Device for controlling a three-phase rotating electrical machine to reduce the peak value of the phase current

Abstract

A control device for driving a three-phase rotating electrical machine (80) having two three-phase winding groups (801, 802), wherein the three-phase rotating electrical machine (80) generates a torque which is a sum of magnetic torque and reluctance torque when alternating current with mutually different phases is passed through the two winding groups (801, 802), the control device comprising: two power converters (601, 602), each provided for a corresponding one of the two winding groups (801, 802), each power converter (601, 602) supplying each of the two winding groups (801, 802) with alternating current, the alternating current supplied to the two winding groups (801, 802) having the same amplitude and the mutually different phases being defined as 30 ± 60 x n [degrees], where n is an integer;anda control unit (65) configured to calculate the d-axis current and the q-axis current of the 6(2k+1)-th order component superimposed on a fundamental wave component on a dq coordinate, where k=0 or a higher integer, to reduce a peak value of a first-order component in the phase current conducted into the two winding groups (801, 802), thereby controlling conduction of the three-phase rotating electric machine (80), wherein the control unit (65) is configured to calculate current such that an amplitude of the q-axis current of the 6(2k+1)-th order component is greater than an amplitude of the d-axis current of the 6(2k+1)-th order component.;

Description

TECHNICAL FIELDThe present invention relates to a control device for a three-phase rotating electric machine.PRIOR ARTIn essence, in polyphase rotating electric machines, a relationship between an amount of current conducted into the windings and the drive torque is such that the greater the line current, the greater the drive torque. However, when a large current flows through the windings, a large amount of heat is generated, and damage to the rotating electric machine or malfunction in the rotating electric machine may be caused.In this regard, a method is known in which a harmonic current component superimposes a fundamental wave, which is the first-order phase current component, thereby reducing the peak value of the phase current and the heat caused by the current peak value.For example, JP 5 672 278 B2 discloses a method used for a three-phase rotating machine having two winding groups, wherein current having predetermined phase differences is supplied to the two winding groups, whereby 6-th-order torque ripple caused by 5-th-order harmonic components is cancelled and the peak value of the phase current is reduced. In this method, a 6th-order dq axis current for superposing 0th-order components on the dq coordinate system is calculated, thereby generating 5th-order harmonic components for superposing the first-order components of the phase current.The technique in the above-mentioned method of JP 5 672 278 B2 is to reduce torque ripple of the magnetic torque, assuming that a surface magnetic type rotating electric machine is mainly used. However, a permanent magnet embedded type rotating electric machine in which the d-axis inductance and the q-axis inductance have different values generates reluctance torque based on the difference. Also, in the rotating electric machine having a surface magnet, a small reluctance torque can be generated.Thus, in the case where the method of JP 5 672 278 B2 described above is applied to a three-phase rotating electric machine that generates torque including magnetic torque and reluctance torque, there arises a problem that torque ripple of the reluctance torque cannot be cancelled out.From EP 2 958 228 A1 a motor drive device is further known, wherein a brushless motor comprises an armature iron core wound with armature windings including a plurality of winding sets, each of which includes polyphase windings. A voltage application device applies voltages to the plurality of winding sets. A controller calculates a voltage command for the plurality of winding sets and controls the voltage application device based on the voltage command. The controller controls the voltage application device such that the induced voltages generated in the plurality of winding sets by rotation of the brushless motor have a trapezoidal waveform, and a phase difference between the voltages for the plurality of winding sets becomes a value that reduces torque ripple caused by the induced voltages.U.S. Pat. No. 6,710,495 B2 relates generally to the field of electric motors and energy conversion systems for driving electric motors.SUMMARYIt is an object of the present invention to provide a control device for driving a three-phase rotating electric machine, which is capable of reducing a peak value of the phase current and the torque ripple, in a range of control devices of three-phase rotating electric machines that generate torque including the magnetic torque and the reluctance torque.The object is achieved by the subject matter of the main claim. Advantageous refinements are specified in the dependent claims.Note that the 6(2k+1)-th order component corresponds to 6-th, 18-th, and 30-th order components.In the three-phase rotating electric machine that generates reluctance torque, torque ripple is generated from a 12 (2k+1)-th order component based on a product of the amplitudes of the d-axis current of 6 (2k+1)-th order components and the q-axis current of 6 (2k+1)-th order components.In order to reduce the torque ripple, the product of the amplitudes of the q-axis current and the d-axis current of the 6(2k+1)-th order component needs to become smaller. However, in the case where an amplitude of the d-axis current is set larger and an amplitude of the q-axis current is set smaller, a reduction effect on the peak phase current is small.According to the present invention, an amplitude of the q-axis current of the 6(2k+1)-th order component is set larger than an amplitude of the d-axis current of the 6(2k+1)-th order component.Thus, in the three-phase rotating electric machine that generates torque including magnetic torque and reluctance torque, torque ripple can be reduced and the peak value of the phase current can be reduced accordingly. As a result, vibrations and noise and also heat and power loss caused by the peak value of the phase current can be reduced.Preferably, the control unit sets the amplitude of the d-axis current of the 6(2k+1)-th order component to 0 and transmits only the q-axis current of the 6(2k+1)-th order component. Thus, torque ripple of the 12(2k+1)-th order component can be set to 0. Thereby, the influence of the torque ripple can be eliminated.BRIEF DESCRIPTION OF THE DRAWINGSThe drawings show: FIG. 1 is a diagram showing an overall configuration of an electric power steering apparatus to which a control device of a three-phase rotating electric machine according to the embodiments of the present invention is applied; FIG. 2 is a diagram showing an overall configuration of a control device of a three-phase rotating electric machine; FIG. 3 is a block diagram showing a control unit according to a first embodiment; FIG. 4 is a block diagram showing a peak reduction current target value calculation unit