current component i d_det , i q_det contained in the current detection value i dh_det , i qh_det ​Calculate (referred to as ")".

[0053] Next, the phase / rotation speed estimation unit 19 calculates the in-phase component (real part) and the quadrature component (imaginary part) of the high-frequency current components i dh_det , i qh_det according to the following equations (9) and (10). p , s´ p and the quadrature component (image component) [c´ n , s´ n .

[0054]

Equation

[0055] Furthermore, the phase / rotation speed estimation unit 19 performs a predetermined filter process (hereinafter referred to as "latter half filter process") on the obtained in-phase component [c´ p , s´ p and quadrature component [c´ n , s´ n to obtain the final in-phase component [c p , s p and quadrature component [c n , s n .

[0056] The details of the above-mentioned former half filter process and latter half filter process will be described later.

[0057] The in-phase component [c p , s p and quadrature component [c n , s n obtained as described above have a symmetric positional relationship (mirror image relationship) with respect to the major axis phase θ re (phase error estimated value θ γ ′) in the γδ-axis coordinate system. Therefore, the phase / rotation speed estimation unit 19 can calculate the phase error estimated value θ p , s p and quadrature component [c n , s n according to the following equation (11). γ ′.

[0058] [Numerical]

[0059] Also, the phase / rotation speed estimation unit 19 calculates an estimated electrical angular velocity ω′, which is an estimated value of the electrical angular velocity ω of the motor 200, by PI control of the following equation (12) based on the estimated phase error θ γ ′.

[0060] [Numerical]

[0061] Note that "K" in Equation (12) p3 " is the proportional gain, and "K" i3 " is the integral gain.

[0062] Furthermore, the phase / rotation speed estimation unit 19 obtains an estimated phase value θ′ by executing the operation of the following equation (13) (integration of the estimated electrical angular velocity ω′).

[0063] [Numerical]

[0064] Also, the phase / rotation speed estimation unit 19 calculates an estimated rotation speed N m ′ [rpm] by unit-converting the estimated electrical angular velocity ω′ [rad / sec] obtained by the above equation (13).

[0065] On the other hand, the high-frequency voltage command generation unit 20 takes the torque command value T * as an input and generates and outputs the above-described high-frequency voltages V dh * , V qh * .

[0066] More specifically, the high-frequency voltage command generation unit 20 generates a high-frequency voltage V represented by, for example, the following equation (14) dh *, V qh * is generated.

[0067]

Number

[0068] Note that "V h " in Equation (14) is the fundamental amplitude of the high-frequency voltage V dh * , V qh * ; "ω h " is the angular frequency. Also, "K h " (-1 ≤ K h < 0, 0 < K h ≤ 1) is the coefficient that determines the locus of the high-frequency voltage V γ in the two-axis control coordinate system (γδ-axis coordinate system) having a predetermined phase difference (phase error θ dh * , V qh * with respect to the dq-axis coordinate system (the ratio of the amplitude of the q-axis high-frequency voltage V dh * to the d-axis high-frequency voltage V qh * ). Note that the fundamental amplitude V h and the coefficient K h are set to appropriate values in advance.

[0069] The high-frequency signal removal unit 25 calculates and outputs the corrected current detection values i´ d_det , i´ q_det by performing a filter process described later on the current detection values i d_det , i´ q_det .

[0070] Here, the control in which the angular frequency ω dh * , V qh * of the superimposed high-frequency voltage V h (frequency f h ) is constant is typical. On the other hand, the high-frequency voltage V dh * , V qh *For the purpose of reducing the electromagnetic noise of the motor 200 generated by the application of [the symbol], the angular frequency ω h (frequency f h ) is changed while applying high-frequency voltages V dh * , V qh * . Control may be employed where such application is made. And also in the present embodiment, it is premised that high-frequency voltages V dh * , V qh * are applied while changing the angular frequency ω h (frequency f h ).

[0071] In particular, in the present embodiment, an example will be described in which the angular frequency ω dh * , V qh * of high-frequency voltages V h (frequency f h ) is switched between a predetermined first angular frequency ω h1 (first frequency f1) and a second angular frequency ω h2 (second frequency f2) different therefrom at a predetermined time interval (period). More specifically, the frequency f dh * , V qh * of high-frequency voltages V h is set to the first frequency f1 for a predetermined number of times (number of cycles) of the high-frequency voltages V dh * , V qh * , and then a cycle is repeated in which the frequency f dh * , V qh * of high-frequency voltages V h is set to the second frequency f2 for a predetermined number of times (number of cycles) of the high-frequency voltages V dh * , V qh * . Note that hereinafter, the number of cycles in which the frequency f h is set to the first frequency f1 is described as the first coefficient N1, and the frequency f hThe number of periods set to the second frequency f2 is described as the second coefficient N2.

