JP6740842B2 - Control device for multi-phase rotating machine - Google Patents

Description

The present invention relates to a control device for a multi-phase rotating machine.

Conventionally, in a control device that supplies electric power converted by a power converter such as an inverter to a multi-phase rotating machine and controls driving of the multi-phase rotating machine, an open failure of a switching element that constitutes the power converter or a disconnection failure of a current path. Techniques for detecting the are known.
For example, the device disclosed in Patent Document 1 is disconnected when a line voltage between phases of a three-phase rotating machine is equal to or higher than a reference value and a current detection value of each phase is equal to or less than a threshold value for a predetermined time. To determine.

JP, 2005-213666, A

In the current control that controls the current flowing through the multi-phase rotating machine based on the current command value calculated by the control device, generally, the current flowing through the multi-phase rotating machine has a first-order lag type frequency response characteristic with respect to the current command value. .. Therefore, when the frequency of the current command value is low, the phase of the voltage command value calculated based on the current command value and the current match, but when the frequency of the current command value is high, the phase of the voltage command value is The current phase is delayed.
Therefore, in the technique of Patent Document 1, when the frequency of the current command value is high, it becomes difficult to compare the corresponding voltage command value and current detection value with each reference value and threshold value at the same timing.

Further, in the first-order lag type frequency response characteristic, the gain response decreases when the frequency of the current command value is high. If the gain of the voltage command value is set to be large in order to compensate for this decrease in the gain response, there is a high possibility that the voltage command value will exceed the threshold value, and erroneous determination is likely to occur. As described above, when the frequency of the current command value is high and the responsiveness of the current control is high, it is difficult to correctly distinguish between the normal state and the fault state in which the disconnection fault or the open fault has occurred.
The present invention has been made in view of such a point, and an object thereof is to provide a control device for a multi-phase rotating machine capable of appropriately determining a disconnection failure or an open failure while enhancing the responsiveness of current control. To provide.

The present invention is an invention relating to a control device for controlling the driving of a polyphase rotating machine (801, 802) having one or more sets of polyphase windings (84, 841, 842). This control device for a polyphase rotating machine includes one or more power converters (60, 601, 602), one or more controllers (24, 241, 242, 261, 262), and one or more filters. (49, 441, 442, 461, 462, 491, 492) and one or more failure determiners (501, 502).

The power converter energizes one or more sets of windings by the operation of the plurality of switching elements (61-66, 611-661, 612-662).
The controller outputs the voltage command value calculated by the feedback control based on the current flowing through the multi-phase rotating machine to the power converter to control the current flowing through the multi-phase rotating machine.
The filter corrects the voltage command value output by the controller by a filtering process for amplifying or attenuating according to the frequency, and outputs the voltage command filter value.

The failure determiner, when the absolute value of the voltage command filter value is greater than the voltage threshold value (Vth) for each phase and the absolute value of the current flowing through the multi-phase rotating machine is less than the current threshold value (Ith), the power converter. To the multi-phase winding, it is determined that there is a disconnection failure in the current path or an open failure in the switching element.
Here, the time constant of the filter is set to be larger than the time constant (To, Ta, Tb) of the polyphase rotating machine.

That is, the frequency response characteristic of the output with respect to the input of the filter is about the same as the frequency response characteristic of the current with respect to the current command value, or is a response characteristic shifted to a lower frequency side. As a result, the phase of the voltage command filter value approaches the phase of the current flowing through the multi-phase rotating machine.
As a result, the failure determiner can make a failure determination using information whose phases are similar to each other, and therefore, even when the frequency of the current command value is high, it can be erroneously determined as a failure or an open failure. Can be prevented. Therefore, it is possible to properly determine the disconnection failure of the current path or the open failure of the switching element while improving the responsiveness of the current control.

Preferably, the present invention is a controller for controlling the driving of a polyphase rotating machine having two sets of multiphase windings magnetically coupled to each other, wherein the two sets of multiphase windings can be individually energized. It has two power converters.
If the unit of each winding, the power converter that energizes each winding, and the group of components that control energization to each winding is defined as a "system," a failure determiner is provided for each system and A disconnection failure or an open failure is determined for each phase.

In the two-system control device, when the response of the current control is high due to the influence of the mutual inductance between the two sets of windings, it is more possible to distinguish between the normal state and the fault state as compared with the one-system control device. It will be difficult. Therefore, the effect of the present invention, in which the electric power command value output by the controller is corrected by the filter process and the failure determination is performed based on the voltage command filter value and the current value flowing in the multi-phase rotating machine, is particularly effectively exhibited.

The whole block diagram of the control device of the three-phase rotating machine by a 1st embodiment. 3 is a control block diagram of a control unit according to the first embodiment. FIG. Frequency characteristic diagram of first-order lag response. The figure which shows the phase difference of a voltage command value and current when (a) the frequency of a current command value is low, and (b) the frequency of a current command value is high. The flowchart of a failure determination process. The whole block diagram of the control apparatus of the 3-phase rotary machine by 2nd-4th embodiment. The control block diagram of the control part by 2nd Embodiment. A motor model considering the effect of mutual inductance between two sets of windings. The control block diagram of the control part by 3rd Embodiment. FIG. 3A is a block diagram of a sum and difference model of a motor, and FIG. The control block diagram of the control part by 4th Embodiment.

