JP2008236990A - Motor control apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a motor control apparatus capable of contributing to improvement of accuracy in motor control, by suitably setting the initial value of an M-serial signal to be overlapped to a motor-driving signal of two coordinate axes of two-phase coordinate system. <P>SOLUTION: The motor control apparatus 5 includes a current sensor 9 to detect a motor current passing through an electric motor 3, an indicating-voltage generator 13 for generating a directing voltage applied to the electric motor 3; a positional estimator 21 for calculating the rotational position of the electric motor 3 based on the motor current and a motor voltage; a parameter identifier 22 for obtaining a motor parameter used for estimating position; and an M-serial signal generator 23. The M-serial signal generator 23 generates a pair of M-serial signals that should be overlapped to a two-phase indicating voltage Vαβ. The initial values of the pair of M-serial signals are differentially set in the range between 1/4 cycle and 1/2 cycle for signal generation. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a motor control device for sensorless driving of a brushless motor. The brushless motor is used, for example, as a source for generating a steering assist force in an electric power steering apparatus.

  A motor control device for driving and controlling a brushless DC motor is generally configured to control the supply of motor current according to the output of a position sensor for detecting the rotational position of the rotor. However, the environmental resistance of the position sensor becomes a problem, and the expensive position sensor and the wiring associated therewith hinder the cost reduction and the size reduction. Therefore, a sensorless driving method for driving a brushless DC motor without using a position sensor has been proposed. The sensorless driving method is a method for estimating the phase of the magnetic pole (electrical angle of the rotor) by estimating the induced voltage accompanying the rotation of the rotor.

  For the estimation of the induced voltage, an inverse motor model (mathematical model) that is an inverse model of the motor model is used. The inverse motor model includes inductance and armature winding resistance which are motor parameters. These motor parameters also vary during motor drive due to winding temperature, current and other variable factors. For this reason, it has been proposed to identify motor parameters based on current and voltage information when the motor is driven sensorlessly, and to apply the identified motor parameters to an inverse motor model (Non-Patent Document 1 below). .

For system identification, a technique using a white input signal that satisfies PE (persistently exciting) properties is common. Therefore, it is possible to generate a whiteness signal from the white signal generation source, superimpose it on the motor drive signal, and identify the motor parameter based on the current and voltage information at that time.
A typical white signal generator is an M-sequence signal generator. The M-sequence (Maximum-length linear shift resister sequence) is a random number sequence in which “0” and “1” are randomly arranged. A pseudo white binary signal (pseudo white noise signal) can be generated from the M-sequence signal. The M-sequence signal generator includes a shift register that delays and shifts a signal by one sample time, and an exclusive OR circuit. Any initial value other than 0 can be given to the shift register.

A pair of such M-sequence signal generators is used to generate, for example, an M-sequence signal to be applied to the α-axis and an M-sequence signal to be applied to the β-axis of the two-phase fixed coordinate system (α-β). What is necessary is just to superimpose on the motor drive signal of a beta axis.
Japanese Patent Laying-Open No. 2005-151714 (paragraph [0090]) Akihisa Iwata et al., “Sensorless control of synchronous reluctance motors using on-line parameter identification method independent of position estimation accuracy”, IEEJ Transactions D, 124, 12, 2004

  However, a method for determining an appropriate initial value of the M-sequence signal to be superimposed on the α-axis and β-axis motor drive signals is not known. For example, if the initial values of the M-sequence signals of the α axis and the β axis are the same, the motor parameters cannot be correctly identified, and therefore the rotor position cannot be correctly estimated. Even when the initial values of the M-sequence signals input to the α-axis and the β-axis are different, the parameter identification accuracy deteriorates if the initial value deviation is small, and the position estimation accuracy deteriorates. However, there is no known guideline that suggests how much the initial value is shifted to obtain the required identification accuracy.

  Accordingly, an object of the present invention is to provide a motor control that can contribute to an improvement in motor control accuracy by appropriately determining an initial value of an M-sequence signal to be superimposed on a motor drive signal of two coordinate axes in a two-phase coordinate system. Is to provide a device.

  In order to achieve the above object, the invention according to claim 1 is an M-sequence signal generating means for generating an M-sequence signal to be added to a motor drive signal of two coordinate axes of a two-phase fixed coordinate system or a two-phase rotary coordinate system. (61, 62) and an initial value for setting the initial value of the M-sequence signal generated by the M-sequence signal generating means to be different between the two coordinate axes in the range of ¼ cycle to ½ cycle. A motor control device (5) including setting means (60). The alphanumeric characters in parentheses indicate corresponding components in the embodiments described later. The same applies hereinafter.