[0072] Hereinafter, under the above premises, the details of the processing in the phase / rotation speed estimation unit 19 and the processing in the high-frequency signal removal unit 25 will be described.

[0073] (Phase / Rotation Speed Estimation Unit) The phase / rotation speed estimation unit 19 of the present embodiment defines a band-pass filter for executing the above-described first half filter processing.

[0074] Specifically, the phase / rotation speed estimation unit 19 calculates the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2 from the current detection values i dh * , V qh * indicating the current value that responds when superimposed. For example, the phase / rotation speed estimation unit 19 applies a frequency analysis algorithm such as FFT (Fast Fourier Transform) to the current detection values i d_det , i q_det to obtain the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2. d_det , i q_det Next, the phase / rotation speed estimation unit 19 calculates the maximum peak frequency f

[0075] from the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2 according to the following formula (15). max

[0076]

Equation

[0077] Here, the maximum peak frequency f max is the frequency f (hereinafter referred to as the "peak frequency f nd ") at which the intensity of the response current (hereinafter referred to as the "response current intensity I k ") takes a peak (maximum value) among a plurality of frequencies f at which the response current intensity I ndmeans the one that takes the largest value. Also, the peak frequency f in this embodiment k (k = 1, 2, 3 ···) is a term that includes the maximum peak frequency f max and the frequencies of the sideband waves having peaks other than the maximum peak. In particular, the peak frequency f k (k = 1, 2, 3 ···) can be determined by the following equation (16).

[0078]

Equation

[0079] That is, according to equation (16), in addition to the maximum peak frequency f max , each peak frequency f nd having a peak of the response current intensity I other than the maximum peak as a sideband wave can be obtained. k

[0080] On the other hand, in this embodiment, the phase / rotation speed estimation unit 19 defines a band-pass filter that extracts the current component corresponding to the maximum peak frequency f max . Then, by processing the current detection values i d_det , i q_det with the band-pass filter, the high-frequency current components i dh_det , i qh_det are calculated. More specifically, the phase / rotation speed estimation unit 19 executes the operation represented by the following equation (17) to obtain the high-frequency current components i d_det , i q_det from the current detection values i dh_det , i qh_det .

[0081]

Equation

[0082] Note that "s" in equation (17) represents the Laplace operator, and "a" is the maximum peak frequency f max ​Indicates a coefficient (positive real number) representing the width of the passband centered around. In particular, in the band-pass filter of Equation (17), the larger the coefficient a, the wider the passband centered around the maximum peak frequency f max becomes (the change in gain near the maximum peak frequency f max becomes gentle), and the smaller the coefficient a, the narrower the passband (the change in gain near the maximum peak frequency f max becomes steep).

[0083] Therefore, in the first half of the filter processing, by using the band-pass filter of Equation (17), the high-frequency current components i d_det , i q_det from the current detection values can extract the components corresponding to the maximum peak frequency f max to determine the high-frequency current components i dh_det , i qh_det .

[0084] Next, the phase / rotation speed estimation unit 19 determines a filter for performing the latter half of the filter processing, and processes the positive-phase components [c´ p , s´ p and the negative-phase components [c´ n , s´ n with the filter to obtain the final positive-phase components [c p , s p and the negative-phase components [c n , s n . More specifically, the phase / rotation speed estimation unit 19 executes the operation represented by the following Equation (18) to obtain the final positive-phase components [c p , s p and the negative-phase components [c n , s n .

[0085]

Equation

[0086] Note that "b0" in Equation (18) represents a predetermined coefficient (positive real number). Also, "b1" is the maximum peak frequency f maxIndicates a coefficient (positive real number) representing the width of the stopband centered around twice the frequency. In particular, in the filter of Equation (18), the larger the coefficient b1, the narrower the stopband width (the change in gain near the center frequency becomes steeper), and the smaller the coefficient b1, the wider the stopband width (the change in gain near the center frequency becomes gentler).

[0087] Therefore, in the latter half filter process, by using the filter represented by the above Equation (18), the in-phase components [c´ p ,s´ p and the anti-phase components [c´ n ,s´ n remove the components corresponding to the maximum peak frequency f max to determine the in-phase components [c p ,s p and the anti-phase components [c n ,s n .

[0088] Note that, for example, by the process using a low-pass filter, the in-phase components [c´ p ,s´ p and the anti-phase components [c´ n ,s´ n can sufficiently attenuate the frequency gain of the applied high-frequency voltage V dh * ,V qh * at a frequency f h (i.e., twice or more the first frequency f1 or the second frequency f2). In the latter half filter process, the low-pass filter may be used instead of the filter represented by Equation (18).