shown in FIG. 3 ; FIG. 5 is a graph showing a relationship between a rotational speed of a three-phase rotating electric machine and a d-axis current limiting gain; FIG. 6 is a graph showing a relationship between a current amplitude limit value and a current amplitude limit value; FIG. 7A is a graph showing a relationship between a rotational speed of the three-phase rotating electric machine and a current amplitude gain; FIG. 7B is a graph showing a relationship between a rotational speed of the three-phase rotating electric machine and a current amplitude gain; FIG. 8 is a graph showing a relationship between a rotational speed of the three-phase rotating electric machine and a phase compensation amount; FIG. 9 is a graph showing a relationship between a 6-th-order q-axis current ratio and a phase current peak reduction ratio; FIG. 10A is a graph showing a phase current waveform with 4% of the 6-th-order q-axis current ratio; FIG. 10B is a graph showing an enlarged view of the tip portion of the phase current waveform shown in FIG. 10A ; FIG. 11A is a graph showing a phase current waveform with 4.4% of the 6th-order q-axis current ratio; FIG. 11B is a graph showing an enlarged view of the tip portion of the phase current waveform shown in FIG. 11A ; FIG. 12A is a graph showing a phase current waveform with 5% of the 6-th-order q-axis current ratio; FIG. 12B is a graph showing an enlarged view of the tip portion of the phase current waveform shown in FIG. 12A ; FIG. 13 is a block diagram showing a control unit according to a second embodiment; FIG. 14A is a graph showing a relationship between a d-axis current amplitude ratio and a phase current peak reduction ratio; FIG. 14B is a graph showing a relationship between a phase current peak reduction ratio and a torque ripple index; FIG. 15 is a block diagram showing a control unit according to a third embodiment; and FIG. 16 is a block diagram showing a control unit according to a fourth embodiment.DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTSHereinafter, a plurality of embodiments of a control device of a three-phase rotating electric machine will be described with reference to the drawings. In each of the embodiments, an ECU is used as a control device of a three-phase rotating electric machine for an electric power steering device of a vehicle that controls conduction of a motor (i.e., a three-phase rotating electric machine) that generates the steering assist torque. Referring to FIGS. 1 and 2, a common configuration will be described first by the respective embodiments.[Configuration of Electric Power Steering Apparatus]FIG. 1 shows an overall configuration of a steering system 100 including an electric power steering device 90. it is to be noted that the electric power steering device 90 shown in FIG. 1 is configured as a steering column assist type, but may also be configured as a rack assist type.The steering system 100 includes a steering wheel 91, a steering shaft 92, a pinion gear 96, a rack 97, wheels 98, and the electric power steering apparatus 90.The steering wheel 91 is connected to the steering shaft 92. The pinion gear 96, which is disposed at a rear end portion of the steering shaft 92, meshes with the rack 97. At both ends of the rack 97, wheels 98 are provided via a tie rod or the like. When the driver rotates the steering wheel 91, the steering shaft 92 connected to the steering wheel 91 is rotated.A rotational movement of the steering shaft 92 is converted into a linear movement of the rack 97 by the pinion gear 96, and the wheels 98 are steered at an angle in response to a displacement of the rack 97.The power steering apparatus 90 includes a steering torque sensor 93, an ECU 10, a motor 80, and a reduction gear 94. the steering torque sensor 93 is provided in a portion of the steering shaft 92, whereby the driver's steering torque Ts is detected. The ECU 10 controls driving of the motor 80 based on the steering torque Ts so that the motor 80 generates a desired assist torque. The assist torque output by the motor 80 is transmitted to the steering shaft 92 via the reducer 94.[Configuration of Control Device]FIG. 2 shows an overall configuration of the control device. The motor 80 as a three-phase rotating electric machine is configured as a brushless motor having two three-phase winding groups 801 and 802. Each phase coil 821, 822, and 823of the second winding group 802 is arranged to have a positional relationship, i.e., an electrical angle (electrical angle) of 30 degrees with respect to the phase coils 811, 812, 813of the first winding group 801. The rotation angle sensor 85 detects the phase angle θ of the motor 80 and outputs the angle θ to a control unit 65.Further, the motor 80 is configured as a permanent magnet embedded type synchronous motor (i.e., IPSM). In essence, a permanent magnet embedded type rotating electric machine generates torque which is a sum of magnet torque and reluctance torque. As described above, the control device according to the embodiments is configured to control the drive of the motor 80 that generates a sum of the magnet torque and the reluctance torque. The configuration shown in FIG. 2 is different from the configuration disclosed in the above-mentioned JP 5 672 278 B2 in that the motor 80 is clearly referred to as an IPMSM.The ECU 10 as a control device includes inverters 601 and 602, current sensors 701 and 702, and a control unit 65.The first inverter 601 as a first power converter and the second inverter 602 as a second power converter 602 are provided corresponding to two winding groups 801 and 802, and each of the two winding groups 801 and 802 is conducted with alternating current. The alternating current that is supplied to the respective two winding groups has a phase difference (i.e., 30±60×n) between two winding groups, where n is an integer. In addition, the alternating current that is conducted into the corresponding two winding groups has the same amplitude between two winding groups.Hereinafter, a "system" is defined as a unit including a winding group and an inverter corresponding to the winding group. In the reference numerals of the elements, a third digit, i.e., 1 or 2, represents a system number. Also, in symbols representing physical quantities such as current or voltage, the last digit such as 1 or 2 represents the system number. In the case where a physical quantity needs to be designated for two systems, no last location, 1 or 2, is used.For the inverters 601 and 602, each inverter has a configuration in which 6 switching elements (i.e., 611 to 616, 621 to 626) are connected between the high potential line Lp and the low potential line Lg to form a bridge connection. The inverters 601 and 602 are each controlled by the control unit 65, so that a drive signal of a control circuit 68 is output to the inverters 601 and 602 to perform switching operation, thereby converting direct current of a battery 51 to alternating current, which is supplied to the winding groups 801 and 802.The input portion of each inverter 601 and 602 includes a power relay 521, 522 and a smoothing capacitor 53 according to the respective system. The input voltage thereof can be detected by detecting the divided voltages Vr1 and Vr2.The current sensors 701 and 702 are configured of the current