JP7055547B2 - Motor control device - Google Patents

Description

The present invention relates to a motor control device that estimates the position of the rotor based on the neutral point voltage of the stator.

In Patent Document 1, the neutral point voltage of a permanent magnet motor is detected in synchronization with the PWM waveform of the inverter, and the rotor position of the permanent magnet motor is estimated from the fluctuation of the neutral point voltage. The system is disclosed.

Japanese Unexamined Patent Publication No. 2010-0748998

Here, when a low-pass filter process is applied to the detection data of the neutral point voltage to remove the noise component and the rotor position is estimated based on the detection data after passing through the low-pass filter, the energization mode is switched and the position is estimated. Therefore, if the neutral point voltage changes sharply when the voltage vector for sampling the neutral point voltage is switched, a discrepancy occurs between the detection data before passing through the low-pass filter and the detection data after passing through the low-pass filter, resulting in rotation. There was a problem that the estimation accuracy of the child position was lowered.

The present invention has been made in view of the conventional circumstances, and an object of the present invention is to provide a motor control device capable of suppressing a decrease in the estimation accuracy of the rotor position due to low-pass filter processing of the detection data of the neutral point voltage. To do.

According to the present invention, in one aspect thereof , the position of the rotor is determined without using the neutral point voltage that has passed through the first low-pass filter for a predetermined period from the switching of the energization mode between the plurality of energization modes. The position of the rotor was estimated by using the neutral point voltage that passed through the first low-pass filter during the period other than the predetermined period .

According to the present invention, it is possible to suppress a decrease in the estimation accuracy of the rotor position due to the low-pass filter processing of the detection data of the neutral point voltage.

It is a system schematic diagram of the motor control device which concerns on this invention. It is a block diagram which shows the structure of a motor control means. It is a vector figure which shows the output voltage of an inverter circuit. It is a time chart which exemplifies the correlation of a PWM pulse, a voltage vector, and a neutral point voltage. It is a figure which exemplifies the correlation between the rotor position and the neutral point voltage Vnn1 and Vnn2. It is a block diagram which shows the structure of the rotor position estimation means in 1st Embodiment. It is a figure which illustrates the storage value of the change characteristic of a neutral point voltage. It is a block diagram which shows the structure of the LPF processing part in 2nd Embodiment. It is a block diagram which shows the structure of the rotor position estimation means in 3rd Embodiment. It is a block diagram which shows the structure of the rotor position estimation means in 4th Embodiment. It is a system schematic diagram of the electric power steering apparatus in 5th Embodiment.

Hereinafter, embodiments of the motor control device according to the present invention will be described with reference to the drawings.
FIG. 1 is a configuration diagram of a motor control device 10 for driving an electric motor 1.
The electric motor 1 includes a first stator winding (U-phase coil) Lu, a second stator winding (V-phase coil) Lv, and a third stator winding (W-phase coil) Lw, which are connected to each other in a star shape. It is a three-phase permanent magnet synchronous motor having a stator 1a provided with a stator 1a and a permanent magnet rotor 1b rotatably provided in a space formed in the central portion of the stator.
The electric motor 1 is not limited to the permanent magnet synchronous motor, and may be another AC motor that can obtain magnetic saturation characteristics with respect to the rotor position.

The inverter circuit 2 is a three-phase bridge circuit provided with three sets of switching elements for each of the three phases (U phase, V phase, and W phase) of the electric motor 1, and is used for the electric motor 1 and the DC power supply 3. Be connected. As a plurality of switching elements of the inverter circuit 2, a semiconductor switching element such as a MOSFET is used.
The DC power supply 3 is a DC power supply that supplies electric power to the inverter circuit 2.
The DC bus current sensor 4 is a current detector that detects the supply current I _DC to the inverter circuit 2 and outputs it to the motor control means 5.

The motor control means 5 outputs a PWM (Pulse Width Modulation) signal to the inverter circuit 2 in order to generate a motor torque that matches the torque command value.
The motor control means 5 switches the plurality of energization modes corresponding to the combination of the on and off states of the plurality of switching elements of the inverter circuit 2 based on the position of the rotor 1b of the electric motor 1 to switch the inverter circuit 2 It is an inverter control unit that drives and controls.

The rotor position estimation means 6 is a rotor position estimation unit that estimates the rotor position by utilizing the fact that the neutral point voltage Vn0 of the electric motor 1 changes under the influence of the rotor position.
When the inductances of the stator windings Lu, Lv, and Lw are equal, the neutral point voltage Vn0 becomes zero, but in reality, the magnetic flux of the magnet of the rotor 1b affects the winding, so that the inductance moves to the rotor position. Shows the corresponding changes. Therefore, when the neutral point voltage Vn0 is observed with a constant voltage vector applied to the electric motor 1, the neutral point voltage Vn0 shows a change according to the rotor position.

Therefore, the rotor position estimation means 6 utilizes the fact that the neutral point voltage Vn0 changes depending on the rotor position, and the rotor 1b is based on the neutral point voltage Vn0 and the switching state of a plurality of energization modes. Position estimation is performed.
The rotor position estimation means 6 has a virtual neutral point potential Vnc, which is the potential of the neutral point of the virtual neutral point circuit 7, and a three-phase winding connection point potential Vn of the electric motor 1 (neutral point of the stator 1a). The neutral point voltage Vn0 of the stator 1a is obtained as the difference from the potential of the stator neutral point potential), and the rotor position (phase angle) θdest (deg) of the electric motor 1 is based on the neutral point voltage Vn0. Is estimated and output to the motor control means 5.

The virtual neutral point circuit 7 has resistors Ru, Rv, and Rw connected to each other in a star shape, and is a circuit that generates a virtual neutral point potential Vnc with respect to the output voltage of the inverter circuit 2.
Here, one end of the resistance Ru is connected to the U-phase drive line of the electric motor 1, one end of the resistance Rv is connected to the V-phase drive line of the electric motor 1, and the resistance Rw is the W-phase drive of the electric motor 1. One end is connected to the line, and the other end of the resistance Ru, the other end of the resistance Rv, and the other end of the resistance Rw are connected in a star shape.

Then, the potentials of the connection points of the resistors Ru, Rv, and Rw are output to the rotor position estimation means 6 as virtual neutral point potentials Vnc.
That is, the resistance Ru is the first stator potential output unit having the same potential as the first stator winding (U-phase coil) Lu, and the resistance Rv is the second stator winding (V-phase coil) Lv. It is a second stator potential output unit having the same potential, and the resistor Rw is a third stator potential output unit having the same potential as the third stator winding (W phase coil) Lw.
Here, the virtual neutral point circuit 7 is arranged so that the magnet magnetic flux of the rotor 1b does not affect it, and outputs a reference potential to the rotor position estimation means 6.

