JP5281339B2 - Synchronous motor drive system and control device used therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To estimate the position of a magnetic pole even if a current flowing to a synchronous motor increases. <P>SOLUTION: A drive system is equipped with: a power converter 2 which supplies the synchronous motor 1 with AC power; a controller 100 which controls its rotation into the target value by transmitting a control signal to the power converter 2; and a current detecting means 3 which detects the current of the motor. The controller 100 is equipped with: a high frequency signal generating means 13, which generates a high frequency signal synchronized with the current of the motor; control signal generating means 8, 9, 10, 11, 12, and 14, which generate control signals that are modulated by using the high frequency signal; high frequency current extracting means 13 and 505, which extract high frequency currents from the current that is detected in the motor by the current detecting means 3; and magnetic pole position estimating means 5, 6, and 7, which estimate the position of the magnetic pole in the synchronous motor, using the high frequency extracted current extracted with the high frequency current extracting means. The phase of the high frequency signal shifts in synchronism with the rotating angle of the rotor in the synchronous motor, according to the load torque of the synchronous motor. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a drive system for a synchronous motor that realizes motor control by estimating a rotation angle of a rotor of the synchronous motor, and a control device used therefor.

A magnetic pole position estimation method for estimating the magnetic pole position of a synchronous motor without using an angle sensor that detects an electrical angle position is disclosed.
Patent Document 1 estimates the magnetic pole position using the saliency of a permanent magnet synchronous motor (hereinafter referred to as “PM (Permanent Magnetic) motor”). An alternating magnetic field is generated on the estimated magnetic pole axis (dc axis) of the PM motor, the pulsating current (or voltage) of the estimated torque axis (qc axis) component orthogonal to the dc axis is detected, and based on this, the PM Estimate and calculate the magnetic pole position inside the motor. This technique uses the feature that there is an inductance interference term from the dc axis to the qc axis when there is an error between the actual magnetic pole axis and the estimated magnetic pole axis.
Patent Document 2 estimates and calculates the magnetic pole position using the magnetic saturation characteristics of the PM motor. Based on the magnitude of current generated by applying a voltage to the PM motor in a certain direction, Estimate and calculate the magnetic pole position.
JP 7-245981 A JP 2002-78392 A

Since the technique of Patent Document 1 uses the rotor saliency of the PM motor, the application is restricted by the structure of the PM motor, and when the primary current of the PM motor increases, the saliency is lost and the magnetic pole is lost. There is a problem in that it is difficult to estimate the position.
The technique of Patent Document 2 utilizes a phenomenon in which a high-frequency voltage is applied to a stator coil of a PM motor, and the waveform of the high-frequency current superimposed on the primary current by this applied voltage is distorted by magnetic saturation. Although this technique is not limited by the structure of the PM motor, there is a problem that it is difficult to estimate the magnetic pole position when the primary current increases, because the influence of distortion due to magnetic saturation changes depending on the primary current.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a synchronous motor drive system capable of estimating the magnetic pole position even when the current flowing through the synchronous motor increases, and a control device used therefor.

In order to solve the above-described problems, the means of the present invention includes a non-salient-pole synchronous motor (1), a power converter (2) for supplying AC power to the non-saliency-type synchronous motor, and the power converter. current detecting a control device for controlling dq vector to the target value of the rotation by transmitting a PWM control signal the non-salient-pole type synchronous motor (100), a motor current flowing through the non-salient pole synchronous motor as a motor detection current the detection means (3) and the drive system of the synchronous motor having a (200), wherein the control device includes a high-frequency signal generating means (13) for generating a high-frequency signal to match the phase d c-axis, the target value , and based on the value of the motor detected current, the PWM control signal proportional to the vector operation voltage command value (Vdq0 ^*) and the voltage value and the added voltage value obtained by adding the high-frequency signal (Vdq ^*) The control signal generating means (8, 9, 10, 11, 12, 14) to be formed, the high frequency current extracting means (13, 505) for extracting the high frequency current from the motor detected current, and the high frequency current extracting means Magnetic pole position estimating means (21 , 506 ) for estimating the magnetic pole position (Δθc) of the non-salient-pole synchronous motor using a high-frequency extraction current with reference to the dc axis, and the motor detected current is qc The dc-axis current is decomposed into a shaft current and a dc-axis current. The dc-axis current is magnetically saturated based on the command value of the qc-axis current vector-calculated based on the target value. It is set to the value to be. The parentheses are examples.