Hereinafter, a plurality of embodiments of a control device for a multi-phase rotating machine will be described with reference to the drawings. In a plurality of embodiments, substantially the same configurations are given the same reference numerals and the description thereof will be omitted. Further, the following first to fourth embodiments are collectively referred to as "the present embodiment".
In the present embodiment, a control device for a three-phase rotating machine is shown as a representative of the multi-phase rotating machine. This control device for a three-phase rotating machine is used as a control device for driving a steering assist motor in, for example, an electric power steering device.

(First embodiment)
The first embodiment will be described with reference to FIGS.
FIG. 1 shows the overall configuration of the ECU 101 as a “control device” that controls driving of a motor 801 as a “three-phase rotating machine”. The motor 801 is a three-phase brushless motor having a set of three-phase windings 84 including a U-phase coil 81, a V-phase coil 82, and a W-phase coil 83.
The ECU 101 includes an inverter 60, a current sensor 70, a control unit 651 and the like.

The power supply voltage Vr from the battery 11 is input to the inverter 60 as a “power converter”, and the DC power of the battery 11 is converted into 3-phase AC power by the operation of the six switching elements 61-66 to generate 3-phase AC power. The winding 84 is energized. A power supply relay 12 and a smoothing capacitor 13 are provided at the input part of the inverter 60.

The switching elements 61 to 66 are, for example, MOSFETs and are bridge-connected. Switching elements 61, 62 and 63 are upper arm switching elements of U phase, V phase and W phase, respectively, and switching elements 64, 65 and 66 are lower arm switching of U phase, V phase and W phase, respectively. It is an element.
The current sensor 70 detects the respective phase currents Iu, Iv, Iw flowing through the motor 801 by the current detection elements 71, 72, 73 and feeds them back to the control unit 651.
The rotation angle sensor 85 detects the electrical angle θ of the motor 801 and notifies the control unit 651 of it.

The control unit 651 includes a microcomputer 67, a drive circuit (or pre-driver) 68, and the like. In the case of the ECU 101 applied to the electric power steering device, information on the steering torque trq from a steering torque sensor (not shown) is input to the control unit 651. The control unit 651 operates the switching elements 61-66 of the inverter 60 based on the steering torque trq, the phase currents Iu, Iv, Iw, and the feedback information of the electrical angle θ, and controls the energization of the motor 801.

Further, the control unit 651 acquires information on the power supply voltage Vr and the rotation speed ω obtained by differentiating the electrical angle θ with respect to time. The power supply voltage Vr is calculated, for example, by converting the partial voltage detected between the high potential line and the low potential line of the inverter 60. Further, in the present specification, the rotation speed obtained by multiplying the electrical angular velocity ω [deg/s], which is the time differential value of the electrical angle θ, by a predetermined proportional constant is referred to as “revolution speed ω”.

Here, the winding 84 of the motor 801, the inverter 60 that energizes the winding 84, and the group of constituent elements that control the energization of the winding 84 in the control unit 651 are defined as a “system”. The ECU 101 of the first embodiment is a single-system control device. Note that the second to fourth embodiments are two-system control devices including two inverters 601 and 602 for energizing a motor 802 having two sets of windings 841 and 842, as shown in FIG. 6. ..

By the way, in the system in which the drive of the motor 801 is controlled by the ECU 101, disconnection failure of the current path from the inverter 60 to the winding 84 or open failure of the switching elements 61-66 of the inverter 60 may occur.
The control unit 651 of the present embodiment presupposes that the power supply voltage Vr input to the inverter 60 is normal and the rotation speed ω is close to 0, and the voltage command value of each phase and the current flowing through the motor 801. It has a function of determining a disconnection failure or an open failure based on the above.

Next, FIG. 2 shows a control block diagram of the control unit 651 of the first embodiment.
The control unit 651 includes a controller 24, a two-phase/three-phase conversion unit 28, a three-phase/two-phase conversion unit 31, a current deviation calculator 33, and the like as a general current feedback control configuration. Further, a filter 49 and a failure determiner 50 are provided as a configuration unique to this embodiment.
In FIG. 2, the d-axis and q-axis currents and voltages are collectively represented by one line, and the three-phase currents and voltages are collectively represented by one line. Further, the input illustration of the electrical angle θ used for coordinate conversion calculation in the two-phase/three-phase conversion unit 28 and the three-phase/two-phase conversion unit 31 is omitted. Similarly, in the control block diagrams of the two systems of the second to fourth embodiments, the input illustration of the electrical angles θ and θ+30 [deg] is omitted.

The three-phase/two-phase converter 31 feeds back the phase currents Iu, Iv, and Iw detected by the current sensor 70 as dq-axis currents Id and Iq, which are three-phase to two-phase conversion.
The current deviation calculator 33 calculates current deviations ΔId, ΔIq between the dq axis current command values Id ^* , Iq ^* and the fed back dq axis currents Id, Iq.
The controller 24 calculates dq axis voltage command values Vd and Vq by PI control so that the current deviations ΔId and ΔIq converge to zero. Specifically, the controller 24 multiplies the command currents Id ^** and Iq ^** after PI control by the impedance term (R+Ls) to calculate the dq axis voltage command values Vd and Vq.
In addition, although the symbol "PI" is shown in the controller 24 of FIG. 2, only I (integration) control may be performed.