detectors 711, 712, and 713 and the current detectors 721, 722, and 723, respectively, which detect phase current corresponding to the respective system, the phase current including lu1, Iv1, Iw1, and Iu2, Iv2, Iw2, and transmit the detected phase current to the control unit 65 as feedback information. The control unit 65 is configured of a microprocessor 67, a control circuit (or pre-driver (pre-driver)) 68, or the like, and controls the conduction of the motor 80 based on feedback information including the steering torque Ts detected by the steering torque sensor 93, the phase currents Iu 1, Iv 1, Iw 1, Iu 2, Iv 2, and Iw 2, and the phase angle θ.Next, configurations of the control unit 65 according to the respective embodiments for each embodiment will be described. In the reference numerals applied to the control unit of the first to fourth embodiments, the third digit after 65 is the number of the embodiment. Each of the control units 651 to 654 corresponding to the first to fourth embodiments performs feedback control in which sum and subtraction of the dq-axis actual current flowing through the winding groups 801 and 802 is returned to the sum and subtraction of the dq-axis current target value between the two systems. According to this method, the calculation can be simplified compared to a configuration in which the feedback is performed for each system. In the embodiments, the same reference numerals are applied to substantially the same elements.(First Embodiment)Next, the first embodiment will be described with reference to Figs. 3 to 12.As shown in FIG. 3, the control unit 651 includes a current command summation subtraction unit 41, a peak reduction current command calculation unit 20, a summation subtraction unit 28, and a current feedback calculation unit 40.In the current command value summation subtraction unit 41, the current command value Id*, Iq* for the d-axis and the q-axis is calculated so as to generate the summation of the current command values, i.e., Id sum* and Iq sum*, and the difference between the current command values, i.e., Id diff* and Iq diff*. Since the electrical characteristics between the two systems are equivalent to each other, Id sum* and Iq sum* correspond to twice the values of Id* and Iq*, and Id diff* and Iq diff* correspond to 0. Note that, optionally, no current command summation subtraction unit 41 may be provided, and a configuration setting "Id sum=2xld*, Iq sum=2xlq*, Id diff*=0, Iq diff*=0" may be used instead.The peak reduction current target value calculation unit 20 calculates harmonic components superimposed on the phase current in the dq coordinate system so as to reduce the peak of the first-order component of the phase current flowing through the winding groups 801 and 802, i.e., the peak of the fundamental wave. According to the first to fourth embodiments, as the harmonic component, 5th and 7th order components are generated, the 5th order component having a frequency that is 5 times the first order component of the phase current, and the 7th order component having a frequency that is 7 times the first order component of the phase current. Further, when harmonic components are superimposed on the fundamental waves of the 0-th order component in the dq coordinate, dq axis current of the 6-th order component having a frequency 6 times the first order component of the phase current is generated.The detailed configuration of the peak reduction current target value calculation unit 20 will be described later.The 6th-order dq axis current described in the present specification refers to a current command value that overlaps the fundamental waves to reduce the peak of the phase current. The amplitude of the 6th-order d-axis current relative to the amplitude of the fundamental waves in the first and second systems is referred to as Id1 6* and Id2 6* respectively, and the difference between two systems is referred to as Id diff 6* respectively.Similarly, the amplitude of the 6-th order q-axis current relative to the amplitude of the fundamental waves in the first and second systems is referred to as Iq 1 6* and Iq 2 6* respectively, and the difference between the two systems is referred to as Iq diff 6* respectively.The 6th-order dq-axis currents Id 1 6*, Id 2 6*, Iq 1 6* and Iq 2 6* calculated for each system by the peak reduction current target calculation unit 20 are calculated in the summation-subtraction unit 28 to perform summation or subtraction for each of the d-axis and the q-axis. Since the 6th-order components between two systems having a current phase shifted by 30 degrees can be canceled by summation with each other, each of the Id sum 6* and the Iq sum 6*, 0.Specifically, according to the first embodiment, the amplitudes of the 6th-order q-axis current Iq 1 6* and Iq 2 6* are defined as positive values, and the amplitudes of the d-axis current Id 1 6* and Id 2 6* are defined as 0. The technical features thereof will be described later.The amplitudes of the 6th-order q-axis current Iq 1 6* and Iq 2 6*, which are positive values, are larger than the amplitudes of the 6th-order d-axis current Id 1 6* and Id 2 6*, i.e., apparently larger than 0. Also, since Id diff 6*, which is a difference of the 6th-order d-axis current between two systems, is 0, only Iq diff 6*, which is a difference of the 6th-order q-axis current between two systems, is output from the summation-subtraction unit 28.Note that a configuration may be used in which Iqdiff 6 is directly calculated in the peak reduction current command value calculation unit 20 without providing the summation-subtraction unit 28.The value Iqdiff 6* calculated by the summation-subtraction unit 28 is added to Iqdiff*, which is a difference of the q-axis current command value between two systems and calculated by the current command value summation-subtraction unit 41. In other words, the value Iqdiff 6* is added to 0. Thus, the Id sum*, Iq sum*, and Id diff* calculated by the current command summation subtraction unit 41 are directly output to the current feedback calculation unit 40. The value Iqdiff 6* is added to Iqdiff*, and the Iqdiff* is output to the current feedback calculation unit 40.The current feedback calculation unit 40 includes a summation control unit 421, a difference control unit 422, a system voltage calculation unit 43, and a feedback current summation subtraction unit 46. Accordingly, the coordinate conversion units 341, 342, 351, and 352 are shown outside the current feedback calculation unit.The summation control unit 421 receives deviations between Id sum*, Iq sum* and Id sum, Iq sum calculated by the feedback current summation subtraction unit 46. The summation control unit 421 calculates, by a proportional-integral-regulator calculation, Vdsum and Vqsum which are voltage set points for two systems to converge the deviations.The difference control unit 422 receives Id diff* (=0), a value in which Iq diff 6* has been added to Iq diff* (=0), and deviations of Id diff and Iq diff have been calculated by the feedback current summation subtraction unit 46. The difference control unit 422 calculates, by the proportional-integral-controller calculation, Vddiff and Vqdiff, which are the differences of the voltage command values between two systems, to converge the deviations to 0.For