FIG. 2 is a block diagram of the motor control means 5.
The motor control means 5 includes a vector controller 20, a dq / 3-phase coordinate converter 21, a pulse width modulator 22, a 3-phase / dq coordinate converter 23, a current reproducer 24, and a speed calculator 25.
The vector controller 20 inputs data of a torque command value, a q-axis current Iqc, a d-axis current Idc, and a rotation speed ωr of the electric motor 1, and dq so that the electric motor 1 generates a torque corresponding to the torque command value. The voltage commands Vd * and Vq * on the axis are calculated, and the calculated voltage commands Vd * and Vq * are output to the dq / 3-phase coordinate converter 21.

The dq / 3-phase coordinate converter 21 calculates the three-phase AC voltage commands Vu *, Vv *, and Vw * based on the voltage commands Vd * and Vq * on the dq axis and the estimated rotor phase θdest, and calculates 3 The phase AC voltage commands Vu *, Vv *, and Vw * are output to the pulse width modulator 22.
The pulse width modulator 22 generates a PWM signal for turning on / off a plurality of switching elements of the inverter circuit 2 based on the three-phase AC voltage commands Vu *, Vv *, and Vw *, and outputs the PWM signal to the inverter circuit 2. ..

Based on the three-phase AC currents Iuc, Ivc, Iwc and the estimated rotor phase θdest, the three-phase / dq coordinate converter 23 has a current Iqc (q-axis current component) that contributes to torque and a current Idc (d) that contributes to magnetic flux. (Shaft current component) is obtained and output to the vector controller 20.
The current reproducer 24 reproduces and reproduces the U-phase, V-phase, and W-phase phase currents Iuc, Ivc, and Iwc based on the supply current I _DC to the inverter circuit 2 detected by the DC bus current sensor 4. The phase currents Iuc, Ivc, and Iwc of the phase are output to the 3-phase / dq coordinate converter 23.

The rotor position estimation means 6 obtains the estimated rotor phase θdest of the electric motor 1 based on the neutral point voltage Vn0, and the obtained estimated rotor phase θdest is used as the dq / 3 phase coordinate converter 21 of the motor control means 5. It is output to the 3-phase / dq coordinate converter 23 and the speed calculator 25.
The speed calculator 25 calculates the rotation speed ωr (motor rotation speed rpm) of the electric motor 1 based on the estimated rotor phase θdest, and outputs the calculation to the vector controller 20.

Next, the basic operation of the motor control means 5 will be described.
The motor control means 5 drives and controls the electric motor 1 by a known vector control which is a method of linearizing the torque of the electric motor 1.

The vector controller 20 obtains a current command Iq * that contributes to torque and a current command Id * that contributes to magnetic flux based on the torque command value.
However, the current command Id * is usually set to zero if the electric motor 1 is a non-slip-type permanent magnet synchronous motor. On the other hand, when the electric motor 1 is a permanent magnet synchronous motor having a salient pole structure, or when field weakening control and efficiency maximization control are performed, the current command Id * is set to other than zero.

The 3-phase / dq coordinate converter 23 uses the 3-phase AC currents Iuc, Ivc, and Iwc, which are the AC current detection values of the electric motor 1, as the currents Iqc (q-axis current component) that contributes to torque based on the estimated rotor phase θdest. ) And the current Idc (d-axis current component) that contributes to the magnetic flux, and output to the vector controller 20.
The vector controller 20 controls the current so that the current detection values match the current command Iq * that contributes to torque and the current command Id * that contributes to magnetic flux, so that the voltage on the dq axis, which is the rotational coordinate axis, is matched. Calculate the commands Vd * and Vq *.

Then, the dq / 3-phase coordinate converter 21 converts the voltage commands Vd * and Vq * into the three-phase AC voltage commands Vu *, Vv * and Vw * based on the estimated rotor phase θdest.
The vector controller 20 can calculate the voltage commands Vd * and Vq * on the dq axis by combining the result of the current control and the result of the non-interference control that compensates for the interference term of the dq axis.

The pulse width modulator 22 applies a voltage corresponding to the three-phase AC voltage commands Vu *, Vv *, and Vw * to each phase of the electric motor 1 by pulse width modulation that turns on / off a plurality of switching elements of the inverter circuit 2. It is applied to U phase, V phase, W phase).
Then, the motor control means 5 sequentially switches the energization of each phase of the electric motor 1 to supply a current to each phase to rotationally drive the electric motor 1.
It should be noted that the system can detect each of the three-phase alternating currents without detecting the direct current bus current.

Next, the output voltage of the inverter circuit 2 will be described.
The output voltage of the inverter circuit 2 has a total of eight patterns according to the switch state of each of the three-phase switching elements.

FIG. 3 is a vector representation of the output voltage of the inverter circuit 2 on the αβ coordinates.
Each vector is expressed as, for example, V (1, 0, 0), and the sequence of numbers in parentheses represents the switching state in the order of "U phase, V phase, W phase", and the upper switching element of the inverter circuit 2 The on state of the (upper arm) is expressed as "1", and the on state of the lower switching element (lower arm) is expressed as "0".

Therefore, for example, in the voltage vector _VA = V (1, 0, 0), the upper switching element of the U phase is on (the lower switching element of the U phase is off), and the lower switching element of the V phase and the W phase is on. It represents the state (the upper switching element of V phase and W phase is off).
The voltage vectors are V ₀ = V (0, 0, 0), _VA = V (1, 0, 0), V _B = V (1, 1, 0), Vc = V (0, 1, 0). 0), V _D = V (0, 1, 1), _VE = V (0, 0, 1), _VF = V (1, 0, 1), V ₇ = V (1, 1, 1) There are 8 ways.

The motor control means 5 creates a sinusoidal pulse pattern by the combination of the above voltage vectors V ₀ to V ₇ , and applies the pulse pattern to the electric motor 1.
For example, when the voltage command V * shown in FIG. 3 is given, the motor control means 5 has a voltage vector _VA = V (1, 0, 0) surrounding the voltage command V *, V _B = V (1, 1). , 0), and the zero vector V ₀ = V (0, 0, 0), V ₇ = V (1, 1, 1) are combined to create a voltage corresponding to the voltage command V *.
That is, the motor control means 5 switches the energization mode, which is a combination pattern of voltage vectors, every time the rotor 1b rotates by 60 deg, and creates a voltage corresponding to the voltage command V *.

FIG. 4 illustrates the relationship between the PWM pulse signals PVu, PVv, PVw, the voltage vector V ₀ −V ₇ , and the neutral point voltage VnA −VnF in one energization mode.
As shown in FIG. 4, when the voltage vectors _VA - _VF excluding the zero vector V ₀ = V (0, 0, 0) and V ₇ = V (1, 1, 1) are applied, respectively. It is possible to detect the neutral point voltage VnA-VnF, and the change characteristic of the neutral point voltage Vn0 according to the rotor position is different for each voltage vector.
Therefore, the rotor position estimation means 6 detects the neutral point voltage VnA- _VnF when the voltage vector _VA -VF is applied to the electric motor 1, and which is the detected neutral point voltage VnA-VnF. The rotor position can be estimated based on whether the value is at the rotor position.