The meaning of the symbol of the impedance term (R+Ls) is as follows.
R: resistance of winding 84 L: self-inductance of winding 84 s: differential operator (Laplace variable)
The impedance term (R+Ls) is a transfer function of the motor inverse model corresponding to the impedance reciprocal term “1/(R+Ls)” in the motor model. The motor model will be described later in the second embodiment.

The 2-phase/3-phase converter 28 converts the dq-axis voltage command values Vd, Vq output by the controller 24 into 2-phase/3-phase to generate 3-phase voltage command values Vu, Vv, Vw, and outputs them to the inverter 60. ..
The inverter 601 converts the DC power of the battery 11 by operating the switching elements 61-66 by the PWM signal generated based on the three-phase voltage command values Vu, Vv, Vw, and the motor 801 receives the three-phase voltage Vu_i. , Vv_i, Vw_i are applied.

In such a current feedback control, the frequency response of the current actually flowing in the motor 801 with respect to the current command value is represented by the equation (1) of the transfer function of the first-order lag system of the time constant T.

As shown in FIG. 3, in the frequency response of the first-order lag system, the gain is 0 [dB] and the phase is 0 [deg] in the region where the frequency is sufficiently lower than (1/T), and the frequency is (1/T). When it exceeds, the gain becomes gradually smaller as the frequency increases, and the phase becomes a negative value. That is, a phase delay occurs. In a region where the frequency is sufficiently higher than (1/T), the phase converges to -90 [deg].

Here, assuming that the phase of the voltage command value output by the controller 24 is equal to the phase of the current command value, the current flowing through the motor 801 when the frequency of the current command value is low, as shown in FIG. The phase of is in agreement with the phase of the voltage command value.
On the other hand, as shown in FIG. 4B, when the frequency of the current command value is high, the phase of the current flowing through the motor 801 lags behind the phase of the voltage command value. In addition, in order to compensate for the decrease in the gain response of the current to the voltage command value, the gain of the voltage command value must be set to a large value.

By the way, in Patent Document 1 (JP-A-2005-213666), a state in which the current detection value of each phase is equal to or less than a threshold value is predetermined when the line voltage between each phase of the three-phase rotating machine is equal to or higher than a reference value. A technique for determining a disconnection when the time has continued is disclosed.
Assuming that the technique of Patent Document 1 is applied, FIG. 4 illustrates a voltage threshold value ±Vth and a current threshold value ±Ith in failure determination. When the absolute value of the voltage command value is larger than the absolute value |Vth| of the voltage threshold value and the absolute value of the current is smaller than the absolute value |Ith| of the current threshold value, it is determined as a failure. Further, in both cases of FIGS. 4A and 4B, it is assumed that the circuit is actually normal and no disconnection failure or open failure has occurred.

In the example of FIG. 4A, the absolute value of the voltage command value does not exceed the absolute value |Vth| of the voltage threshold value during one cycle, and the failure is not determined. Even if the absolute value of the peak of the voltage command value exceeds the absolute value of the voltage threshold value |Vth|, the current also peaks at that timing, and in a normal state, the absolute value of the current is the absolute value of the current threshold value. Since it is never smaller than Ith|, it is not determined as a failure.

On the other hand, in the example of FIG. 4B, the absolute value of the peak of the voltage command value exceeds the absolute value |Vth| of the voltage threshold value because the gain of the voltage command value is set large. Further, the phase of the current is delayed with respect to the phase of the voltage command value, and the peak timing is deviated. Therefore, in the period JX between the time t1 and the time t2, the voltage command value is larger than the positive voltage threshold +Vth, and the current is smaller than the positive current threshold +Ith, and it is determined that there is a failure. Similarly, in the period JY between time t3 and time t4, the voltage command value is smaller than the negative voltage threshold −Vth, and the current is larger than the negative current threshold −Ith, and it is determined that there is a failure.

As described above, when the conventional technique of Patent Document 1 is used, in a situation where the frequency of the current command value is high and the phase delay of the current detection value with respect to the voltage command value occurs, it is possible to correctly distinguish between the normal state and the failure state. Have difficulty. Therefore, even when the disconnection failure or the open failure does not actually occur, there is a possibility that the failure is erroneously determined.
However, increasing the responsiveness of the current control is a need for speeding up the control, and it is not preferable to purposely reduce the responsiveness of the current control in order to avoid erroneous determination of failure.
Therefore, the present embodiment has an object to properly determine a disconnection failure or an open failure even when the frequency of the current command value is high and the responsiveness of the current control is high.

As a configuration for that, the control unit 651 of the first embodiment includes a filter 49 and a failure determiner 50.
The filter 49 corrects the three-phase voltage command values Vu, Vv, and Vw output by the controller 24 and converted by the two-phase/three-phase conversion unit 28 into two-phase/three-phase by a filtering process for amplifying or attenuating according to the frequency, The voltage command filter values Vu_f, Vv_f, and Vw_f are output.
Specifically, the filter 49 is a first-order lag filter having the transfer function of Expression (1). Here, the time constant T ₁ of the filter 49 is set to be larger than the time constant (L/R) of the motor 801. When “To=(L/R)” is established, the relationship of the equation (2) is established.
T ₁ >To (=(L/R)) (2)

The transfer function Go(s) of the filter 49 under the boundary condition of Expression (2) is represented by Expression (3) corresponding to an expression obtained by dividing the winding resistance R by the impedance term (R+Ls) of the controller 24.

