Electric motor drive device

Abstract

To provide a drive device that eliminates the need for constants that depend on an electric motor and improves the controllability of the electric motor.SOLUTION: A vector control unit 11 of a drive device includes a unit vector generation unit 35 that generates a unit vector in the same direction as the voltage command vector whose components include the voltage command value to an inverter 10, a vector rotation angle calculation unit 50 that calculates the rotation angle Δθu of the unit vector per control period of the vector control unit 11, a low-pass filter 52 that generates an estimated rotor rotation angle Δθr by removing noise from the rotation angle Δθu of the unit vector, and a sampling processing unit 55 that samples the unit vector and generates an output vector (<uα>, <uβ>) based on the estimated rotor rotation angle Δθr.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a drive device that drives an electric motor such as a synchronous motor or an induction motor, and more particularly to a drive device that performs vector control based on the output current of an inverter. [Background technology]

[0002] Conventional commonly used motor control methods include V / F control, which keeps the motor magnetic flux constant by outputting a voltage corresponding to a command frequency, and vector control, which breaks down the inverter output current into excitation current and torque current, and controls the excitation voltage and torque voltage so that a motor current appropriate for the load flows.

[0003] V / F control does not require high-speed calculations and can control motors with a simple configuration. However, this V / F control has little feedback information, so it is not possible to expect highly efficient control that matches the characteristics of individual motors. In addition, since the position of the motor rotor is not detected, in the case of a synchronous machine, the motor rotor may lose synchronization.

[0004] On the other hand, there is a control method called sensorless vector control, which can prevent a synchronous machine from losing synchronism and control the synchronous machine without using an expensive position sensor. This sensorless vector control is a control method that estimates the rotor position from the motor current fed back without using a position sensor.

[0005] Patent Document 1 discloses a sensorless vector control capable of controlling a motor without using a rotor position sensor. Fig. 7 is a block diagram showing a conventional sensor vector control. As shown in Fig. 7, a vector control unit 511 includes a 3 / 2 phase conversion unit 517 that converts three-phase currents Iu, Iv, and Iw detected by a current detector 512 into two-phase currents Iα and Iβ on a stationary coordinate system, and a stationary / rotating coordinate conversion unit 518 that converts the two-phase currents Iα and Iβ on the stationary coordinate system into two-phase currents Iγ and Iδ on a rotating coordinate system. The δ axis is an axis of the vector direction of the motor terminal voltage (inverter output voltage), and the γ axis is an axis perpendicular to the δ axis.

[0006] The vector control unit 511 further calculates a voltage command value e in the rotating coordinate system for minimizing the deviation between the current Iγ and the γ-axis current command value Iγ* (=0). INV -γ-axis current control unit 521 that generates γ* and a voltage command value e INV -δ-axis current control unit 522 that generates -δ* and a voltage command value e INV -γ*,e INV -δ* is the voltage command value e INV -α*,e INV -β*, and a two-phase voltage command value e INV -α*,e INV The inverter 522 includes a 2 / 3 phase converter 525 that converts −β* into three-phase voltage command values ​​Vu*, Vv*, and Vw*. The γ-axis current controller 521 and the δ-axis current controller 522 are PI controllers.

[0007] The vector control unit 511 further calculates a voltage command value e INV -α*,e INV Vector e with -β* as its component → INV *Length|e → INV A vector length calculation unit 529 calculates ||*|, a rotation angle estimator 531 calculates an estimated value Δθe* of the rotation angle [rad] per one control period, and a voltage command value e INV -α*,e INV-β* is a unit vector in the same direction as the voltage command vector (u α ,u β ), and a unit vector generating unit 535 for generating a unit vector (u α ,u β ) is outside the preset update range, the unit vector (u α ,u β ), and an update stop unit 537 that stops updating the unit vector (u α ,u β The output of the transition limiting unit 538 is α >, β > is sent to the stationary / rotating coordinate conversion unit 518 and the rotating / stationary coordinate conversion unit 524.