  The two-phase fixed coordinate system is a two-dimensional coordinate system that defines two fixed coordinate axes (coordinate axes fixed without rotation of the rotor) orthogonal to the rotor rotation plane, with the rotation center of the rotor as the origin. For example, the αβ coordinate system is applicable. The two-phase rotational coordinate system refers to two rotational coordinate axes (coordinate axes that rotate with the rotor. In other words, coordinate axes that are stationary relative to the rotor with the rotation center of the rotor as the origin. .) Is a two-dimensional coordinate system. For example, in addition to the dq coordinate system, the γδ coordinate system, which is an estimated coordinate system for sensorless control, is also a two-phase rotating coordinate system.

  The case where the initial values of the M-sequence signals are different from each other by 1/2 cycle (half cycle) is the case where the distance between the two M-sequence signals input to the two coordinate axes is the largest, and sufficient parameter identification accuracy can be ensured. On the other hand, even if it is difficult to vary the initial value by 1/2 period due to system limitations, the required parameter identification accuracy can be improved by setting the initial value of the M-sequence signal to be different by at least 1/4 period. It can be secured. As described above, since the required parameter identification accuracy can be ensured by making the initial value of the M-sequence signal different from 1/4 cycle to 1/2 cycle, the motor control accuracy can be improved. In addition, since the initial value of the M-sequence signal can be easily determined, the design man-hour for the motor control device can be shortened.

  According to a second aspect of the present invention, there is provided a parameter identifying means for identifying a motor parameter based on a response of a motor when a motor drive signal to which the M series signal generated by the M series signal generating means is added is applied to the motor. The motor control device according to claim 1, further comprising: 22) and position estimating means (21) for estimating the rotational position of the motor using the motor parameter identified by the parameter identifying means.

  With this configuration, the motor parameter is identified based on the response when the motor drive signal on which the M-sequence signal is superimposed is applied to the motor, and the rotational position of the motor is estimated using the identified motor parameter. Since the initial values of the M series signals respectively input to the two-phase coordinate system are appropriately set, the motor parameters can be identified with high accuracy, and therefore the rotational position of the motor can be estimated with good accuracy. By using this estimated motor rotation position, the motor can be controlled with high accuracy.

The position estimation unit may estimate the rotational position of the motor using a motor current and a motor voltage and an inverse motor model including the motor parameter identified by the parameter identification unit.
The M-sequence signal generating means may generate the M-sequence signal at all times during startup of the motor control device, but generates an M-sequence signal by setting an initial value when identifying the motor parameter. Also good. For example, when motor parameters are identified every predetermined control cycle, an initial value is set in each control cycle, and an M-sequence signal is generated based on the set initial value. Good.

  The invention according to claim 3 further includes signal processing means (24) for performing low-pass filter processing for removing high-frequency components at a predetermined cutoff frequency for the M-sequence signal generated by the M-sequence signal generating means. The motor control device according to claim 1, wherein an M-sequence signal that has been low-pass filtered by the signal processing means is added to a motor drive signal.

The M-sequence signal has a relatively flat frequency characteristic over a wide frequency band. By superimposing such a signal on the motor drive signal, the high frequency characteristics of the motor may be excited. For example, motors for electric power steering devices tend to have characteristics in the high frequency range that hinder identification of motor parameters and estimation of rotational positions.
Therefore, by applying low-pass filter processing to the M-sequence signal and superimposing the filtered M-sequence signal on the motor drive signal, excitation of high-frequency characteristics of the motor can be suppressed or prevented. Thereby, the identification accuracy of the motor parameter is increased and the motor rotational position can be estimated with high accuracy, so that the motor control accuracy can be improved.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
FIG. 1 is a block diagram for explaining an electrical configuration of an electric power steering apparatus to which a motor control apparatus according to an embodiment of the present invention is applied. The electric power steering apparatus includes a torque sensor 1 that detects a steering torque applied to a steering wheel of a vehicle, an electric motor 3 that applies a steering assist force to a steering mechanism 2 of the vehicle, and a motor control that drives and controls the electric motor 3. And a device 5. The motor control device 5 drives the electric motor 3 in accordance with the steering torque detected by the torque sensor 1, thereby realizing appropriate steering assistance according to the steering situation.