  According to the present invention, the magnetic pole position can be estimated even when the current flowing through the synchronous motor increases. Thereby, even with a PM motor having no saliency, it is possible to estimate the magnetic pole position from the stop, thereby realizing high-efficiency driving by position sensorless control in the entire speed range.

(First embodiment)
A configuration of an electric motor drive system according to an embodiment of the present invention will be described.
An electric motor drive system 200 of FIG. 1 includes an electric motor 1 (ACM), a power converter 2, a current detector 3, and a control device 100. The control device 100 uses a load torque command value τ ^* as a target value. Thus, the dq vector control of the electric motor 1 is performed.
The electric motor 1 is a three-phase magnet synchronous motor (PM motor), and is configured such that a rotor having a permanent magnet rotates inside the stator. The stator is provided with a plurality of slots in the core, and a stator coil (primary winding) is wound around the slots. By supplying three-phase power to the stator coil, a current magnetic flux Φi is generated, and the permanent magnet generates a magnetic magnetic flux Φm. The current detector 3 detects three-phase currents (motor currents) Iu, Iv, and Iw flowing through the motor 1 using a Hall element or the like.

  The definition of the coordinate system of the electric motor 1 and the axis error Δθ will be described with reference to FIG. The d-axis is an axis with the direction of the N pole of the permanent magnet inside the rotor as the positive direction. Further, a direction whose phase is advanced by 90 degrees in electrical angle with respect to the positive direction of the d-axis is defined as a positive q-axis direction. Further, the dc axis is an estimated axis in which the d axis is estimated inside the control apparatus 100, and the qc axis is an estimated axis in which the q axis is estimated. The axis error Δθ is defined as the phase difference angle of the advance direction of the dc axis with respect to the d axis.

  In FIG. 1, the power converter 2 is configured to include six IGBTs (Insulated Gate Bipolar Transistors) and a plurality of commutation diodes. Convert to electricity. The current detector 3 detects a motor current (phase current) Iuvw flowing from the power converter 2 to the motor 1 and functions as a current detection unit.

The control device 100 includes a ROM (Read Only Memory), a RAM (Random Access Memory), and a CPU (Central Processing Unit), and includes a dq converter 4, an axis error estimator 5, a speed estimator 6, and an integration. 7, a command value generator 8, a current controller 9, a vector controller 10, a dq inverse converter 11, a pulse width modulator 12, and a high frequency voltage command generator 13 that functions as a high frequency signal generating means. And an adder 14, a temperature estimator 15, and a high-frequency current amplitude determiner 16. The axis error estimator 5, the speed estimator 6, and the integrator 7 function as the magnetic pole position estimating means 21, and the command value generator 8 functions as the average current control means. That is, the control device 100 controls the average value of the motor current.

  The dq converter 4 converts the detected value of the motor current Iuvw into a dc-axis current detection value Idc and a qc-axis current detection value Idc as values on the dc axis and the qc axis, which are reference coordinate axes in the control device. . In the drawing, the dc-axis current detection value Idc and the qc-axis current detection value Idc are described as Idqc. The axis error estimator 5 uses the dc-axis current detection value Idc and the qc-axis current detection value Iqc, which are the outputs of the dq converter 4, to estimate the axis error Δθc that is the phase difference angle between the dc axis and the d axis. Is calculated.