[0089] (High-frequency signal removal unit) The high-frequency signal removal unit 25 determines a filter that removes the high-frequency current components i max using the maximum peak frequency f d_det ,i q_det corresponding to the maximum peak frequency f max from the current detection values i dh_det ,i qh_det . Then, the high-frequency signal removal unit 25 uses the current detection values id_det , i q_det By processing the current detection value i with the filter, a corrected current detection value i' d_det , i' q_det is calculated. More specifically, the high-frequency signal removal unit 25 calculates the corrected current detection value i' d_det , i q_det by performing the operation represented by the following equation (19) on the current detection value i d_det , i' q_det .

[0090]

Equation

[0091] Note that "c0" and "c1" in Equation (19) represent coefficients indicating the width of the stop band centered on the maximum peak frequency f max . In particular, the coefficient c0 is defined as a real number greater than or equal to 0, the coefficient c1 is defined as a positive real number, and the relationship c0 < c1 holds between them. Furthermore, in the filter of Equation (19), the larger the coefficient c1 is with respect to the coefficient c0, the narrower the stop band width becomes (the change in gain near the maximum peak frequency f max becomes steeper), and the closer the coefficient c1 is to the coefficient c0, the wider the stop band width becomes (the change in gain near the maximum peak frequency f max becomes gentler).

[0092] In this way, by processing the current detection value i d_det , i q_det with the filter represented by the above Equation (19), a corrected current detection value i' d_det , i q_det from which the component corresponding to the maximum peak frequency f max is removed can be determined. That is, in the first voltage command generation unit 12, a corrected current detection value i' d_det , i' q_det from which the component (high-frequency current component i max corresponding to the maximum peak frequency f dh , i qh ) is removed can be obtained. d_det , i' q_detBased on this, the first voltage command value V d1 * , V q1 * can be determined.

[0093] Next, a series of processes in the phase / rotation speed estimation unit 19 and the high-frequency signal removal unit 25 will be described together.

[0094] FIG. 2 is a flowchart showing the flow of the process in the phase / rotation speed estimation unit 19. Note that each of the following processes is repeatedly executed at a predetermined operation cycle.

[0095] As shown in the figure, first, in step S10, the current detection values i d_det , i q_det , the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2 are acquired.

[0096] In step S20, from the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2, the operation represented by Equation (15) is executed to calculate the maximum peak frequency f max .

[0097] In step S30, the current detection values i d_det , i q_det are processed by the band-pass filter of Equation (17) to calculate the high-frequency current components i dh_det , i qh_det .

[0098] In step S40, for the high-frequency current components i dh_det , i qh_det , the operations represented by Equation (9) and Equation (10) are executed to obtain the in-phase components [c´ p , s´ p and the quadrature-phase components [c´ n , s´ n , and the in-phase components [c´ p , s´ p and the quadrature-phase components [c´ n , s´ n are processed by the filter of Equation (18) to obtain the in-phase components [c p , s pand the reverse-phase component [c n , s n are calculated.

[0099] In step S50, for the positive-phase component [c p , s p and the reverse-phase component [c n , s n , the operation represented by equation (11) is executed to calculate the phase error estimated value θ γ ′.

[0100] In step S60, for the phase error estimated value θ γ ′, the operation represented by equation (12) is executed to calculate the phase estimated value θ′. Further, for the obtained phase estimated value θ′, the operation represented by equation (13) is executed to calculate the rotation speed estimated value N m ′.

[0101] Figure 3 is a flowchart showing the processing flow in the high-frequency signal removal unit 25. Each of the following processes is repeatedly executed at a predetermined operation cycle.

[0102] As shown in the figure, first in step T10, the current detection value i d_det , i q_det , the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2 are acquired.

[0103] In step T20, from the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2, the operation represented by equation (15) is executed to calculate the maximum peak frequency f max .

[0104] In step T30, the current detection value i d_det , i q_det is processed by the filter of equation (19) to calculate the corrected current detection value i′ d_det , i′ q_det . (An example of the control result)

[0105] FIG. 4 is a timing chart showing an example of control results by the motor control methods of Comparative Example 1 and Example 1. In FIG. 4, a constant torque command value T * is given, and the frequencies f dh * , V qh * of the high-frequency voltages V h are made variable, and the time changes of the phase error estimated value θ m ′, the rotational speed N γ ′, and the torque T are shown when the rotational speed N m is swept.

[0106] Specifically, FIG. 4(a) shows the control results (Comparative Example 1) when the processing in the high-frequency signal removal unit 25 is not executed. On the other hand, FIG. 4(b) shows the control results (Example 1) when the component corresponding to the maximum peak frequency f d_det , i q_det is removed from the current detection values i max in the high-frequency signal removal unit 25.

[0107] In Comparative Example 1 shown in FIG. 4(a), after time t1, the phase error estimated value θ γ ′ hunts, the estimation error of the rotational speed N m becomes large, and the actual output torque (estimated torque) of the motor 200 does not respond (cannot be driven) to the torque command value T * .

