JP6183285B2 - Power converter control device - Google Patents

Description

  The present invention relates to a power converter control device.

  2. Description of the Related Art Conventionally, a power converter control device that controls driving of an AC motor by controlling on / off of switching elements that constitute the power converter is known. For example, in Patent Document 1, in order to reduce noise generated by an electric motor, a specific frequency band is set as a noise range, and a pulse pattern that suppresses harmonic components generated in the noise range is calculated.

JP 2013-215041 A

In Patent Document 1, for example, when the number of times of switching in one cycle is 7, two frequency components on fixed coordinates can be set to zero. However, in Patent Document 1, there is a limit to the number of harmonics that can be suppressed according to the number of times of switching, and when more frequency components are to be suppressed, the number of simultaneous equations becomes larger than the unknowns, and all simultaneous equations are satisfied. I cannot find a solution. Accordingly, since the harmonic components of the order that cannot be removed are determined by the course, the calculated pulse pattern is not always optimal in terms of noise and vibration.
This invention is made | formed in view of the above-mentioned subject, The objective is to provide the power converter control apparatus which can reduce the noise and vibration resulting from a harmonic component.

A power converter control device of the present invention controls a power converter that converts power supplied to a load, and includes a pulse pattern calculation unit, a harmonic calculation unit, a sound calculation unit, and a pattern selection unit. And comprising.
The pulse pattern calculation unit calculates a pulse pattern according to the load request.
The harmonic calculation unit is a voltage harmonic that is a harmonic component of the voltage applied to the load when the power converter is controlled by a pulse pattern, and a current harmonic that is a harmonic component of the current that is applied to the load. At least one of the waves is calculated for each order for each pulse pattern.
The sound calculation unit estimates the sound pressure harmonic, which is a harmonic component of the sound generated when the power converter is controlled with a pulse pattern, for each order based on the voltage harmonic or current harmonic. An estimation unit and a sound pressure level calculation unit that calculates a sound pressure level generated when the power converter is controlled by a pulse pattern based on the sound pressure harmonics.
A pattern selection part selects the output pulse pattern which is a pulse pattern used for control of a power converter using an evaluation function based on at least one of a voltage harmonic and a current harmonic.

  In the present invention, since a pulse pattern is selected using an evaluation function based on at least one of voltage harmonics and current harmonics, vibration and noise caused by harmonic components can be reduced compared to other control methods. Can do. Moreover, even when the order of harmonic components that affect vibration and noise is large, vibration and noise can be appropriately reduced.

It is a block diagram which shows the structure of the drive system by one Embodiment of this invention. It is a block diagram explaining the modulator of the power converter control apparatus by one Embodiment of this invention. It is explanatory drawing explaining the pulse pattern calculation in the pulse pattern calculating part by one Embodiment of this invention. It is a block diagram explaining the sound calculating part and gain characteristic derivation | leading-out part by one Embodiment of this invention. It is explanatory drawing explaining the map which linked | related the voltage harmonic and sound pressure harmonic by one Embodiment of this invention. It is explanatory drawing explaining the map which linked | related the current harmonic and sound pressure harmonic by one Embodiment of this invention. It is a flowchart explaining the pulse pattern selection process by one Embodiment of this invention. It is a flowchart explaining the pulse pattern selection process by one Embodiment of this invention. It is explanatory drawing explaining the noise reduction effect by pulse pattern control by one Embodiment of this invention. It is explanatory drawing explaining the noise reduction effect by pulse pattern control by one Embodiment of this invention.

Hereinafter, a power converter control device according to the present invention will be described with reference to the drawings.
(One embodiment)
The power converter control apparatus by one Embodiment of this invention is demonstrated based on FIGS.
As shown in FIG. 1, the power converter control device 20 is applied to a drive system 1 that drives a motor generator (hereinafter referred to as “MG”) 10 as a load.

The drive system 1 includes an MG 10, an inverter 15 as a power converter, a power converter control device 20, and the like.
The MG 10 is a so-called “main motor” that generates torque for driving drive wheels of an electric vehicle, for example. The MG 10 of this embodiment is a permanent magnet type synchronous three-phase AC motor. The MG 10 has a function as an electric motor that generates torque for driving drive wheels (not shown), and a function as a generator that can be driven by the kinetic energy of a vehicle transmitted from an engine and drive wheels (not shown) to generate electric power. Hereinafter, the case where MG10 functions as an electric motor will be mainly described.

A power supply voltage Vdc, which is a voltage of a DC power supply, is applied to the inverter 15 according to the driving state of the MG 10 and vehicle requirements. Further, for example, a voltage transformed by a boost converter or the like may be applied to the inverter 15.
The inverter 15 has six switching elements (not shown) that are bridge-connected. As the switching element, for example, an insulated gate bipolar transistor (IGBT), a metal oxide semiconductor (MOS) transistor, a bipolar transistor, or the like can be used. On / off of the switching element is controlled based on an output pulse pattern that is a pulse pattern output from the power converter control device 20. Thereby, the inverter 15 controls the three-phase AC voltages Vu, Vv, Vw applied to the MG 10. The drive of MG 10 is controlled by applying three-phase AC voltages Vu, Vv, and Vw generated by inverter 15.

  The power converter control device 20 is configured by a microcomputer or the like, and includes a CPU, a ROM, a RAM, an I / O, a bus line that connects these, and the like, all not shown. The power converter control device 20 acquires a current detection value from a current detection unit that detects a current passed through a winding constituting the MG 10, an electrical angle of the MG 10 from a resolver (not shown), and the like, and these detection values Based on the above, the operation of the MG 10 is controlled by software processing by executing a program stored in advance by the CPU or hardware processing by a dedicated electronic circuit.