That is, the frequency response characteristic of the output with respect to the input of the filter 49 is about the same as the frequency response characteristic of the current with respect to the current command value, or a response characteristic shifted to a lower frequency side. As a result, the phases of the voltage command filter values Vu_f, Vv_f, Vw_f at least approach the phases of the currents Iu, Iv, Iw flowing through the motor 801, and ideally match each other.
Instead of the configuration in which the filter 49 corrects the three-phase voltage command values Vu, Vv, Vw, after the dq-axis voltage command values Vd, Vq output by the controller 24 are corrected by the filter, the dq-axis voltage command filter values are obtained. A configuration in which Vd_f and Vq_f are converted into two phases and three phases may be used. The filter of that example is shown in 2nd Embodiment regarding a control system of two systems.

The failure determiner 50 detects the power supply voltage Vr of the inverter 60, the motor rotation speed ω, the voltage command filter values Vu_f, Vv_f, Vw_f output by the filter 44, and the current sensor 70, for example, several times or more during one electrical cycle. The obtained currents Iu, Iv, and Iw are acquired. Then, the failure determiner 50 executes the failure determination process shown in the flowchart of FIG. In the description of FIG. 5, the symbol S represents “step”.

The failure determiner 50 determines that the power supply voltage Vr is larger than the power supply voltage threshold Vrth in S1 and determines that the motor rotation speed ω is lower than the rotation speed threshold ωth in S2 as a premise for starting the failure judgment. .. The power supply voltage threshold Vrth is set to the lower limit of the normal range, and the rotation speed threshold ωth is set to the minimum rotation speed that can ignore the influence of the back electromotive force.
When the normal power supply voltage Vr is not applied to the inverter 60, the voltage according to the voltage command value is not output to the motor 801. Further, in the high rotation region where the influence of the back electromotive force becomes large, it becomes difficult to make a normal determination. Therefore, in the case of NO in S1 or S2, the failure determiner 50 does not start the failure determination and ends the routine.

If YES in S1 and S2, the failure determiner 50 starts the failure determination of each phase in S3. Here, the symbol “#” represents any one of “u, v, w”. FIG. 5 shows a process in which the failure determiner 50 sequentially makes a determination for each phase. However, depending on the processing capability of the failure determiner 50, the determination of three phases may be performed simultaneously in parallel.
In S4, the failure determiner 50 determines whether or not the absolute value of the voltage command filter value Vu_f is larger than the voltage threshold value Vth and the absolute value of the current Iu is smaller than the current threshold value Ith for the U phase, for example.

When YES is determined in S4, the failure determiner 50 determines that the U-phase disconnection failure or the open failure is S5, and outputs a stop signal to the inverter 60. It should be noted that the stop signal may not be output immediately after YES is determined in S4, but the stop signal may be output when the predetermined integrated value is reached a predetermined number of times or within a fixed time.
When NO is determined in S4, the failure determiner 50 determines in S6 that the U phase is not in failure.
In S7, it is determined whether the determination of all phases is completed. If the V phase has not been determined, the process returns to S4, and the V phase is determined next. Further, if the W phase has not been determined, the process returns to S4 again, and then the W phase is determined. When the determination of all phases is completed in this way, the routine ends.

In the case of aiming at fail-safe, the routine may be ended when a failure is determined in any one phase and a stop signal is output.
As for the measures to be taken when the stop signal is output, in addition to turning off the drive signals of all the switching elements 61 to 66, the measures for the failure of well-known techniques such as the interruption of the power relay 12 and the warning to the driver are taken. It can be implemented as appropriate. In the two-system control device according to the second to fourth embodiments, the drive of the system in which the failure is detected may be stopped and switched to the normal one-system drive.

As described above, in the failure determination process of the first embodiment, the voltage command values output by the controller 24 at the same timing are not used for the currents Iu, Iv, and Iw flowing through the motor 801, but are corrected by the filter 49. The voltage command filter values Vu_f, Vv_f, and Vw_f are used.
As a result, the failure determiner 50 can make a failure determination using information whose phases are similar to each other. Therefore, even when the frequency of the current command value is high, the failure determiner 50 can erroneously determine a failure or an open failure. Can be prevented. Therefore, it is possible to properly determine the disconnection failure of the current path or the open failure of the switching element while improving the responsiveness of the current control.

(Second embodiment)
The second embodiment will be described with reference to FIGS. 6 to 8. As shown in FIG. 6 of the overall configuration, the ECU 102 of the second embodiment is a two-system control device that drives a motor 802 having two sets of three-phase windings 841 and 842 that are magnetically coupled to each other. The ECU 102 includes inverters 601, 602, current sensors 701, 702, a control unit 652, and the like.
In the following, with respect to the configuration of the first embodiment, the configuration unique to the two-system control device will be mainly described. Other than the above, the description common to the first embodiment will be omitted.

The first winding 841 of the motor 802, the first inverter 601 that energizes the first winding 841, and the group of components that control energization of the first winding 841 in the control unit 652 are the “first system”. Make up. In addition, the second winding 842 of the motor 802, the second inverter 602 that energizes the first winding 842, and the group of constituent elements that control the energization of the second winding 842 in the control unit 652 are the “second System".