the coordinate conversion units 341, 342, 351, and 352 shown in FIG. 3, the suffix "first" is given to the conversion unit for the first system, and the suffix "second" is given to the conversion unit for the second system. However, since functions of the respective control blocks are the same between two systems, the suffixes "first" and "second" are omitted accordingly. In the coordinate conversion calculation, θ is used as a phase angle for the first system, and θ-30 is used for the second system, the phase for the second system being shifted by 30 degrees. Hereinafter, the unit of θ is defined as [degree] throughout the specification.The system voltage calculation unit 43 converts Vd sum, Vq sum, Vd diff, and Vq diff to the voltage command values for the first and second systems, i.e., Vd 1, Vq 1, Vd 2, and Vq 2, and outputs them to the 2-to-3-phase conversion units 341 and 342.The 2-to-3-phase conversion units 341 and 342 perform coordinate conversion of the dq-axis voltage target values Vd 1, Vq 1, Vd 2, and Vq 2 into three-phase voltage target values Vu 1, Vv 1, Vw 1, Vu 2, Vv 2, and Vw 2.Subsequently, PWM control or the like is performed based on the three-phase voltage command values Vu 1, Vv 1, Vw 1, Vu 2, Vv 2, and Vw 2 so as to generate drive signals for the inverters 601 and 602.The 3-to-2-phase conversion units 351 and 352 perform coordinate conversion of the current currents Iu 1, Iv 1, Iw 1, Iu 2, Iv 2, and Iw 2 detected by the current sensor into the dq axis currents Id 1, Iq 1, Id 2, and Iq 2, thereby performing feedback control.The feedback current summation subtraction unit 46 performs summation and subtraction for the current detection values Id 1, Iq 1, Id 2, and Iq 2 transmitted by the 3-to-2 phase conversion units 351 and 352 to calculate Id sum, Iq sum, Id diff, and Iq diff.Next, with reference to FIGS. 4 to 8, a configuration of the peak reduction current target value calculation unit 20 that is commonly used for each embodiment will be described.As shown in FIG. 4, the peak reduction current command value calculation unit 20 is provided with a d-axis current command value limiting unit 21, a current amplitude calculation unit 22, a current amplitude limiting unit 23, a current amplitude gain adjustment unit 24, a current phase calculation unit 25, a phase compensation amount calculation unit 26, and a final current command value calculation unit 27. The peak reduction current command value calculation unit 20 calculates the dq axis currents Id 1 6*, Id 2 6*, Iq 1 6* and Iq 2 6* 6- of order as peak reduction current command values based on the dq axis current command values Id* and Iq*. According to the first embodiment, the amplitudes of the d-axis currents Id 1 are 6* and Id 2 are 6* 0.The peak of the phase current is reduced, so that the heat generated at the inverters 601 and 602 or the winding groups 801 and 802 can be reduced particularly when the motor 80 is stopped or at a low speed.The peak reduction current command value calculation unit 20 accepts a phase angle θ detected by the rotation angle sensor 85 and an angular velocity ω obtained by differentiating the phase angle θ with time, in addition to the dq axis current command values Id* and Iq*. The electrical angular velocity (electrical angular velocity) ω is converted into an engine speed [rpm] by multiplication with the proportionality constant. In the following description and drawings, "rotation speed obtained by converting the phase angle speed ω" is abbreviated as "rotation speed ω", respectively. Also, a positive or negative direction of the rotational speed ω, i.e., the rotational direction of the motor 80, is defined based on the positive or negative direction of the phase angle θ.The peak reduction current target value calculation unit 20 calculates dq 6-order axis current in the dq coordinate so as to reduce the peak of the component of the first-order phase current. The calculations performed by the respective control blocks will be described below. Each calculation may be performed with reference to a map or by using equations.The d-axis current command value limiting unit 21 limits the d-axis current command value Id* based on the rotational speed ω of the motor 80 and outputs the d-axis current command value limiting value Id*_lim. Specifically, as shown in FIG. 5, when the absolute value of the rotational speed ω is greater than or equal to wd 1 the d-axis current command value Id* is multiplied by 1 as the d-axis current limiting gain Kd. When the absolute value of the rotational speed ω is less than ωd 0 the d-axis current command value Id* is multiplied by 0 as the d-axis current limiting gain Kd. Even if the absolute value of the rotational speed ω is in a range of ωd 0 to ωd 1 the gain Kd gradually increases from 0 to 1.When the absolute value of the rotational speed ω is greater than or equal to ωd 1 the d-axis current command value Id* is maintained, and when the absolute value of the rotational speed ω is less than ωd 0 the d-axis current command value Id* is set to 0 to fix the current phase θi to 0 degrees. Even when the absolute value of the rotational speed ω is in a range from ωd 0 to ωd 1 the d-axis current target value Id* gradually changes. Thus, in a low rotation speed range in which the absolute value of the rotation speed ω is less than ωd 0 the calculation of the dq-axis current phase θi can be omitted.The current phase θi corresponds to an angle relative to the +q axis of the current vector, where Id* and Iq* in the dq axis coordinate are defined as d-axis and q-axis components, respectively. The current phase θi is defined such that the counterclockwise direction relative to the +q axis is a positive direction.The current amplitude calculation unit 22 calculates the current amplitude reference value Ip 0 of the 6th order component of the dq axis current.The current amplitude limiting unit 23 limits the value of the current amplitude reference value Ip 0 and outputs the value as the current amplitude limiting value IP LIM. Specifically, as shown in FIG. 6, when the absolute value of the current amplitude reference value Ip 0 is greater than or equal to Ip_neg and less than or equal to Ip_grd, the current amplitude reference value Ip 0 is maintained. On the other hand, when the absolute value of the current amplitude reference value Ip 0 is smaller than Ip_neg, the current amplitude limiting value IP LIM is set to 0. Further, when the absolute value of the current amplitude reference value Ip exceeds 0 Ip_grd, the current amplitude limiting value IP LIM is limited to a protection value ±Ip_grd.When the absolute value of the current amplitude reference value Ip 0 is smaller than Ip_neg, it is not necessary to reduce the peak because the peak of the first-order component of the phase current is relatively low, thereby generating less heat.Therefore, the calculation of the peak reduction current target value can be omitted when the current amplitude limit value IP LIM is set to 0 and the peak reduction target value output by the peak reduction current target value calculation unit 20 is set to 0.The current amplitude gain setting unit 24 sets the current amplitude gain Kp based on the rotational speed ω of the motor 80. The current amplitude Ip is output to the final current target value