  In this embodiment, the electric motor 3 is a three-phase brushless DC motor. As shown schematically in FIG. 2, the electric motor 3 has a rotor 50 as a field, and stator windings 51 of U phase, V phase and W phase, 52, 53. The electric motor 3 may be an inner rotor type in which a stator is disposed outside the rotor, or may be an outer rotor type in which a stator is disposed inside a cylindrical rotor.

The motor control device 5 includes a microcomputer 7, a drive circuit (inverter circuit) 8 that is controlled by the microcomputer 7 and supplies electric power to the electric motor 3, and currents that flow through the stator windings of each phase of the electric motor 3. And a current sensor 9 for detection.
The microcomputer 7 includes a CPU and a memory (such as a ROM and a RAM), and functions as a plurality of function processing units by executing a predetermined program. The plurality of function processing units include a current command value generation unit 11, a PI (proportional integration) control unit 12, an instruction voltage generation unit 13, a γδ / αβ coordinate conversion unit 14, and an αβ / UVW coordinate conversion unit 15. A PWM control unit 16, a UVW / αβ coordinate conversion unit 17, an αβ / γδ coordinate conversion unit 18, a deviation calculation unit 19, a signal processing unit 20, a position estimation unit 21, a parameter identification unit 22, An M-sequence signal generator 23, a low-pass filter 24, and an adder 28 are provided.

The current command value generation unit 11 generates a command value I _d ^* of a d-axis current component along the rotor magnetic pole direction of the electric motor 3 and a command value I _q ^* of a q-axis current component orthogonal to the d axis. Hereinafter, these are collectively referred to as “current command value I _dq ”. However, the dq coordinate plane is a plane along the rotation direction of the rotor 50, and the d axis and the q axis define a two-phase rotation coordinate system (dq) that rotates together with the rotor 50 (see FIG. 2).

When the current command value I ^* representing the amplitude of the current (sine wave current) to be applied to the U phase, V phase and W phase of the electric motor 3 is used, the d axis current command value I _d ^* and the q axis current command value I _q ^* Is expressed by the following equations (1) and (2).

したがって、電流指令値生成部１１は、ｄ軸電流指令値Ｉ_d ^*＝０を生成する一方で、トルクセンサ１によって検出される操舵トルクに応じたｑ軸電流指令値Ｉ_q ^*を生成する。より具体的には、操舵トルクに対応したｑ軸電流指令値Ｉ_q ^*を記憶したマップ（テーブル）を用いてｑ軸電流指令値Ｉ_q ^*が生成されるようになっていてもよい。電動モータ３が発生するトルクは、モータ電流に対応するから、電流指令値Ｉ_dqは、電動モータ３から発生させるべきトルクを指令するための「トルク指令値」と言い換えることもできる。

Therefore, the current command value generation unit 11 generates the d-axis current command value I _d ^* = 0, while generating the q-axis current command value I _q ^* corresponding to the steering torque detected by the torque sensor 1. More specifically, the q-axis current command value I _q ^* may be generated using a map (table) that stores the q-axis current command value I _q ^* corresponding to the steering torque. Since the torque generated by the electric motor 3 corresponds to the motor current, the current command value I _dq can be rephrased as a “torque command value” for commanding the torque to be generated from the electric motor 3.

The current sensor 9 detects the U-phase current I _U , the V-phase current I _V and the W-phase current I _w of the electric motor 3 (hereinafter, collectively referred to as “three-phase detection current I _UVW ”). The detected value is given to the UVW / αβ coordinate converter 17.
The UVW / αβ coordinate converter 17 converts the three-phase detection current I _UVW into the currents I _α and I _β on the two-phase fixed coordinate system (α-β) (hereinafter referred to as “two-phase detection current I _The coordinates are converted to “ _αβ ”. The two-phase fixed coordinate system (α-β) is a fixed coordinate system in which the rotation center of the rotor 50 is the origin and the α axis and the β axis orthogonal to the rotation axis of the rotor 50 are defined (see FIG. 2). ). The coordinate-converted two-phase detection current I _αβ is given to the αβ / γδ coordinate conversion unit 18.