The PWM pulse signals PVu, PVv, and PVw in FIG. 4 are PWM signals in the energization mode that create a voltage corresponding to the voltage command V * in FIG. 3, and the voltage vector _VA = V in the latter half of the PWM cycle. PWM to secure the detection period of the neutral point voltage V _nA at (1, 0, 0) and the detection period of the neutral point voltage V _nB at the voltage vector V _B = V (1, 1, 0). A state in which pulse shift, which is a process of shifting the center of the pulse signal from the valley of the triangular wave carrier, is performed is shown.
Further, FIG. 4 illustrates a case where the rotor position estimation means 6 samples the neutral point voltage Vn0 in the latter half of the PWM cycle, and the rotor position estimation means 6 exemplifies the case where the neutral point voltage Vn0 is sampled in the latter half of the PWM cycle. The voltage Vn0 can be sampled.

5 (A) and 5 (B) show changes in the neutral point voltages Vnn1 and _Vnn2 when six voltage vectors V _A -VF are applied to the electric motor 1 and the rotor position is changed by one cycle. show.
Here, the neutral point voltages Vnn1 and Vnn2 shown in FIG. 5 are values after low-pass filter processing for noise reduction.

In the present embodiment, the neutral point voltage Vn0 is sampled at the time of two different voltage vectors in one cycle of PWM, and the neutral point voltage Vn0 sampled first on the time axis is used as the neutral point voltage. Let Vnn1 be, and let the neutral point voltage Vn0 sampled next to the neutral point voltage Vnn1 be the neutral point voltage Vnn2.

For example, in the switching state of FIG. 4 corresponding to the interval between 330 deg and 30 deg of FIG. 5, the neutral point voltage VnB at the voltage vector V _B = V (1, 1, 0) is detected as Vnn1 and the voltage vector V _A. The neutral point voltage VnA at = V (1, 0, 0) is detected as the neutral point voltage Vnn2.
Similarly, when the rotor position is between 30 deg and 90 deg, the neutral point voltage VnB at the voltage vector V _B = V (1, 1, 0) is detected as Vnn1, and the voltage vector Vc = V (0, 1). , 0), the neutral point voltage VnC is detected as Vnn2.

Further, when the rotor position is between 90 deg and 150 deg, the neutral point voltage VnD in the voltage vector V _D = V (0, 1, 1) is detected as Vnn1, and the voltage vector Vc = V (0, 1, 0). ), The neutral point voltage VnC is detected as Vnn2.
Further, when the rotor position is between 150 deg and 210 deg, the neutral point voltage VnD in the voltage vector V _D = V (0, 1, 1) is detected as _Vnn1 , and the voltage vector VE = V (0, 0, The neutral point voltage VnE in 1) is detected as Vnn2.

Further, when the rotor position is between 210 deg and 270 deg, the neutral point voltage _VnF at the voltage vector VF = V (1, 0, 1) is detected as _Vnn1 , and the voltage vector VE = V (0, 0, The neutral point voltage VnE in 1) is detected as Vnn2.
Further, when the rotor position is between 270 deg and 330 deg, the neutral point voltage _VnF at the voltage vector VF = V (1, 0, 1) is detected as Vnn1, and the voltage vector _VA = V (1, 0, The neutral point voltage VnA at 0) is detected as Vnn2.

As shown in FIG. 5, the neutral point voltages Vnn1 and Vnn2 change according to the rotor position, and show different values for each voltage vector V _A − V _F. Therefore, the rotor position estimation means 6 estimates the rotor position by utilizing the characteristic that the neutral point voltage _Vn0 changes depending on the rotor position in a pattern corresponding to the voltage vectors _VA -VF.

Next, the rotor position estimation process in the rotor position estimation means 6 will be described in detail.
"First embodiment"
FIG. 6 is a block diagram showing a first embodiment of the rotor position estimation means 6.
The rotor position estimation means 6 shown in FIG. 6 has a neutral point voltage calculation unit 300, an LPF processing unit 301, and a phase estimation unit 302.

The neutral point voltage calculation unit 300 calculates the neutral point voltage Vn0 as the difference between the virtual neutral point potential Vnc, which is the output of the virtual neutral point circuit 7, and the three-phase winding connection point potential Vn of the electric motor 1. ..
In the neutral point voltage calculation unit 300, a voltage vector _VA - _VF excluding V ₀ = V (0, 0, 0) and V ₇ = V (1, 1, 1) is applied to the electric motor 1. When is the sampling timing of the neutral point voltage Vn0, the neutral point voltage VnA-VnF is detected at the sampling timing, and the neutral point voltages Vnn1 and Vnn2 are output in the order of detection on the time axis.

That is, as shown in FIG. 5A, the neutral point voltage calculation unit 300 outputs VnB, VnD, and VnF as the neutral point voltage Vnn1 every 60 deg rotation of the rotor 1b, and outputs VnB, VnD, and VnF in this order, and FIG. As shown in (B), VnA, VnC, and VnE are output in this order as the neutral point voltage Vnn2 every time the rotor 1b rotates by 60 deg.
The signal of the neutral point voltage Vnn1 and the signal of the neutral point voltage Vnn2 output from the neutral point voltage calculation unit 300 are cut off by attenuating the frequency signal higher than a specific threshold value in the LPF processing unit 301, respectively. , Low-pass filter processing is performed to pass only low frequency as a signal.

The LPF processing unit 301 includes a first LPF unit 31 (first low-pass filter), a second LPF unit 32 (second low-pass filter), and an LPF output switching unit 33.
The first LPF unit 31 and the second LPF unit 32 input the signal of the neutral point voltage Vnn1 and the signal of the neutral point voltage Vnn2 output from the neutral point voltage calculation unit 300, respectively, and input the neutral point voltage Vnn1 and Vnn2. Low-pass filter processing is applied to the signal of.

The cutoff frequency (first time constant τ) of the first LPF unit 31 is set so that the noise components of the neutral point voltages Vnn1 and Vnn2 are removed and the change according to the rotor position is captured.
On the other hand, the second LPF unit 32 is set to a cutoff frequency (second time constant τ) different from that of the first LPF unit 31.
The cutoff frequency (Hz) of the second LPF unit 32 is higher than the cutoff frequency (Hz) of the first LPF unit 31, and the time constant τ (sec) of the second LPF unit 32 is higher than the time constant τ (sec) of the first LPF unit 31. The filter characteristics of the first LPF unit 31 and the second LPF unit 32 are set so as to be shorter.