calculation unit 27, in which the current amplitude Ip is obtained by multiplying the current amplitude limit value IP LIM output by the current amplitude limit unit 23 by the current amplitude gain Kp.As shown in FIGS. 7A and 7B, when the absolute value of the rotational speed ω is larger than ωp, the current amplitude gain Kp is set to 0. In other words, in a high-speed range in which the absolute value of the rotational speed ω is larger than ωp, the peak reduction current target value is set to 0 and the peak of the 1st-order component of the phase current is not reduced. Therefore, assuming that a relationship between a rotation speed threshold ωp of the current amplitude gain setting unit 24 and a rotation speed threshold ωd 0 of the d-axis current command value limiting unit 21 is ωd 0 ≥ ωp, the calculation of the current phase θi in all the rotation speed ranges can be omitted.On the other hand, when the absolute value of the rotational speed ω is less than or equal to ωp, as shown in an example in FIG. 7A, the current amplitude gain Kp is set such that as the rotational speed ω approaches 0, the current amplitude gain Kp is linearly increased. In an example shown in FIG. 7B, the current amplitude gain Kp is set so as to compensate an amount of decreased amplitude due to a response delay of the current control. For example, the current amplitude gain Kp increases when the absolute value of the rotational speed decreases from ωp, and decreases when the absolute value of the rotational speed ω is 0.According to this configuration, harmonic components are superimposed on the current when rotated at a speed of 0 and at a low speed, respectively, the generated heat is likely to affect the peak value, and no harmonic components are superimposed on the current when rotated at a high speed, the root mean square value is likely to affect heat generation.Thus, effective control can be performed depending on various requirements.In the case where a relationship between the rotation speed threshold ωp of the current amplitude gain setting unit 24 and the rotation speed threshold ωd is 0 of the d-axis current command limiting unit 21 ωd is 0< ωp, the current phase calculating unit 25 calculates, based on the dq-axis current command values Id* and Iq*, the dq-axis current phase θi in a range of the absolute value of the rotation speed ω defined as ωd 0< | ω|<ωp, and outputs the dq-axis current phase θi to the final current command calculating unit 27.The phase compensation amount calculation unit 26 calculates a phase compensation amount θc that is set depending on the rotational speed ω of the motor 80. The higher the rotational speed ω, the higher the frequency of the current to be fed, so that an amount of phase delay due to a response delay of the current control must be compensated. Thus, as shown in FIG. 8, the phase compensation amount calculation unit 26 calculates a phase compensation amount θc having a positive correlation with the rotational speed ω in a range from the lower limit value θc_min to the upper limit value θc_max, and outputs the phase compensation amount θc to the final current command value calculation unit 27.The final current target value calculation unit 27 optionally takes the current amplitude Ip, the phase compensation θc, and the dq-axis current phase θi. In addition, the final current command value calculation unit 27 acquires a phase angle θ from the rotation angle sensor 85, and calculates the dq 6*, Id 2 6*, Iq 1 6* and Iq 2 6* 6- th-order.In a configuration without summation-subtraction unit 28, the final current set point calculation unit 27 calculates Id diff 6* and Iq diff 6*, which are the differences of the 6th-order dq axis currents between two systems.According to the first embodiment configured as above, as the 6th-order dq-axis currents for the phase current spike reduction, only the 6* and 6* 6-th-order q-axis currents Iq 1 and Iq 2 flow through the winding groups 801 and 802, and not the 6* and 6* 6-th-order d-axis currents Id 1 and Id 2. Next, technical advantages according to the above-described configuration will be described.The torque Tm produced by the motor is expressed by the equation (1), here the torque is composed of a sum of the magnet torque and the reluctance torque. where Kt represents the magnet torque constant and Ld and Lq represent the d-axis inductance and the q-axis inductance.In essence, since a permanent magnet embedded type rotating electric machine has the characteristic Ld≠Lq, reluctance torque corresponding to the second term is generated in the right side of the equation (1).For the current passed through the winding groups 801 and 802 of the two systems having a positional relationship with phase angles shifted by 30 degrees from each other, a configuration is used in which dq 6-order axis current is superimposed on the 0th-order fundamental waves in the dq coordinate. When it is assumed that the two systems are in a normal state, the amplitudes of the alternating current of the two systems outputted by the inverters 601 and 602 are equal, and the phase difference between two systems is 30 degrees.In this configuration, the winding torque Tm 1 of the first system and the winding torque Tm 2 of the second system are expressed by the following equations (2.1) and (2.2). For example, Iq1 0 refers to the 0th order q-axis current amplitude in the first system and Iq2 6 refers to the 6th order q-axis current amplitude in the second system.The electrical characteristics of the two systems are the same. Therefore, the following equations (3.1) to (3.4) define the 0th-order dq-axis currents Id 0, Iq 0 and the 6th-order dq-axis currents Id 6, Iq 6 which are common to the two systems.When both sides in the equations (2.1) and (2.2) are summed using the equations (3.1) to (3.4), the total winding torque Tm of the two systems is represented by the equation (4).Specifically, in the second term of the right side of equation (4), the 6th-order torque ripple for the two systems may be canceled, but the 12th-order torque ripple may not be canceled.Therefore, when either an amplitude of the 6th-order d-axis current Id 6 or the 6th-order q-axis current Iq 6 is set to 0, the 12th-order torque ripple becomes 0 with these products. However, in the case where the amplitude of the 6-order q-axis current Iq 6 is set to 0 and only the 6-order d-axis current Id 6 flows, it was determined by a test that no peak reduction effect of the phase current can be obtained. Accordingly, in the first embodiment, a configuration is adopted in which the amplitude of the 6th-order d-axis current Id 6 is set to 0 and only the 6th-order q-axis current Iq 6 flows.In this case, when substituting Id 6= 0 in the equation (4), the following equation (5) is obtained. Specifically, the reluctance torque is defined only by the 0th-order dq axis currents Id 0 and Iq 0 which are direct current components, and no 12th-order torque ripple is generated.Next, the peak reduction effect of the phase current according to the first embodiment will be described with reference to FIGS. 9 to 12.Hereinafter, the amplitude ratios for the 6th-order d-axis current Id 6 and the 6th-order q-axis current Iq 6 with respect to the base amplitude are respectively referred to as the 6th-order d-axis current amplitude