[0008] The rotation angle estimator 531 estimates the control period Tc [s] of the vector control unit 11, the back electromotive force constant Ke [Vs / rad] (one-phase electromotive force per electrical angular velocity), and the vector length |e → INV *| is used to calculate an estimated value Δθe* of the rotation angle [rad] per control cycle.

[0009] The unit vector generating unit 535 calculates the voltage command value e INV -α*,e INV -β* is a voltage command vector e → INV * its length |e → INV *| divides the unit vector (u α ,u β ) The unit vector generating unit 535 generates a unit vector for each control period Tc.

[0010] ​​When the unit vector generated by the unit vector generation unit 535 is within the update range, the update stop unit 537 replaces the previously generated unit vector with a newly generated unit vector (i.e., updates the unit vector), and on the other hand, when the unit vector generated by the unit vector generation unit 535 is outside the update range, the update stop unit 537 does not update the unit vector and outputs the previously generated unit vector. An example of the update stop unit 537 having such a function is a sample-and-hold circuit.

[0011] The vector control unit 511 further calculates each component u of the unit vector. α ,u β a speed converter 541 that calculates a frequency command value f* from a speed command value [rpm] input from the outside to the vector control unit 511, and a speed control unit 542 that calculates a δ-axis current command value Iδ* for minimizing the deviation between the rotation frequency f of the motor M and the frequency command value f*. The speed control unit 542 is a PI controller.

[0012] This configuration eliminates the need for an integral element used in general sensorless control to obtain a rotation angle signal from an angular velocity signal, which eliminates the need for parameters due to the moment of inertia of the rotating system, and eliminates the need for parameter adjustments, while improving transient response during sudden acceleration and deceleration. [Prior art documents] [Patent documents]

[0013] [Patent Document 1] JP 2022-117302 A Summary of the Invention [Problem to be solved by the invention]

[0014] In the conventional sensorless vector control shown in Figure 7, instantaneous values ​​are used except for the current control system controller (PI controller). This is to improve the speed transient response to the maximum extent possible. However, it is necessary to set the back electromotive force constant Ke to obtain the estimated value Δθe* of the rotation angle [rad] per control cycle. This back electromotive force constant Ke is an inherent value that depends on the electric motor M, and varies depending on the model of the electric motor M. To realize ideal operation of the electric motor M over the entire range from no load to full load, it is necessary to adjust the specific value of this back electromotive force constant Ke by trial and error.

[0015] In Fig. 7, the terminal voltage of the motor M is |e → INV The estimated value Δθe* of the rotation angle [rad] per control period is calculated by dividing |* by the back electromotive force constant Ke. However, the terminal voltage of the motor M contains not only the electromotive force (φω) generated by the rotation of the permanent magnet magnetic flux, but also voltage components (especially harmonic components) that depend on the winding current (Id, Iq). For this reason, errors occur in the estimated value Δθe* depending on the motor current and load state. As a result, it was sometimes impossible to drive the motor M with high accuracy.

[0016] Therefore, the present invention provides a drive device that does not require constants that depend on the electric motor and improves the controllability of the electric motor. [Means for solving the problem]

[0017] In one aspect, there is provided a drive device for an electric motor including an inverter, a current detector that detects an output current of the inverter, and a vector control unit that determines a voltage command value to the inverter, wherein the vector control unit includes a unit vector generation unit that generates a unit vector in the same direction as a voltage command vector whose components include the voltage command value to the inverter, a vector rotation angle calculation unit that calculates the rotation angle of the unit vector per control period of the vector control unit, a low-pass filter that generates an estimated rotor rotation angle by removing noise from the rotation angle of the unit vector, and a sampling processing unit that samples the unit vector and generates an output vector based on the estimated rotor rotation angle.