[0108] In contrast, in Example 1 shown in FIG. 4(b), hunting of the phase error estimated value θ γ ′ is not seen, the estimation accuracy of the rotational speed N m is maintained, and the estimated torque is also maintained at the torque command value T * . That is, the magnetic pole position (phase θ) is estimated using the high-frequency current components i d_det , i q_det extracted from the current detection values i max corresponding to the maximum peak frequency f dh_det , i qh_det , and at the same time, the current detection values i d_det , i q_det from the maximum peak frequency f maxThe corrected current detection value i´ from which the component corresponding thereto has been removed d_det , i´ q_det is used to determine the final voltage command value V d * , V q * By doing so, it can be seen that sensorless driving can be appropriately performed even when the frequency f dh * , V qh * of the high-frequency voltage V h is changed.

[0109] The effects of the motor control method of the present embodiment described above will be collectively described.

[0110] According to the present embodiment, a high-frequency signal (high-frequency voltage V d1 * , V q1 * or the second voltage command value V d2 * , V q2 * ) is superimposed on a predetermined basic power command value (first voltage command value V dh * , V qh * ) to determine a power command value (final voltage command value V d * , V q * ), the detection value of the response current with respect to the input of the power command value (current detection value i d_det , i q_det ) is obtained, the magnetic pole position (θ) of the motor 200 is estimated based on the obtained current detection value i d_det , i q_det , and a motor control method for controlling the motor 200 based on the estimated magnetic pole position (θ) and the power command value (V d * , V q * ) is provided.

[0111] In this motor control method, a high-frequency voltage V dh * , V qh *is superimposed on the power command value, and the intensity of the response current (response current intensity I nd ) represents the frequency f at which the peak occurs. k (In this embodiment, the maximum peak frequency f max ) and calculate the current detection value i d_det ,i q_det to maximum peak frequency f max The high-frequency current component i dh_det ,i qh_det The extracted high-frequency current component i dh_det ,i qh_det The magnetic pole position (θ) is estimated based on the current detection value i d_det ,i q_det From the high frequency current component i dh_det ,i qh_det Corrected current detection value i' d_det ,i´ q_det Calculate the estimated magnetic pole position (θ) and the corrected current detection value i´ d_det ,i´ q_det Based on this, the basic power command value (particularly the first voltage command value V d1 * ,V q1 * ) is calculated.

[0112] This allows the superimposed high-frequency voltage V dh * ,V qh * Even if the current detection value i d_det ,i q_det to maximum peak frequency f max The high-frequency current component i dh_det ,i qh_det Since the magnetic pole position (θ) is estimated by appropriately extracting the current detection value i d_det ,i q_det to the corresponding high frequency current component i dh_det ,i qh_det The corrected current detection value i' with d_det ,i´ q_det Since the power command value of the motor 200 can be controlled by this, it is possible to prevent a decrease in the control stability of the motor 200 (such as the occurrence of torque fluctuations).

[0113] In particular, in this embodiment, the high-frequency voltage V dh * , V qh * of frequency f h is switched while being superimposed on the basic power command value (the first voltage command value V d1 * , V q1 * or the second voltage command value V d2 * , V q2 * ). Then, the peak frequency f k is calculated as the maximum peak frequency f dh * , V qh * at which the response current intensity I nd becomes maximum, based on the frequencies (f1, f2) to be switched and the times (N1, N2) during which the high-frequency voltages V max corresponding to the respective frequencies (f1, f2) are superimposed. Then, the components corresponding to the maximum peak frequency f d_det , i q_det are removed from the current detection values i max to calculate the corrected current detection values i´ d_det , i´ q_det .

[0114] Thereby, when changing the frequency f dh * , V qh * of the superimposed high-frequency voltages V h it is possible to accurately remove the frequency component that has the greatest influence on the drive control of the motor 200 from the response current (i.e., the current detection values i d_det , i q_det ). Therefore, it is possible to more reliably prevent a decrease in the control stability of the motor 200.

[0115] More specifically, in this embodiment, each frequency f to be switched includes a predetermined first frequency f1 and a second frequency f2 different from the first frequency f1. And the high-frequency voltages V dh * , Vqh * The frequency f h is switched between the first frequency f1 and the second frequency f2. In particular, the high-frequency voltage V of the first frequency f1 to be superimposed (f1) dh * , V (f1) qh * is set to the first coefficient N1 indicating the number of periods (times), and the high-frequency voltage V of the second frequency f2 to be superimposed (f2) dh * , V (f2) qh * is set to the second coefficient N2 indicating the number of periods (times). Further, the maximum peak frequency f max is calculated based on the first frequency f1, the second frequency f2, the first coefficient N1, and the second coefficient N2.

[0116] Thereby, when the frequency f of the high-frequency voltage V dh * , V qh * to be superimposed is periodically switched between the first frequency f1 and the second frequency f2, a more specific logic for calculating the maximum peak frequency f h can be realized. max dh

[0117] Furthermore, in the present embodiment, a motor control device 100 suitable for executing the above motor control method is provided.