The speed estimator 6 calculates an estimated electrical angular speed value ω1c of the electric motor 1 based on Δθc. The integrator 7 integrates the electrical angular velocity estimation value ω1c to calculate the dc axis position θdc. The command value generator 8 generates a dc-axis current command value Id ^* , a qc-axis current command value Iq ^* , and the phase of the high-frequency voltage applied to the motor 1 based on the given load torque command value τ ^*. A command value γ ^* is generated .

The current controller 9 generates a second dc-axis current command value Id ^** based on the dc-axis current command value Id ^* , the current command value Iq ^* on the qc-axis, and the current detection values Idc and Iqc. A second qc-axis current command value Iq ^** is generated.
The vector controller 10 generates a dc-axis voltage command value to be given to the motor 1 based on the second dc-axis current command value Id ^** , the second qc-axis current command value Iq ^** , and the electrical angular velocity estimation value ω1c. Vd0 ^* and qc-axis voltage command value Vq0 ^* are calculated.
The high-frequency voltage command generator 13 receives the phase command value γ ^* and outputs a dc-axis high-frequency voltage command Vdh ^* and a qc-axis high-frequency voltage command Vqh ^* to be applied to the motor 1.

The adder 14 adds the dc-axis voltage command value Vd0 ^* and dc axis high-frequency voltage command Vdh ^*, to generate a second dc-axis voltage command value Vd ^*, qc axis voltage command value Vq0 ^* and qc axis The high frequency voltage command Vqh ^* is added to generate a second qc-axis voltage command value Vq ^* .
The dq inverse converter 11 converts the three-phase voltage command value Vuvw ^{* of the} electric motor 1 using the second dc-axis voltage command value Vd ^* and the second qc-axis voltage command value Vq ^* . The pulse width modulator 12 performs PWM pulse width modulation by comparing the three-phase voltage command value Vuvw ^* and the triangular wave signal, and generates a control signal for controlling the power converter 2. The temperature estimator 15 estimates the temperature of the electric motor 1. The high frequency current amplitude determiner 16 determines whether or not the amplitude of the high frequency current is within an appropriate range, and outputs a determination signal Sig.

FIG. 3 is an internal configuration diagram of the current controller 9.
The current controller 9 includes differentiators 901 and 903 and PI controllers 902 and 904. The PI controllers 902 and 904 are automatic current regulators (ACRs) that perform PI control. The difference unit 901 calculates a difference between Id ^* and Idc on the dc axis, and the PI controller 902 performs a proportional integration calculation (PI calculation) on the output of the difference unit 901. The differentiator 903 calculates the difference between Iq ^* and Iqc on the qc axis, and the PI controller 904 performs a proportional-integral operation on the output of the differencer 903.

Next, the principle of magnetic pole position estimation in this embodiment will be described.
In the present embodiment, the high frequency voltage generated by the high frequency voltage command generator 13 is applied (superposed) to the stator coil of the electric motor 1 and the position is estimated based on the generated high frequency current component.

FIG. 4 is a schematic diagram showing an estimation principle using magnetic saturation, which is one of the principles of conventional magnetic pole position estimation.
FIG. 4A is a schematic diagram of the inside of the rotor of the PM motor. In the d-axis direction (dc axis direction) inside the PM motor, there is a magnet magnetic flux Φm by a permanent magnet. FIG. 4B is a schematic diagram showing the magnetic saturation characteristics of the iron core at this time. FIG. 4C shows a high-frequency current waveform that flows when a high-frequency voltage in the d-axis direction is applied to the stator coil during magnetic saturation. FIG. 4D shows a high-frequency current waveform that flows when a high-frequency voltage in the q-axis direction is applied to the stator coil during magnetic saturation.