The power converter control device 20 includes a controller 21, a modulator 30, and a gain characteristic deriving unit 40 (see FIG. 4).
The controller 21 performs a d-axis current command value Id ^* and a q-axis current command value Iq ^* calculated by a current command value calculation unit (not shown) based on a torque command value from a host ECU (not shown) and dq conversion (not shown). The d-axis voltage command value Vd ^* and the q-axis voltage command value Vq ^* are calculated based on the d-axis current detection value Id_sns and the q-axis current detection value Iq_sns, which are current detection values dq-converted by the unit. Specifically, the current deviation ΔId between the d-axis current command value Id ^* and the d-axis current detection value Id_sns and the current deviation ΔIq between the q-axis current command value Iq ^* and the q-axis current detection value Iq_sns are calculated, In order to make the detected values Id_sns and Iq_sns follow the current command values Id ^* and Iq ^* , the voltage command values Vd ^* and Vq ^* are calculated so that the current deviations ΔId and ΔIq converge to zero.

As shown in FIG. 2, the modulator 30 includes a modulation factor calculation unit 31, a pulse pattern calculation unit 32, a harmonic calculation unit 33, a sound calculation unit 34, and a pattern selection unit 35.
The modulation factor calculator 31 calculates a command modulation factor m _ref based on the d-axis voltage command value Vd ^* and the q-axis voltage command value Vq ^* . The command modulation rate m _ref is calculated by equation (1).

The pulse pattern calculation unit 32 includes a first calculation unit 321 and a second calculation unit 322, and calculates a pulse pattern according to the command modulation rate m _ref .
When the command modulation rate m _ref is less than the determination threshold Mth, the first calculation unit 321 calculates a pulse pattern by pulse width modulation (PWM) control using a triangular wave comparison method. Further, instead of the pulse width modulation, a pulse pattern may be calculated by space vector modulation (SVM) control. In PWM control or SVM control, since the pulse pattern is uniquely determined according to the command modulation rate m _ref , the pulse pattern calculated by the first calculation unit 321 is selected as the output pulse pattern by the pattern selection unit 35, It is output to the inverter 15.

Further, the pulse pattern calculated by the first calculation unit 321 is output to the harmonic calculation unit 33. Note that, for example, the command modulation rate m _ref is equal to or greater than a predetermined value, and only the pulse pattern in a region that may shift to pulse pattern control described later is output to the harmonic calculation unit 33, and the command modulation rate m The output of the pulse pattern in the region where _ref is less than the predetermined value and does not directly shift to the pulse pattern control to the harmonic calculation unit 33 may be omitted.

When the command modulation rate m _ref is greater than or equal to the determination threshold value Mth, the second calculation unit 322 calculates a plurality of pulse patterns based on the command modulation rate m _ref and the switching count k. The switching frequency k is acquired from, for example, a host ECU (not shown). In the present embodiment, the control based on the pulse pattern calculated based on the command modulation rate m _ref and the switching frequency k is referred to as “pulse pattern control”. In the present embodiment, the determination threshold Mth is set to 1, and “pulse pattern control” is employed in a so-called “overmodulation region” in which the command modulation rate m _ref is equal to or greater than the determination threshold Mth (= 1). In the pulse pattern control, an optimum pulse pattern is selected by the pattern selection unit 35 using an evaluation function based on the calculation results of the harmonic calculation unit 33 and the sound calculation unit 34 described later, and the selected pulse pattern is converted to the inverter 15. Output to.
Hereinafter, the description will focus on pulse pattern control.

The pulse pattern calculation based on the command modulation rate m _ref and the switching frequency k will be described with reference to FIG. In the present embodiment, the switching number k is the number of switchings in an electrical angle half cycle. FIG. 3 shows a pulse pattern related to on / off of the U-phase upper arm element when the number of times of switching is k = 5. The U-phase upper arm element is turned on when the pulse pattern is 1, and is turned off when the pulse pattern is 0. The
First, when the number of times of switching is k, the number of on / off times n in the region of an electrical angle of 0 [°] to 90 [°], which is a quarter cycle, is expressed by Expression (2). “Floor” in formula (2) means truncation. Here, the “number of on / off times n” is the number of sets when switching in the order of on / off is one set.
n = floor (k / 2) (2)

For example, as shown in FIG. 3, when the number of times of switching k = 5, the number of on / off times n of an electrical angle ¼ period is n = 2. For example, at an electrical angle of 0 [°] to 90 [°], the first off time is Except for the period and the last on-period, the pulse has two off-periods and two on-periods. Here, the center phase of the off period is θ _n , the width from the center phase θ _n to being turned on (hereinafter referred to as “the width of the off period” as appropriate) is δ _{n + 1} , the command modulation rate m _ref and the switching frequency k The initial values of the center phases θ _{1 to} θ _n that satisfy the above and the widths δ _{1 to} δ _{n + 1} of the off period are determined. Hereinafter, the center phases θ _{1 to} θ _n and the off-period widths δ _{1 to} δ _{n + 1} are simply referred to as “center phases θ _n and δ _{n + 1} ”. If the width δ _{n + 1} and the center phase θ _{n of the} off period are determined, a pulse pattern having an electrical angle of 0 [°] to 90 [°] is determined. The pulse pattern with electrical angles of 90 [°] to 180 [°] is obtained by inverting the pulse pattern with electrical angles of 0 [°] to 90 [°] around the electrical angle of 90 [°]. The pulse pattern of [°] to 360 [°] is obtained by inverting the pulse pattern of the electrical angle 0 [°] to 180 [°] with the electrical angle 180 [°] as the center.