[0118] This motor control device 100 includes a superimposing unit (20) that superimposes a high-frequency voltage V dh * , V qh * including a plurality of frequency components on a power command value, and a peak frequency f representing the frequency f at which the intensity of the response current (response current intensity I nd ) reaches a peak k (in this embodiment, particularly the maximum peak frequency f max ) is calculated by a frequency calculation unit (19, 25), and the maximum peak frequency f is obtained from the current detection values i d_det , i q_det ​max The high-frequency current component i corresponding thereto dh_det , i qh_det is extracted, and the magnetic pole position (θ) is estimated based on the high-frequency current component i dh_det , i qh_det by an estimation unit (19), and the correction current detection value i d_det , i q_det is obtained from the current detection value i dh_det , i qh_det by removing the high-frequency current component i d_det , i q_det to calculate a correction current calculation unit (25), and the estimated magnetic pole position (θ) and the correction current detection value i´ d_det , i´ q_det Based on this, a basic power command value (especially the first voltage command value V d1 * , V q1 * ) is calculated by a power command value calculation unit (12).

[0119] [Second Embodiment] Hereinafter, the second embodiment will be described. In this embodiment, a part of the arithmetic logic in the high-frequency signal removal unit 25 is different from that in the first embodiment. In particular, the high-frequency signal removal unit 25 of this embodiment extracts the current detection value i d_det , i q_det from the maximum peak frequency f max and removes components corresponding to a plurality of peak frequencies f k including the maximum peak frequency f d_det , i´ q_det to obtain a correction current detection value i´

[0120] More specifically, the high-frequency signal removal unit 25 uses the maximum peak frequency f max calculated by the formula (15) to calculate the maximum peak frequency f max and the peak frequency f k (k = 1, 2...) consisting of the frequency of the sideband wave having a peak other than the maximum peak according to the formula (16). Further, the high-frequency signal removal unit 25 refers to the relationship between each peak frequency f k determined in advance by experiments or the like or determined by sequential calculations and the response current intensity I nd to obtain the response current intensity I ndThe peak frequency f at which it becomes equal to or higher than the intensity threshold value A k (hereinafter referred to as "removal target frequency f r ") is extracted. Note that the intensity threshold value A is determined in advance by experiments or the like as a value for discriminating the intensity of the frequency component that affects the drive control of the motor 200.

[0121] FIG. 5 shows an example of the relationship determined between each peak frequency f k and the response current intensity I nd . Note that in FIG. 5, it is assumed that the first coefficient N1 is set to 1 and the second coefficient N2 is set to 2. In the example shown in FIG. 5, as the peak frequency f k , f1 = 1 / 3f max , f2 = 2 / 3f max , f3 = 3 / 3f max = f max , f4 = 4 / 3f max , f5 = 5 / 3f max , and f6 = 6 / 3f max = 2f max occur. Then, when the intensity threshold value A is determined as shown in FIG. 5, the peak frequencies f2 (= 2 / 3f nd ), f3 (= f max ), and f4 (= 4 / 3f max ) at which the response current intensity I max becomes equal to or higher than the intensity threshold value A are extracted as the removal target frequency f r .

[0122] Then, the high-frequency signal removal unit 25 determines a filter (notch filter) that removes the current components corresponding to each of the removal target frequencies f r (that is, 2 / 3f max , f max , and 4 / 3f max ), and processes the current detection values i d_det , i q_det by the filter to calculate the corrected current detection values i´ d_det , i´ q_det . More specifically, the high-frequency signal removal unit 25 executes the operation of the following equation (20) on the current detection values i d_det , i q_det to obtain the corrected current detection values i´d_det , i´ q_det is calculated.

[0123]

Number

[0124] Note that "c0" and "c1" in Equation (20) represent coefficients indicating the width of the stopband centered around the maximum peak frequency f max . Also, "c2" and "c3" represent coefficients indicating the width of the stopband centered around the peak frequency f2 (= 2 / 3f max ). Furthermore, "c4" and "c5" represent coefficients indicating the width of the stopband centered around the peak frequency f4 (= 4 / 3f max ). In particular, the coefficient c0 is defined as a real number greater than or equal to 0, the coefficient c1 is defined as a positive real number, and the relationship c0 < c1 holds between them. Also, the coefficient c2 is defined as a real number greater than or equal to 0, the coefficient c3 is defined as a positive real number, and the relationship c2 < c3 holds between them. Furthermore, the coefficient c4 is defined as a real number greater than or equal to 0, the coefficient c5 is defined as a positive real number, and the relationship c4 < c5 holds between them.