That is, each cutoff frequency (time constant τ) is set so that the response speed of the second LPF unit 32 is faster than that of the first LPF unit 31, and the noise component removal performance is ensured in the first LPF unit 31. The 2LPF unit 32 is designed so that the transient changes of the neutral point voltages Vnn1 and Vnn2 can be captured with better response than the first LPF unit 31.

The LPF output switching unit 33 has passed the output of the first LPF unit 31 (Vnn1 signal and Vnn2 signal that have passed through the first LPF unit 31) and the output of the second LPF unit 32 (passed through the second LPF unit 32) based on the switching signal of the energization mode. Either one of the Vnn1 signal and the Vnn2 signal) is output to the phase estimation unit 302.

The phase estimation unit 302 includes a neutral point voltage data memory 34 and a phase estimation calculation unit 35.
The neutral point voltage data memory 34 is a memory that stores change characteristics according to the rotor positions of the neutral point voltages Vnn1 and Vnn2.
The neutral point voltage data memory 34 stores a function for deriving the value of the neutral point voltage VnA-VnF with the rotor position as a variable as a change characteristic according to the rotor positions of the neutral point voltages Vnn1 and Vnn2. In addition, the value of the neutral point voltage VnA-VnF for each rotor position can be stored.

FIG. 7 is a diagram illustrating the change characteristics of the neutral point voltages Vnn1 and Vnn2 stored in the neutral point voltage data memory 34.
The neutral point voltage data memory 34 is set to switch and specify either the neutral point voltage Vnn1 or Vnn2 as the voltage data used for position estimation every time the rotor 1b rotates by 60 deg, and use the neutral point voltage Vnn1. In the position range of, the change characteristic according to the rotor position of the neutral point voltage VnA-VnF detected as the neutral point voltage Vnn1 is stored, and in the position range of the setting using the neutral point voltage Vnn2, the neutral point voltage Vnn2 The change characteristic according to the rotor position of the neutral point voltage VnA-VnF detected as is stored.

In other words, in a configuration in which two different voltage vectors are applied and the neutral point voltage Vn0 at each is sampled as Vnn1 and Vnn2, the rotor 1b rotates every 60 deg rotation cycle (every energization mode switching cycle). The voltage vector that can easily capture the change in the neutral point voltage Vn0 depending on the position of the child 1b is selected, and the change characteristic according to the rotor position of the neutral point voltage VnA-VnF in the selected voltage vector is the neutral point. It is stored in the voltage data memory 34.

For example, as shown in FIG. 4, in the energization mode (rotor position is between 330deg and 30deg) in which the neutral voltage VnB is detected as Vnn1 and the neutral voltage VnA is detected as Vnn2, the rotor position. When the estimation is performed based on the neutral point voltage VnB (Vnn1), the change characteristic of the neutral point voltage VnB (between 330deg and 30deg) detected as Vnn1 in the corresponding position range of the neutral point voltage data memory 34. The value of VnB in) is saved.
However, the phase estimation unit 302 is not limited to the configuration in which the neutral point voltage VnA-VnF used for the position estimation of the rotor 1b is switched every 60 deg, in other words, every energization mode.

Further, the phase estimation unit 302 can estimate the rotor position using only one of the neutral point voltages Vnn1 and Vnn2. However, if the configuration is such that the position estimation is performed by selecting the neutral point voltage Vnn1 or Vnn2 that is suitable for the position estimation, the position estimation result with high accuracy can be obtained.
Further, the phase estimation unit 302 can estimate the position of the rotor 1b at a rate of once for each PWM cycle or once for each of a plurality of PWM cycles.

The phase estimation calculation unit 35 performs an estimation calculation of the rotor position by comparing the neutral point voltage Vnn1 or the neutral point voltage Vnn2 with the change characteristic stored in the neutral point voltage data memory 34.
For example, the phase estimation unit 302 has a neutral point voltage Vnn1 (VnB) that has passed through the LPF processing unit 301 and a neutral point voltage between the estimation of the rotor position 330deg and the estimation of the rotor position 30deg. The position of the rotor 1b at that time is estimated by comparing with the change characteristic (see FIG. 7) of the neutral point voltage Vnn1 (VnB) stored in the data memory 34 corresponding to the range from 330 deg to 30 deg. do.
Here, when the neutral point voltage data memory 34 functionalizes and stores the change of the neutral point voltage VnA—VnF with respect to the rotor position, the phase estimation calculation unit 35 uses the inverse function to make the neutral point voltage data memory 34 neutral. The rotor position can be calculated from the point voltages Vnn1 and Vnn2.

Next, the switching operation by the LPF output switching unit 33 will be described.
In the characteristics of the neutral point voltages Vnn1 and Vnn2 shown in FIG. 5, when either the neutral point voltage Vnn1 or Vnn2 is specified as the data used for position estimation for each energization mode as shown in FIG. 7, for example. Assuming that the rotor position is within the range of 0deg-30deg, the phase estimation calculation unit 35 is stored in the corresponding ranges of the neutral point voltage Vnn1 (neutral point voltage VnB) and the neutral point voltage data memory 34. The neutral point voltage VnB (see FIG. 7) is compared with the neutral point voltage, and the rotor position at that time is estimated and calculated.

Then, when 30 deg, which is the switching position of the energization mode, is detected based on the neutral point voltage Vnn1 (neutral point voltage VnB), the energization mode is switched and the data used for position estimation is the neutral point voltage Vnn1 (neutral). The point voltage VnB) is switched to the neutral point voltage Vnn2 (neutral point voltage VnC).
Here, as shown in FIG. 5, for example, at the rotor position 30 deg where the energization mode is switched, the neutral point voltage Vnn2 is switched from the neutral point voltage VnA to the neutral point voltage VnC, and is medium at the rotor position 30 deg. Due to the deviation between the neutral point voltage VnA and the neutral point voltage VnC, the neutral point voltage Vnn2 before the low pass filter processing (before passing through the low pass filter) changes stepwise.

On the other hand, the neutral point voltage Vnn2 after the low-pass filter processing (after passing through the low-pass filter) used for estimating the rotor position is delayed in response from the level of the neutral point voltage VnA to the level of the neutral point voltage VnC. Follows and changes with.
On the other hand, the neutral point voltage data memory 34 shown in FIG. 7 stores a steady value of the neutral point voltage VnA-VnF according to the rotor position, and is neutral with the rotor position 30 deg as a boundary. It will change step by step from the point voltage VnB to the neutral point voltage VnC.

Therefore, immediately after switching the energization mode at the rotor position 30 deg, the neutral point voltage Vnn2 after passing through the LPF processing unit 301 where the response delay occurs, and the stored value of the neutral point voltage data memory 34 without the response delay. Will be compared, and the estimation accuracy of the rotor position will decrease.
In the characteristic example shown in FIG. 5, the decrease in estimation accuracy due to the response delay due to such low-pass filter processing is caused by the rotor position 270deg in which the neutral point voltage Vnn2 changes sharply, and further, the neutral point voltage Vnn1. Will occur even at the rotor positions 90deg and 210deg where the voltage changes sharply.