  In the example of FIG. 3, the width δ1 of the first off period, the width δ2 of the second off period, the width δ3 of the third off period, the center phase θ1 of the second off period, and the width of the third off period The center phase is θ2. Note that the width δ1 of the first off period can be regarded as the width from the center phase until it is turned on, similarly to the widths δ2 and δ3, when the electrical angle 0 [°] is regarded as the center phase.

In the present embodiment, the widths δ1 to δ3 of the off period and the initial values of the center phases θ1 and θ2 are determined based on the command modulation rate m _ref and the switching frequency k. Also, the widths δ1 to δ3 of the off period are shifted by a pulse width shift amount Δδ, which is a value corresponding to the resolution (LSB) of the resolver that detects the electrical angle of the MG 10, and the center phases θ1 and θ2 are changed according to the resolution of the resolver. A plurality of pulse patterns are calculated by shifting the phase shift amount Δθ as a value.
In the present embodiment, the center phase θ _n corresponds to the “pulse position”, the width δ _{n + 1} of the off period corresponds to the “pulse width”, and the phase shift amount Δθ and the pulse width shift amount Δδ are “predetermined intervals”. ".

Further, the second calculation unit 322 calculates a modulation factor m for each calculated pulse pattern, and uses a pulse pattern in which the command modulation factor m _ref matches the calculated modulation factor m as a candidate pulse pattern, and generates a harmonic. The result is output to the calculation unit 33.
The modulation factor m is calculated by equation (3).

The harmonic calculation unit 33 includes a voltage harmonic calculation unit 331 and a current harmonic calculation unit 332.
The voltage harmonic calculation unit 331 calculates a voltage harmonic V _h that is a harmonic voltage of each order of the phase voltage for the candidate pulse pattern output from the pulse pattern calculation unit 32 by Fourier series expansion. Voltage harmonics V _h of order h is calculated by Equation (4). Note that harmonic components having an order h that is a multiple of 6 ± 1 (that is, h = 6 × ± 1) that is likely to cause noise in the overmodulation region may be calculated.

The current harmonic calculation unit 332 calculates the current harmonic I _h of the dq axis current based on the voltage harmonic V _h calculated by the voltage harmonic calculation unit 331. Note that a harmonic component whose order is likely to cause noise in the overmodulation region is a multiple of 6 (6 × order) harmonic components may be calculated.
First, the dq-axis voltage equation related to the h-order voltage harmonic V _h is shown in Equation (5).

The characters in the formula are as follows.
V _dh , V _qh : d-axis component and q-axis component of h-order voltage harmonic V _h I _dh , I _qh : d-axis component and q-axis component of h-order current harmonic I _h s: Laplace operator ω ₁ : Electrical angular velocity L _d , L _q : d-axis component of steady inductance, q-axis component L _dd , L _qq : d-axis component of differential inductance, q-axis component e _dh , e _qh : d-axis component of induced voltage, q Axial component

また、微分インダクタンスＬ_dd、Ｌ_qqは、磁束λおよび電流Ｉの偏微分に基づき、式（７−１）、（７−２）で表される。

The steady-state inductances L _d and L _q are based on magnetic flux λ (d-axis component λ _d , q-axis component λ _q ) and current I (d-axis component I _d , q-axis component I _q ), -1) and (6-2).

The differential inductances L _dd and L _qq are expressed by equations (7-1) and (7-2) based on the partial differentiation of the magnetic flux λ and the current I.

Further, when the Laplace operator s = jω ₁ (j is an imaginary unit) is substituted and the inverse matrix of Equation (5) is solved, the current harmonic I _h is expressed by Equation (8-1). Further, h = 6x, and the 6x-order current harmonic I _6x is expressed by Expression (8-2). _{Incidentally,} I d_6x, I q_6x in the formula (8-2) is, d-axis component of 6x next current harmonics I _6x, a q-axis component.

The voltage harmonic V _h and current harmonic I _h calculated by the harmonic calculation unit 33 are output to the sound calculation unit 34 and the pattern selection unit 35.
In the sound calculation unit 34, based on the voltage harmonic V _h or the current harmonic I _h calculated by the harmonic calculation unit 33, the sound pressure harmonic S is obtained for each candidate pulse pattern by mathematical model approximation or map calculation. _h is calculated for each order.

As shown in FIG. 4, the sound calculation unit 34 includes a gain characteristic storage unit 341, a sound pressure harmonic estimation unit 342, and a sound pressure level calculation unit 343.
The gain characteristic storage unit 341, a mathematical model voltage harmonics derived by the gain characteristic deriving unit 40 V _h or current harmonics I _h and the sound pressure high harmonic S _h is associated, or map the gain characteristics Is remembered as

Gain characteristic deriving unit 40, the relationship between the voltage harmonics V _h or current harmonics I _h and sound pressure high harmonic S _h, is derived based on the numerical analysis or actual measurement data.
And a voltage harmonics V _h and the sound pressure high harmonic S _h explaining the derivation of a mathematical model associated in the gain characteristic deriving unit 40.
The MG 10 is driven with a plurality of pulse patterns for each rotation speed and torque of the MG 10, and the noise (or vibration) at that time is measured. Further, frequency analysis is performed on the pulse pattern and the measured noise (or vibration), and the amplitude of the h-order harmonic component in the audible range is extracted.