  When a high-frequency voltage having a phase in the d-axis direction is applied, the current in the d-axis positive direction generates a magnetic flux in the same direction as the magnet magnetic flux Φm and strengthens the magnetic saturation, thereby reducing the winding inductance Lds of the stator coil (L small) ), The peak value ΔIdp of the high-frequency current increases. Conversely, the d-axis negative current generates a magnetic flux in a direction that cancels out the magnet magnetic flux Φm and weakens the magnetic saturation. Therefore, the winding inductance Ld0 is slightly increased (L is large). For this reason, the current in the d-axis negative direction slightly reduces the peak value ΔIdn of the high-frequency current.

Further, when a high frequency voltage is applied in the q-axis direction, a magnetic flux is generated in a direction orthogonal to the magnet magnetic flux Φm, so that magnetic saturation is hardly increased, and the peak value of the high frequency current is q as shown in FIG. Axes are equal in the positive and negative directions. At this time, if the high-frequency voltage is a rectangular wave voltage, the high-frequency current becomes a triangular wave that linearly increases and decreases.
Therefore, if the phase in which the high frequency voltage is applied is the positive / negative direction of the dc axis, the smaller the absolute value of the axial error Δθ, the more the high frequency current has an asymmetry in which the peak value in the positive direction of the dc axis increases due to the change in winding inductance. By utilizing this, the magnetic pole position can be estimated.

  FIG. 4A shows a case where no current other than the high-frequency current flows, but the decrease in winding inductance due to magnetic saturation is actually caused by a magnetic flux Φd in the d-axis direction as shown in FIG. 4B. At a certain value larger than the magnet magnetic flux Φm, it becomes prominent to have a break point. This break point is defined as a magnetic saturation point Φ10.

  The influence of magnetic saturation when the primary current flows through the electric motor 1 will be described using the schematic diagram of FIG. In addition to the magnet flux Φm, there is also a current flux Φi created by the primary current in the electric motor 1. Therefore, the magnetic flux inside the electric motor 1 is a vector sum of the magnet magnetic flux Φm and the current magnetic flux Φi as shown in FIG. That is, the current magnetic flux Φi also affects the magnetic saturation of the iron core. It is noted that what is actually observed is the influence of the magnetic flux interlinked with the primary winding of the electric motor 1, and hereinafter, the magnetic flux interlinking with the primary winding will be referred to as the interlinking magnetic flux Φ1, and the magnetic flux inside the electric motor 1 will be described below. It shall be equivalent to Hereinafter, the direction of the interlinkage magnetic flux Φ1 is defined as the φ axis.

Due to the influence of the current magnetic flux Φi, the φ axis has a phase difference angle γ with respect to the d axis as shown in FIG. 5A, and the magnetic saturation characteristics of the iron core with respect to the high frequency current also change as shown in FIG. 5B. At this time, the asymmetry of the crest value of the high-frequency current due to the magnetic saturation of the iron core appears most strongly in the φ axis direction. That is, when a high frequency voltage is applied in the φ axis direction, the high frequency current has a peak value in the positive direction of the φ axis as shown in FIG. On the other hand, when a high-frequency voltage is applied in the z-axis direction, which is an axis orthogonal to the φ-axis, the peak value of the high-frequency current in the z-axis direction becomes equal in the positive direction and the negative direction, as shown in FIG. However, γ is not a constant value, but varies depending on the load state (load torque) of the electric motor 1. Therefore, when the axis error Δθ is estimated based on the concept of FIG. 4, γ is included as an estimation error. Since the phase of the high frequency voltage is applied in the φ axis direction, it moves in synchronization with any of the rotation angle of the rotor, the φ axis, and the z axis.