In the figure, "first" is written before the name in the components of the first system, and "second" is written before the name in the components of the second system. In addition, "1" is attached to the third digit of the reference number for the component of the first system, and "2" is attached to the third digit of the reference symbol for the component of the second system. Similarly, the numbers “1” and “2” at the end of the symbols of current and voltage indicate that they are the current and voltage of the first system or the second system. The same applies to FIG. 7 and the subsequent figures.

The phase coils 812, 822, 832 of the second winding 842 of the motor 802 are arranged in a positional relationship of, for example, an electrical angle of 30 deg with respect to the phase coils 811, 821, 831 of the first winding 841. Such a configuration of the windings 841 and 842 is disclosed in, for example, FIG. 3 of Japanese Patent No. 5555645 (hereinafter referred to as “reference document”).

The first inverter 601 includes six switching elements 611, 621, 631, 641, 651, 661 (hereinafter referred to as “611-661”). The second inverter 602 includes six switching elements 612, 622, 632, 642, 652, 662 (hereinafter referred to as “612-662”).
Under normal conditions, the power supply voltage Vr1 is input to the first inverter 601 and the power supply voltage Vr2 is input to the second inverter 602. The first inverter 601 applies the three-phase voltages Vu1_i, Vv1_i, and Vw1_i to the first winding 841 of the motor 802. The second inverter 602 applies the three-phase voltages Vu2_i, Vv2_i, and Vw2_i to the second winding 842.
The power supply relays 121 and 122 of each system and the smoothing capacitor 13 are provided at the input parts of the first inverter 601 and the second inverter 602 that are connected in parallel.

The current sensor 701 of the first system detects the three-phase currents Iu1, Iv1, and Iw1 of the first system by the current detection elements 711, 721, and 731, and feeds them back to the control unit 652.
The second system current sensor 702 detects the second system three-phase currents Iu2, Iv2, and Iw2 by the current detection elements 712, 722, and 732, and feeds them back to the control unit 652.
The control unit 652 controls energization of the motor 802 based on the steering torque trq, the phase currents Iu1, Iv1, Iw1, Iu2, Iv2, Iw2, feedback information of the electrical angle θ, and the like. The control unit 652 also acquires information about the power supply voltages Vr1 and Vr2 and the rotation speed ω.

The control unit 652 of the second embodiment shown in FIG. 7 feedback-controls the currents of the two systems for each system, and includes two sets of substantially the same constituent elements.
The current deviation calculator 331 of the first system calculates the current deviations ΔId1, ΔIq1 between the dq-axis current command values Id ^* , Iq ^* and the dq-axis currents Id1, Iq1 fed back from the three-phase/two-phase conversion unit 311. .. The controller 241 calculates the dq axis voltage command values Vd1 and Vq1 by PI control so that the current deviations ΔId1 and ΔIq1 converge to zero.
The current deviation calculator 332 of the second system calculates the current deviations ΔId2 and ΔIq2 between the dq-axis current command values Id ^* and Iq ^* and the dq-axis currents Id2 and Iq2 fed back from the three-phase/two-phase conversion unit 312. .. The controller 242 calculates the dq axis voltage command values Vd2 and Vq2 by PI control so that the current deviations ΔId2 and ΔIq2 converge to zero.
Although the symbols “PI” are shown on the controllers 241 and 242 in FIG. 7, only I (integration) control may be performed.

However, the control unit 652 further performs control in consideration of intersystem interference due to the influence of mutual inductance generated between the two sets of three-phase windings 841 and 842 of the motor 802. Since inter-system interference is disclosed in detail in the above references, only the main points will be described here. Further, as described above, it is premised that the operation is performed in the low rotation region where the influence of the back electromotive force is not generated, that is, in the region of “revolution speed ω≈0”.

FIG. 8 shows a motor model considering the influence of mutual inductance between the two sets of windings.
In this motor model, the distinction of the dq axes is omitted, and the voltage and current of the first system are described as V1 and I1, and the voltage and current of the second system are described as V2 and I2. The voltages V1 and V2 are voltages applied from the inverters 601 and 602 to the windings 841 and 842, and the currents I1 and I2 are currents flowing through the windings 841 and 842. The self-inductance L of each winding of the first system and the second system and the mutual inductance M between the two sets of windings 841 and 842 are equal to each other.

The voltage equation including the mutual inductance M term is expressed by equations (4.1) and (4.2) when ω≈0.

When formulas (4.1) and (4.2) are arranged for the currents I1 and I2, formulas (4.3) and (4.4) are obtained. The motor model in FIG. 8 expresses this equation.

As shown in the motor model, the product Ms of the mutual inductance M and the Laplace variable s becomes the interference term. The value obtained by subtracting "the interference voltage obtained by multiplying the current of the partner system by the interference term Ms" from the input voltages V1 and V2 of each system is multiplied by "1/(R+Ls)" which is the reciprocal term of the impedance, and the current It is output as I1 and I2.