[0125] In this way, by processing the current detection value i d_det , i q_det with the filter represented by the above Equation (20), in addition to the component corresponding to the maximum peak frequency f d_det , i q_det , the components corresponding to the surrounding peak frequencies f2 (= 2 / 3f max ) and the peak frequency f4 (= 4 / 3f max ) are also removed to determine the corrected current detection value i´ max . d_det , i´ q_det can be determined.

[0126] Next, a series of processes in the high-frequency signal removal unit 25 of the present embodiment will be collectively described.

[0127] FIG. 6 is a flowchart showing the processing flow in the high-frequency signal removal unit 25 of the present embodiment.

[0128] As shown in the figure, in the present embodiment, the processing in steps T21 and T22 is different from the processing of the high-frequency signal removal unit 25 shown in FIG. 3 in that it is executed.

[0129] More specifically, in step T21, an operation represented by Equation (16) is executed from the maximum peak frequency f max obtained in step T20, the first coefficient N1, and the second coefficient N2 to calculate the peak frequency f k (k = 1, 2,...).

[0130] In step T22, the removal target frequencies f k (k = 1, 2,...) are extracted. More specifically, each peak frequency f r (f2, f3, and f4) is extracted. Using the relationship between k and the response current intensity I nd (see FIG. 5), the peak frequencies f2 (= 2 / 3f nd ), f3 (= f max ), and f4 (= 4 / 3f max ) at which the response current intensity I max is equal to or greater than the intensity threshold A are extracted as the removal target frequencies f r .

[0131] In step T30, the current detection values i d_det , i q_det , and the removal target frequencies f r (2 / 3f max , f max , 4 / 3f max ) are processed by the filter represented by Equation (20) to calculate the corrected current detection values i' d_det , i' q_det .

[0132] (An example of the control result) ​FIG. 7 is a timing chart showing an example of control results by the motor control methods of Comparative Example 2 and Example 2. In FIG. 7, a constant torque command value T larger than the example shown in FIG. 4 * is given, and the frequencies f dh * , V qh * of the high-frequency voltages V h are made variable, and the rotational speed N m is swept, and the change over time of the phase error estimated value θ γ ′, the rotational speed N m , and the torque T are shown.

[0133] Specifically, FIG. 7(a) shows the control result (Comparative Example 2) when only the component corresponding to the maximum peak frequency f d_det , i q_det is removed from the current detection values i max in the high-frequency signal removal unit 25. On the other hand, FIG. 7(b) shows that, in addition to the component corresponding to the maximum peak frequency f d_det , i q_det in the current detection values i max in the high-frequency signal removal unit 25, the components of relatively strong sidebands (i.e., I nd ≥ A) (the component corresponding to the peak frequency f2 and the component corresponding to the peak frequency f4) are removed, and the control result (Example 2) is shown.

[0134] In Comparative Example 2 shown in FIG. 7(a), after time t2, the phase error estimated value θ γ ′ hunts, the estimation error of the rotational speed N m becomes large, and the actual output torque (estimated torque) of the motor 200 does not respond to the torque command value T * (it cannot be driven).

[0135] In contrast, in Example 2 shown in FIG. 7(b), even after time t2, hunting of the phase error estimated value θ γ ′ is not seen, the estimation accuracy of the rotational speed N m is maintained, and the estimated torque is also maintained at the torque command value T * . That is, the current detection values i d_det , iq_det from the maximum peak frequency f max to obtain the high-frequency current component i dh_det , i qh_det is used to estimate the magnetic pole position (phase θ), and the current detection value i d_det , i q_det from the maximum peak frequency f max In addition to the component corresponding to, the sideband component is also removed to obtain the corrected current detection value i´ d_det , i´ q_det is used to determine the final voltage command value V d * , V q * dh * qh , V * h of the high-frequency voltage V k It can be seen that sensorless driving can be appropriately performed even when the frequency f

[0136] The effects of the motor control method of the present embodiment described above will be summarized and described.

[0137] In this embodiment, the peak frequency f k included in, and the response current intensity I nd is the removal target frequency f which is all frequencies f at which the threshold value A or more is satisfied r (2 / 3f max , f max , 4 / 3f max ) is extracted. Then, the corrected current detection value i´ d_det , i´ q_det is calculated by removing the component corresponding to the removal target frequency f d_det , i q_det from the current detection value i r .

[0138] Thereby, when the frequency f dh * , V qh * of the superimposed high-frequency voltage V h is changed, the current detection value i d_det , i q_detAmong the high-frequency components generated in [context], components that have a relatively large impact on the drive control of the motor 200 can be comprehensively removed, so that a decrease in the control stability of the motor 200 can be more reliably prevented.

[0139] Also, in the present embodiment, based on the maximum peak frequency f max , the first coefficient N1, and the second coefficient N2, a plurality of peak frequencies f max including the maximum peak frequency f k (k = 1, 2 ···) are calculated, and among the plurality of peak frequencies f k , those for which the response current intensity I nd is equal to or greater than the intensity threshold A are extracted as the frequencies f r to be removed.