ここでγoffsetは位相指令値γ^＊とφｃ軸との相差角であり、任意に決定することができる。非突極型のＰＭモータに適用した場合、トルクを発生するのは主にｑ軸電流Ｉｑとなり、重負荷時は電動機１の電流はＩｑが殆どを占めると考えられる。このとき、鎖交磁束Φ１はｑ軸電流Ｉｑによる磁束Φｉｑが支配的となり、φ軸はｑ軸の近傍に存在すると考えられる。仮に、高周波電圧を印加する位相をφ軸とすると、発生する高周波電流は殆どがｑ軸成分となるので、電動機１の電流値は、トルク分を示すｑ軸電流Ｉｑに、高周波電流を加算した値に略等しくなる。 Therefore, in the present embodiment, the phase in which the high frequency voltage is applied is changed according to the load state of the electric motor 1. Specifically, the phase to which the high frequency voltage is applied is determined with reference to the φc axis that is the estimated axis inside the φ axis control device, and the phase difference angle γ between the φ axis and the d axis is increased according to the load state. Move the voltage phase command value γ ^* . Since the phase difference angle between the φ axis and the φc axis is equal to the axis error Δθ, the estimation of the magnetic pole position based on the asymmetry of the peak value of the high frequency current can be easily realized regardless of the load state. Note that γ ^* only needs to maintain a constant phase difference angle from the φc axis, and the phase command value γ ^* is given by the following equation.

Here, γoffset is a phase difference angle between the phase command value γ ^* and the φc axis, and can be arbitrarily determined. When applied to a non-salient pole type PM motor, torque is mainly generated by the q-axis current Iq, and it is considered that Iq accounts for most of the current of the motor 1 under heavy load. At this time, the flux linkage Φ1 is dominated by the magnetic flux Φiq caused by the q-axis current Iq, and the φ-axis is considered to exist in the vicinity of the q-axis. Assuming that the phase for applying a high-frequency voltage is φ axis, most of the generated high-frequency current is a q-axis component. Therefore, the current value of the motor 1 is obtained by adding the high-frequency current to the q-axis current Iq indicating the torque component. Approximately equal to the value.

However, since the upper limit of the output current of the power converter 2 is determined, the maximum value of the output torque of the electric motor 1 is limited by the amount of high-frequency current flowing. Therefore, it is desirable that the phase command value γ ^{* be} in the direction of the zc axis, which is an estimated axis inside the z axis control device, from the viewpoint of securing output torque. The definition of each coordinate axis is shown in FIG. As can be seen from FIG. 6, the phase difference angle between the φ axis and the φc axis and the phase difference angle between the z axis and the zc axis are equal to the axis error Δθ. The position can be estimated.

Next, attention is paid to a change in the magnitude of the flux linkage Φ1. The decrease in winding inductance due to magnetic saturation is significant when the flux linkage Φ1 is larger than the magnetic saturation point Φ10. For example, when a high-frequency voltage is applied in the d-axis direction when no load is applied as shown in FIG. 4, when the magnitude of the interlinkage magnetic flux Φ1 including the magnetic flux fluctuation due to the generated high-frequency current is less than the magnetic saturation point Φ10, The asymmetry of the peak value of the high-frequency current due to the decrease in the line inductance hardly appears, and a problem that the magnetic pole position cannot be estimated may occur.
Therefore, in the present embodiment, the primary current is such that the magnitude of the flux linkage Φ1 when no high-frequency voltage is applied is maintained at or above the magnetic saturation point Φ10 as shown in FIG. (Motor current) is applied to the motor 1.

  FIG. 7 is a schematic diagram illustrating a non-salient PM type motor as an example. FIG. 7A shows a state in which the d-axis current Id is applied so that the sum of the d-axis current magnetic flux Φid and the magnet magnetic flux Φm becomes the magnetic saturation point Φ10 even when there is no load. FIG. 7B shows the d-axis current Id so that the magnitude of the linkage flux Φ1 is equal to or greater than the magnetic saturation point Φ10 at the same time as the q-axis current Iq necessary for torque generation at the time of light load. There is a magnetic flux Φiq due to Φid and q-axis current. FIG. 7C shows Φiq generated at Iq necessary for torque generation, and when Φ1 is Φ10, Id is zero. In FIG. 7D, the combined size of Φiq and magnet magnetic flux Φm exceeds Φ10.