Therefore, the LPF output switching unit 33 replaces the output of the first LPF unit 31 (the Vnn1 signal and the Vnn2 signal that have passed through the first LPF unit 31) with the output of the second LPF unit 32 (the second LPF unit 32) in a predetermined period from the switching of the energization mode. 2 The Vnn1 signal and the Vnn2 signal that have passed through the LPF unit 32) are input to the phase estimation calculation unit 35.
That is, the phase estimation calculation unit 35 estimates the rotor position based on the Vnn1 signal and the Vnn2 signal that have passed through the second LPF unit 32 in a predetermined period from the switching of the energization mode, and otherwise passes through the first LPF unit 31. The rotor position is estimated based on the Vnn1 signal and Vnn2 signal.

In other words, the phase estimation calculation unit 35 uses the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 31 between one of the plurality of energization modes, while the phase estimation calculation unit 35 is between the plurality of energization modes. When the energization mode is switched with, the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF section 31 are not used, and instead the neutral point voltages Vnn1 and Vnn2 that have passed through the second LPF section 32 are used to use the rotor 1b. Estimate the position of.

As described above, the cutoff frequency of the second LPF section 32 is set higher than the cutoff frequency of the first LPF section 31, and the time constant τ (sec) of the second LPF section 32 is shorter than that of the first LPF section 31. The response delay of the second LPF section 32 when the voltages Vnn1 and Vnn2 change suddenly, in other words, the difference between the neutral point voltages Vnn1 and Vnn2 before the low-pass filter and the neutral point voltages Vnn1 and Vnn2 after the low-pass filter processing. Is smaller than that of the first LPF unit 31.
Therefore, if the output of the second LPF unit 32 is input to the phase estimation calculation unit 35 instead of the output of the first LPF unit 31 within a predetermined period from the switching of the energization mode, the estimation accuracy due to the response delay due to the low-pass filter processing is a factor. The decrease is suppressed.

For example, when the rotor position is within the range of 0deg-30deg, the phase estimation calculation unit 35 inputs the Vnn1 signal and the Vnn2 signal output from the first LPF unit 31, and among these, the Vnn1 signal (neutral). The rotor position is estimated based on the point voltage VnB).
Then, when the rotor position 30deg is estimated and the energization mode is switched, the phase estimation calculation unit 35 switches to the state of inputting the Vnn1 signal and the Vnn2 signal output from the second LPF unit 32, and Vnn2 of these is switched to. The rotor position is estimated based on the signal (neutral point voltage VnC).

Here, the Vnn2 signal output from the second LPF unit 32 changes from the level of the neutral point voltage VnA to the level of the neutral point voltage VnC with a faster response than the Vnn2 signal output from the first LPF unit 31. Therefore, the estimation error can be made smaller than the case where the position is estimated based on the Vnn2 signal output from the first LPF unit 31.
Then, when a predetermined period elapses from the switching of the energization mode, the phase estimation calculation unit 35 returns to the state of inputting the Vnn1 signal and the Vnn2 signal output from the first LPF unit 31, and the estimation process in which the influence of noise is sufficiently suppressed. Can be done.

The LPF output switching unit 33 causes the phase estimation calculation unit 35 to input the output of the second LPF unit 32 instead of the output of the first LPF unit 31. The predetermined period from the switching of the energization mode is, for example, the first LPF unit 31. The time corresponding to the time constant τ of, that is, the predetermined time required for the difference between the Vnn1 and Vnn2 signals before passing through the first LPF section 31 and the Vnn1 and Vnn2 signals after passing through the first LPF section 31 becomes sufficiently small. Set.

Further, the LPF output switching unit 33 switches the energization mode in which the output of the second LPF unit 32 is input to the phase estimation calculation unit 35 instead of the output of the first LPF unit 31, and the energization accompanied by a steep change of the Vnn1 and Vnn2 signals. It can be limited to mode switching (rotor positions 30 deg, 90 deg, 210 deg, 270 deg in the example of FIG. 5).
However, if the LPF output switching unit 33 performs the switching operation when the energization mode is switched without determining whether or not the energization mode is switched with a sudden change in the Vnn1 and Vnn2 signals, control is performed. Can be simplified.

Further, in the present embodiment, as shown in FIG. 7, one of the neutral point voltages Vnn1 and Vnn2 is designated as the data used for position estimation in each energization mode, but is neutral regardless of the energization mode. When the position is estimated using the point voltage Vnn1 (or the neutral point voltage Vnn2), the position is estimated by using the output of the second LPF unit 32 within a predetermined period from the switching of the energization mode. It has an effect.

Further, the LPF processing unit 301 can have a first LPF unit 31 and a second LPF unit 32 having a lower cutoff frequency (longer time constant τ (sec)) than the first LPF unit 31. In the case of such a configuration, the LPF output switching unit 33 causes the phase estimation calculation unit 35 to input the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 31 in a predetermined period from the switching of the energization mode, and in other periods. , The neutral point voltages Vnn1 and Vnn2 that have passed through the second LPF unit 32 are input to the phase estimation calculation unit 35.

"Second embodiment"
FIG. 8 is a block diagram showing the configuration of the LPF processing unit 301A in the second embodiment.
In the second embodiment, the LPF processing unit 301 in the configuration of the rotor position estimation means 6 shown in FIG. 6 is changed to the LPF processing unit 301A shown in FIG. 8, and the neutral point voltage is used. The calculation unit 300 and the phase estimation unit 302 are provided in the same manner as in the first embodiment. Therefore, detailed description of the neutral point voltage calculation unit 300 and the phase estimation unit 302 will be omitted.

The LPF processing unit 301A shown in FIG. 8 includes a first LPF unit 41 (first low-pass filter) for low-pass filtering the neutral point voltages Vnn1 and Vnn2, which are outputs of the neutral point voltage calculation unit 300, and an initial value setting unit. It has 42 and.
The first LPF unit 41 is set to a cutoff frequency (time constant τ) that can sufficiently remove the noise components of the signals of the neutral point voltages Vnn1 and Vnn2.

Further, when the energization mode is switched, the initial value setting unit 42 sets the values before the low-pass filter processing (before passing through the first LPF unit 41) before the switching timing of the energization mode for each of the neutral point voltages Vnn1 and Vnn2. The average value with the value before the low-pass filter processing (before passing through the first LPF unit 41) after the switching timing of the energization mode is calculated, and the calculated average value is set as the initial value of the low-pass filter processing in the first LPF unit 41.