The gain characteristic deriving unit 40 derives a relational expression (formula (9)) between the sound pressure harmonic S _h and the voltage harmonic V _h . Incidentally, a _{P_11} in the _formula, a p_12, a p_13 is a coefficient of each term (constant), with different values and the coefficient a _{2_11} when, for example, p = coefficient when the 1 a _{1_11} and p = 2 It means that there is. b ₁ is an intercept (constant). R is the degree of the polynomial to be approximated. That is, for example, when approximating a quadratic polynomial, r = 2.

Similarly, the gain characteristic deriving unit 40 derives a relational expression (formula (10)) between the sound pressure harmonic S _h and the current harmonic I _h . In the formula, a _{p — 21} , a _{p — 22} , a _{p — 23} are coefficients (constants) in the respective orders, and b ₂ is an intercept (constant).

Further, instead of the relationship between the sound pressure high harmonic S _h and voltage harmonics V _h shown in equation (9), and a voltage harmonics V _h and the sound pressure high harmonic S _h, the coordinate system shown in FIG. 5 You may map it. Further, instead of the relationship between the sound pressure high harmonic S _h and current harmonics I _h shown in equation (10), and a current harmonics I _h and sound pressure high harmonic S _h, the coordinate system shown in FIG. 6 You may map it.

The gain characteristic storage unit 341 includes (i) a relational expression (formula (9)) between the sound pressure harmonic S _h and the voltage harmonic V _h, and (ii) a sound pressure harmonic S _h and a current harmonic I _h . And (iii) a map in which the voltage harmonic V _h and the sound pressure harmonic S _h are related (FIG. 5), (iv) the current harmonic I _h and the sound pressure harmonic S _At least one of the maps (FIG. 6) associated with _h is stored.
The current harmonics I _h has a strong interphase between the electromagnetic noise than the voltage harmonics V _h, it is preferable to use the current harmonics I _h can be estimated more appropriately sound pressure high harmonic S _h.

In the sound pressure harmonic estimation unit 342, sound generated by the voltage harmonic V _h or the current harmonic I _h calculated by the harmonic calculation unit 33 based on the map or the mathematical model stored in the gain characteristic storage unit 341. the pressure high harmonic S _h, estimated for each pulse pattern.
In the sound pressure level calculating unit 343, the sound pressure high harmonic sounds pressure high harmonic is computed by the estimation unit 342 S _h, the sound pressure level L which is estimated to occur [dB] by a formula (11) in MG10 The calculation is performed for each pulse pattern. S _o in equation (11) is a reference value. The calculated sound pressure level L is output to the pattern selection unit 35.
The voltage harmonic V _h , the current harmonic I _h , the sound pressure harmonic S _h , and the sound pressure level L are similarly applied to the pulse pattern in the PWM control or the pulse pattern in the SVM control. And the sound calculation unit 34 can calculate.

The pattern selector 35, and the voltage harmonic V _h and current harmonics I _h, which is calculated by the harmonic calculation unit 33, a sound pressure high harmonic S _h and sound pressure level L, which is calculated by the sound operation section 34 Based on the evaluation functions f1 to f5. Details of the evaluation functions f1 to f5 will be described later.

Next, the pulse pattern selection processing according to the present embodiment will be described based on the flowcharts shown in FIGS. The pulse pattern selection process is executed by the modulator 30 at predetermined intervals.
In the first step S101 (hereinafter, “step” is omitted and simply indicated by the symbol “S”), the modulation factor calculation unit 31 is based on the d-axis voltage command value Vd ^* and the q-axis voltage command value Vq ^* . The command modulation factor m _ref is calculated.

In S102, the pulse pattern calculation unit 32 determines whether the command modulation rate m _ref is less than the determination threshold Mth. When it is determined that the command modulation rate m _ref is equal to or greater than the determination threshold Mth (S102: YES), the process proceeds to S104 to perform pulse pattern control. When it is determined that the command modulation rate m _ref is less than the determination threshold Mth (S102: YES), the process proceeds to S103.
In S103, the first calculation unit 321 sets the control method to PWM control, determines a pulse pattern based on the PWM control, and proceeds to S121 in FIG. Note that SVM control may be used instead of PWM control.

In S104, when the command modulation rate m _ref is determined to be equal to or greater than the determination threshold Mth (S102: NO), the second calculation unit 322 sets the control method to pulse pattern control, and switches the number of electrical angle half cycles. k is obtained, and the number of on / off times n in the region of the electrical angle ¼ period is determined.

In S105, the pulse pattern calculation unit 32 determines whether or not the control method is switched from the PWM control. When it is determined that the control method is switched from PWM control (S105: YES), that is, the command modulation rate m _ref in the previous calculation is less than the determination threshold value Mth, and the command modulation rate m _ref in the current calculation is greater than or equal to the determination threshold value Mth. Yes, when switching from PWM control to pulse pattern control, the process proceeds to S109. When it is determined that the control method is not switching from PWM control (S105: NO), that is, both the command modulation rate m _ref in the previous calculation and the current calculation are greater than or equal to the determination threshold Mth, and the pulse pattern control is continued. In the case, the process proceeds to S106.

  In S106, the pulse pattern calculation unit 32 determines whether or not the switching count k has been changed. When it is determined that the switching count k has been changed (S106: YES), the process proceeds to S109. If it is determined that the switching count k has not been changed (S106: NO), the process proceeds to S107.

In S107, the pulse pattern calculation unit 32 determines whether or not the command modulation rate m _ref has been changed. When it is determined that the command modulation factor m _ref has been changed (S107: YES), the process proceeds to S109. When it is determined that the command modulation rate m _ref has not been changed (S107: NO), the process proceeds to S108.
In S108, the pulse pattern calculation unit 32 continues the current pulse pattern, and proceeds to S121 in FIG.