[0018] In one aspect, the sampling processing unit includes an update stopping unit that stops updating the unit vector when the unit vector is outside an update range, and a transition limiting unit that limits an amount of movement of the unit vector output from the update stopping unit to within a limited range, and the update range and the limited range are determined based on the estimated rotor rotation angle. In one aspect, the sampling processing unit includes a vector stabilizer configured to calculate a first hold vector by advancing a past output vector by one control cycle using the estimated rotor rotation angle. In one aspect, the vector stabilizer is configured to output the primary hold vector as a current output vector when a value obtained by multiplying the absolute value of the estimated rotor rotation angle by a predetermined coefficient is smaller than the absolute value of the rotation angle of the unit vector, and to output a hybrid output vector obtained by combining the unit vector and the primary hold vector at a predetermined ratio as a current output vector when a value obtained by multiplying the absolute value of the estimated rotor rotation angle by a predetermined coefficient is equal to or greater than the absolute value of the rotation angle of the unit vector. Effect of the Invention

[0019] By introducing the vector rotation angle calculation unit and the low-pass filter into the vector control unit, the back electromotive force constant that depends on the electric motor can be eliminated. In addition, since the low-pass filter removes noise from the rotation angle of the unit vector, the estimation accuracy of the estimated rotor rotation angle can be improved. As a result, the drive device can drive the electric motor with improved controllability. [Brief description of the drawings]

[0020] [Figure 1] FIG. 2 is a block diagram illustrating one embodiment of a drive device. [Diagram 2] 2(a) to 2(c) are diagrams showing the state of a unit vector in the same direction as the command value of the motor terminal voltage (inverter output voltage) before and after one control period. [Diagram 3] FIG. 13 is a diagram illustrating an update range used to determine whether or not to update a unit vector. [Figure 4] FIG. 13 is a diagram showing a limit range in which the movement of a unit vector is limited. [Diagram 5] FIG. 13 is a diagram showing a unit vector whose movement is restricted by a transition restriction unit. [Figure 6] FIG. 13 is a block diagram showing another embodiment of the drive device. [Figure 7] FIG. 1 is a block diagram showing a conventional sensor vector control. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a block diagram showing an embodiment of a drive device. This drive device is an inverter device (power conversion device) that drives an electric motor M, and is composed of a plurality of elements including an inverter 10 and a vector control unit 11 as shown in Fig. 1. That is, the drive device includes the inverter 10 that generates a voltage to be applied to the electric motor M, the vector control unit 11 that determines a voltage command value to the inverter 10, and a current detector (ammeter) 12 that detects a current flowing from the inverter 10 to the electric motor M. The electric motor M is a permanent magnet synchronous motor (PMSM) with a permanent magnet in its rotor. The vector control unit 11 can be composed of a processing unit such as a CPU (central processing unit) and a computer equipped with a memory.

[0022] The current detector 12 measures three-phase currents Iu, Iv, Iw flowing from the inverter 10 to the motor M. The measured values ​​are input to the vector control unit 11. The three-phase currents Iu, Iv, Iw may be measured by measuring the currents of any two phases and determining the remaining current from the equation Iu+Iv+Iw=0. The vector control unit 11 generates three-phase voltage command values ​​Vu*, Vv*, Vw* based on the three-phase currents Iu, Iv, Iw and a speed command value input from the outside. The vector control unit 11 sends these three-phase voltage command values ​​Vu*, Vv*, Vw* to the inverter 10. The inverter 10 generates voltages based on the three-phase voltage command values ​​Vu*, Vv*, Vw* and applies them to the motor M.

[0023] The vector control unit 11 includes a 3 / 2 phase conversion unit 17 that converts the three-phase currents Iu, Iv, Iw detected by the current detector 12 into two-phase currents Iα, Iβ on a stationary coordinate system, and a stationary / rotating coordinate conversion unit 18 that converts the two-phase currents Iα, Iβ on the stationary coordinate system into two-phase currents Iγ, Iδ on a rotating coordinate system. The δ-axis is the axis of the vector direction of the motor terminal voltage (inverter output voltage), and the γ-axis is the axis perpendicular to the δ-axis.