The control unit 652 of the second embodiment includes the decoupling processing unit 25 that decouples the outputs of the controllers 241, 242 in order to compensate for this interference voltage. The decoupling processing unit 25 performs the post-PI control command currents Id2 ^** , Iq2 ^** , and Id1 ^* calculated by the controllers 242 and 241 of the partner system with respect to the outputs of the controllers 241 and 242 of the own system ^{. The} compensation voltage obtained by multiplying ^* , Iq1 ^** by the interference term Ms is added. The value after the compensation voltage is added is input to the two-phase/three-phase conversion units 281 and 282 as the dq axis voltage command values Vd1, Vq1, Vd2, and Vq2.

The filters 441 and 442 of the second embodiment correct the dq-axis voltage command values output from the controllers 241, 242 before the decoupling process, that is, before the compensation voltage by the interference term Ms is added. That is, the dq axis voltage command values Vd1_0, Vq1_0, Vd2_0, Vq2_0 obtained by multiplying the command currents Id1 ^** , Iq1 ^** , and Id2 ^** , Iq2 ^** after PI control by the impedance term (R+Ls) are obtained. It is input to the filters 441 and 442.
The dq-axis voltage command values Vd1_0, Vq1_0, Vd2_0, Vq2_0 before the inter-system decoupling process correspond to the voltage values after the interference voltage due to the mutual inductance between the two sets of windings 841 and 842 in the motor 802 is removed. To do.

The first filter 441 corrects the dq-axis voltage command values Vd1_0 and Vq1_0 before the decoupling process output by the first controller 241, and outputs the voltage command filter values Vd1_f and Vq1_f.
The second filter 442 corrects the dq axis voltage command values Vd2_0, Vq2_0 before the decoupling process output by the second controller 242, and outputs the voltage command filter values Vd2_f, Vq2_f.

Here, the time constant T ₂ of the filters 441 and 442 is set to be larger than the motor time constant To according to the equation (5), as in the first embodiment in which intersystem interference is not considered.
T ₂ >To (=L/R) (5)
The two-phase/three-phase conversion units 481 and 482 coordinate-convert the voltage command filter values Vd1_f and Vq1_f and Vd2_f and Vq2_f of the respective systems, and three-phase voltage command filter values Vu1_f, Vv1_f, Vw1_f, and Vu2_f, Vv2_f, Calculate Vw2_f.

The first failure determiner 501, for each phase of the first system, based on the flowchart of FIG. 5, disconnection failure of the current path from the first inverter 601 to the first winding 841, or the switching element of the first inverter 601. The open failure of 611-661 is determined. When the first failure determiner 501 determines that there is a failure, the first failure determiner 501 sends a stop signal to the first inverter 601.
The second failure determiner 502 determines, for each phase of the second system, a disconnection failure of the current path from the second inverter 602 to the second winding 842 or a switching element of the second inverter 602 based on the flowchart of FIG. Determine the open failure of 612-662. When the second failure determiner 502 determines that there is a failure, the second failure determiner 502 transmits a stop signal to the second inverter 602.
Regarding S2 of the flowchart regarding the motor rotating machine ω, the determination result of one failure determiner may be shared by the other failure determiner.

As described above, generally, in a two-system control device, due to the influence of mutual inductance between the two sets of three-phase windings 841 and 842, when the response of current control is high, the normal state and the fault state are correctly distinguished. Becomes more difficult than a single-system control device.
On the other hand, in the second embodiment, by correcting the voltage command value before the decoupling process by the filters 441 and 442, the failure determination can be performed under the condition that the phase difference from the current flowing through the motor 802 is reduced. .. Therefore, similarly to the one-system first embodiment, it is possible to appropriately determine the disconnection failure or the open failure while improving the responsiveness of the current control.

(Third Embodiment)
The third embodiment will be described with reference to FIGS. 9 and 10. Note that the overall configuration diagrams of the third and fourth embodiments use FIG. 6 of the second embodiment, and only the reference numeral of the control unit 652 of FIG. 6 is the reference numeral 653 of the control unit of the third and fourth embodiments. Replaced with 654.
The control unit 653 of the third embodiment shown in FIG. 9 performs feedback control of the sum and difference of the currents of the two systems. The control of the sum and the difference is disclosed in the above-mentioned references, and is effective for suppressing torque ripple and improving thermal characteristics.

The control unit 653 includes a current command value addition/subtraction unit 22 and a feedback (“F/B” in the figure) current addition/subtraction unit 32 in addition to the configuration of the control unit 652 of the second embodiment. Further, a sum controller 261, a difference controller 262, and a system voltage calculator 27 are provided instead of the controllers 241 and 242 of each system. Further, instead of the filters 441 and 442 of each system, a sum filter 461, a difference filter 462, and a system voltage calculation unit 47 are provided.

The current command value addition/subtraction unit 22 performs addition/subtraction of the current command values Id ^* and Iq ^* , and the Id sum ^* and Iq sum ^* that are the sum of the current command values, and the Id difference ^* and the Iq difference that are the difference of the current command values. Calculate ^* . Since the two systems have the same electrical characteristics, the Id sum ^* and Iq sum ^* correspond to twice the Id ^* and Iq ^* , and the Id difference ^* and the Iq difference ^* correspond to “0”.
Note that the current command value addition/subtraction unit 22 may not be provided, and “Id sum ^* =2×Id ^* , Iq sum ^* =2×Iq ^* , Id difference ^* =0, Iq difference ^* =0” may be set. ..