In S109, when the control method is switched from PWM control (S105: YES), the switching frequency k is changed (S106: YES), or the command modulation rate m _ref is changed (S107: YES). The pulse pattern calculation unit 32 sets an initial value of the off-period width δ _{n + 1} that satisfies the command modulation rate m _ref and the switching count k, and the center phase θ _n of the off-period.

In S110, the pulse pattern calculation unit 32 calculates the modulation factor m for the pulse pattern determined based on the width δ _{n + 1} of the off period and the center phase θ _n of the off period.
In S111, it is determined whether or not the modulation factor m calculated in S110 matches the command modulation factor _mref . When it is determined that the modulation factor m does not match the command modulation factor m _ref (S111: NO), the process proceeds to S113 without performing the process of S112. When it is determined that the modulation factor m matches the command modulation factor m _ref (S111: YES), the process proceeds to S112.
In S112, a pulse pattern in which the command modulation rate m _ref matches the calculated modulation rate m is stored as a candidate pattern. The storage of the candidate pulse pattern here is temporary, and is reset when the processing up to S121 is completed.

In S113, the pulse pattern calculation unit 32 determines whether or not extraction of all pulse patterns that satisfy the command modulation rate m _ref and the switching count k is completed. If it is determined that extraction of all pulse patterns has been completed (S113: YES), the candidate pulse pattern stored in S112 is output to the harmonic calculation unit 33, and the process proceeds to S115 in FIG. When it is determined that the extraction of all the pulse patterns is not completed (S113: NO), the process proceeds to S114.
In S114, the width δ _{n + 1} of the off period is shifted by the pulse width shift amount Δδ. Further, the center phase θ _n of the off period is shifted by the phase shift amount Δθ, and the process returns to S110.

When it is determined that extraction of all pulse patterns satisfying the command modulation rate m _ref and the switching frequency k has been completed (S113: YES), the processing of S115 to S118 in FIG. Is executed for each pulse pattern.

In S115, the voltage harmonics calculation unit 331, a Fourier series expansion, calculates the voltage harmonics V _h.
In S116, the current harmonics calculation unit 332, based on the voltage harmonic V _h and voltage equation is calculated in S115, and calculates the current harmonics I _h.
In S117, the sound pressure high harmonic estimation unit 342, based on the voltage harmonics V _h or current harmonics I _h, calculates the sound pressure high harmonic S _h.
In S118, the sound pressure level calculating unit 343, based on the sound pressure high harmonic S _h, calculates the sound pressure level L.

For all candidate pulse patterns, in S119, after the voltage harmonics V _h , current harmonics I _h , sound pressure harmonics S _h , and sound pressure levels L are calculated, the pattern selection unit 35 selects the pulse patterns. Select an evaluation function to be used for selection.
In S120, the pattern selection unit 35 selects an optimal pulse pattern from the pulse patterns stored in S112 based on the evaluation function selected in S119.
In S <b> 121, the pattern selection unit 35 outputs the selected pulse pattern to the inverter 15.

Here, selection of an evaluation function and selection of a pulse pattern in the pattern selection unit 35 will be described.
The pattern selection unit 35 stores a first evaluation function f1, a second evaluation function f2, a third evaluation function f3, a fourth evaluation function f4, and a fifth evaluation function f5. In the pattern selection unit 35, any one of the first evaluation function f1, the second evaluation function f2, the third evaluation function f3, the fourth evaluation function f4, and the fifth evaluation function f5 is selected according to, for example, the rotational speed or torque of the MG 10. The evaluation function used for selecting the pulse pattern is selected.

The first evaluation function f1 selects a pulse pattern that minimizes the sound pressure level L, and is represented by Expression (12).
f1 = min (L) (12)
The second evaluation function f2 selects a candidate pulse pattern that minimizes the noise difference from the current pulse pattern, and is represented by Expression (13). Note that the sound pressure level of the candidate pulse pattern is L (z + 1), and the sound pressure level of the current pulse pattern used for control of the inverter 15 is L (z).
f2 = min | L (z) -L (z + 1) | (13)

Third evaluation function f3 is for selecting a candidate pulse pattern each order component of the current pulse patterns and sound pressure high harmonic S _h is the most similar, the formula (14). It is assumed that the sound pressure harmonic S _h (z + 1) of the candidate pulse pattern and the sound pressure harmonic of the current pulse pattern used for the control of the inverter 15 are S _h (z).

Fourth evaluation function f3 is for selecting a candidate pulse pattern each order component of the current pulse pattern and voltage harmonics V _h is the most similar, represented by the formula (15). Note that the voltage harmonic V _h (z + 1) of the candidate pulse pattern and the voltage harmonic of the current pulse pattern used for the control of the inverter 15 are V _h (z).

Fifth evaluation function f5 is for selecting the pulse patterns each order component of the current pulse pattern and current harmonics I _h is most similar, the formula (16). Note that the current harmonic of the candidate pulse pattern is I _h (z + 1), and the current harmonic of the current pulse pattern used for controlling the inverter 15 is I _h (z).

The output pulse selected from the candidate pulse patterns by the pattern selection unit 35 based on the first evaluation function f1, the second evaluation function f2, the third evaluation function f3, the fourth evaluation function f4, or the fifth evaluation function f5 The pattern is output to the inverter 15 and on / off of the switching elements constituting the inverter 15 is controlled based on the output pulse pattern.