In other words, when the energization mode is switched between the plurality of energization modes, the phase estimation unit 302 does not pass through the first LPF unit 41 (arbitrary filter), and the neutral point voltages Vnn1 and Vnn2 (before and after the switching timing). The position of the rotor 1b is estimated using (the average value of), and then the position of the rotor 1b is estimated using the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 41.
According to this configuration, when the signals of the neutral point voltages Vnn1 and Vnn2 after passing through the first LPF section 41 (after low-pass filter processing) are switched between the energization modes, the above initial value setting is not performed. In comparison, it approaches the value before low-pass filter processing after switching.

Therefore, the discrepancy between the data before the low-pass filter processing and the data after the low-pass filter processing due to the switching of the energization mode can be reduced as compared with the case where the initial value setting unit 42 does not set the initial value, and the energization mode can be switched. Immediately after, the estimation accuracy of the rotor position can be improved.
Further, in the configuration in which the output of the first LPF unit 31 and the output of the second LPF unit 32 are switched as in the first embodiment, the cutoff frequency in the second LPF unit 32 is achieved so as to achieve both sufficient responsiveness and noise reduction performance. However, in the second embodiment, the adjustment of the cutoff frequency becomes unnecessary, and the development cost and the adjustment man-hours can be reduced.

Further, in the case of the first embodiment, in a situation where the energization mode is frequently switched (at high rotation speed), the response delay may increase in the second LPF unit 32 as well, and an estimation error may occur. If the initial value of the low-pass filter processing is set based on the value before the low-pass filter processing as in the above, the responsiveness of the value after the low-pass filter processing can be maintained high, and the estimation error can be suppressed to a small value.
The initial value setting unit 42 obtains the average value of the values before and after the switching of the energization mode and sets it as the initial value of the first LPF unit 41. However, the value before the low-pass filter processing immediately after the switching of the energization mode is used as the first LPF unit 41. Can be the initial value of.

However, since it is unknown whether the position of the rotor 1b is in the phase before the switching of the energization mode or in the phase after the switching depending on the rotation state of the rotor 1b, the initial value setting unit 42 switches. By setting the average value before and after as the initial value of the first LPF unit 41, it is possible to prevent the estimated value of the rotor position from deviating significantly.
Further, when the energization mode is switched, the initial value setting unit 42 sets the neutral point voltages Vnn1 and Vnn2 that have passed through a low-pass filter having a higher cutoff frequency (shorter time constant τ) than the first LPF unit 41 to the first LPF. It can be the initial value of the unit 41.

As described above, in the second embodiment, the phase estimation unit 302 estimates the rotor position by using the neutral point voltages Vnn1 and Vnn2 that have not passed through an arbitrary low-pass filter when the energization mode is switched. After that, the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 41 are used to estimate the rotor position.
Then, the phase estimation unit 302 uses the neutral point voltages Vnn1 and Vnn2 that have passed through the low-pass filter by using the neutral point voltages Vnn1 and Vnn2 that have not passed through the low-pass filter when the energization mode is switched. It is possible to improve the followability of the neutral point voltage Vnn1 and Vnn2 changes due to the switching of the energization mode. Further, since the phase estimation unit 302 returns to the state of using the neutral point voltages Vnn1 and Vnn2 that have passed through the low-pass filter after switching the energization mode, a noise reduction effect can be obtained.

"Third embodiment"
FIG. 9 is a block diagram showing the configuration of the rotor position estimation means 6A in the third embodiment.
The rotor position estimation means 6A of FIG. 9 has a neutral point voltage calculation unit 401, an LPF processing unit 402, and a phase estimation unit 403. Here, the neutral point voltage calculation unit 401 of FIG. 9 has the same function as the neutral point voltage calculation unit 300 of FIG. 6, and detailed description thereof will be omitted.

The LPF processing unit 402 has a first LPF unit 51 (first low-pass filter) that performs low-pass filter processing on the neutral point voltages Vnn1 and Vnn2 output by the neutral point voltage calculation unit 401.
The first LPF unit 51 is set to a cutoff frequency (time constant τ) capable of removing noise components of signals of neutral point voltages Vnn1 and Vnn2.
The phase estimation unit 403 includes a neutral point voltage data memory 52 and a phase estimation calculation unit 53.

The neutral point voltage data memory 52 stores a map having data on the positions of the rotors 1b corresponding to the neutral point voltages Vnn1 and Vnn2.
Then, the phase estimation calculation unit 53 correlates the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 51 with the neutral point voltages Vnn1 and Vnn2 stored in the neutral point voltage data memory 34 and the rotor position ( Compare with the map) to estimate the rotor position.

The neutral point voltage data memory 52 in FIG. 9 changes the change characteristics of the neutral point voltages Vnn1 and Vnn2 output to the phase estimation calculation unit 53 according to the motor rotation speed ωr (rotational speed of the rotor 1b). However, it is different from the neutral point voltage data memory 34 in FIG.
The neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF section 51 cause a response delay with respect to the neutral point voltages Vnn1 and Vnn2 before passing, but the neutral point voltage data memory 52 passes through the first LPF section 51. Therefore, if the change characteristics of the neutral point voltages Vnn1 and Vnn2 that cause the response delay are stored in advance, it is possible to estimate the rotor position while suppressing the influence of the response delay.

That is, the map of the neutral point voltage data memory 52 in FIG. 9 shows the change characteristic shown in FIG. 7 in the first LPF section without the neutral point voltages Vnn1 and Vnn2 changing sharply when the energization mode is switched. The characteristic changed so as to show a change close to the response change of the neutral point voltages Vnn1 and Vnn2 that have passed through 51 is stored.
In other words, the map data of the neutral point voltage data memory 52 is the neutral point voltage Vnn1 and Vnn2 that have passed through the first LPF section 51 rather than the change of the neutral point voltage Vnn1 and Vnn2 before passing through the first LPF section 51. It is designed to approach the change of.

However, the change characteristics of the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 51 change depending on the motor rotation speed ωr, and the faster the motor rotation speed ωr, the larger the response delay to the change in the rotor position.
Therefore, the neutral point voltage data memory 52 changes the rotor position according to how much the response delay occurs in the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 51 according to the motor rotation speed ωr. It is configured so that the data of the change characteristics of the neutral point voltages Vnn1 and Vnn2 with respect to the neutral point can be adjusted.

That is, in the neutral point voltage data memory 52, as the motor rotation speed ωr is faster and the response delay of the first LPF unit 51 becomes larger, the neutral point voltages Vnn1 and Vnn2 referred to by the phase estimation calculation unit 53 when the energization mode is switched. Delay the change of.
For example, the neutral point voltage data memory 52 has a map that stores changes in the neutral point voltages Vnn1 and Vnn2 with characteristics corresponding to the response delay (time constant τ) of the first LPF unit 51 at a predetermined motor rotation speed ωr. A plurality of maps having different motor rotation speeds ωr and rotation speed conditions close to the motor rotation speed ωr at that time can be selected, and the change characteristics of the selected maps can be output to the phase estimation calculation unit 53.

