JP6665745B2 - Rotating machine control device - Google Patents

Description

  The present invention relates to a rotating machine control device that performs position sensorless control.

2. Description of the Related Art Conventionally, in a control device that drives a multiple winding synchronous rotating machine with a plurality of inverters, a control device that performs position sensorless control for estimating a magnetic pole position from voltage and current information is known.
For example, in a position sensorless control device for a synchronous machine disclosed in Patent Literature 1, an average value is calculated after an estimated position is calculated by each of a master inverter and a slave inverter.

Japanese Patent No. 5527025

  In this specification, a unit of a group of components that controls the energization of a winding set of a rotating machine is defined as a “system”. The configuration including the master inverter and the slave inverter disclosed in FIG. 1 of Patent Document 1 and the like is a two-system configuration. Since a motor constant such as a winding resistance and an inductance and a power supply voltage may have an imbalance among a plurality of systems, an error is included in an extended induced voltage estimated based on a voltage and a current of each system. Then, as a result of performing the PI control to reduce the position error calculated from the extended induced voltage to zero, a position error occurs at the estimated magnetic pole position.

In the device of Patent Literature 1, the position is estimated based on the average value of the extended induced voltages of the two systems, that is, the sum (hereinafter, “system sum induced voltage”). If the sign of the error of the extended electromotive force of the first system and the error of the error of the extended electromotive voltage of the second system are opposite and the absolute values are equal, the error of the two systems is canceled by calculating the average value. The error of the magnetic pole position calculated from the sum induced voltage is reduced.
However, when the error of the extended induced voltage of the two systems has the same sign, or when the error of the extended induced voltage exists only in one system and is almost zero in the other system, the error remains in the system sum induced voltage. Therefore, a position error occurs based on the error of the system sum induced voltage.

  The present invention has been made in view of such a point, and an object thereof is to reduce a position error due to imbalance between systems in a position sensorless control of a rotating machine having a plurality of multi-phase winding sets. An object of the present invention is to provide a control device for a rotating machine.

The control device for a rotating machine of the present invention controls the energization of a plurality of winding sets by a position sensorless control for a rotating machine (80) having a plurality of multi-phase winding sets of three or more phases in a stator (84). . The control device for the rotating machine includes a plurality of power converters (61, 62), a system sum induced voltage estimation unit (41, 42), a system difference induced voltage estimation unit (43, 44, 47), a magnetic pole position. An estimating unit (56).

The plurality of power converters convert the input power into AC power and supply the AC power to the rotating machine.
When the unit of a group of components that controls the energization to the specific winding set is defined as “system”, the system sum induced voltage estimation unit calculates the sum of the extended induced voltages of two specific systems selected from a plurality of systems. Estimate a certain system sum induced voltage (e _γδ w). Strain differences induced voltage estimation unit estimates the difference in a strain differences induced voltage of extension induction voltage of a particular two systems _(e γδ s).
Here, the "sum" of the extended induced voltage includes the average value obtained by dividing the added value by 2, and the "difference" includes the half difference value obtained by dividing the subtracted value by 2.
The magnetic pole position estimating unit calculates the minimum position error (Δθ) from the system sum induced voltage, a value obtained by adding the system difference induced voltage to the system sum induced voltage, and a value obtained by subtracting the system difference induced voltage from the system sum induced voltage. The magnetic pole position (θγ) of the rotor (85) of the rotating machine is estimated based on one of the values selected so that

  As described above, the present invention does not always use the system sum induced voltage as in the related art of Patent Literature 1, but also uses the addition / subtraction value of the system sum induced voltage and the system difference induced voltage as occasion demands. The magnetic pole position can be estimated. The position error estimated by the magnetic pole position estimating unit can be reduced by appropriately selecting a value that minimizes the position error from the three types of calculated values of the extended induced voltage.

The control device for a rotating machine of the present invention preferably further includes a determination unit (53). The determination unit selects the minimum value of the system sum induced voltage, the value obtained by adding the system difference induced voltage to the system sum induced voltage, and the value obtained by subtracting the system difference induced voltage from the system sum induced voltage, and estimates the magnetic pole position. It outputs the selected extended induced voltage (e _γδ #) used for estimating the magnetic pole position by the section.
Specifically, the determination unit selects the minimum value among the three types of calculation values that are input each time the controller of the rotating machine is operating, and outputs the selected value as the extended induced electromotive force to the magnetic pole position estimation unit. Thus, even when the error pattern is not known or changes with time, the position error can be minimized in each cycle of the position estimation by the magnetic pole position estimation unit.

FIG. 3 is a control block diagram of a control device for the rotating machine according to the first embodiment. (A) Schematic diagram of a field circuit. (B) Partial sectional view of the stator of the two-system rotating machine. The schematic diagram of the winding of the three-phase two-system rotating machine. The figure explaining the relationship of the axis | shaft of a fixed coordinate system, a rotating coordinate system, and an estimated rotating coordinate system. FIG. 4 is a control block diagram of a control device for a single rotating machine for explaining estimation of a magnetic pole position. The vector diagram of the induced voltage vector. 9 is a table showing a relationship between an error pattern of estimated induced voltages of two systems, an error of a sum, and a difference. 5 is a flowchart (1) of a calculation value selection process by a determination unit. 9 is a flowchart (2) of a calculation value selection process by the determination unit. 9 is a flowchart (3) of a calculation value selection process by the determination unit. The control block diagram of the control device of the rotating machine by a 2nd embodiment. The control block diagram of the control device of the rotating machine by a 3rd embodiment. The control block diagram of the control device of the rotating machine by a 4th embodiment. The figure explaining the effect | action of (a) 6th harmonic and (b) 12th harmonic by calculation of the sum of 3 phase 2 system electric currents. 6A is a diagram illustrating a relationship between a speed (motor rotation speed) and a cutoff frequency, and FIG. 6B is a diagram illustrating a relationship between a cutoff frequency of a system sum / system difference induced voltage estimating unit. FIG. 4 is a control block diagram of a control device for a rotating machine of a comparative example.

  Hereinafter, a plurality of embodiments of a rotating machine control device will be described with reference to the drawings. In a plurality of embodiments, substantially the same configuration is denoted by the same reference numeral, and description thereof is omitted. In addition, the matters described in the above-mentioned embodiment are applied in principle in common, except for matters that are described differently in the following embodiment.