[0024] The vector control unit 11 further calculates a voltage command value e in the rotating coordinate system for minimizing the deviation between the current Iγ and the γ-axis current command value Iγ* (=0). INV-γ-axis current control unit 21 that generates γ* and a voltage command value e in a rotating coordinate system for minimizing the deviation between the current Iδ and the δ-axis current command value Iδ* INV -δ-axis current control unit 22 that generates -δ* and a voltage command value e INV -γ*,e INV -δ* is the voltage command value e INV -α*,e INV -β*, and a two-phase voltage command value e INV -α*,e INV The controller 20 includes a 2 / 3 phase converter 25 that converts −β* into three-phase voltage command values ​​Vu*, Vv*, and Vw*. The γ-axis current controller 21 and the δ-axis current controller 22 are PI controllers.

[0025] The vector control unit 11 further calculates a voltage command value e INV -α*,e INV Vector e with -β* as its component → INV *Length|e → INV A vector length calculation unit 29 calculates a voltage command value e INV -α*,e INV -β* is a unit vector in the same direction as the voltage command vector (u α ,u β ), and a unit vector generating unit 35 for generating a unit vector (u α ,u β A vector rotation angle calculation unit 50 calculates the rotation angle Δθu of the unit vector (u α ,u β A low-pass filter 52 generates an estimated rotor rotation angle Δθr by removing noise from the rotation angle Δθu of the unit vector (u α ,u β ) and outputs the output vector ( α >, β The sampling processor 55 generates a sampling signal.

[0026] ​​In this embodiment, the sampling processing unit 55 includes an update stopping unit 37 and a transition limiting unit 38. The update stopping unit 37 detects a unit vector (u α ,u β ) is outside the preset update range, the unit vector (u α ,u β The transition limiting unit 38 is configured to stop updating the unit vector (u α ,u β The transition limiting unit 38 is configured to limit the amount of movement of the output vector ( α >, β >) is sent to the stationary / rotating coordinate conversion unit 18 and the rotating / stationary coordinate conversion unit 24.

[0027] The unit vector generating unit 35 calculates the voltage command value e INV -α*,e INV -β* is a voltage command vector e → INV * its length |e → INV *| divides the unit vector (u α ,u β The unit vector generating unit 35 generates a unit vector for each control period Tc of the vector control unit 11.

[0028] The vector rotation angle calculation unit 50 calculates the current unit vector (u α [n],u β [n]) and the past unit vector (u α [n-1],u β [n-1]), the rotation angle Δθu of the unit vector is calculated using the following equation (1). Δθu=u α [n]u β [n-1]-u α [n-1]u β [n] (1) Here, [n] is the unit vector (u α ,u β ​​) represents the current sampling time point (i.e., the current control cycle), and [n-1] represents the previous sampling time point (i.e., the previous control cycle).

[0029] The low-pass filter 52 generates an estimated rotor rotation angle Δθr by removing noise such as harmonic components from the rotation angle Δθu of the unit vector. The estimated rotor rotation angle Δθr is input to both the update stop unit 37 and the transition limit unit 38. In order to improve the accuracy of the estimated rotor rotation angle Δθr, it is necessary to set the cutoff frequency of the low-pass filter 52 low, while in order to deal with a sudden change in the motor speed, it is necessary to set the cutoff frequency high. However, in normal motor driving, it is considered that the speed signal hardly contains a signal exceeding several tens of Hz, so the cutoff frequency of the low-pass filter 52 is set to about several tens of Hz. In one example, the cutoff frequency of the low-pass filter 52 is 20 Hz.

[0030] By introducing the vector rotation angle calculation unit 50 and the low-pass filter 52 into the vector control unit 11 in this way, the back electromotive force constant Ke that depends on the electric motor M can be made unnecessary. In addition, since the low-pass filter 52 removes noise from the rotation angle Δθu of the unit vector, the estimation accuracy of the estimated rotor rotation angle Δθr can be improved. As a result, the drive device can drive the electric motor M with improved controllability.

[0031] When the unit vector generated by the unit vector generation unit 35 is within the update range, the update stop unit 37 replaces the previously generated unit vector with a newly generated unit vector (i.e., updates the unit vector), and on the other hand, when the unit vector generated by the unit vector generation unit 35 is outside the update range, the update stop unit 37 does not update the unit vector and outputs the previously generated unit vector. An example of the update stop unit 37 having such a function is a sample-and-hold circuit.