ratio R(Id 6) and the 6th-order q-axis current amplitude ratio R(Iq 6) respectively. FIG. 9 shows a change in the phase current peak reduction ratio when the 6-order q-axis current amplitude ratio R(Iq 6) is changed from 0% to 10%. The phase current waveform corresponding to the 6-order q-axis current amplitude ratio R(Iq 6) around the maximum peak phase reduction ratio (shown in FIG. 9 ), i.e., 4%, 4.4%, and 5%, is shown in FIGS. 10, 11, and 12, respectively.FIGS. 10A, 10B, 11A, 11B, 12A, and 12B show phase current waveforms for one electric period. The respective three-phase waveforms in the same system have the same waveform contour, with phases shifted 120 degrees from each other. Also, waveforms between two systems corresponding to one of the three phases are shifted by 30 degrees from each other. The waveforms in FIGS. 10A, 10B, 11A, 11B, 12A, and 12B may occur in each system and each of the three phases, and a position on the phase axis indicated by 0 degrees is not particularly important. In FIGS. 10A, 10B, 11A, 11B, 12A, and 12B, the fundamental wave of the phase current is expressed by sin(θ+165).In other words, a positive peak and a negative peak appear in the fundamental wave at the phase θ=-75 and 105.In each of FIGS. 10B, 11B, and 12B, a range around the positive peak of the phase current shown in each of FIGS. 10A, 11A, and 12A is enlarged with respect to the phase θ=-75 degrees as the central phase. As shown in FIGS. 10B, 11B, and 12B, in the phase current waveform in which the 6-th-order q-axis current Iq 6 is superimposed, the current value at the central phase decreases compared to the current values at the both sides, and has peak values at both sides of the central phase. Here, a phase current peak reduction ratio is defined as a difference between the peak value and the 100% phase current. The phase current peak reduction ratio is about 3.4% when the 6th-order q-axis current amplitude ratio R(Iq 6) is in the range of 4% to 5%.According to the method disclosed in the above-mentioned JP 5 672 278 B2, the 5th order harmonic is superimposed on the fundamental wave, thereby reducing the phase current peak. At this time, current having a phase shifted by 30 degrees flows through the winding groups of the two systems, thereby cancelling the 6-th-order torque ripple. However, according to JP 5 672 278 B2, there is no designation of the 12th-order reluctance torque which has not been canceled out. Accordingly, the method disclosed by the above-mentioned JP 5 672 278 B2 is effective for a surface magnetic type rotary electric machine. However, according to the above-mentioned JP 5 672 278 B2, when applied to a rotating electric machine that generates torque consisting of the magnetic torque and the reluctance torque, some torque ripple cannot be cancelled out.However, according to the present invention, the control unit 651 of the first embodiment is configured such that the amplitude of the 6th-order d-axis current Id 6 for the peak reduction current superimposed on the fundamental wave of the dq-axis coordinate is set to 0 so as to allow only the 6th-order q-axis current Iq 6 to flow. Thus, in motor 80 which generates torque from a sum of the magnet torque and the reluctance torque, the phase current peak can be reduced, thereby avoiding torque ripple of the reluctance torque. Therefore, effects such as reduction of heat and power loss by the phase current peak can be obtained, and further influence by noise or vibration due to the torque ripple can be prevented.In particular, according to an electric power steering apparatus of vehicles, a small ECU 10 for supplying a large amount of current is required due to a limited installation space. Therefore, reduction of heat and dissipation by reduction of the phase current peak is urgent. On the other hand, since noise or vibration due to torque ripple significantly affects the sense of the driver operating the steering wheel and also affects the performance, reduction of the torque ripple is highly desirable.Therefore, according to respective embodiments of the present invention, both an effect of reducing the torque ripple and a peak reduction effect of the phase current can be effectively achieved. Further, according to the first embodiment, since the 12-th order torque ripple can be 0, the configuration of the present invention can be suitably adapted to a system in which reduction of noise and vibration is highly desirable.(Second Embodiment)Next, the second embodiment will be described with reference to FIGS. 13, 14A, and 14B.The control unit 652 according to the second embodiment sets the amplitudes of the 6th-order d-axis currents Id 1 6* and Id 2 6* to a non-zero value, and the d-axis currents Id 1 6* and Id 2 6* are calculated by the peak reduction current target value calculation unit 20. In addition, the amplitudes of the 6th-order q-axis currents Iq 1 6* and Iq 2 6* are set to values larger than the amplitudes of the 6th-order d-axis currents Id 1 6* and Id 2 6*. The differences Id diff 6* and Iq diff 6* calculated by the summation-subtraction unit 28 are added, i.e., added to 0, to the differences Id diff* and Iq diff* calculated by the current set point summation-subtraction unit 41. Then, the summed current value is output to the current feedback calculation unit 40.FIGS. 14A and 14B show a comparison of characteristics between the second embodiment and the comparative example. In the second embodiment, the 6-order q-axis current amplitude ratio R(Iq 6) is set to a value larger than the 6-order d-axis current amplitude ratio R(Id 6).In the comparative example, the 6-order q-axis current amplitude ratio R(Iq 6) and the 6-order d-axis current amplitude ratio R(Id 6) are set to the same value. According to an example shown in FIGS. 14A and 14B, the maximum value of the 6-order q-axis current amplitude ratio R(Iq 6) is set to 4.6%. The 6th-order d-axis current amplitude ratio R(Id 6) and the 6th-order q-axis current amplitude ratio R(Iq 6) at the points A, B, and C shown in FIGS. 14A and 14B are as follows.According to the comparative example represented by black dots and broken lines, the 6th-order d-axis current amplitude ratio R(Id 6) is changed between the dots A and C, maintaining the relationship R(Id 6) = R(Iq 6).The amplitude ratio (I 5) of the 5thharmonic of the phase current and the amplitude ratio (I 7) of the harmonics of the 7th-order phase current are expressed by the following equations (6.1) and (6.2).Therefore, when a condition R(Id 6) = R(Iq 6) is satisfied, the amplitude ratio R(I 7), which corresponds to the 7th order harmonic, becomes 0. In other words, the comparative example is based on a configuration in which only the component of the 5th-order phase current overlaps the fundamental wave.Meanwhile, according to the second embodiment, which is represented by white points and solid lines, a condition R(Iq 6) = 4.6% is set, and the 6th-order d-axis current amplitude ratio R(Id 6) is changed between the points B and C.According to the second embodiment, the 6th-order d-axis current amplitude ratio R(Id 6) is not 0 and the 6th-order q-axis current amplitude ratio R(Iq 6) is larger than the 6th-order d-axis current amplitude ratio R(Id 6), so that the points B and C are not included. In other words, the second embodiment