The αβ / γδ coordinate conversion unit 18 converts the two-phase detection current I _αβ into a two-phase rotation coordinate system (hereinafter referred to as “estimated rotation position θ ^”) estimated by the position estimation unit 21 (hereinafter referred to as “estimated rotation position θ ^”). Coordinates are converted into currents I _γ and I _δ on γ−δ) (hereinafter, collectively referred to as “two-phase detection current I _γδ ”). The two-phase rotational coordinate system (γ−δ) is a rotational coordinate system defined by a γ axis along the rotor magnetic pole direction and a δ axis orthogonal to the γ axis when the rotor 50 is at the estimated rotational position θ ^. (See FIG. 2). When the estimated rotational position θ ^ has no error and coincides with the actual rotor rotational position, the two-phase rotational coordinate system (dq) and the two-phase rotational coordinate system (γ-δ) coincide.

The two-phase detection current I _γδ is supplied to the deviation calculation unit 19. The deviation calculator 19 calculates a deviation of the γ-axis current I _{γ with} respect to the d-axis current command value I _d ^* and a deviation of the δ-axis current I _{δ with} respect to the q-axis current command value I _q ^* . These deviations are given to the PI control unit 12 and are each subjected to PI calculation processing. Then, in accordance with these calculation results, the command voltage generator 13 causes the γ-axis command voltage V _γ ^* and the δ-axis command voltage V _δ ^* (hereinafter referred to as “two-phase command voltage V _γδ ” when collectively referred to). ) Is generated and provided to the γδ / αβ coordinate converter 14.

The γδ / αβ coordinate converter 14 converts the γ-axis command voltage V _γ ^* and the δ-axis command voltage V _δ ^* into α-axis command voltages V _α ^* and β that are command voltages of the two-phase fixed coordinate system (α-β). Coordinates are converted to an axis command voltage V _β ^* (hereinafter, collectively referred to as “two-phase command voltage V _αβ ”). The two-phase instruction voltage V _αβ is given to the αβ / UVW coordinate conversion unit 15.
The αβ / UVW coordinate converter 15 converts the α-axis command voltage V _α ^* and the β-axis command voltage V _β ^* into the command voltages of the three-phase fixed coordinate system, that is, the command voltages V _U ^{* of the} U phase, V phase, and W phase ^. , V _V ^* , V _W ^* (hereinafter collectively referred to as “three-phase indicating voltage V _UVW ”).

The PWM control unit 16 generates a drive signal having a duty ratio controlled according to the three-phase instruction voltages V _U ^* , V _V ^* , and V _W ^* and supplies the drive signal to the drive circuit 8. Thus, a voltage is applied to each phase of the electric motor 3 with a duty ratio corresponding to the instruction voltages V _U ^* , V _V ^* , and V _W ^* of the corresponding phase.
With such a configuration, when a steering torque is applied to a steering wheel (not shown) as an operation member coupled to the steering mechanism 2, this is detected by the torque sensor 1. Then, the current command value generation unit 11 generates a current command value I _dq corresponding to the detected steering torque. A deviation between the current command value I _dq and the two-phase detection current I _γδ is obtained by the deviation calculation unit 19, and PI calculation is performed by the PI control unit 12 so as to lead this deviation to zero. A two-phase command voltage V _γδ corresponding to the calculation result is generated by the command voltage generation unit 13, and this is converted into a three-phase command voltage V _UVW through the coordinate conversion units 14 and 15. Then, the drive circuit 8 operates at a duty ratio corresponding to the three-phase instruction voltage V _UVW by the action of the PWM control unit 16, whereby the electric motor 3 is driven, and an assist torque corresponding to the current command value I _dq is generated. The steering mechanism 2 will be given. Thus, steering assistance can be performed according to the steering torque. The three-phase detection current I _UVW detected by the current sensor 9 passes through the coordinate conversion units 17 and 18 and is represented by the two-phase rotational coordinate system (γ−δ) so as to correspond to the current command value I _dq. After being converted to the detected current I _γδ , the deviation is given to the deviation calculating unit 19.

In order to perform coordinate conversion between the rotating coordinate system and the fixed coordinate system, a phase angle (electrical angle) θ representing the rotational position of the rotor 50 is required. An estimated rotational position θ ^ representing this phase angle is generated by the position estimation unit 21 and is given to the γδ / αβ coordinate conversion unit 14 and the αβ / γδ coordinate conversion unit 18.
The position estimation unit 21 is based on an inverse model (inverse motor model) of a motor model that is a mathematical model of the electric motor 3, and a disturbance observer 25 that estimates an induced voltage of the electric motor 3 as a disturbance, and the disturbance observer 25 outputs the disturbance observer 25. An estimated rotational position of the rotor 50 based on an estimated value filter 26 configured by a low-pass filter that removes high-frequency components from the estimated induced voltage and an estimated induced voltage (value after filtering) output from the estimated value filter 26. and an estimated position generation unit 27 for generating θ ^.