[0032] The update range is determined based on the estimated rotor rotation angle Δθr. Specifically, the update range is determined as follows. Δθrenew=MrenewΔθr (2) The coefficient Mrenew is a parameter given as a constant greater than or equal to 1. This allows an update range that is adapted to the operating speed to be given.

[0033] The vector control unit 11 further calculates each component u of the unit vector. α ,u β a speed converter 41 that calculates a frequency command value f* from a speed command value [rpm] input from the outside to the vector control unit 11, and a speed control unit 42 that calculates a δ-axis current command value Iδ* to minimize the deviation between the rotation frequency f of the motor M and the frequency command value f*. The speed control unit 42 is a PI controller.

[0034] As described above, the vector control unit 11 incorporates a current control system including αβ-γδ transformation and γδ-αβ transformation, and controls the actual current of the motor M to follow the current command values ​​Iγ* and Iδ*. Note that the δ axis is the axis of the vector direction of the motor terminal voltage (inverter output voltage), and the γ axis is an axis perpendicular to the δ axis. The αβ-γδ transformation and γδ-αβ transformation are not given a coordinate angle, but the voltage command value e to the inverter 10 is INV -α*,e INV -β* is a unit vector in the same direction as the voltage command vector (u α ,u β ) is given by each component of the unit vector u α ,u β are given as cosine and sin values, respectively, and a rotation matrix operation is performed in each block. Since sin and cosine functions are not used, the amount of calculation (calculation load) of the vector control unit 11 can be reduced.

[0035] In this embodiment, the voltage command value e INV -α*,e INV -β* converted to a unit vector, u α ,u β Instead of using it as is, we use u according to given conditions. α ,u βand a transition limiting unit 38 that limits the amount of movement of the unit vector per control period. α >, β In particular, the update stop unit 37 performs coordinate conversion using the unit vector (u α ,u β ) is deviated from the direction of the back electromotive force caused by the magnetic flux of the permanent magnet of the rotor of the motor M, this prevents the γδ axes from losing synchronization.

[0036] Next, the update stop unit 37 will be described. FIG. 2 shows a unit vector (u α ,u β ) before and after one control cycle. [n] is a unit vector (u α ,u β ) represents the current sampling time point (i.e., the current control cycle), and [n-1] represents the previous sampling time point (i.e., the previous control cycle). Tc is the control cycle of the vector control unit 11, and ωe is the rotational angular velocity (electrical angle) [rad / s].

[0037] When the rotor rotation and the motor terminal voltage are synchronized, the unit vector (u α ,u β ) advances the rotation angle by Δθe = Tcωe per control period, as shown in Figure 2(a). When the unit vector follows the transition shown in Figure 2(a), the phase of the motor terminal voltage and the back electromotive force caused by the rotor rotation are considered to be almost the same, and the command value of the current control system and the actual current value are considered to be almost the same. Therefore, the unit vector (u α ,u β ) the rotation angle information of the rotor of the electric motor M can be obtained.

[0038] ​​On the other hand, Fig. 2(b) shows a case where the unit vector is significantly ahead of the rotation angle Δθe per control cycle, and Fig. 2(c) shows a case where it is significantly behind. In these cases, it is considered that a large voltage is generated in the impedance of each phase of the motor, especially in the self-inductance, and it is not possible to simply find the phase of the back electromotive force caused by the rotation from the phase of the terminal voltage. In other words, u α ,u β I can't get the rotation angle information from

[0039] Therefore, in this embodiment, an update range including the rotation angle Δθe is provided (Δθrenew>Δθe) as shown by the thick line in FIG. 3. The update range (-Δθrenew to +Δθrenew) is the unit vector (u α [n-1],u β [n-1]) is an angle range that includes the rotation angle Δθe from the estimated rotor rotation angle Δθr [n-1]. Δθrenew, which determines the update range, is calculated from the estimated rotor rotation angle Δθr and the coefficient Mrenew (a constant equal to or greater than 1) as shown in the above formula (2).