The feedback current addition/subtraction unit 32 adds/subtracts the two-system feedback currents Id1, Iq1, Id2, and Iq2 coordinate-converted by the three-phase/two-phase conversion units 311 and 312 to obtain the sum of feedback currents Id sum, Iq sum, and The Id difference and Iq difference, which are the differences in the feedback currents, are calculated.
Current deviation calculator 341, Id sum ^* and Iq sum ^*, calculates Id sum deviation ΔId sum of Iq sum, the ΔIq sum.
The current deviation calculator 342 calculates deviations ΔId difference and ΔIq difference between Id difference ^* (=0) and Iq difference ^* (=0) and Id difference and Iq difference.

The sum controller 261 calculates the Vd sum and Vq sum, which are the sums of the dq axis voltage command values of the two systems, by PI control so that the deviations ΔId sum and ΔIq sum of the current sums converge to zero.
The difference controller 262 calculates the Vd difference and the Vq difference which are the differences between the dq axis voltage command values of the two systems by PI control so that the deviations ΔId and ΔIq of the current difference converge to zero.
In FIG. 10B described later, the symbol “PI” is shown for the sum controller 261 and the difference controller 262, but only I (integration) control may be performed.

The system voltage calculator 27 uses the equations (6.1) to (6.4) to calculate the dq-axis voltage command values Vd1, Vq1, Vq1, Vq1, Vq1, Vq1, Vq1, Vq1, Also, Vd2 and Vq2 are calculated.
Vd1=(Vd sum+Vd difference)/2 (6.1)
Vq1=(Vq sum+Vq difference)/2 (6.2)
Vd2=(Vd sum-Vd difference)/2 (6.3)
Vq2=(Vq sum−Vq difference)/2 (6.4)

The sum filter 461 corrects the Vd sum and Vq sum that are the sums of the dq axis voltage command values output by the sum controller 261 and outputs the voltage command filter values Vd sum_f and Vq sum_f.
The difference filter 462 corrects the Vd difference and the Vq difference which are the differences between the dq axis voltage command values output by the difference controller 262, and outputs the voltage command filter values Vd difference_f and Vq difference_f.
The system voltage calculation unit 47, like the system voltage calculation unit 27, calculates the dq-axis voltage command filter values Vd1_f, Vq1_f, and Vd2_f, Vq2_f for each system.

Next, with reference to FIG. 10A, the “model of motor sum and difference” will be described.
By adding and subtracting both sides of the equations (4.1) and (4.2), the equations (7.1) and (7.2) are obtained.

When the equations (7.1) and (7.2) are arranged with respect to the current sum (I1+I2) and the current difference (I1-I2), the equations (7.3) and (7.4) are obtained. The “motor sum/difference model” in FIG. 10A represents this equation.

Also, the sum controller 261 and the difference controller 262 are represented as shown in FIG. 10B by the corresponding motor inverse model.
In the sum controller 261, the post-PI control command currents Id sum ^** and Iq sum ^** are multiplied by the impedance term “R+(L+M)s)”.
In the difference controller 262, the post-PI control command currents Id difference ^** and Iq difference ^** are multiplied by the impedance term “R+(LM)s”.

The coefficient of the differential operator s when each impedance term is divided by the winding resistance R is defined by equations (8.1) and (8.2).
Ta=(L+M)/R... (8.1)
Tb=(LM)/R... (8.2)
This Ta is called "interference system sum time constant", and Tb is called "interference system difference time constant". In the third embodiment, the interference system sum time constant Ta or the interference system difference time constant Tb is treated as the “time constant of the motor 802”.

The time constant T ₃₊ of the sum filter 461 is set to be larger than the interference system sum time constant Ta as shown in equation (9.1).
T ₃₊ >Ta (=(L+M)/R)... (9.1)
The time constant T ₃₋ of the difference filter 462 is set to be larger than the interference system difference time constant Tb as shown in equation (9.2).
T _{3 -} >Tb (=(L−M)/R) (9.2)
Here, since Ta>Tb, the time constant T ₃₋ of the difference filter 462 may be set larger than the interference system sum time constant Ta in common with the time constant T ₃₊ of the sum filter 461.

The transfer function Ga(s) of the sum filter 461 under the boundary condition of the equation (9.1) and the transfer function Gb(s) of the difference filter 462 under the boundary condition of the equation (9.2) are respectively expressed by the equation ( It is represented by 10.1) and (10.2).

Accordingly, the phase of the three-phase voltage command filter value obtained by coordinate conversion of the dq axis voltage command filter values Vd1_f, Vq1_f, Vd2_f, and Vq2_f corrected by the sum filter 461 and the difference filter 462 is the phase of the current flowing through the motor 802. It approaches the phase.
Therefore, the third embodiment has the same effects as the above-described embodiment.

(Fourth Embodiment)
The fourth embodiment will be described with reference to FIG.
The control unit 654 of the fourth embodiment is provided with filters 491 and 492 for correcting the three-phase voltage command value of each system in the control system of two systems, in contrast to the control unit 651 of the control device of one system according to the first embodiment. .. The first filter 491 corrects the three-phase voltage command values Vu1, Vv1, Vw1 of the first system, and the second filter 492 corrects the three-phase voltage command values Vu2, Vv2, Vw2 of the second system.