[0140] Thereby, when the frequency f dh * ,V qh * of the superimposed high-frequency voltages V h is periodically switched between the first frequency f1 and the second frequency f2, a more specific logic for appropriately determining the frequencies f r to be removed, which have a relatively large impact on the drive control of the motor 200, can be realized.

[0141] Note that in the present embodiment, the first coefficient N1 and / or the second coefficient N2 are preferably set to a predetermined time or less. That is, it is preferable that the period for switching the frequency f dh * ,V qh * of the high-frequency voltages V h between the first frequency f1 and the second frequency f2 is set to a constant value or less (the switching frequency is made faster than a certain value).

[0142] Thereby, the response current intensity I k of each peak frequency f nd can be reduced, and the number of peak frequencies f nd for which the response current intensity I k exceeds the intensity threshold A (that is, the frequencies f rThe number) can be reduced. Therefore, the filter represented by Equation (20) can be simplified to reduce the computational load.

[0143] [Third Embodiment] Hereinafter, the third embodiment will be described. In this embodiment, for the processing in the high-frequency signal removal unit 25 described in the second embodiment, further processing for suppressing the influence on the drive control of the motor 200 is executed to correct the current detection value i' d_det ,i' q_det is calculated.

[0144] FIG. 8 is a flowchart showing the processing flow in the high-frequency signal removal unit 25 of this embodiment.

[0145] As shown in the figure, the high-frequency signal removal unit 25 of this embodiment, similar to the second embodiment, calculates the maximum peak frequency f max (step T20), calculates the peak frequency f k (step T21), and extracts the frequency f to be removed r (step T22).

[0146] Then, in step T23, it is determined whether there is a frequency smaller than a predetermined frequency threshold B among the extracted frequencies f to be removed r (hereinafter referred to as "fundamental wave near frequency f d "). The frequency threshold B is determined in advance by experiments or the like as a value close to the fundamental wave frequency of the motor 200 to the extent that it affects the drive control of the motor 200.

[0147] If the fundamental wave near frequency f d does not exist, in step T24, a filter of Equation (20) in which the coefficients c0, c1, c2, c3, c4, and c5 are determined in the same manner as in the second embodiment (hereinafter referred to as the "first filter") is determined. Then, in step T30, the current detection values i d_det ,i q_det are processed by the first filter to calculate the corrected current detection values i' d_det ,i' q_det is calculated.

[0148] On the other hand, when there is a frequency f near the fundamental wave d , in step T25, a second filter is defined. Here, the second filter is defined such that the change in gain centered on the frequency f near the fundamental wave d becomes steeper (the stopband width becomes narrower) than that of the first filter. For example, assuming that the frequency f near the fundamental wave d is the peak frequency f2 = 2 / 3f shown in FIG. 5 max , the second filter can be set by changing the coefficient c2 and the coefficient c3 that determine the width of the stopband centered on the peak frequency f2 = 2 / 3f max with respect to the first filter. More specifically, the second filter can be obtained by making the value of the coefficient c2 smaller or the value of the coefficient c3 larger, so that the ratio of the coefficient c3 to the coefficient c2 is larger than that of the first filter.

[0149] Then, in step T30, the current detection values i d_det , i q_det are processed by the second filter to calculate the corrected current detection values i' d_det , i' q_det .

[0150] According to the control of the present embodiment described above, when there is a frequency less than the frequency threshold B (the frequency f near the fundamental wave r ) included in the frequency f to be removed, by using the second filter with a narrower stopband width centered on the frequency f near the fundamental wave d than that of the first filter, it is possible to more locally exclude (avoiding the surrounding frequency components) the gain of the component corresponding to the frequency f near the fundamental wave from the current detection values i d d_det , i q_det . d

[0151] The effects of the motor control method of the present embodiment described above will be collectively described.

[0152] ​ In this embodiment, when the extracted frequency f to be removed r is equal to or higher than a predetermined frequency threshold B, the current detection values i d_det , i q_det are processed by a first filter to calculate corrected current detection values i' d_det , i' q_det When the frequency f to be removed r (in this embodiment, the peak frequency f2 = 2 / 3f max ) is lower than the frequency threshold B, the current detection values i d_det , i q_det are processed by a second filter to calculate corrected current detection values i' d_det , i' q_det In particular, the second filter is set to have a narrower stopband width centered on the frequency f to be removed r (the peak frequency f2 = 2 / 3f max ).

[0153] As a result, when there is a frequency f to be removed r (a frequency f near the fundamental wave d ) that is lower than the frequency threshold B, the components corresponding to the frequency f near the fundamental wave d_det , i q_det (for example, the peak frequency f2 = 2 / 3f d ) in the current detection values i max can be excluded more locally in terms of gain to determine the corrected current detection values i' d_det , i' q_det . Therefore, it is possible to prevent a situation in which components relatively close to the fundamental wave frequency in the drive control of the motor 200 are unintentionally excluded from the current detection values i d_det , i q_det , and it is possible to more reliably prevent a decrease in control stability by reducing the influence on the fundamental wave component in the drive control.