[0040] The unit vector (u α ,u β ) within this update range, each component of the unit vector u α ,u β Since the rotation angle can be obtained from u α ,u β If it is outside the update range, the update stop unit 37 samples u α ,u β Without sampling, the unit vector (u α ,u β ) is output by the update stop unit 37. α ,u β In the following description, the unit vector output from the update stop unit 37 is referred to as (u' α ,u' β ) is expressed as

[0041] In the embodiment shown in FIG. 3, the update range (−Δθrenew to +Δθrenew) is the unit vector (u α [n-1],u β [n-1]), but in one embodiment, the update range (-Δθrenew to +Δθrenew) is the unit vector (u α [n-1],u β [n-1]) may have an asymmetric angle.

[0042] Next, the transition limiting unit 38 will be described. As described above, if a unit vector in the same direction as the motor terminal voltage is acquired only under conditions where the rotation angle can be acquired, erroneous rotation angle information can be ignored, and operation of the electric motor M can be maintained by performing γδ-αβ transformation using this information. However, if the suspension of unit vector updates is maintained, the rotor of the electric motor M continues to rotate during that time, so the phase of the back electromotive force caused by the rotation continues to advance, and when updating is resumed, the angle difference with the previously sampled unit vector becomes large. In other words, in one control cycle immediately after recovery from the suspension of updates, the unit vector (u' α ,u' β ) fluctuates significantly, causing the angle of the γδ axes to fluctuate significantly. This causes a large step to be input into the current control system. This causes a large voltage to be generated in the self-inductance of each phase of the motor, making it impossible to detect the back electromotive force caused by the rotation, and the current control system also enters a transient state in which the current command value and the actual current diverge. At this point, the state shown in Figure 2(b) or Figure 2(c) is reached, and sampling is stopped again. For this reason, stopping the update of the 1-degree unit vector causes sampling to be stopped and restarted repeatedly, which may make it impossible to obtain continuous rotation angle information.

[0043] Therefore, in this embodiment, a limit is placed on the angle of the unit vector that can move in one degree, and the transition limiting unit 38 operates to limit the amount of movement of the angle of the unit vector to a specified value if a fluctuation exceeding a preset limit range occurs in one control cycle.

[0044] Fig. 4 is a diagram showing a limit range (indicated by a thick line) in which the movement of a unit vector is limited. Fig. 4 is very similar to Fig. 3, but the angle that defines the limit range is given by Δθlimit, and the unit vector is the unit vector (u' α ,u' β The limit range is given as ±Δθlimit, but like the update range (-Δθrenew to +Δθrenew), it does not have to be given as an angle symmetrical with respect to the unit vector in the previous control cycle.

[0045] However, for the rotation direction, since it is necessary to catch up with the unit vector that has advanced after the update has stopped, a limit range that basically includes the rotation angle Δθe is set (Δθlimit>Δθe). In other words, the limit range (-Δθlimit to +Δθlimit) is the unit vector (u' α [n-1],u' β [n-1]) is an angle range that includes the rotation angle Δθe from

[0046] The limit range is determined based on the estimated rotor rotation angle Δθr. Specifically, the limit range is determined as follows. Δθlimit=MlimitΔθr (3) The coefficient Mlimit is a parameter given as a constant greater than or equal to 1. This makes it possible to provide a limit range that is adapted to the operating speed.

[0047] The unit vector (u') that exceeds the limit range given in Figure 4 α [n],u' β When the next control cycle is reached, the amount of movement (or rotation) of the unit vector is limited to the upper limit value +Δθlimit or the lower limit value -Δθlimit of the limited range. α >[n], β >[n]) is the output vector of the transition limiter 38. In this case, multiple control periods are used to α [n],u' β In reality, the output vector changes to catch up with (u' [n]) every control period.​​α ,u' β ) is updated, and the output vector ( α >, β That is, since Δθlimit>Δθe as described above, the output vector ( α >, β >) will eventually (u' α ,u' β ) can be caught up.