The control device of the rotating machine is applied to a system that controls the energization of an ISG (Integrated Starter Generator) in which the functions of a starter and an alternator are integrated in a vehicle having an engine as a driving force source, for example.
In the following description of the embodiments, the “rotating machine” in the claims is referred to as “motor”, and the “rotating machine control device” is referred to as “motor control device”. The first to fourth embodiments are collectively referred to as “the present embodiment”.

  The motor control device of each embodiment is a device that controls energization from a plurality of power converters to a corresponding winding set for a motor having a plurality of multi-phase winding sets of three or more phases in a stator. Hereinafter, a unit of a group of constituent elements that controls energization to a specific winding set is defined as “system”. Specifically, each of the motors of the first to fourth embodiments has three phases and two systems. Assuming such a configuration, the motor control device of each embodiment controls the output voltage of the power converter of each system by position sensorless control.

  The reference numeral of the motor control device according to each embodiment is the number of the embodiment in the third digit following “10”. In addition, the reference numerals of the two-system motor of each embodiment are “2” in the third digit following “80”. The power converter, the current sensor, and the like, which are components of each system, use the last number of the two- or three-digit code as the system number. For example, the first system component has a code suffix “1”, and the second system component has a code suffix “2”.

(1st Embodiment)
A three-phase two-system motor control device according to the first embodiment will be described with reference to FIGS. As shown in FIG. 1, the motor control device 101 is applied to a system that supplies AC power from two power converters 61 and 62 to a motor 802 having two sets of three-phase windings. The motor 802 is a permanent magnet synchronous type three-phase AC motor, for example. The motor 802 is not provided with a rotation angle sensor for detecting the magnetic pole position of the rotor.

As shown in FIG. 2A, near the motor 802, a field circuit 87 having a field winding 88 having an inductance Mf and flowing a field current if is provided. Note that a controller for controlling the field current if is not shown.
As shown in FIG. 2B, in a permanent magnet synchronous motor 802, a rotor 85 having magnetic poles 86 arranged in a circumferential direction is rotatably supported inside a stator 84. FIG. 2B shows an example in which the rotor 85 is provided with two pairs of N poles and S poles (ie, a total of four poles).
Regarding the windings wound around the stator 84, for example, "U1" means a first system U-phase winding, and "V2" means a second system V-phase winding. In-phase windings of the first system and the second system are alternately arranged in the circumferential direction of the stator 84.

In FIG. 3, the first system winding 81 is represented by a solid line, and the second system winding 82 is represented by a broken line. In each phase winding of the three-phase two-system motor 802, the phase between the first system and the second system is shifted by an electrical angle of 30 (that is, 60/2) °.
The αβ axis of the two-phase fixed coordinate system is defined such that the α axes α1 and α2 coincide with the U-phase axes U1 and U2.

Returning to FIG. 1, the first and second power converters 61 and 62 correspond to the first and second system windings 81 and 82, respectively. The three diagonal lines attached to the input signal lines and the output power paths of the power converters 61 and 62 represent three phases. Typically, power converters 61 and 62 are inverters that convert DC power input from a DC power source such as a battery into AC power by the operation of a plurality of switching elements.
In FIG. 1, illustration of the input-side power paths of the power converters 61 and 62 is omitted. A DCDC converter or the like may be provided between the DC power supply and the power converters 61 and 62.
Also, the configuration of a three-phase AC inverter including six switching elements of three-phase upper and lower arms is well known in the art, and is not shown.

  The triangular-wave marks illustrated in the power path between the power converters 61 and 62 and the motor 802 indicate the current sensors 71 and 72 that detect the phase current of each system. The current sensors 71 and 72 may be provided for all three phases, or may be provided for two of the three phases, and the remaining one-phase current may be calculated according to Kirchhoff's law.

  The motor control device 101 is based on a vector control configuration based on position sensorless control, which is a well-known technique disclosed in, for example, Japanese Patent No. 3411878. Here, FIG. 4 shows the relationship between axes of each coordinate system used for vector control. The αβ axes are axes orthogonal to each other in a two-phase fixed coordinate system, and the dq axes are axes orthogonal to each other in a two-phase rotating coordinate system. Due to the rotation of the rotor 85, the electrical angle θ, that is, the phase of the dq axis with respect to the αβ axis changes with time.

  In the position sensorless control that does not directly detect the actual electrical angle θ, the γδ axis of the two-phase estimated rotational coordinate system is defined separately from the dq axes. The γ axis is the estimated d axis, and the δ axis is the estimated q axis. The difference between the estimated magnetic pole position θγ on the γδ axis and the actual electrical angle θ is represented as “position error Δθ”. In the position sensorless control, an electrical angular velocity ω is calculated by PI control or the like so as to converge the position error Δθ to 0, and the control calculation is executed by regarding the estimated magnetic pole position θγ as the actual electrical angle θ.

In the following description and drawings, the symbol “i _γδ ” means γ-axis current i _γ and δ-axis current i _δ . In the equations, the γ-axis current i _γ and the δ-axis current i _δ are separately represented in a matrix form. The same applies to “ _vγδ ” and “ _eγδ ”. The dq-axis extended induced voltage is also described as “ _edq ”.
As for “θγ”, γ is described in normal characters instead of subscript characters.

Returning to FIG. 1, the motor control device 101 of this embodiment controls the energization of the motor 802 having two sets of three-phase windings by position sensorless control. Here, the configuration up to calculating the two-system γδ-axis voltage command values v ^* _γδ1 and v ^* _γδ2 is _assumed to be applicable to any known technique, and is not illustrated. For example, the specified γδ axis voltage values v ^* _γδ1 and v ^* _γδ2 may be calculated by current feedback control that matches the detected actual current with the _specified current value. Alternatively, it may be calculated by torque feedback control that matches the detected actual torque or the estimated torque estimated from the actual current with the torque command value.

The motor control device 101 includes a first two-phase / three-phase converter 171, three-phase two-phase converter 181, and a second two-phase / three-phase converter 172, three-phase two-phase converter 182.
Further, the motor control device 101 includes a voltage sum calculator 31, a current sum calculator 32, a voltage difference calculator 33, a current difference calculator 34, a system sum induced voltage estimation unit 41, a system difference induced voltage estimation unit 43, a sum difference It includes an adder 51, a sum-difference subtractor 52, a determining unit 53, a magnetic pole position estimating unit 56, and the like.
The portion I indicated by the dashed line is referred to in the description of the fourth embodiment.