The position estimation unit 21 is supplied with the two-phase command voltage V _αβ and the two-phase detection current I _αβ after the signal processing (filtering) by the signal processing unit 20. The signal processing unit 20 includes a voltage filter 31 configured with a low-pass filter that removes a high-frequency component of the two-phase instruction voltage V _αβ and a low-pass filter that removes a high-frequency component of the two-phase detection current I _αβ. Current filter 32. Therefore, the two-phase indication voltage V _αβ filtered by the voltage filter 31 and the two-phase detection current I _αβ filtered by the current filter 32 are input to the disturbance observer 25 of the position estimation unit 21. .

The parameter identification unit 22 calculates a motor parameter used for the reverse motor model of the disturbance observer 25 based on the two-phase detection current I _γδ and the two-phase command voltage V _γδ . In this embodiment, the motor parameters include the armature winding resistance R and the αβ axis inductance L.
The M sequence signal generation unit 23 generates an M sequence signal as a sensing signal to be superimposed on the two-phase instruction voltage V _αβ . The M-sequence signal is superposed on the two-phase instruction voltage V _αβ by the adding unit 28 after high frequency components (for example, components of 2000 Hz or more) are removed by the low-pass filter 24.

FIG. 3 is a block diagram for explaining a configuration example of the M-sequence signal generation unit 23. The M-sequence signal generator 23 includes an initial value setting unit 60, an α-axis M-sequence signal generator 61, and a β-axis M-sequence signal generator 62. Each of the M series signal generators 61 and 62 includes a shift register 65 and an exclusive OR operation unit (EXOR) 66.
The shift register 65 includes a plurality of stages (six stages in this embodiment) of registers 67, and shifts held data to the next stage register 67 every one sample time (for example, 0.001 second). Therefore, each register 67 functions as an operator Z ⁻¹ representing a delay of one sample time.

The first-stage register 67 is supplied with the output of the exclusive OR operation unit 66. The exclusive OR calculator 66 calculates the exclusive OR of the output of the first-stage register 67 and the output of the final-stage register 67 and inputs the result to the first-stage register 67.
The initial value setting unit 60 stores initial values x (0) and x (x (0)) in a plurality of stages of registers 67 constituting the shift registers 65 of the α-axis M-sequence signal generator 61 and the β-axis M-sequence signal generator 62, respectively. 1),..., X (k) (k = 5 in the example of FIG. 3) is set. However, different initial values x (0), x (1),..., X (n) are set in the α-axis M-sequence signal generator 61 and the β-axis M-sequence signal generator 62. In the following, the initial values x (0), x (1),..., X (n) set in the shift register 65 of the α-axis M-sequence signal generator 61 are referred to as “α-axis initial value x _α ”, and β The initial values x (0), x (1),..., X (n) set in the shift register 65 of the axis M series signal generator 62 are referred to as “β-axis initial value x _β ”.

With this configuration, alpha axis M-sequence signal generator 61 generates a M-sequence signal starting from alpha axial initial value x _alpha (alpha axis M-sequence signal), the beta axis M-sequence signal generator 62 beta axial initial An M-sequence signal (β-axis M-sequence signal) starting from the value x _β is generated. Then, alpha axis M-sequence signal is superimposed on the alpha -axis command voltage V _alpha, beta-axis M-sequence signal is superimposed on the beta-axis command voltage V _beta.
FIG. 4 is a diagram illustrating a sixth-order M-sequence signal. 6-bit values of “first” to “63th” are sequentially held in the shift register 65, and with this as one cycle, the hold value of the shift register 65 changes cyclically. Then, the final value (binary value) is output as an M-sequence signal.

The initial value setting unit 60 sets the α-axis initial value x _α and the β-axis initial value x _β to values separated by ¼ or more of one cycle of the M-sequence signal. Specifically, when the α-axis initial value x _α is set to the “first” value of the sixth-order M-sequence signal, the β-axis initial value x _β is “16th to the sixth-order M-sequence signal”. It is set to a value within the “48th” range. At maximum, two values separated by a half cycle can be set as an α-axis initial value x _α and a β-axis initial value x _β , respectively.