[0048] The transition limiting unit 38 limits excessive movement of the unit vector, and when a fluctuation that exceeds a preset limit range occurs, limits the amount of movement of the unit vector within the limit range. Therefore, the difference between the angle of the previously updated unit vector and the actual rotor angle falls within the limit range. As a result, rotor angle information can be obtained continuously. In the embodiment shown in FIG. 1, the unit vector (u' α ,u' β ) is limited by the transition limiting unit 38, but in one embodiment, the transition limiting unit 38 functions effectively even in a configuration without the update stopping unit 37. In other words, the transition limiting unit 38 limits excessive movement of the unit vector generated by the unit vector generating unit 35, thereby making it possible to continuously acquire rotor angle information and allowing the vector control unit 11 to stably control the electric motor M.

[0049] Next, another embodiment of the vector control unit 11 will be described with reference to Fig. 5. The configuration and operation of this embodiment that are not specifically described are the same as those of the embodiment described with reference to Figs. 1 to 4, so that the overlapping description will be omitted.

[0050] In the embodiment described with reference to FIGS. 1 to 4, the update stop process, i.e., the unit vector (u α ,u β ​​​​The hold operation of the current feedback circuit is the previous value hold (zero-order hold). In addition, a movement limiting process is introduced to prevent a step change from being input to the current feedback system and causing a transient state. However, some transient states may occur.

[0051] Therefore, in the embodiment described below, a primary hold is adopted in which the estimated rotor rotation angle Δθr is used to advance the unit vector by one control cycle. In addition, in order to mitigate a jump that may occur in the unit vector when transitioning from a hold operation to a sampling operation, a process is executed to seamlessly connect the unit vector generated in the current control cycle with the unit vector advanced by one control cycle (primary hold vector). This makes the movement amount limiting process unnecessary, and realizes a simpler algorithm.

[0052] As shown in Fig. 5, the sampling processing unit 55 of this embodiment includes a vector stabilizer 60 configured to calculate a first hold vector by advancing a past output vector by one control cycle using an estimated rotor rotation angle Δθr. The vector stabilizer 60 is connected to the unit vector generating unit 35, the vector rotation angle calculating unit 50, and the low-pass filter 52. In this embodiment, the update stopping unit 37 and the transition limiting unit 38 shown in Fig. 1 are not provided.

[0053] The unit vector (u α ,u β ), the rotation angle Δθu of the unit vector per control period calculated by the vector rotation angle calculation unit 50, and the estimated rotor rotation angle Δθr generated by the low-pass filter 52 are input to a vector stabilizer 60. The vector stabilizer 60 outputs an output vector (u" α [n],u” β [n]) and generate the output vector (u” α [n],u” β [n]) is sent to the stationary / rotating coordinate conversion unit 18 and the rotating / stationary coordinate conversion unit 24.

[0054] If the value obtained by multiplying the absolute value of the estimated rotor rotation angle Δθr by a predetermined coefficient Mrenew is smaller than the absolute value of the rotation angle Δθu of the unit vector, the vector stabilizer 60 calculates the past output vector (u" α [n-1],u” β The primary hold vector, which is the current output vector (u” [n-1]) advanced by one control period, is α [n],u” β Specifically, the vector stabilizer 60 is configured to output the current output vector (u" [n]) using the following equation (4): α [n],u” β [n]).

number

[0055] Furthermore, if the value obtained by multiplying the absolute value of the estimated rotor rotation angle Δθr by a predetermined coefficient Mrenew is equal to or greater than the absolute value of the rotation angle Δθu of the unit vector, the vector stabilizer 60 stabilizes the current unit vector (u α [n],u β The hybrid output vector, which combines the current output vector (u" [n]) and the first hold vector in a predetermined ratio, is α [n],u” β Specifically, the vector stabilizer 60 is configured to output the current output vector (u" [n]) using the following equation (5): α [n],u” β [n]).

number

[0056] As can be seen from the above equation (5), the current unit vector (u α [n],u βThe ratio of the current unit vector (u [n]) to the first hold vector is expressed as 1-μ:μ using the forgetting factor μ. The forgetting factor μ is a predetermined constant that is greater than 0 and less than 1 (0<μ<1). For example, when the forgetting factor μ is 0.8, the current unit vector (u α [n],u β A hybrid output vector is generated by combining 20% ​​of the vectors (u” [n]) and 80% of the first hold vector. This hybrid output vector is the output vector (u” α [n],u” β [n]) is output from the vector stabilizer 60.