  As shown by the broken lines, the magnetic pole position θγ estimated by the magnetic pole position estimating unit 56 is input to the first two-phase three-phase converter 171 and the three-phase two-phase converter 181. Here, the magnetic pole position θγ indicates the magnetic pole position of the first system according to the definition of the axis reference. On the other hand, a phase difference (see FIG. 3) of an electrical angle of 30 ° is added to the magnetic pole position of the second system with respect to the magnetic pole position of the first system. Therefore, the position obtained by adding the electrical angle of 30 ° to the magnetic pole position θγ by the angle adders 272 and 282 is input to the two-phase three-phase converter 172 and the three-phase two-phase converter 182 of the second system.

The two-phase / three-phase converters 171 and 172 and the three-phase / two-phase converters 181 and 182 of each system perform the above-described coordinate conversion operation using the input position information.
The two-phase / three-phase converters 171 and 172 of each system convert the γδ axis voltage command values v ^* _γδ1 and v ^* _γδ2 into three-phase voltage command values v ^* _UVW1 and v ^* _UVW2. Output.
The three-phase / two-phase converters 181 and 182 of each system acquire the phase currents i _UVW1 and i _UVW2 detected by the current sensors 71 and 72, and convert them into the detected γδ-axis current values i _γδ1 and i _γδ2 .
Hereinafter, `` γδ axis '' of `` γδ axis voltage command value v ^* _γδ1 , v ^* _γδ2 '' and `` γδ axis current detection value _iγδ1 , _iγδ2 '' is appropriately omitted except for the description regarding conversion with phase voltage or phase current. , “Voltage command values v ^* _γδ1 , v ^* _γδ2 ” and “current detection values _iγδ1 , _iγδ2 ”.

Voltage sum calculator 31 calculates the voltage sum v ^* _{the ??} w of the voltage command value _v ^* γδ2 between the voltage command value _v ^* γδ1 of the first system a second system, and outputs to the system sum the induced voltage estimation unit 41 .
Current sum calculator 32 includes a current detection value i _Ganmaderuta1 of the first system to calculate the sum of currents i _{the ??} w of the current detection value i _Ganmaderuta2 of the second system, and outputs to the system sum the induced voltage estimation unit 41.
Voltage difference calculator 33 calculates the voltage difference v ^* _{the ??} s from the voltage command value _v ^* γδ1 of the first system by subtracting the voltage command value _v ^* γδ2 of the second system, the output to the strain differences induced voltage estimation unit 43 I do.
The current difference calculator 34 calculates a current difference i _γδ s _obtained by subtracting the second system current detection value i _γδ2 from the first system current detection value i _γδ1 , and outputs the result to the system difference induced voltage estimating unit 43.

Here, in the present specification, “the sum of A and B” is used not only literally for “A + B” but also for “(A + B) / 2”, that is, “mean value of A and B”. Can be In addition, “the difference obtained by subtracting B from A” literally means not only the meaning of “AB”, but also “(AB) / 2”, that is, “half the value obtained by subtracting B from A divided by 2”. It is used in the meaning of "difference value."
Voltage sum calculator 31 and the current sum calculator 32 calculates a voltage sum v ^* _{the ??} w and the current sum i _{the ??} w means an average value, a voltage difference calculator 33 and a current difference calculator 34, the semi-difference value _Is calculated, and the voltage difference v ^* _γδs and the current difference _iγδs are calculated. Note that the suffix “w” means “sum”, and the suffix “s” means “difference”. “W” and “s” are written in ordinary characters in the specification, and are written as subscripts in figures and formulas.

In short, the voltage and current values handled by each block of the motor control device 101 are basically “values converted per system”. The operation of dividing by 2 for the conversion is merely an adjustment. Rather, in the present embodiment, the difference of whether to add or subtract the two systems of voltage and current values is more important.
Due to such technical ideas, in this specification, emphasis is placed on the distinction between “sum” and “difference”. Therefore, "sum" is essentially interpreted as a concept including an average value, and "difference" is interpreted as a concept including a half difference value.

The system sum induced voltage estimating unit 41 estimates the system sum induced voltage e _γδ w which is the sum of the extended induced voltages of the two systems based on the voltage sum v ^* _γδw and the current sum i _γδ w.
The system difference induced voltage estimating unit 43 estimates a system difference induced voltage e _γδ s, which is a difference between the extended induced voltages of the two systems, based on the voltage difference v ^* _γδs and the current difference i _γδ s.
The sum difference adder 51 adds the system sum induced voltage e _γδ w and the system difference induced voltage e _γδ s and outputs the result to the determination unit 53.
The sum difference subtracter 52 subtracts the system difference induced voltage e _γδ s from the system sum induced voltage e _γδ w and outputs the result to the determination unit 53.

The determining unit 53 directly obtains the system sum induced voltage e _γδ w from the system sum induced voltage estimating unit 41, and also adds the sum result (e _γδ w + e _γδ s) of the sum difference adder 51 and the sum difference subtracter 52. The subtraction result ( _eγδw - _eγδs ) is obtained. That is, the determination unit 53 acquires three types of calculated values regarding the extended induced voltage e _γδ .
Hereinafter, the system sum induced voltage e _γδ w will be simply referred to as “sum” as appropriate, for example, when it is obvious from the context. The addition result ( _eγδw + _eγδs ) of the sum-difference adder 51 is referred to as “sum + difference”, and the subtraction result ( _eγδwe - _eγδs ) of the sum-difference subtractor 52 is referred to as “sum−difference”. Write.

The judging unit 53 selects an optimum value for the magnetic pole position estimating unit 56 to use for estimating the magnetic pole position θγ among the three types of calculated values of “sum”, “sum + difference”, and “sum−difference”. , And “selected extended induced voltage e _γδ #” to the magnetic pole position estimating unit 56. The selection method will be described later.
Here, assuming that “sum” is a reference value in principle, the calculated value used by the magnetic pole position estimating unit 56 is the reference value “sum” when “sum + difference” or “sum−difference” is used. Can be understood to be executed by the “difference”. According to this concept, the “post-selection expansion induced voltage” may be rephrased as the “correction expanded induction voltage”.