In the present embodiment, the sound pressure level can be minimized by selecting a pulse pattern based on the first evaluation function f1.
Further, by selecting a pulse pattern based on the second evaluation function f2, it is possible to minimize the sound pressure level difference associated with the switching of the pulse pattern, so that the change in sound associated with the switching of the pulse pattern is minimized. Can be suppressed. Similarly, by selecting a pulse pattern based on the third evaluation function f3, it is possible to minimize a change in sound accompanying switching of the pulse pattern.
Furthermore, by selecting a pulse pattern based on the fourth evaluation function f4 or the fifth evaluation function f5, changes in vibration, noise, etc. associated with the switching of the pulse pattern are minimized, and the user feels uncomfortable. Can be suppressed.

  Here, the noise reduction effect of the power converter control apparatus 20 of this embodiment is demonstrated based on FIG. 9 and FIG. 9 and 10, the MG 10 is rotated at a modulation rate of 1.1 and a rotation speed N = 1860 [rpm]. At this time, it is assumed that the 18th harmonic is about 4.5 [kHz] and the mechanical resonance frequency of MG10 is 4.5 [kHz]. In FIGS. 9 and 10, the 6 × harmonic components are shown, and other orders are omitted.

In FIG. 9, (a) to (c) show a case where overmodulation PWM control is performed, and (d) to (f) show a case where pulse pattern control is performed. In the pulse pattern control of (d) to (f), the pulse pattern that minimizes the sound pressure level L is selected by the first evaluation function f1. Further, FIG. 9 (a), (d) shows the sound pressure level L, (b), (e ) shows a d-axis component I _dh of current harmonics _{I h, (c), (} f) the current The q-axis component I _qh of the harmonic Ih is shown.
9A to 9F, the bar graphs corresponding to the respective orders indicate the switching times k = 5, 7, and 9 in order from the left.

  In FIG. 10, (a) is the sound pressure level L when the switching frequency k = 5, (b) is the sound pressure level L when the switching frequency k = 7, and (c) is the switching frequency k =. 9 shows the sound pressure level L. In FIGS. 10A to 10C, the left side shows a case where overmodulation PWM control is performed, and the right side shows a case where pulse pattern control is performed. The sound pressure level L in FIG. 10 is a result of calculation using only the 6 × order component.

As shown in FIG. 9, by performing the pulse pattern control of the present embodiment, the d-axis component I _dh and the q-axis component I _{qh of the} 6 × _-order current harmonic I _h are compared with the case where overmodulation PWM control is performed. Are equivalent or reduced. Further, as shown in FIGS. 9 and 10, the sound pressure level L caused by the harmonic component of each order is also equalized or reduced. Particularly, since the d-axis component I _dh and the q-axis component I _{qh of the} 18th-order current harmonic I _h close to the mechanical resonance frequency are suppressed, vibration and noise of the MG 10 due to resonance can be reduced.

As described above in detail, the power converter control device 20 controls the inverter 15 that converts the power supplied to the MG 10, and includes a pulse pattern calculation unit 32, a harmonic calculation unit 33, and a pattern selection. Part 35.
The pulse pattern calculation unit 32 calculates a pulse pattern according to the request of the MG 10.

The harmonic calculation unit 33 is a voltage harmonic V _h that is a harmonic component of a voltage applied to the MG 10 when the inverter 15 is controlled by a pulse pattern, and a harmonic component of a current that is passed through the MG 10. At least one of the current harmonics I _h is calculated for each order for each pulse pattern.
Pattern selection unit 35 based on at least one of the voltage harmonic V _h and current harmonics I _h, selects an output pulse pattern is a pulse pattern used for control of the inverter 15 using an evaluation function f1 to f5.

In the present embodiment, based on at least one of the voltage harmonic V _h and current harmonics I _h, so selecting the pulse pattern by using the evaluation function f1 to f5, in comparison with other control methods, due to the harmonic components Vibration and noise can be reduced. Further, even when the order of harmonic components that affect vibration and noise is large, vibration and noise can be appropriately reduced.

The power converter control device 20 further includes a sound calculation unit 34. The sound calculation unit 34 includes a sound pressure harmonic estimation unit 342 and a sound pressure level calculation unit 343. Sound pressure high harmonic estimation unit 342, based on the sound pressure high harmonic S _h is a harmonic component of the sound generated when controlling the inverter 15 with pulse pattern voltage harmonics V _h or current harmonics I _h Estimate for each order. Sound pressure level calculating section 343 calculates on the basis of the sound pressure level L that occur when controlling the inverter 15 with pulse pattern to the sound pressure high harmonic S _h.
Thereby, the noise generated due to the harmonic component can be appropriately estimated.

The power converter controller 20 further includes voltage harmonics V _h or current harmonics I _h and sound pressure high harmonic S _h and are associated map, or the gain characteristic deriving unit 40 for deriving a mathematical model.
Sound pressure high harmonic estimation unit 342, based on the derived map or mathematical model with a gain characteristic deriving unit 40, estimates the sound pressure high harmonic S _h.
Thus, the sound pressure high harmonic S _h can be appropriately estimated.

The pattern selection unit 35 can select an output pulse pattern as follows by selecting an evaluation function according to, for example, the rotation speed or torque of the MG 10.
The evaluation function f1 selects a pulse pattern having the smallest sound pressure level L as an output pulse pattern. Thereby, the noise which generate | occur | produces in MG10 can be reduced.

The evaluation function f2 selects a pulse pattern having the smallest difference from the current sound pressure level L in the MG 10 as an output pulse pattern when switching the output pulse pattern.
Evaluation function f3, when switching the output pulse pattern, and selects as an output pulse pattern of the current sound pressure high harmonic S _h and pulse pattern each order component is the most similar in MG 10.
Thereby, the change of the sound accompanying switching of a pulse pattern can be suppressed to the minimum, and the discomfort given to a user can be reduced.