If the dimension of the M-sequence signal is n-th order (n = k + 1), one period of the M-sequence signal is 2 ⁿ −1 (63 in the case of the 6th order). When the α-axis initial value x _α and the β-axis initial value x _β are set to values separated by a half cycle, the “i-th” of the M-sequence signal (i is a natural number of 1 to 2 ⁿ −1) ) Is the α-axis initial value x _α , and the “i + 2 ⁿ⁻¹ −1” or “i + 2 ⁿ⁻¹ ” is the β-axis initial value x _β . For example, the value of "first" in the M-sequence signal and alpha axial initial value x _alpha, the value of "2 ^n-1 th" or "2 ^n-1 +1 th" may be a beta axis initial value x _beta . Of course, there are other combinations of initial values separated by a half cycle.

In general, in order to set two values separated by more than a quarter cycle of the M-sequence signal as the α-axis initial value x _α and the β-axis initial value x _β , The “i-th” value of the signal is the α-axis initial value x _α, and the value in the range of “i + 2 ⁿ⁻² −1” to “i + 3 · 2 ⁿ⁻² −1” is the β-axis initial value x _β . do it. Specifically, for example, the “first” value of the M-sequence signal is set to the α-axis initial value x _α, and the value in the range of “2 ⁿ⁻² ” to “3 · 2 ⁿ⁻² ” is set to the β axis. The initial value x _β may be set.

FIG. 5 shows an α axis generated when two values separated by a half period are set as an α axis initial value x _α and a β axis initial value x _β in a 6th-order M-sequence signal, respectively. It is a figure which shows the waveform of an M series signal and a beta-axis M series signal. Since the correlation of the M-sequence signal between the two axes of the α-axis and the β-axis is minimized, by using such an α-axis M-sequence signal and β-axis M-sequence signal, sufficient parameter identification accuracy can be obtained. . Further, by setting two values separated by at least a quarter cycle of the M-sequence signal as an α-axis initial value x _α and a β-axis initial value x _β , respectively, the α-axis M-sequence signal and the β-axis M Since the correlation of the sequence signal can be made sufficiently small, necessary parameter identification accuracy can be obtained.

FIG. 6 is a block diagram for explaining an example of the disturbance observer 25 and a configuration related thereto. A motor model that is a mathematical model of the electric motor 3 can be expressed as, for example, (R + pL) ⁻¹ . However, as described above, R is the armature winding resistance, and L is the αβ axis inductance. P is a differential operator. It can be considered that the two-phase command voltage V _αβ and the induced voltage E _αβ (α-axis induced voltage E _α and β-axis induced voltage E _β ) are applied to the electric motor 3.

The disturbance observer 25 receives the two-phase detection current I _αβ as an input and estimates an inverse motor model (inverse model of the motor model) 35 for estimating the motor voltage, and the motor voltage estimated by the inverse motor model 35 and the two-phase indication voltage V _αβ. And a voltage deviation calculation unit 36 for obtaining the deviation. The voltage deviation calculation unit 36 obtains a disturbance with respect to the two-phase indication voltage V _αβ , and as is apparent from FIG. 6, this disturbance is an α-axis induced voltage estimated value E ^ _α and β corresponding to the induced voltage E _αβ. The shaft induced voltage estimated value E ^ _β is obtained. Hereinafter, these are collectively referred to as “estimated induced voltage E ^ _αβ ”.

The reverse motor model 35 is represented by R + pL, for example. The armature winding resistance R and the αβ-axis inductance L constituting the reverse motor model are obtained by the parameter identification unit 22, and the reverse motor model 35 is created using them.
The estimated value filter 26 can be configured by, for example, a low-pass filter represented by a / (s + a). a is a design parameter, and the cutoff frequency of the estimated value filter 26 can be set to an appropriate value by appropriately determining the design parameter a.

The induced voltage E _αβ can be expressed by the following equation (3). However, K _E is an induced voltage constant, θ is a rotor rotational position, and ω is a rotor rotational speed.

したがって、推定誘起電圧Ｅ^_αβが求まれば、次の(4)式に従って、ロータ５０の推定回転位置θ^が求まる。この演算が、推定位置生成部２７によって行われるようになっている。

Therefore, when the estimated induced voltage E ^ _αβ is obtained, the estimated rotational position θ ^ of the rotor 50 is obtained according to the following equation (4). This calculation is performed by the estimated position generation unit 27.