The magnetic pole position estimating unit 56 estimates the magnetic pole position θγ and the electrical angular velocity ω based on the position estimation principle described later in detail, based on the selected expanded induced voltage e _γδ # acquired from the determining unit 53. That is, the “magnetic pole position estimating unit” functions correctly as a “magnetic pole position and speed estimating unit”. The detailed configuration of the magnetic pole position estimating unit 56 conforms to the configuration shown in FIG. 5 described below.
Hereinafter, the electrical angular velocity ω is appropriately referred to as “speed ω”. Further, since the motor rotation speed is obtained by multiplying the speed ω by the proportional constant, “ω” is also used in the meaning of the motor rotation speed.

Here, a general position estimation principle in the position sensorless control will be described with reference to FIGS.
FIG. 5 shows a control block diagram of the motor control device 100 that controls the driving of the three-phase one-system motor 801. Regarding the two-phase three-phase converter 171, the three-phase two-phase converter 181, the power converter 61, and the current sensor 71, reference numerals of the components of the first system in the motor control device 101 shown in FIG. 1 are used.

In addition, a current subtractor 14 and a current controller 15 are shown as configurations of the current feedback control. The phase current i _UVW detected by the current sensor 71 is converted into a γδ-axis current detection value i _γδ by the three-phase / two-phase conversion unit 181 and fed back to the _specified γδ-axis current value i ^* _γδ . The current controller 15 calculates the voltage command value v ^* _γδ by PI control or the like so that the difference between the current command value i ^* _γδ calculated by the current subtractor 14 and the current detection value _iγδ converges to 0. .

The voltage equation of the estimated rotation coordinate system is represented by Expression 1. The following symbols in Equation 1 are motor constants. Extended induced voltage estimation unit 40 obtains the voltage command value v ^* _{the ??} and the current detection value i _{the ??,} from Equation 1, calculates the extended electromotive force e _{the ??.}
R: resistance L _d , L _q : d-axis, q-axis self-inductance K _E : back electromotive force constant

また、図６のγδ軸ベクトル図に示すように、拡張誘起電圧ｅ_γδのδ軸成分ｅ_δの大きさを１としたときのγ軸成分ｅ_γの大きさの絶対値をベクトル比率Ｘと表すと、位置誤差Δθは、「Δθ＝ｔａｎ^-1｜Ｘ｜」と表すことができる。 The magnetic pole position estimating unit 56 includes an arc tangent calculating unit 57, a PI controller 58, and an integrator 59.
The arc tangent calculation unit 57 calculates the position error Δθ by Expression 2 based on “−sin △ θ, cos △ θ” in Expression 1.

6, the absolute value of the magnitude of the γ-axis component e _γ when the magnitude of the δ-axis component e _δ of the extended induced voltage e _γδ is set to 1, and the vector ratio X When expressed, the position error Δθ can be expressed as “Δθ = tan ⁻¹ | X |”.

積分器（図中、「１／ｓ」）５９は、ＰＩ制御器５８が演算した速度ωを積分し、磁極位置θγを演算する。 The PI controller 58 calculates a target value of the speed ω by PI control so that the position error Δθ becomes zero. When the position error Δθ is 0, the extended induced voltage e _γδ is represented by Expression 3. That is, when the position error Δθ is 0, the γ-axis component e _γ of the extended induced voltage e _γδ is “e _γ = 0”.

An integrator (“1 / s” in the figure) 59 integrates the speed ω calculated by the PI controller 58 and calculates the magnetic pole position θγ.

Incidentally, it is assumed that there is an error between the voltage command value v ^* _γδ and the actual voltage, an error between the current detection value _iγδ and the actual current, or an error in the motor constant. Then, as shown in Equation 4, extended electromotive force e _{the ??} expansion induced voltage estimation unit 40 has estimated, compared extended electromotive force e _dq real dq axes that containing the error .DELTA.e _{the ??.}

For example, assuming that the resistance R, which is a motor constant, includes an error ΔR, the error Δe _γδ of the extended induced voltage due to this is expressed by Expression 5.

As described above, even when the extended induced voltage estimated by the extended induced voltage estimation unit 40 includes the error Δe _{γδ, the operation} is controlled by the PI controller 58 so that the position error Δθ becomes zero. That is, since the speed ω and the magnetic pole position θγ are calculated so as to compensate for the estimated error Δe _γδ of the extended induced voltage, a position error Δθ corresponding thereto is generated.

Next, returning to FIG. 1, the position estimation in the two-system motor control device 101 will be described.
The voltage equations in the two systems include d-axis and q-axis mutual inductances M _d and M _q in consideration of inter-system interference as motor constants. Further, as a constant corresponding to the back electromotive force constant K _E in equation 1, the equation 6, using the value of the product of inductance M _f and the field current i _f the field circuit 87 shown in FIG. 2 (a) .

The voltage equation of the estimated rotational coordinate system of the first system and the extended induced voltage e _γδ1 are represented by Expression 6. The suffix “1” of each symbol indicates that it is a value of the first system.

The voltage equation of the estimated rotational coordinate system of the second system and the extended induced voltage e _γδ2 are expressed by Expression 7. The suffix “2” of each symbol indicates a value of the second system.

First, a description will be given of position estimation using only the information of the “sum” of the voltages and currents of the first system and the second system. The estimation configuration based on only the “sum” information is executed by the voltage sum calculator 31, the current sum calculator 32, and the system sum induced voltage estimation unit 41 in FIG. That is, the voltage difference calculator 33, the current difference calculator 34, the system difference induced voltage estimation unit 43, the sum difference adder 51, the sum difference subtracter 52, and the determination unit 53 in FIG. 1 are substantially excluded. Will be considered.
In other words, the configuration of the motor control device 109 of the comparative example shown in FIG. 16 corresponds to an estimation configuration using only the information of “sum”. The configuration of the comparative example shown in FIG. 16 basically corresponds to the configuration of Patent Document 1 (Japanese Patent No. 5527025), which is a conventional technique.

The “sum” (that is, the average value) of each of the voltage, current, and extended induced voltage of the first system and the second system is defined by Expression 8.

It is also assumed that the relationship of Equation 9 holds for the two systems of self-inductance.