defines a range from a point immediately after point B to point C. Therefore, the outlines of the white points B are represented by broken lines.Note that the first embodiment defines a range represented by a broken line in which the points A and B are included and the 6th-order d-axis current amplitude ratio R(Id is 6) 0. When a condition R(Id is 6) = 0 according to equations (6.1) and (6.2), a condition R(I 5) = - R(I 7) is satisfied. In other words, components of the 5-th and 7-th order phase current are superimposed, with the absolute value of the amplitude ratios being equal and the phases being opposite to each other. Therefore, according to the first embodiment, the amplitudes of the 5-th and 7-th order phase current components are combined.However, according to the second embodiment in which only a relationship of 0≠R(Id 6) < R(Iq 6 is defined, both the 5th-order phase current component and the 7th-order phase current component are superimposed on the fundamental wave, and amplitude combinations of the 5th-order component and the 7th-order phase current component can be largely selected.FIG. 14A shows a relationship between the 6th-order d-axis current amplitude ratio R(Id 6) and the phase current peak reduction ratio. In the comparative example, when the 6-order d-axis current amplitude ratio R(Id 6) increases from 0% to 4.6%, the phase current peak reduction ratio increases from 0% to about 4.5%. In the second embodiment, when the 6th-order d-axis current amplitude ratio R(Id 6) increases from 0% to 4.6%, the phase current peak reduction ratio increases from about 3.4% to about 4.5%.FIG. 14B shows a relationship between the torque ripple index and the phase current peak reduction ratio. The torque ripple index is a dimensionless number calculated by the equation (7), and the torque ripple index is converted into a torque value by multiplication by a proportionality constant based on the characteristics of the motor 80. Here, an index value is used to make a relative comparison.A smaller torque ripple index is preferably used in terms of reducing noise and vibration of the motor 80. The torque ripple index is in a range 0 in which the 6th-order d-axis current amplitude ratio R(Id 6) between the points A and B is 0%.In a comparative example, the torque ripple index increases approximately in proportion to the square of the phase current peak reduction ratio. In other words, in order to reduce the phase current peak, the 6th-order dq axis current is superimposed on the fundamental wave, whereby the 12th-order torque ripple increases in a parabolic curve.According to the second embodiment, the 6th-order d-axis current amplitude ratio R(Id 6) from point B toward point C is increased from 0% to 4.6%. As a result, the phase current peak reduction ratio of 3.4% can be further increased, but the torque ripple also increases.Accordingly, a target value is preferably determined in consideration of a balance of requirements, reduction of the phase current peak, and reduction of the torque ripple in a realistic system to which the present invention is applied. For example, if peak reduction of phase current is more required to reduce heat and power dissipation, point C is preferably used as a target, and if torque ripple reduction is more required to reduce noise and vibration, point B is preferably used as a target.Further, when a fixed value of the 6-th-order q-axis current amplitude ratio R(Iq 6) is set to be smaller than 4.6%, the characteristic line connecting the points B and C moves toward the point A side. Also in this case, the operation range of the second embodiment corresponds to a smaller range than a range defined by the broken line shown in the comparative example. In particular, according to the second embodiment, as compared with the comparative example, torque ripple corresponding to the same phase current peak reduction ratios can be reduced.(Third Embodiment)Next, the third embodiment will be described with reference to FIG. 15.The control unit 653 according to the third embodiment further includes a torque ripple compensation amount calculation unit 29 that calculates a torque ripple compensation amount of the 12th-order component of the reluctance torque generated based on the product of the amplitudes of the 6th-order dq axis current. In a configuration shown in FIG. 15 in which feedback control for the summation and subtraction of the current command values is performed between the two systems, the torque ripple compensation amount calculation unit 29 acquires Id diff 6* and Iq diff 6*, which are differences of the dq 6-order axis currents. The compensation amount calculated by the torque ripple compensation amount calculation unit 29 is added to Iq sum*, which is a sum of the q-axis current command values in the two systems.Specifically, the torque ripple in the one system is calculated based on the product of (Id diff 6 / 2) and (Iq diff 6 / 2). Subsequently, a value that is a double compensation amount for the one system is added to Iq sum* that is a sum of the q-axis current command values in the two systems.As a realistic target, a product of the 6th-order d-axis current and the 6th-order q-axis current for the one system may preferably be set to 1 / 10 or less of a 0th-order component of the q-axis current.According to the third embodiment, the 6th-order dq-axis currents Id 6 and Iq 6 are calculated to achieve a sufficient phase current peak reduction effect, and the q-axis current is passed to compensate for the 12th-order torque ripple generated by the dq-axis currents Id 6 and Iq 6. Thus, both a peak reduction effect of the phase current and an effect of reducing the torque ripple can be favorably achieved.Even in a case of the condition Id diff 6= 0, as described in the first embodiment, the torque ripple compensation amount calculation unit 29 can calculate a compensation amount equal to 0 using a usual calculation algorithm.(Fourth Embodiment)Next, the fourth embodiment will be described with reference to FIG. 15.The control unit 654 according to the fourth embodiment further includes, in addition to the configuration of the control unit 651 of the first embodiment, a configuration for compensating a dead time of the three-phase voltage command values Vu 1, Vv 1, Vw 1, Vu 2, Vv 2, and Vw 2. Dead time compensation is a method in which a voltage command value is compensated to cancel the influences of the dead time, whereby a voltage utilization rate can be improved and distortion of the line voltage can be reduced.The dead time compensation units (abbreviated as DT compensation units in the drawings) 391 and 392 compensate the three-phase voltage command values Vu 1, Vv 1, Vw 1, Vu 2, Vv 2, and Vw 2 to cancel influences of the dead time, and output the compensated voltages Vu_dt 1, Vv_dt 1, Vw_dt 1, Vu_dt 2, Vv_dt 2, and Vw_dt 2. In the calculation, the compensation units 391 and 392 are required to determine the polarities of the phase current.The dead time compensation units 391 and 392 are configured to determine the polarities of the phase current based on the three-phase superposition current target value which is a current value consisting of the sum of the components of the first, 5th, and 7th orders of phase current. This configuration is shown by two-dot lines.According to