The time constant T ₄ of the filters 491 and 492 is set to be larger than the interference system sum time constant Ta (=(L+M)/R) defined in the third embodiment, as shown in Expression (11).
T ₄ >Ta (=(L+M)/R) (11)
Compared to the dq-axis voltage command value, the analysis of inter-system interference for the three-phase voltage command value is complicated. Therefore, in the fourth embodiment, the interference system sum time constant Ta is used as the maximum value of the time constant that can eliminate intersystem interference. That is, by setting the time constants of the filters 491 and 492 to be larger than at least the interference system sum time constant Ta, the phase of the voltage command filter value of each system is set to the phase of the current flowing through the motor 802 while considering the interference between the systems. It can approach the phase.
Therefore, the fourth embodiment has the same effects as the above embodiment.

(Other embodiments)
(A) In the above embodiment, the control device for controlling the drive of a motor having one set or three sets of three-phase windings is described, but the present invention has a multi-phase winding of four or more phases. The same applies to a motor or a motor having three or more sets of multiphase windings.
However, in a motor having three or more sets of multiphase windings, analysis of mutual inductance becomes complicated. Even in that case, by setting the time constant of the filter larger than the theoretically and experimentally estimated maximum motor time constant, the same effect as that of the above embodiment can be obtained.

(B) In the failure determination process of the above embodiment, the three-phase voltage command filter values Vu_f, Vv_f, Vw_f and the three-phase currents Iu, Iv, Iw are compared with the voltage threshold Vth and the current threshold Ith, respectively. Instead of this, the failure determination may be performed based on the dq axis voltage command filter values Vd_f, Vq_f and the dq axial flows Id, Iq.

(C) The control device for a multi-phase rotating machine of the present invention is not limited to the electric power steering device, and may be applied as a control device for any other rotating machine.
As described above, the present invention is not limited to such an embodiment, and can be implemented in various forms without departing from the spirit of the invention.

101, 102... ECU (control device),
24, 241, 242, 261, 262... Controller,
49, 441, 442, 461, 462, 491, 492... Filter,
50, 501, 502... Failure determination device,
60, 601, 602... Inverter (power converter),
61-66, 611-661, 612-662... Switching elements,
651-654... Control unit,
801, 802... Motors (three-phase rotary machine, multi-phase rotary machine),
84, 841, 842... Three-phase winding.

Claims (5)

A controller for controlling driving of a multi-phase rotating machine (801, 802) having one or more sets of multi-phase windings (84, 841, 842),
One or more power converters (60, 601, 602) for energizing the one or more windings by the operation of the plurality of switching elements (61-66, 611-661, 612-662);
One or more controllers (24, 241, 242) for outputting the voltage command value calculated by the feedback control based on the current flowing through the multi-phase rotating machine to the power converter to control the current flowing through the multi-phase rotating machine. , 261, 262),
One or more filters (49, 441, 442, 461, 462, 491, which correct the voltage command value output by the controller by a filtering process for amplifying or attenuating according to the frequency, and output the voltage command filter value, 492),
For each phase, when the absolute value of the voltage command filter value is larger than the voltage threshold value (Vth) and the absolute value of the current flowing through the multi-phase rotating machine is smaller than the current threshold value (Ith), the power converter outputs the One or more failure determiners (50, 501, 502) for determining a disconnection failure of a current path to the multi-phase winding or an open failure of the switching element;
Equipped with
A control device for a multi-phase rotating machine, wherein a time constant of the filter is set to be larger than a time constant (To, Ta, Tb) of the multi-phase rotating machine. A controller for controlling the driving of a multi-phase rotating machine having two sets of multi-phase windings magnetically coupled to each other,
Comprising two power converters capable of individually energizing the two sets of polyphase windings,
When each winding, the power converter that energizes each winding, and a group of constituent elements that control energization to each winding are defined as a system,
The control device for a multi-phase rotating machine according to claim 1, wherein the failure determiner is provided for each system and determines the disconnection failure or the open failure for each phase of each system. The time constant of the filter is
3. The interference system sum time constant (Ta) calculated by dividing the sum of the self-inductance of each winding and the mutual inductance between two sets of windings by the resistance of each winding. Control device for multi-phase rotating machine. The filters (441, 442) are
Decoupling the voltage command value of each system output from the controller (241, 242) corresponding to the voltage value after removing the interference voltage due to the mutual inductance between the two sets of windings in the multiphase rotating machine. The control device for a multi-phase rotating machine according to claim 2, wherein the voltage command value before processing is corrected. As the controller, a sum controller (261) for controlling the sum of currents of the two systems, and a difference controller (262) for controlling the difference of the currents of the two systems are provided,
As the filter, a sum filter (461) for correcting the sum of the voltage command values output by the sum controller, and a difference filter (462) for correcting the difference between the voltage command values output by the difference controller are provided,
The time constant of the sum filter is
It is set larger than the interference system sum time constant (Ta) calculated by dividing the sum of the self-inductance of each winding and the mutual inductance between two sets of windings by the resistance of each winding,
The time constant of the difference filter is
3. The interference system difference time constant (Tb) calculated by dividing the difference between the self-inductance of each winding and the mutual inductance between the two windings by the resistance of each winding. Control device for multi-phase rotating machine.

Images