Incidentally, the sound pressure high harmonic S _h and sound pressure level L, because they are calculated based on voltage harmonics V _h or current harmonics I _h, based on the sound pressure high harmonic S _h or sound pressure level L output Selecting the pulse pattern is included in the concept of “selecting the output pulse pattern based on the voltage harmonic V _h or the current harmonic I _h ”.

Evaluation function f4 when switching the output pulse pattern, and selects as an output pulse pattern of the current voltage harmonics V _h and pulse pattern each order component is the most similar in MG 10.
Evaluation function f5, when switching the output pulse pattern, and selects as an output pulse pattern of the current of the current harmonics I _h and pulse pattern each order component is the most similar in MG 10.
Thereby, the change of the sound and vibration accompanying switching of a pulse pattern can be suppressed to the minimum, and the discomfort given to a user can be reduced.

  Here, “switching the output pulse pattern” means changing to a pulse pattern different from the pulse pattern used for the current control of the MG 10, and switching from PWM control or SVM control to pulse pattern control. (S105: YES in FIG. 7), or a case where the pulse pattern is switched while the pulse pattern control is continued (S106: YES or S107: YES).

The voltage harmonic calculation unit 331 of the harmonic calculation unit 33 calculates the voltage harmonic V _h by Fourier series expansion. Thus, it is possible to appropriately calculates voltage harmonics V _h.
Moreover, the current harmonic calculation unit 332 of the harmonic calculation unit 33 calculates the current harmonic I _h based on the voltage harmonic V _h and the voltage equation. Thus, it is possible to appropriately calculating the current harmonics I _h.

The second calculation unit 322 of the pulse pattern calculation unit 32 sets initial values of the center phase θ _n of the off period and the width δ _{n + 1} of the off period, which are pulse positions according to the command modulation rate m _ref and the switching frequency k. Then, the pulse pattern is calculated by changing the center phase θ _n and the off period width δ _{n + 1} at predetermined intervals. Specifically, the center phase θ _n of the off period is shifted by the phase shift amount Δθ, and the width δ _{n + 1} of the off period is shifted by the pulse width shift amount Δδ. Thereby, it is possible to appropriately calculate a pulse pattern that satisfies the command modulation rate m _ref and the switching frequency k.

In addition, the pulse pattern calculation unit 32 includes a first calculation unit 321 and a second calculation unit 322.
When the command modulation factor m _ref is less than the determination threshold value Mth, the first calculation unit 321 calculates one pulse pattern by PWM control or SVW control.
When the command modulation rate m _ref is equal to or greater than the determination threshold value Mth, the second calculation unit 322 calculates a plurality of pulse patterns according to the command modulation rate m _ref and the switching count k.

When the command modulation factor m _ref is less than the determination threshold Mth, the pattern selection unit 35 selects the pulse pattern calculated by the first calculation unit 321 as the output pulse pattern. When the command modulation rate m _ref is equal to or greater than the determination threshold value Mth, an output pulse pattern is selected from the pulse patterns calculated by the second calculation unit 322 using the evaluation functions f1 to f5.
Thereby, an appropriate pulse pattern can be selected according to the command modulation rate m _ref . In this embodiment, by setting the determination threshold Mth to 1 and performing pulse pattern control in the overmodulation region, vibration and noise in the overmodulation region can be reduced.

(Other embodiments)
(A) Pulse pattern calculation unit In the above embodiment, the determination threshold is set to 1, and pulse pattern control is performed in the overmodulation region. In other embodiments, the determination threshold is not limited to 1 and may be any value. Further, the determination threshold value may not be provided, and pulse pattern control may be performed regardless of the modulation rate.
(A) Harmonic Calculation Unit In the above embodiment, the harmonic calculation unit calculates a voltage harmonic and a current harmonic. In other embodiments, the computation of voltage harmonics or current harmonics may be omitted. For example, by omitting the calculation of the current harmonics, the calculation for solving the inverse matrix of the voltage equation can be omitted, so that the calculation load can be reduced.

In the above embodiment, the voltage harmonic is calculated by Fourier series expansion, and the current harmonic is calculated based on the voltage harmonic and the voltage equation. In other embodiments, the voltage and current harmonics may be computed in any manner.
In the above-described embodiment, an example of mainly calculating a 6x-order harmonic component has been described. However, in other embodiments, it is not limited to 6x-order, and at least one order of harmonic component may be calculated. The harmonic components of all orders below the order may be calculated, or the harmonic components of a specific order different from the 6x order may be calculated.

(C) Estimation of sound pressure harmonics In the above embodiment, sound pressure harmonics are estimated based on a map or a mathematical model. In another embodiment, the gain characteristic deriving unit may derive a map or a mathematical model related to the estimation of the sound pressure harmonic for each operating region of the load. The sound calculation unit may estimate the sound pressure harmonic using a map or a mathematical model corresponding to the operating region of the load.

(D) Pattern Selection Unit In the above embodiment, the pattern selection unit includes the first evaluation function f1, the second evaluation function f2, the third evaluation function f3, the fourth evaluation function f4, and the fifth evaluation function f5. In other embodiments, some evaluation functions may be omitted. Note that the pattern selection unit may have one evaluation function and select an output pulse pattern based on the evaluation function.