this configuration, the four values Id sum*, Iq sum*, Id diff 6*(=0) and a sum of Iq diff*(=0) and Iq diff 6*, which have been generated based on the outputs of the current command summation subtraction unit 41 and the summation subtraction unit 28, are output to a system voltage calculation unit 47 in addition to the summation control unit 421 and the difference control unit 422.The system voltage calculation unit 47 converts the four values into the dq-axis superimposed current command values Id_sup 1, Iq_sup 1, Id_sup 2, and Iq_sup 2 for the first and second systems, and outputs the converted values to the current 2-to-3 phase conversion units 381 and 382.The dq-axis superposition current command values Id_sup 1, Iq_sup 1, Id_sup 2, and Iq_sup 2 are expressed by the equations (8.1)-(8.4).The current 2-to-3-phase conversion units 381 and 382 perform 2-to-3-phase conversion using the phase angles θ and θ-30, converting the dq-axis superimposed current command values Id_sup 1, Iq_sup 1, Id_sup 2, and Iq_sup 2 into the three-phase superimposed current command values Iu_sup 1, Iv_sup 1, Iw_sup 1, lu_sup 2, Iv_sup 2, and Iw_sup 2, and outputs the converted values to the dead time compensation units 391 and 392. The dead time compensation units 391 and 392 determine the polarities of the phase current based on the obtained three-phase superposition current command values.In FIG. 16, in order to distinguish the current 2-to-3-phase conversion units 381 and 382 and the 2-to-3-phase conversion units 341 and 342, the 2-to-3-phase conversion units 341 and 342 are referred to as voltage 2-to-3-phase conversion units 341 and 342.In FIG. 16, input-output signals are represented by broken lines in the configuration in which the dead time compensation units 391 and 392 determine the polarities of the phase current based on the current current.In this configuration, the current current detected by the current sensors 701 and 702, i.e., lu1, Iv1, Iw1, Iu2, Iv2, and Iw2, is transmitted to the 3-to-2-phase conversion units 351 and 352 and also to the dead time compensation units 391 and 392.The control unit 654 according to the fourth embodiment is provided with dead time compensation units 391 and 392 that compensate the voltage based on the polarities of the phase current to cancel the influence of the dead time on the voltage applied to the motor 80. Thus, the voltage utilization rate is improved, and the distortion of the line voltage can be reduced.(Other Embodiments)(a) According to the above-described embodiments, a surface magnetic type rotating electric machine is used as a typical three-phase rotating electric machine that generates torque by summing magnetic torque and reluctance torque. However, since even the surface magnetic type rotating electric machine sometimes generates a small reluctance torque, the configuration of the present invention can be applied to a controller of a surface magnetic type rotating electric machine.(b) In the above-described embodiments, control for current amplitudes of the 6th-order dq-axis component is described, wherein the 6th-order dq-axis component is converted into 5th- and 7th-order components of the phase current in the three-phase rotating electric machine. This control can be applied to components of different order which are expressed substantially as 6(2k+1)-th order, where k=0 or a higher integer, such as 18-th or 30-th order components, except for dq-axis 6-th order components. The order number of the generated reluctance torque ripple is based on a product of the dq axis current amplitudes of the 6(2k+1)th-order components, which corresponds to 12-th, 36-th, and 60-th-order components, which are expressed substantially as 12(2k+1)th-order.(c) According to the above-described embodiments, a configuration is employed in which feedback control is performed for sum and subtraction between the dq-axis current command values. However, it is not limited to this configuration. For example, feedback control is performed for each current command value of each system.(d) Specific configurations for the control device of the three-phase rotating electrical machine are not limited to the configurations exemplified in FIGS. 2 and 3 of the above-described embodiments. For example, switching elements of the inverters may be configured of field effect transistors other than MOSFETs or IGBTs or the like.(e) The control device of the three-phase rotating electric machine according to the present invention is not limited to a control device of a steering assist motor of an electric power steering device, but may be applied to a control device for other three-phase motors or a generator.As described, the present invention is not limited to the above-described embodiments, but can be modified in various ways without departing from the scope of the present invention.

Claims

A control device for driving a three-phase rotating electric machine (80) having two three-phase winding groups (801, 802), the three-phase rotating electric machine (80) generating a torque which is a sum of magnetic torque and reluctance torque when alternating current having mutually different phases is passed through the two winding groups (801, 802), the control device comprising: two power converters (601, 602) each provided for a corresponding one of the two winding groups (801, 802), each power converter (601, 602) supplying alternating current to each of the two winding groups (801, 802), the alternating current supplied to the two winding groups (801, 802) having the same amplitude and the mutually different phases being defined as 30±60 x n [degrees], where n is an integer; A control unit (65) configured to calculate the d-axis current and the q-axis current of the 6(2k+1)-th order component that superimposes a fundamental wave component on a dq coordinate, where k=0 or a higher integer, to reduce a peak value of a first-order component in the phase current that is supplied to the two winding groups (801, 802), thereby controlling the supply of the three-phase rotating electric machine (80), wherein the control unit (65) is configured to calculate current such that an amplitude of the q-axis current of the 6(2k+1)-th order component is larger than an amplitude of the d-axis current of the 6(2k+1)-th order component.The control device according to claim 1, wherein the control unit (65) is configured to set an amplitude of the d-axis current of the 6(2k+1)-th order component to 0.The control device according to any one of claims 1 or 2, wherein the control unit (65) further comprises a torque ripple compensation amount calculation unit (29) that calculates a torque ripple compensation amount of the 12(2k+1)-th order component generated based on a product of the amplitudes of the d-axis current and the q-axis current of the 6(2k+1)-th order component; and the control unit (65) adds the torque ripple compensation amount to the q-axis current, the torque ripple compensation amount being calculated by the torque ripple compensation amount calculation unit (29).The control device according to any one of claims 1 to 3, wherein the control device is configured for an electric power steering device (90) of a vehicle and controls driving of a steering assist motor as the three-phase rotating electric machine (80).

Images