[0057] In one embodiment, the forgetting factor μ may be a variable. Specifically, the forgetting factor μ may be increased according to the rotation angle Δθu of the unit vector. For example, the forgetting factor μ is determined by the following formula: μ = |Δθu-Δθr|÷|Δθr| The above formula corresponds to normalizing the deviation between the rotation angle Δθu of the unit vector and the estimated rotor rotation angle Δθr by using the estimated rotor rotation angle Δθr. However, the forgetting factor μ is limited to 0≦μ≦1.

[0058] In calculating the primary hold vector, a trigonometric function is used as the rotation matrix. However, if the control period is sufficiently short in proportion to the rotation speed of the motor M, the above equations (4) and (5) can be expressed by the following approximate equations (6) and (7).

number

number

[0059] According to the above formulas (6) and (7), trigonometric functions are eliminated from all processing, so that the calculation load for control can be significantly reduced.

[0060] The above-described embodiments have been described for the purpose of enabling a person having ordinary skill in the art to practice the present invention. Various modifications of the above-described embodiments are naturally possible for a person skilled in the art, and the technical idea of ​​the present invention can be applied to other embodiments. Therefore, the present invention is not limited to the described embodiments, but is to be interpreted in the broadest scope according to the technical idea defined by the claims. [Explanation of symbols]

[0061] 10 Inverter 11 Vector control unit 12 Current detector 17 3 / 2 phase converter 18 Static / rotating coordinate conversion section 21 γ-axis current control section 22 δ-axis current control section 24 Rotation / stationary coordinate conversion section 25 2 / 3 phase converter 29 Vector Length Calculation Unit 31 Rotation Angle Estimator 35 Unit Vector Generator 36 Filters 37 Update stop part 38 Transition Restriction Section 50 Vector rotation angle calculation unit 52 Low-pass filter 55 Sampling processing section 60 Vector Stabilizer Medium motor

Claims

1. A drive device for an electric motor including an inverter, a current detector that detects an output current of the inverter, and a vector control unit that determines a voltage command value for the inverter, The vector control unit includes: a unit vector generating unit that generates a unit vector in the same direction as a voltage command vector whose components include a voltage command value to the inverter; a vector rotation angle calculation unit that calculates a rotation angle of the unit vector per one control cycle of the vector control unit; a low-pass filter for generating an estimated rotor rotation angle by removing noise from the rotation angle of the unit vector; a sampling processor that samples the unit vector and generates an output vector based on the estimated rotor rotation angle.

2. The sampling processing unit includes: an update stopping unit that stops updating the unit vector when the unit vector is outside an update range; a transition limiting unit that limits a movement amount of the unit vector output from the update stopping unit to a limited range; The drive device according to claim 1 , wherein the update range and the limit range are determined based on the estimated rotor rotation angle.

3. 2. The drive device according to claim 1, wherein the sampling processing unit includes a vector stabilizer configured to calculate a primary hold vector by advancing a past output vector by one control period using the estimated rotor rotation angle.

4. The vector stabilizer comprises: If a value obtained by multiplying the absolute value of the estimated rotor rotation angle by a predetermined coefficient is smaller than the absolute value of the rotation angle of the unit vector, the primary hold vector is output as a current output vector; 4. The drive device according to claim 3, wherein when a value obtained by multiplying an absolute value of the estimated rotor rotation angle by a predetermined coefficient is equal to or greater than an absolute value of the rotation angle of the unit vector, a hybrid output vector obtained by combining the unit vector and the primary hold vector at a predetermined ratio is output as a current output vector.

Images