The voltage equation using the “sum” of the voltages and currents of the two systems and the extended induced voltage e _γδ w are represented by Expression 10. In the motor control device 109 of the comparative example, the magnetic pole position estimating unit 56 performs PI control so that the position error Δθ estimated by Expression 10 is set to 0, and estimates the speed ω and the magnetic pole position θγ.

As described above in the example of the single-system motor control device, the estimated extended induced voltage e _γδ may have an error Δe _γδ based on errors in voltage, current, and motor constant. Further, in the two-system motor configuration, it is considered that errors Δe _γδ1 and Δe _γδ2 occur in the induced voltages e _γδ1 and e _γδ2 of the respective systems due to imbalance in the voltage, current, and motor constant between the systems. There can be various patterns in the relationship between the two systems of errors Δe _γδ1 and Δe _γδ2 .
Next, the configuration of “position estimation based on sum and difference information” according to the first embodiment will be described in comparison with the configuration of “position estimation based on sum information only” according to a comparative example.

First, the error of the extended induced voltage due to the imbalance between the systems in the two-system motor control device will be described using mathematical expressions. Hereinafter, the estimated value of the extended induced voltage is referred to as “estimated induced voltage”.
Estimated induced voltage e _Ganmaderuta2 the estimated induced voltage e _{Ganmaderuta1,} and a second system of the first system is represented by Equation 11.

Further, Equation 11 in the form of a matrix is simplified, and the estimated induced voltages on the γδ axis are collectively expressed as Equation 12.

The “sum” (that is, the average value) e _dq w of the dq-axis induced voltages of the two systems is defined as in Expression 13.

Also, as in Expression 14, it is assumed that the dq-axis induced voltages of the respective systems are equal.

The “sum” (that is, the average value) e _γδ w of the estimated induced voltages of the two systems is represented by Expression 15.

Equation 15 in the form of a matrix is simplified, and the sum of the estimated induced voltages on the γδ axis is collectively expressed as equation 16. A summary of the second term, _"Δe γδ w" refers to the "error of the sum" of the estimated induced voltage.

_Further , a difference (that is, a half difference value) e _γδ s _obtained by subtracting the estimated induced voltage e _γδ2 of the second system from the estimated induced voltage e _γδ1 of the first system is expressed by Expression 17.

Equation 17 in the form of a matrix is simplified, and the difference between the estimated induced voltages on the γδ axis is collectively expressed as equation 18.

  Next, refer to FIG. Here, three patterns are assumed as error patterns of the estimated induced voltages of the two systems. Hereinafter, the “error of the estimated induced voltage of each system” is simply referred to as “error of each system”. The “error of the sum of the estimated induced voltages of the two systems” is simply called “sum error”, and the “difference of the estimated induced voltages of the two systems” is simply called “difference”. The “sum error” is as shown in Expression 16, and the “difference” is as shown in Expression 18.

Pattern [A] is “when the error Δe _γδ1 of the first system is equal to a value obtained by inverting the sign of the error Δe _γδ2 of the second system”. In other words, the sign of the error .DELTA.e _Ganmaderuta2 the error .DELTA.e _Ganmaderuta1 of the first system a second system is reversed _{(Δe γδ1> 0, Δe γδ2} <0, _{or, Δe γδ1 <0, Δe γδ2} > 0) is, and the absolute values are equal _{(| Δe γδ1 | = | Δe} γδ2 |) is the case.
In this case, since the error .DELTA.e _Ganmaderuta1 of the first system and the error .DELTA.e _Ganmaderuta2 of the second system is canceled, the error .DELTA.e _{the ??} w of the sum is zero. Note that the difference e _γδ s is equal to the error Δe _γδ1 of the first system.

Pattern [B] is “when the error Δe _γδ1 of the first system is 0”. At this time, the sum error Δe _γδ w is equal to half of the error of the second system (= Δe _γδ2 / 2), and the difference e _γδ s is the sign of half the error of the second system. equal to the inverted value ₍₌ -Δe γδ2 / 2).
Therefore, the "error + difference of the sum _{(Δe γδ w + e γδ s} ) " is zero.
Pattern [C] is “when the error Δe _γδ2 of the second system is 0”. At this time, both the sum error Δe _γδ w and the difference e _γδ s are equal to one half of the error of the first system (= Δe _γδ1 / 2).
Therefore, the "error of the sum - difference _{(Δe γδ w-e γδ s} ) " is zero.

As described above, the calculated value that becomes 0 in each pattern is as follows.
Pattern [A]: Sum error Δe _γδ w
Pattern [B]: sum error + difference (Δe _γδ w + e _γδ s)
Pattern [C]: error of the sum - difference _{(Δe γδ w-e γδ s} )
In the configuration of “position estimation using only sum information” according to the comparative example, in the case of the pattern [A], the error Δe _γδ w of the estimated induced voltage is 0, and thus the position error Δθ can be reduced. However, the configuration of the comparative example cannot respond to the patterns [B] and [C], and may cause an increase in the position error Δθ.

Therefore, as described above in the description of the configuration, the motor control device 101 of the first embodiment calculates the “sum e _γδ w” by the system sum induced voltage estimation unit 41, and also calculates the system sum induced voltage estimation unit 43 The “difference e _γδ s” is calculated. The sum-difference adder 51 and the sum-difference subtractor 52 calculate “sum + difference ( _eγδw + _eγδs )” and “sum−difference ( _eγδw - _eγδs )”. Therefore, it is possible to use three types of calculation values of “sum _eγδw ”, “sum + difference ( _eγδw + _eγδs )”, and “sum−difference ( _eγδw − _eγδs )”. .

Then, the actual error pattern is classified into the closest pattern among [A], [B], and [C], and the magnetic pole position estimating unit 56 uses the minimum value among the three types of calculated values for each pattern. By estimating the magnetic pole position θγ, the position error Δθ can be reduced.
Further, in the present embodiment, the determination unit 53 selects the minimum value among the three types of operation values input each time the motor control device is operating, and sets the selected extended extension voltage e _γδ # as the magnetic pole position estimation unit. Output to 56.

Next, with reference to the flowcharts of FIGS. 8 to 10, three specific examples of the “operation value selection process” executed by the determination unit 53 will be described.
In the description of each flowchart, the symbol S represents “step”. 8 to 10, the common steps are denoted by the same step numbers. Further, steps which are not common but correspond to each other are described by adding branch numbers "-1, -2, -3" after the same step number.