図７は、マイクロコンピュータ７によるロータ回転位置推定演算の流れを説明するためのフローチャートであり、主として、信号処理部２０、位置推定部２１、パラメータ同定部２２、Ｍ系列信号生成部２３およびローパスフィルタ２４によって所定の制御周期毎に繰り返し実行される処理の流れが示されている。

FIG. 7 is a flowchart for explaining the flow of the rotor rotational position estimation calculation by the microcomputer 7. Mainly, the signal processing unit 20, the position estimation unit 21, the parameter identification unit 22, the M-sequence signal generation unit 23, and the low-pass filter Reference numeral 24 denotes a flow of processing that is repeatedly executed every predetermined control cycle.

The initial value setting unit 60 sets an α-axis initial value x _α and a β-axis initial value x _β that are separated by a quarter or more of the period of the M-sequence signal to an α-axis M-sequence signal generator 61 and a β-axis M-sequence signal generator 62. (Step S1). After the initial value is set, the M-sequence signal generators 61 and 62 are operated to generate an α-axis M-sequence signal and a β-axis M-sequence signal from the M-sequence signal generator 23 (step S2).

  The low-pass filter 24 performs signal processing (filter processing) for removing high frequency components of the generated α-axis M-sequence signal and β-axis M-sequence signal (step S3). Although the M-sequence signal has a relatively flat frequency characteristic over a wide frequency band, the high frequency characteristic of the electric motor 3 may be excited by superimposing such a signal on the motor drive signal. . Therefore, by applying low-pass filter processing to the M-sequence signal and superimposing the filtered M-sequence signal on the motor drive signal, excitation of high-frequency characteristics of the electric motor 3 can be suppressed or prevented. Thereby, the identification accuracy of a motor parameter can be improved.

The α-axis M-sequence signal and β-axis M-sequence signal processed by the low-pass filter 24 are respectively superimposed on the α-axis command voltage V _α ^* and β-axis command voltage V _β ^* generated by the γδ / αβ coordinate converter 14. (Step S4).
On the other hand, the parameter identification unit 22 identifies the motor parameter based on the two-phase detection current I _γδ generated by the αβ / γδ coordinate conversion unit 18 and the two-phase command voltage V _γδ generated by the command voltage generation unit 13. (Step S5).

The two-phase indication voltage V _αβ calculated by the γδ / αβ coordinate conversion unit 14 is input to the voltage filter 31, and the two-phase detection current I _αβ calculated by the UVW / αβ coordinate conversion unit 17 is the current filter 32. (Step S6). In the voltage filter 31 and the current filter 32, low-pass filter processing (high-frequency component removal processing) at a predetermined cutoff frequency is performed on the input voltage signal and current signal.

The disturbance observer 25 obtains the estimated induced voltage E ^ _αβ using the voltage signal and current signal after the low-pass filter processing (step S7). At this time, the inverse motor model constituting the disturbance observer 25 is constructed using the motor parameters identified by the parameter identification unit 22. The estimated value filter 26 performs filtering on the estimated induced voltage E ^ _αβ obtained by the disturbance observer 25.

Using the estimated induced voltage E ^ _αβ filtered in this way, the estimated position generation unit 27 performs position estimation to obtain the estimated rotational position θ ^ (step S8).
As described above, according to this embodiment, have at least 1/4 period by the difference between the initial value x _beta of the initial value x _alpha of alpha axis M-sequence signal generator 61 beta axis M-sequence signal generator 62 Thus, the parameter identification unit 22 can identify the motor parameters with sufficient accuracy. Therefore, the estimated induced voltage E ^ _αβ obtained by the disturbance observer 25 can have sufficient accuracy, and as a result, the estimated rotational position θ ^ with high accuracy can be obtained. As a result, the coordinate conversion calculation in the γδ / αβ coordinate conversion unit 14 and the αβ / γδ coordinate conversion unit 18 is accurately performed, and as a result, the electric motor 3 can be sensorlessly controlled with high accuracy.

As mentioned above, although one Embodiment of this invention was described, this invention can also be implemented with another form.
For example, in the above-described embodiment, the position estimation unit 21 constructs the inverse motor model 35 using the parameters identified by the parameter identification unit 22 and estimates by using the disturbance observer 25 using the inverse motor model 35. The rotational position θ ^ is obtained. However, as the position estimation unit 21, a configuration that directly calculates the estimated rotational position θ ^ using the identified parameters can be applied.