[0154] The embodiments of the present invention have been described above. However, the configurations described in the above embodiments and each modification are merely examples of the application examples of the present invention and are not intended to limit the technical scope of the present invention.

Description of Reference Numerals

[0155] 11 Current command generation unit 12 First voltage command generation unit 13 Second voltage command generation unit 14 Final voltage command generation unit 18 Inverter 19 Phase / speed estimation unit 20 High-frequency voltage command generation unit 100 Motor control device 200 Motor

Claims

1. A motor control method for determining a power command value by superimposing a high-frequency signal on a predetermined basic power command value, detecting a response current with respect to the input of the power command value, estimating a magnetic pole position of a motor based on the obtained current detection value, and controlling the motor based on the estimated magnetic pole position and the power command value, comprising: superimposing the high-frequency signal including a plurality of frequency components on the basic power command value; calculating a peak frequency representing a frequency at which the intensity of the response current peaks; extracting a high-frequency current component corresponding to the peak frequency from the current detection value, and estimating the magnetic pole position based on the extracted high-frequency current component; calculating a corrected current detection value obtained by removing the high-frequency current component from the current detection value; calculating the basic power command value based on the corrected current detection value; A motor control method.

2. The motor control method according to claim 1, comprising: superimposing the high-frequency signal on the basic power command value while switching the frequency of the high-frequency signal; calculating a maximum peak frequency at which the intensity of the response current is maximized based on each frequency to be switched and the time for superimposing the high-frequency signal corresponding to each frequency; calculating the corrected current detection value by removing the high-frequency current component corresponding to the maximum peak frequency from the current detection value; A motor control method.

3. The motor control method according to claim 2, comprising: each of the frequencies to be switched includes a predetermined first frequency and a second frequency different from the first frequency; switching the frequency of the high-frequency signal to be superimposed between the first frequency and the second frequency; determining a first coefficient indicating the number of cycles of the high-frequency signal of the first frequency to be superimposed; determining a second coefficient indicating the number of cycles of the high-frequency signal of the second frequency to be superimposed; Calculate the maximum peak frequency based on the first frequency, the second frequency, the first coefficient, and the second coefficient. Motor control method.

4. The motor control method according to claim 1, Among the peak frequencies, extract a removal target frequency at which the intensity of the response current is equal to or greater than a predetermined intensity threshold. Calculate the corrected current detection value by removing the high-frequency current component corresponding to the removal target frequency from the current detection value. Motor control method.

5. The motor control method according to claim 4, Each of the frequencies to be switched includes a predetermined first frequency and a second frequency different from the first frequency. Switch the frequency of the high-frequency signal to be superimposed between the first frequency and the second frequency. Set the time for superimposing the high-frequency signal of the first frequency to a predetermined first coefficient. Set the time for superimposing the high-frequency signal of the second frequency to a predetermined second coefficient. Calculate the maximum peak frequency at which the intensity of the response current is maximum based on the first frequency, the second frequency, the first coefficient, and the second coefficient. Calculate a plurality of the peak frequencies including the maximum peak frequency based on the maximum peak frequency, the first coefficient, and the second coefficient. Among the plurality of the peak frequencies, extract those at which the intensity of the response current is equal to or greater than the intensity threshold as the removal target frequencies. Motor control method.

6. The motor control method according to claim 5, The first coefficient and / or the second coefficient are set to be equal to or less than a predetermined time. Motor control method.

7. The motor control method according to claim 4, When the extracted frequency to be removed is equal to or higher than a predetermined frequency threshold value, the current detection value is processed by a first filter to calculate the corrected current detection value. When the frequency to be removed is less than the frequency threshold value, the current detection value is processed by a second filter to calculate the corrected current detection value. The second filter is set to have a narrower stopband width centered on the frequency to be removed than the first filter. Motor control method. Claim 8 A motor control device that determines a power command value by superimposing a high-frequency signal on a predetermined basic power command value, detects a response current to the input of the power command value, estimates the magnetic pole position of the motor based on the obtained current detection value, and controls the motor based on the estimated magnetic pole position and the power command value, A superimposing unit that superimposes the high-frequency signal including a plurality of frequency components on the power command value, A frequency calculation unit that calculates a peak frequency representing the frequency at which the intensity of the response current peaks, An estimation unit that extracts a high-frequency current component corresponding to the peak frequency from the current detection value and estimates the magnetic pole position based on the extracted high-frequency current component, A corrected current calculation unit that calculates a corrected current detection value obtained by removing the high-frequency current component from the current detection value, And a power command value calculation unit that calculates the basic power command value based on the corrected current detection value. Motor control device.

Images