FIG. 8 shows a flow of a selection process based on the magnitude of the extended induced voltage vector _eγδ .
In S1, the system sum induced voltage estimating unit 41 and the system difference induced voltage estimating unit 43 respectively _{perform a} system sum induced voltage e _γδ w (hereinafter “sum e _γδ w”) and a system difference induced voltage e _γδ s (hereinafter “difference”). e _γδ s ”).
At S2-1, sum and difference adder 51 adds the difference e _{the ??} s the sum e _{the ??} w, and calculates the "sum + difference _{(e γδ w + e γδ s} ) ". Sum and difference subtractor 52 subtracts the difference e _{the ??} s from the sum e _{the ??} w, - it calculates the "sum difference _{(e γδ w-e γδ s} ) ".

In S3-1 and S4-1, the determination unit 53 determines “sum _eγδw ”, “sum + difference ( _eγδw + _eγδs )”, and “sum−difference ( _eγδw − _eγδs )”. Select the smallest value.
First, in S3-1, the “sum _eγδw ” and the smaller value of “sum + difference ( _eγδw + _eγδs ) or sum−difference ( _eγδw − _eγδs )” are obtained. Be compared. If “sum e _γδ w” is small, YES is determined, and the routine shifts to S5.

In S <b> 3-1, if “the sum + difference ( _eγδw + _eγδs ) or the sum−difference ( _eγδw − _eγδs )” is small, it is determined to be NO. Then, at S4-1, the "sum + difference _{(e γδ w + e γδ s} ) " - and the "sum difference _{(e γδ w-e γδ s} ) " are compared.
When “sum + difference ( _eγδw + _eγδs )” is small, it is determined as YES, and the process shifts to S6.
If “sum−difference ( _eγδw − _eγδs )” is small, it is determined as NO, and the process shifts to S7.
The order of the determination is not limited to the above example, and the minimum value may be finally determined.

In S5, “sum e _γδ w” is selected as the selected extended induced voltage e _γδ #.
In S6, “sum + difference ( _eγδw + _eγδs )” is selected as the selected expanded induced voltage _eγδ #.
In S7, “sum−difference ( _eγδw − _eγδs )” is selected as the selected expanded induced voltage _eγδ #.
The selected extended induced voltage e _γδ # thus selected is output to the magnetic pole position estimating unit 56 and used for the arc tangent calculation of position estimation.

Figure 9 shows a flow of a selection process based on the estimated d-axis (gamma-axis) the magnitude of the component e _gamma expansion induced voltage. The extended induced voltage vector e _γδ is originally a vector in the estimated q-axis direction, and the error Δe _δ of the estimated q-axis component is considered to be at a relatively negligible level. The processing in FIG. 9 focuses on this point, and makes a determination using only the error Δe _γ of the estimated d-axis component.
In the steps (S2-2, S3-2, S4-2) with the branch number S "-2" in FIG. 9, "γδ" of the step with the branch number "-1" in FIG. ]. Other than that is the same as FIG. 8, and the description is omitted. The values shown in parentheses in S1, S5, S6, and S7 of FIG. 9 are used in the following second embodiment.
Since only the value of the estimated d-axis component is handled in the process of FIG. 9, the amount of calculation can be reduced as compared with the process of FIG.

FIG. 10 shows a flow of a selection process based on the δ-axis reference phase of the extended induced voltage vector _eγδ .
At S2-3, determination unit 53, based on the system sums the induced voltage e _{the ??} w and strain differences induced voltage e _{the ??} s, sum (sum + difference), (sum - difference) vector ratio Xw triplicate in , Xw + s, and Xw-s. As shown in FIG. 6, the vector ratio X is a value representing the magnitude of the γ-axis component when the magnitude of the δ-axis component of the induced voltage vector e _γδ is set to 1.

In S3-3 and S4-3, the determination unit 53 determines the arc tangent of each vector ratio Xw, Xw + s, Xw-s, tan ^-1 | Xw |, tan ^-1 | Xw + s |, tan ^-1 | Xw-s | Select the smallest value among In this step, the absolute value of the δ-axis reference phase is the minimum value, in other words, the value corresponding to the extended induced voltage vector e _γδ that most closely matches the δ-axis is selected. The concept of the minimum value selection and S5, S6, and S7 for determining the post-selection extended induced voltage e _γδ # are the same as those in FIG. 8, and thus the description thereof is omitted. Thus, it is possible to make a determination reflecting the information of the γ-axis component and the δ-axis component of the extended induced voltage vector e _γδ .

(effect)
The two-system motor control device of this embodiment includes a system difference induced voltage estimating unit 43 in addition to the system sum induced voltage estimating unit 41 in the configuration of the comparative example shown in FIG.
The magnetic pole position estimating unit 56 subtracts the system difference induced voltage from the system sum induced voltage e _γδ w, the value obtained by adding the system difference induced voltage to the system sum induced voltage (e _γδ w + e _γδ s), and the system sum induced voltage. based on one of the values the position error Δθ is selected so as to minimize among the values _{(e γδ w-e γδ s} ), and estimates the magnetic pole position θγ of the rotor. These three types of calculated values correspond to the estimated induced voltage error patterns [A], [B], and [C] shown in FIG. 7 and minimize the position error Δθ in each error pattern.

Therefore, the error pattern of the actual estimated induced voltage caused by the unbalance of the two systems is classified into the closest pattern among [A], [B], and [C] in FIG. The position error Δθ can be reduced by estimating the magnetic pole position θγ using the optimum calculation value for each time.
In addition, the determination unit 53 of the present embodiment selects the minimum value among the three types of calculation values that are input each time the motor control device is operating, and sets the selected extended induction voltage e _γδ # as the magnetic pole position estimation unit 56. Output to Thereby, even when the error pattern is not known or changes with time, the position error Δθ can be minimized for each period of the position estimation by the magnetic pole position estimating unit 56.

(2nd Embodiment)
A motor control device according to a second embodiment will be described with reference to FIG.
The control blocks constituting the motor control device 102 of the second embodiment are the same as those of the motor control device 101 of the first embodiment. However, the difference is that only the system sum induced voltage e _γ w of the estimated d-axis is output from the system sum induced voltage estimation unit 41 to the determination unit 53. The estimated q-axis system sum induced voltage e _δw is directly output to the magnetic pole position estimating unit 56 without being added to or subtracted from the system difference induced voltage e _δs .