Further, the neutral point voltage data memory 52 has a map for storing the change characteristics of the neutral point voltages Vnn1 and Vnn2 before passing through the first LPF unit 51, and when the energization mode is switched, the map data is used. It can be provided to the phase estimation calculation unit 53 by performing a process of causing a response delay according to the motor rotation speed ωr at that time.
As a result, even if the response delay characteristics of the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF section 51 change according to the motor rotation speed ωr, the estimation error can be stably reduced.

"Fourth Embodiment"
FIG. 10 is a block diagram showing the configuration of the rotor position estimation means 6B in the fourth embodiment.
The rotor position estimation means 6B of FIG. 10 has a neutral point voltage calculation unit 501, an LPF processing unit 502, and a phase estimation unit 503, and the estimated rotor position θdest estimated by the phase estimation unit 503 is a motor control means. It is output to 5.

The rotor position estimation means 6B (rotor position estimation unit) and the motor control means 5 (inverter control unit) in FIG. 10 are digital circuits provided in the microprocessor 510.
The neutral point voltage calculation unit 501 and the phase estimation unit 503 have the same configuration and function as the neutral point voltage calculation unit 300 and the phase estimation unit 302 in FIG. 6, and the phase estimation unit 503 has the neutral point voltage data. It has a memory 61 and a phase estimation calculation unit 62.

Here, since the neutral point voltage data memory 61 corresponds to the neutral point voltage data memory 34 of FIG. 6 and the phase estimation calculation unit 62 corresponds to the phase estimation calculation unit 35 of FIG. 6, the neutral point voltage data memory Detailed description of 61 and the phase estimation calculation unit 62 will be omitted.
On the other hand, the LPF processing unit 502 of FIG. 10 has a time constant changing unit 64 for changing the time constant τ (cutoff frequency) of the first LPF unit 63 together with the first LPF unit 63 (first low-pass filter). It is different from the LPF processing unit 301 of.

The rotor position estimation means 6B including the first LPF unit 63 is a digital circuit provided in the microprocessor 510, and the first LPF unit 63 is a so-called digital filter.
The time constant changing unit 64 has a function of switching the time constant τ of the first LPF unit 63 between the first time constant τ1 which is a standard value and the second time constant τ2 shorter than the first time constant τ1. In other words, the first LPF unit 63 is configured in the microprocessor 510 to be adjustable to a second time constant τ2, which is a time constant shorter than the first time constant τ1.

Then, when the energization mode is switched, the time constant changing unit 64 changes the time constant τ of the first LPF unit 63 from the first time constant τ1 to the second time constant τ2, and changes the setting state of the second time constant τ2. After continuing for a predetermined period, the time constant τ of the first LPF unit 63 is returned to the standard value of the first time constant τ1.
That is, the phase estimation unit 503 in FIG. 10 uses the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 63 set in the second time constant τ2 for a predetermined period from the switching of the energization mode to the rotor position. In other periods, the rotor positions are estimated using the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 63 set in the first time constant τ1.

The second time constant τ2 is shorter than the first time constant τ1, and the response changes of the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 63 set in the second time constant τ2 are set in the first time constant τ1. It is faster than the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF section 63.
Therefore, when the energization mode is switched, the phase estimation calculation unit 62 can estimate the rotor position based on the neutral point voltages Vnn1 and Vnn2 in which the response delay is suppressed, and the first LPF set to the first time constant τ1. The estimation error of the rotor position can be made smaller than the case of estimating the rotor position based on the neutral point voltages Vnn1 and Vnn2 that have passed through the unit 63.

Further, since the first LPF unit 63 set to the first time constant τ1 is used except when the energization mode is switched, the rotor position is estimated based on the neutral point voltages Vnn1 and Vnn2 from which the noise component is sufficiently removed. It is possible to suppress the occurrence of estimation error due to noise components.
Further, since the first LPF unit 63 and the time constant changing unit 64 are composed of digital circuits, the time constant τ of the first LPF unit 63 can be easily changed.

The change of the time constant τ by the time constant changing unit 64 is not limited to the stepwise switching between the first time constant τ1 and the second time constant τ2, and is configured to switch to three or more types of time constants τ. The time constant τ can be continuously changed when switching from the first time constant τ1 to the second time constant τ2 or when returning from the second time constant τ2 to the first time constant τ1.

"Fifth Embodiment"
FIG. 11 is a system configuration diagram of the electric power steering device 200 for a vehicle.
The electric power steering device 200 is electrically driven and controlled by the motor control means 5 using the rotor position estimation results by the rotor position estimation means 6, 6A, 6B shown in the first to fourth embodiments described above. This is an application example of the motor 1.

The steering mechanism 210 of the electric power steering device 200 includes a steering wheel 201, a steering shaft (steering shaft) 202, a pinion shaft 203, and a rack shaft 204.
In this steering mechanism 210, when the driver of the vehicle rotates the steering wheel 201 left and right, steering torque is transmitted to the pinion shaft 203 via the steering shaft 202, and the rotational movement of the pinion shaft 203 is a straight line of the rack shaft 204. It is converted into motion and steers the steering wheels (not shown) connected to both ends of the rack shaft 204.

The pinion shaft 203 is provided with a steering torque sensor 206 that detects the steering torque of the steering wheel 201, and generates a torque command to the electric motor 1 that assists the steering torque based on the detected steering torque and vehicle speed information. , Drives the electric motor 1.
The deceleration mechanism 205 transmits the torque generated by the electric motor 1 to the steering wheels to assist the driver in steering.

As shown in FIG. 1, the motor control device 10 includes an inverter circuit 2, a DC bus current sensor 4, a motor control means 5, a rotor position estimation means 6, and a virtual neutral point circuit 7.
The DC power supply 3 is composed of a battery or the like and is connected to the motor control device 10.

In the electric power steering device 200 described above, the torque command value shown in FIG. 1 is a steering assist force calculated based on vehicle speed information, steering torque detected by the steering torque sensor 206, and the like.
Since the motor control device 10 estimates and calculates the rotor position based on the neutral point voltage, the electric power steering device 200 is a so-called sensorless control that eliminates the magnetic pole position sensor necessary for realizing the vector control of the electric motor 1. The electric motor 1 can be driven and controlled.

Further, the motor control device 10 is provided with the rotor position estimation means 6, 6A, 6B shown in the first to fourth embodiments described above, so that the rotor position can be estimated while suppressing the influence of noise. It is possible to suppress the occurrence of estimation error due to response delay when the energization mode is switched, and to control the electric motor 1 (steering assist force) with high accuracy.

The present invention is not limited to the above embodiment, and includes various modifications.
For example, the above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to the one including all the described configurations.
Further, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment.
Further, it is possible to add / delete / replace a part of the configuration of each embodiment with another configuration.