  In the third embodiment, the third evaluation function f3 represents a pulse pattern having a minimum square sum of deviations of each order component of the current sound pressure harmonic in the load and each order component of the pulse pattern, and the current sound pressure in the load. A pulse pattern in which the harmonics and each order component are most similar is used. In another embodiment, the evaluation function relating to the selection of the pulse pattern with the most similar sound pressure harmonics is not limited to the minimum square sum of deviations, and may be any function. The same applies to voltage harmonics and current harmonics.

  The evaluation function may be an evaluation function that selects a pulse pattern that minimizes the sum of squares of each order component of voltage harmonics or current harmonics. Thereby, the loss resulting from a harmonic component can be reduced. Furthermore, the evaluation function is not limited to this, and any function may be used.

(E) Power converter control device In another embodiment, at least a part of the calculations in the pulse pattern calculation unit, harmonic calculation unit, sound calculation unit, and gain characteristic deriving unit are previously calculated offline. The calculation result may be stored in a storage unit or the like.
In the said embodiment, it applies to the MG system which drives MG which is a vehicle main machine. For example, the present invention may be applied to a vehicle other than the vehicle main machine such as a vehicle auxiliary machine. Moreover, you may not have a function as a generator or an electric motor. In the above embodiment, the load is a permanent magnet type synchronous three-phase AC motor. However, in other embodiments, the load may be an induction motor or other synchronous motor. Moreover, it is good also as a multiphase rotating machine more than four phases. Furthermore, the load is not limited to MG and may be any load.
As mentioned above, this invention is not limited to the said embodiment at all, In the range which does not deviate from the meaning of invention, it can implement with a various form.

DESCRIPTION OF SYMBOLS 1 ... Drive system 10 ... Motor generator (load)
DESCRIPTION OF SYMBOLS 15 ... Inverter 20 ... Power converter control apparatus 32 ... Pulse pattern calculation part 33 ... Harmonic calculation part 34 ... Sound calculation part 35 ... Pattern selection part 40 ... Gain characteristic Derived part

Claims (11)

A power converter control device (20) for controlling a power converter (15) for converting power supplied to a load (10),
A pulse pattern calculation unit (32) for calculating a pulse pattern according to the load request;
A voltage harmonic that is a harmonic component of a voltage applied to the load when the power converter is controlled by the pulse pattern, and a current harmonic that is a harmonic component of a current applied to the load. A harmonic calculation unit (33) for calculating at least one of the respective pulse patterns for each order;
Sound pressure harmonic estimation for estimating a sound pressure harmonic, which is a harmonic component of a sound generated when the power converter is controlled by the pulse pattern, for each order based on the voltage harmonic or the current harmonic And a sound calculation unit having a sound pressure level calculation unit (344) that calculates a sound pressure level generated when the power converter is controlled by the pulse pattern based on the sound pressure harmonics (34)
A pattern selection unit (35) that selects an output pulse pattern that is the pulse pattern used to control the power converter using an evaluation function based on at least one of the voltage harmonic and the current harmonic;
A power converter control device comprising: A gain characteristic deriving unit (40) for deriving a map in which the voltage harmonic or the current harmonic and the sound pressure harmonic are associated, or a mathematical model;
The sound pressure high harmonic estimating unit, based on the map or the mathematical model derived by the gain characteristic deriving unit, a power converter according to claim 1, wherein the estimating the sound pressure high harmonics Control device. The evaluation function, the power converter control apparatus according to claim 1 or 2, wherein the sound pressure level is the smallest the pulse pattern which is selected as the output pulse pattern. The evaluation function, when switching the output pulse pattern, according to claim 1, characterized in that the pulse pattern difference between the current of the sound pressure level is the smallest in the load is to select as the output pulse pattern or the power converter control apparatus according to 2. When the output pulse pattern is switched, the evaluation function selects the pulse pattern having the most similar order component as the current sound pressure harmonic in the load as the output pulse pattern. The power converter control device according to claim 1 or 2 . The evaluation function, when switching the output pulse pattern, selects, as the output pulse pattern, the pulse pattern whose order components are most similar to the current voltage harmonics in the load. Item 3. The power converter control device according to Item 1 or 2 . The evaluation function, when switching the output pulse pattern, selects, as the output pulse pattern, the pulse pattern whose order components are most similar to the current harmonics in the load. Item 3. The power converter control device according to Item 1 or 2 . The harmonic calculation section by the Fourier series expansion, the power converter control apparatus according to any one of claim 1 to 7, characterized in that computing the voltage harmonics. The power converter control device according to any one of claims 1 to 8 , wherein the harmonic calculation unit calculates the current harmonic based on the voltage harmonic and a voltage equation. The pulse pattern calculation unit sets an initial value of a pulse position and a pulse width according to a command modulation rate and the number of times of switching, and changes the pulse position and the pulse width at a predetermined interval, thereby generating a plurality of the pulse patterns. power converter control apparatus according to any one of claim 1 to 9, characterized in that computing. The pulse pattern calculator is
A first calculation unit (321) that calculates one pulse pattern by pulse width modulation control or space vector modulation control when the command modulation rate is less than a determination threshold;
When the command modulation rate is equal to or greater than the determination threshold, a second calculation unit (322) that calculates a plurality of the pulse patterns according to the command modulation rate and the number of switching times;
Have
The pattern selection unit
When the command modulation rate is less than the determination threshold, the pulse pattern calculated by the first calculation unit is selected as the output pulse pattern,
When the command modulation rate is equal to or higher than the determination threshold, the output pulse pattern is selected from the plurality of pulse patterns calculated by the second calculation unit using the evaluation function. The power converter control device according to any one of claims 1 to 10 .

Images