The extended induced voltage vector e _γδ is originally a vector in the estimated q-axis direction, and the error Δe _δ of the estimated q-axis component is considered to be at a relatively negligible level. Thus, in the second embodiment, the system sum induced voltage e _γ w and the system difference induced voltage e _γ s are added or subtracted only for the estimated d-axis component.
As shown in parentheses in S1, S5, S6, and S7 of the flowchart of FIG. 9, the determination unit 53 of the second embodiment executes a calculation value selection process only on the estimated d-axis component, e _γ # may be output.
The second embodiment can reduce the calculation amount of the selection process compared to the first embodiment.

(Third embodiment)
A motor control device according to a third embodiment will be described with reference to FIG.
The motor control device 103 according to the third embodiment is different from the first embodiment in the configuration for calculating the system difference induced voltage e _γδ s. That is, the motor control device 103 includes the first system induced voltage estimation unit 45, the second system induced voltage estimation unit 46, and the system difference induced voltage estimation unit 47.

The first system induced voltage estimating unit 45 basically estimates the first system extended induced voltage _eγδ1 based on the first system voltage command value v ^* _γδ1 and the current detection value _iγδ1 .
The second system induced voltage estimating unit 46 basically estimates the extended induced voltage e _γδ2 of the second system based on the voltage command value v ^* _γδ2 of the second system and the detected current value _iγδ2 .
As described above, the first system induced voltage estimating unit 45 and the second system induced voltage estimating unit 46 estimate the extended induced voltage for each system using the voltage and current of each system.

As indicated by the broken line, the first-system voltage command value v ^* _γδ1 and the current detection value _iγδ1 are further input to the second-system induced voltage estimation unit 46, and the second-system voltage command value v ^* _γδ2 and current The detection value i _γδ2 may be further input to the first system induced voltage estimation unit 45. For example, when mutual interference between systems is considered, precise calculation including an interference term can be performed using the voltage and current values of the partner system.

Strain differences induced voltage estimation unit 47 from the extended induced voltage e _Ganmaderuta1 of the first system by subtracting the expanded induced voltage e _Ganmaderuta2 of the second system, to calculate the strain differences induced voltage e _{the ??} s. The system difference induced voltage e _γδ s is added to and subtracted from the system sum induced voltage error Δe _γδ w by the sum difference adder 51 and the sum difference subtracter 52, and is input to the determination unit 53.
Instead of or in addition to the system difference induced voltage estimating unit, the system sum induced voltage estimating unit adds the first system extended induced voltage e _γδ1 and the second system extended induced voltage e _γδ2 to _obtain the system sum induced voltage. e _γδ w may be calculated. With this configuration, the same operation and effect as those of the first embodiment can be obtained.

(Fourth embodiment)
A motor control device according to a fourth embodiment will be described with reference to FIGS.
The motor control device 104 according to the fourth embodiment employs the configuration shown in FIG. 13 instead of the configuration of the portion I in FIG. 1 showing the motor control device 101 according to the first embodiment. That is, in the motor control device 104 of the fourth embodiment, the system sum induced voltage estimation unit 42 and the system difference induced voltage estimation unit 44 are configured as filters that remove high frequency components in the input signal.

The system sum induced voltage estimating unit 42 receives “the value obtained by subtracting the product of the impedance Zw and the sum of the detected current values i _γδ w from the sum v ^* _γδ w of the voltage command values” as an input, and outputs the cutoff frequency f _co w The filter processing is performed using the transfer function Gw (s) of Expression 19 expressed by

The system difference induced voltage estimating unit 44 receives “the value obtained by subtracting the product of the impedance Zs and the sum of the detected current values i _γδ s from the voltage command value difference v ^* _γδ s” and inputs the cutoff frequency f _cos Filtering is performed using the transfer function Gs (s) of Expression 20 represented by

The speed ω calculated by the magnetic pole position estimator 56 is input to the system sum induced voltage estimator 42 and the system difference induced voltage estimator 44.
The impedances Zw and Zs are obtained by summing up the values of the determinants of Equations 6 and 7, and may be “Zw = Zs” depending on the precision of the arithmetic expression. However, when changing the precision of the motor constant considered in the sum calculation and the difference calculation, the actual calculation may be treated as “Zw ≠ Zs”.

Generally, in the control of a three-phase AC rotating machine, the three-phase current detection values i _UVW1 and i _UVW2 include harmonics of the (6k ± 1) order (k is a natural number) such as the fifth, seventh, eleventh, and thirteenth orders. It is known that the components overlap the primary components. The harmonic component is generated due to a non-uniform magnetization of the rotor magnet in the IPM motor, a structural factor due to the shapes of the rotor and the stator, a disturbance during the operation of the motor, and the like. The harmonic component causes a reduction in position estimation accuracy due to pulsation.
The (6k ± 1) -order component superimposed on the first-order component of the phase current becomes the (6k) -order component such as the 6th or 12th order of the γδ-axis current by the coordinate transformation.

FIG. 14A shows two systems of γ-axis currents i _{γ1 — sixth order} , i _{γ2 — sixth order} , and a sum (average value) thereof, in which the sixth order component is superimposed. It is assumed that the values of the zero-order component of the γ-axis current of the two systems and the amplitude of the sixth-order component of the γ-axis current are the same.
For the sixth-order component in which one cycle corresponds to an electrical angle of 60 °, the two systems of γ-axis currents i _{γ1_6} and i _{γ2_6 whose} phases are shifted by an electrical angle of 30 ° have opposite phases in which the peaks and valleys are inverted. . Therefore, when the current sum calculator 32 calculates the “sum” of the γ-axis currents i _{γ1_6} and i _{γ2_6 in} opposite phases, the sixth harmonic components are canceled each other and input to the system sum induced voltage estimation unit 42. .

On the other hand, as shown in FIG. 14 (b), the γ-axis currents i _{γ1_12} and i _{γ2_12 of} the two systems whose phases are shifted by an electrical angle of 30 ° are the same for the 12th-order component whose one cycle corresponds to an electrical angle of 30 °. Phase. Therefore, even when the “sum” is calculated by the current sum calculator 32, the twelfth harmonic component is not canceled and is input to the system sum induced voltage estimation unit 42.
Therefore, it is preferable that the cutoff frequency f _{ow of} the system sum induced voltage estimating unit 42 is set to a frequency that can remove at least the 12th harmonic component.