More specifically, it is known that the α-axis inductance L _α , the β-axis inductance L _β , and the αβ-axis inductance L _αβ depend on the rotational position θ of the rotor and are expressed as follows.
L _α = L ₀ + L ₁ cos 2θ
L _β = L ₀ −L ₁ cos 2θ
L _αβ = L ₁ sin2θ
However, L ₀ and L ₁ are inductance components and are expressed as follows using the d-axis inductance L _d and the q-axis inductance L _q .

L ₀ = (L _d + L _q ) / 2
L ₁ = (L _d −L _q ) / 2
Therefore, the α-axis inductance L _α , β-axis inductance L _β , and αβ-axis inductance L _αβ are identified by the parameter identification unit 22, and using these, the position estimation unit 21 uses the estimated rotational position θ ^ based on the following equation: You can ask for. In this case, since the current and voltage information is not used for the calculation in the position estimation unit 21, it is not necessary to provide the signal processing unit 20.

この手法は、ロータ回転速度が低い場合のように、誘起電圧が十分に得られない場合に有効な方法である。したがって、この手法のための構成と前述の図１に示された構成との両方を位置推定部２１に備え、十分な誘起電圧が得られる状況かどうか（たとえば、ロータ回転速度が所定の閾値以上かどうか）に応じて、前記式(5)により推定回転位置θ^を用いる状態と、外乱オブザーバ２５等を用いて前記式(4)により推定回転位置θ^を用いる状態とを切り換えるようにしてもよい。

This method is effective when the induced voltage cannot be obtained sufficiently, such as when the rotor rotational speed is low. Therefore, the position estimation unit 21 includes both the configuration for this technique and the configuration shown in FIG. 1 described above, and whether or not a sufficient induced voltage is obtained (for example, the rotor rotational speed is equal to or higher than a predetermined threshold value). Depending on whether the estimated rotational position θ ^ is switched according to the equation (5) and the state using the estimated rotational position θ ^ according to the equation (4) using the disturbance observer 25 or the like. Also good.

In the above-described embodiment, the M-sequence signal is superimposed on the two-phase command voltage V _αβ . However, the M-sequence signal is superimposed on the two-phase command voltage V _γδ generated by the command voltage generator 13. Also good.
In the above-described embodiment, the example in which the present invention is applied to the electric motor 3 as a drive source of the electric power steering apparatus has been described. However, the present invention is applicable to the control of the electric motor for uses other than the electric power steering apparatus. It can also be applied to.

  In addition, various design changes can be made within the scope of matters described in the claims.

1 is a block diagram for explaining an electrical configuration of an electric power steering apparatus to which a motor control device according to an embodiment of the present invention is applied. FIG. It is an illustration figure for demonstrating the structure of an electric motor. It is a block diagram for demonstrating the structural example of an M series signal generation part. It is a figure which shows a 6th-order M series signal. The waveforms of the α-axis M-sequence signal and β-axis M-sequence signal generated when two values separated by a half period of the M-sequence signal are set as the α-axis initial value and the β-axis initial value, respectively. FIG. It is a block diagram for demonstrating an example of a disturbance observer and the structure relevant to this. It is a flowchart for demonstrating the flow of a rotor rotational position estimation calculation.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 5 ... Motor control apparatus, 7 ... Microcomputer, 21 ... Position estimation part, 61 ... Alpha axis M series signal generator, 62 ... Beta axis M series signal generator

Claims (3)

M-sequence signal generating means for generating an M-sequence signal to be added to the motor drive signals of the two coordinate axes of the two-phase fixed coordinate system or the two-phase rotational coordinate system;
An initial value setting means for setting an initial value of the M-sequence signal generated by the M-sequence signal generating means so as to be different between the two coordinate axes within a range of ¼ cycle to ½ cycle. Control device. Parameter identification means for identifying a motor parameter based on a response of the motor when a motor drive signal to which the M series signal generated by the M series signal generation means is added is applied to the motor;
The motor control apparatus according to claim 1, further comprising position estimation means for estimating a rotational position of the motor using the motor parameter identified by the parameter identification means. Signal processing means for performing low-pass filter processing for removing high-frequency components at a predetermined cutoff frequency for the M-sequence signal generated by the M-sequence signal generating means;
The motor control device according to claim 1 or 2, wherein an M-sequence signal subjected to low-pass filter processing by the signal processing means is added to a motor drive signal.

Images