In the above embodiment, the rotor position estimation means 6 detects the neutral point voltage Vn0 used for estimating the rotor position with reference to the virtual neutral point potential Vnc as described above, but the reference potential is virtual. It is not limited to the neutral point potential Vnc.
For example, the rotor position estimating means 6 uses the ground level of the DC power supply 3 as a reference potential or the potential obtained by equally dividing the power supply voltage of the DC power supply 3 as a reference, and the reference potential and the electric motor 1 3 The difference from the phase winding connection point potential Vn can be obtained as the neutral point voltage Vn0.

Further, the electric motor 1 is a three-phase permanent magnet synchronous motor, but for example, the electric motor 1 can be a five-phase permanent magnet synchronous motor, and is not limited to three phases.
Further, the LPF processing unit 301 of FIG. 6 may include, for example, an analog filter composed of a capacitor parallel to the input signal and a resistor in series with the input signal as the first LPF unit 31 and the second LPF unit 32. Further, the first LPF unit 31 and the second LPF unit 32 as digital filters can be provided.

Further, the rotor position estimation means 6 has passed through the first phase estimation calculation unit and the second LPF unit 32, which perform the rotation element position estimation calculation based on the neutral point voltages Vnn1 and Vnn2 that have passed through the first LPF unit 31. Energize either the second phase estimation calculation unit that calculates the rotor position based on the sex point voltages Vnn1 and Vnn2, the estimation result by the first phase estimation calculation unit, or the estimation result by the second phase estimation calculation unit. It may have an estimation result switching unit that outputs to the motor control means 5 according to the mode.
In the above configuration, the estimation result switching unit outputs the estimation result by the second phase estimation calculation unit to the motor control means 5 when the energization mode is switched, and then outputs the estimation result by the first phase estimation calculation unit to the motor control means. By outputting to 5, as a result, when the energization mode is switched, the neutral point voltage Vnn1 and Vnn2 that have passed through the first LPF section 31 are not used, and the neutral point voltage Vnn1 that has passed through the second LPF section 32 is not used. This means that the rotor position was estimated using Vnn2.

Further, the first LPF unit 63 in FIG. 10 is used as an analog filter, and the time constant changing unit 64 switches the capacitance of the capacitor constituting the analog filter and the resistance value of the resistor to change the time constant τ of the first LPF unit 63. It can be adjustable. However, if the first LPF unit 63 is a digital filter as in the fourth embodiment, the time constant τ can be easily adjusted with a simple configuration.
Further, the system powered by the electric motor 1 driven and controlled by the motor control device 10 is not limited to the electric power steering device 200, and the electric motor 1 can be, for example, an electric motor for a vehicle pump or an internal combustion engine. It can be used as a power source for the variable valve mechanism.

1 ... electric motor, 1a ... stator, 1b ... rotor, 2 ... inverter circuit, 5 ... motor control means (inverter control unit), 6 ... rotor position estimation means, 7 ... virtual neutral point circuit, 10 ... motor Control device, 31 ... 1st LPF unit (1st low pass filter), 300 ... Neutral point voltage calculation unit, 301 ... LPF processing unit, 302 ... Phase estimation unit (rotor position estimation unit), Lu ... 1st stator winding Wire, Lv ... 2nd stator winding, Lw ... 3rd stator winding

Claims (7)

An electric motor comprising a rotor and a stator, wherein the stator has a first stator winding, a second stator winding, and a third stator winding that are star-connected to each other. In the motor control device for controlling the electric motor
Inverter circuit with multiple switching elements and
The inverter control unit is an inverter control unit that drives and controls the inverter circuit by switching a plurality of energization modes corresponding to combinations of on and off states of the plurality of switching elements based on the position of the rotor. Department and
A first low-pass filter with a first time constant,
The rotor position estimation unit estimates the position of the rotor based on the neutral point voltage of the stator and the switching state of the plurality of energization modes.
During the predetermined period from the switching of the energization mode between the plurality of energization modes, the position of the rotor is estimated without using the neutral point voltage that has passed through the first low-pass filter, and other than the predetermined period. During the period, the position of the rotor is estimated using the neutral point voltage that has passed through the first low-pass filter .
The rotor position estimation unit and
A motor control device characterized by having. The motor control device according to claim 1 has a second low-pass filter having a second time constant having a time constant shorter than that of the first time constant.
The rotor position estimation unit is a motor control device that estimates the position of the rotor by using the neutral point voltage that has passed through the second low-pass filter in the predetermined period . The motor control device according to claim 1 comprises a microprocessor and comprises.
Each of the inverter control unit, the rotor position estimation unit, and the first low-pass filter is a digital circuit provided in the microprocessor.
The first low-pass filter can be adjusted in the microprocessor to a second time constant, which is a time constant shorter than the first time constant.
The rotor position estimation unit passes through the first low-pass filter in which the time constant of the first low-pass filter is adjusted from the first time constant to the second time constant in the predetermined period. A motor control device characterized in that the position of the rotor is estimated using a voltage. In the motor control device according to claim 1, the rotor position estimation unit estimates the position of the rotor by using the neutral point voltage that has not passed through an arbitrary filter in the predetermined period. A motor control device characterized by. An electric motor comprising a rotor and a stator, the stator having a first stator winding, a second stator winding, and a third stator winding that are star-connected to each other. In the motor control device for controlling the electric motor
Inverter circuit with multiple switching elements and
The inverter control unit is an inverter control unit that drives and controls the inverter circuit by switching a plurality of energization modes corresponding to combinations of on and off states of the plurality of switching elements based on the position of the rotor. Department and
A first low-pass filter with a first time constant,
The rotor position estimation unit estimates the position of the rotor based on the neutral point voltage of the stator and the switching state of the plurality of energization modes.
The position of the rotor is estimated using the neutral point voltage that has passed through the first low-pass filter between the energization modes of one of the plurality of energization modes.
When the energization mode is switched between the plurality of energization modes, the value of the neutral point voltage that has not passed through an arbitrary filter before the energization mode switching timing and after the energizing mode switching timing. Estimate the position of the rotor using the average value with the value ,
The rotor position estimation unit and
A motor control device characterized by having. In the motor control device according to claim 5, the rotor position estimation unit estimates the position of the rotor using the average value, and then uses the neutral point voltage that has passed through the first low-pass filter. A motor control device that is used to estimate the position of the rotor. The motor control device according to claim 1 is
It is a virtual neutral point circuit and includes a first stator potential output unit, a second stator potential output unit, and a third stator potential output unit that are connected to each other in a star shape.
The first stator potential output unit has the same potential as the first stator winding.
The second stator potential output unit has the same potential as the second stator winding, and has the same potential.
The third stator potential output unit has the same potential as the third stator winding.
Equipped with the virtual neutral point circuit
The neutral point voltage is the difference between the stator neutral point potential, which is the potential of the neutral point of the stator, and the virtual neutral point potential, which is the potential of the neutral point of the virtual neutral point circuit. A motor control device characterized by.

Images