Conversely, regarding the “difference” of the current detection values input to the system difference induced voltage estimating unit 44, the 12th harmonic component having the same phase is canceled, and the 6th harmonic component having the opposite phase is not canceled. Therefore, it is preferable that the cutoff frequency f _{cos of} the system difference induced voltage estimating unit 44 is set to a frequency that can remove at least the sixth harmonic component.

For these reasons, the cutoff frequencies f _cow and f _cos of the system sum induced voltage estimation unit 42 and the system difference induced voltage estimation unit 44 according to the fourth embodiment are set as shown in FIG.
As shown in FIG. 15A, the cutoff frequencies f _cow and f _cos increase as the speed (motor rotation speed) ω increases in a region where the speed (motor rotation speed) ω is equal to or higher than the lower limit ω _LIM. Is set to
As described above, in the system sum induced voltage estimating unit 42 and the system difference induced voltage estimating unit 44 of the motor control device 104, the cutoff frequency of the filter is variably set according to the motor rotation speed ω. Thereby, the influence of the harmonic component can be appropriately reduced.

The frequency characteristic diagram of FIG. 15B shows an example in which the frequency at which the gain is -3 [dB] is set as the cutoff frequency. The cutoff frequency f _{cos of} the system difference induced voltage estimating unit 44 (difference filter) for removing the sixth harmonic component is different from the system sum induced voltage estimating unit 42 (for removing the twelfth harmonic component). The sum filter is set to a value lower than the cutoff frequency f _ow , for example, about a half.
As described above, in the system sum induced voltage estimating unit 42 and the system difference induced voltage estimating unit 44 of the motor control device 104, the cutoff frequencies f _cow and f _cos of the filters are set to different values. Thereby, the influence of the harmonic component can be appropriately reduced.

(Other embodiments)
(A) The determination unit 53 of the above-described embodiment determines “sum _eγδw ”, “sum + difference ( _eγδw + _eγδs )”, and “sum−difference ( _eγδw− )” each time the motor control device is operating. e _γδ s) ”, and outputs it to the magnetic pole position estimating unit 56 as the selected extended induced voltage e _γδ #.
However, for example, when it is known that the errors Δe _γδ1 and Δe _γδ2 of the extended induced voltage of the first system or the second system are always 0, the determination unit 53 is not provided and “sum + difference (e _γδ w + e _γδ s) )) Or “sum-difference ( _eγδw − _eγδs )” may be selected in advance.
Similarly, when it is known that the error Δe _γδ1 of the first system is always equal to the value obtained by inverting the sign of the error Δe _γδ2 of the second system, the determination unit 53 is not provided and “sum e _γδ w” is used. The position estimation may be selected in advance.

(B) The rotating machine to be controlled in the present invention is not limited to a three-phase rotating machine, but may be a four- or more-phase rotating machine.
The number of winding sets of the rotating machine, in other words, the number of systems of the motor control device may be three or more. However, position estimation according to the present invention is performed on two specific systems arbitrarily selected from three or more systems. A detailed method of fixing two specific systems among the three systems or selecting any two systems depending on the situation may be determined as appropriate.
As described above, the present invention is not limited to the above embodiment, and can be implemented in various forms without departing from the spirit of the invention.

101, 102, 103, 104: motor control device (control device for rotating machine)
41, 42 ... System sum induced voltage estimation unit,
43, 44, 47 ... system difference induced voltage estimation unit,
53 ... judgment unit,
56 magnetic pole position estimating unit
61, 62 ... power converter,
80 ... motor (rotary machine).

Claims (7)

A control device for a rotating machine that controls energization of the plurality of winding sets by a position sensorless control for a rotating machine (80) having a plurality of multi-phase winding sets of three or more phases in a stator (84),
A plurality of power converters (61, 62) for converting input power into AC power and supplying the power to the rotating machine;
If a unit of a group of constituent elements for controlling energization to one set of windings is defined as a system, a system sum induced voltage (e _γδ w) which is a sum of extended induced voltages of two specific systems selected from a plurality of systems. ), A system sum induced voltage estimating unit (41, 42),
A system difference induced voltage estimating unit (43, 44, 47) for estimating a system difference induced voltage (e _γδ s) which is a difference between the specific two systems of extended induced voltages;
The position error (Δθ) is minimized among the system sum induced voltage, a value obtained by adding the system difference induced voltage to the system sum induced voltage, and a value obtained by subtracting the system difference induced voltage from the system sum induced voltage. A magnetic pole position estimating unit (56) for estimating a magnetic pole position (θγ) of the rotor (85) of the rotating machine based on one of the values selected so that
A control device for a rotating machine comprising: Selecting the minimum value among the system sum induced voltage, a value obtained by adding the system difference induced voltage to the system sum induced voltage, and a value obtained by subtracting the system difference induced voltage from the system sum induced voltage, The control device for a rotating machine according to claim 1, further comprising: a determination unit (53) that outputs a post-selection expanded induced voltage ( _eγδ #) used for estimating the magnetic pole position by a position estimation unit. The system sum induced voltage estimating unit (41, 42) estimates the system sum induced voltage based on the voltage sum and the current sum of the specific two systems,
The control device for a rotating machine according to claim 1, wherein the system difference induced voltage estimating unit estimates the system difference induced voltage based on a voltage difference and a current difference between the specific two systems. The system further includes two system induced voltage estimation units (45, 46) for estimating an extended induced voltage for each system using the voltage and current of each system for the specific two systems,
At least one of the system sum induced voltage estimation unit and the system difference induced voltage estimation unit (47) calculates the sum or difference of the extended induced voltages estimated by the two system induced voltage estimation units. 3. The control device for a rotary machine according to 2.   5. The system sum induced voltage estimation section (42) and the system difference induced voltage estimation section (44) are configured as filters that remove high-frequency components in an input signal. 6. Rotary machine control device.   The rotating machine control device according to claim 5, wherein the system sum induced voltage estimating unit and the system difference induced voltage estimating unit variably set a cutoff frequency of a filter according to an electrical angular velocity of the rotating machine.   The control device for a rotating machine according to claim 5, wherein the system sum induced voltage estimating unit and the system difference induced voltage estimating unit have filter cutoff frequencies set to different values.

Images