このように、高周波電圧を印加しない状態での磁束Φ１の大きさが磁気飽和点Φ１０を上回るように、電動機１の一次電流を与えることにより、巻線インダクタンスの減少による高周波電流の波高値の非対称性が発生し易くなり、より好適な磁極位置の推定が実現される。 7C and 7D, either Id may be zero, or Φid may be generated in the d-axis negative direction by negative Id as shown in FIG. 7D. If 1 is a non-salient PM motor, negative Id does not contribute to torque, so Id may be zero unless there is another reason such as field weakening control. In the PM motor with the motor 1 having the reverse saliency, more torque output can be obtained when the magnetic flux Φid is generated in the negative direction of the d-axis, so that the linkage flux Φ1 does not fall below the magnetic saturation point Φ10. If present, negative Id may be passed. That is, the primary current may be applied to the electric motor 1 so that the size of Φ1 satisfies the expression (2).

As described above, by providing the primary current of the motor 1 so that the magnitude of the magnetic flux Φ1 in a state where no high-frequency voltage is applied exceeds the magnetic saturation point Φ10, the asymmetry of the peak value of the high-frequency current due to the decrease in winding inductance. Therefore, the magnetic pole position can be estimated more appropriately.

Next, the command value generator 8, the high-frequency voltage command generator 13, and the axis error estimator 5, which are features of the present embodiment, for realizing the magnetic pole position estimation method described above will be described.
The command value generator 8 receives the torque command τ ^* , and the current commands Id ^* and Iq ^{* at} which the magnitude of the linkage flux Φ1 becomes a magnetic saturation point Φ10 or more when no high frequency voltage is applied. The phase command value γ ^* of the high-frequency voltage is output based on the phase difference angle γ between the d-axis and the φ-axis generated by the above and the equation (1). The input / output relationship of the command value generator 8 may be obtained, for example, by obtaining the characteristics of the motor 1 to be controlled by electromagnetic field analysis. The current controller 9 performs current control based on the current command of the command value generator 8, thereby maintaining the magnitude of the linkage flux Φ1. That is, since the command value generator 8 functions as an average current control means, in the control device 100, the current controller 9 controls the average value of the motor current. Note that the magnitude of the linkage magnetic flux Φ1 in a state where no high-frequency voltage is applied may be arbitrary as long as it is not less than the magnetic saturation point Φ10, and may be changed according to the load state of the electric motor 1.

The configuration of the high-frequency voltage command generator 13 is shown in FIG. A high-frequency voltage command generator 1301 that generates a square-wave high-frequency voltage command Vh ^* and a coordinate converter 1302 that converts Vh ^* as a voltage command on the zc axis into a voltage command value on the dq axis by γ ^*. The A method of operating the determination signal Sig that is an input of the high-frequency voltage command generator 1301 to Vh ^* will be described later. The frequency of Vh ^* is 1 / N (N is a positive integer) of the carrier frequency of the pulse width modulator 12. Thereby, the phase which applies a high frequency voltage can be hold | maintained in a zc-axis direction irrespective of a load state.

9 is a configuration diagram of the axis error estimator 5a, FIG. 10 is a principle diagram for estimating the axis error, and FIG. 11 is a diagram showing an actual operation of the axis error estimator 5a. FIG. 11 shows a case where the frequency of the square wave of the high-frequency voltage command Vh ^* is half of the PWM triangular wave carrier frequency. Note that the calculation cycle of the control device is half of the PWM triangular wave carrier cycle, and the current detection value is obtained at the upper and lower timings of the PWM triangular wave carrier.
11A shows a PWM triangular wave carrier, FIG. 11B shows a high frequency voltage command Vh ^* , FIG. 11C shows a detected current Izc, and FIG. 11D shows a first-order difference value. FIG. 11E shows the current polarity signal Spz, FIG. 11F shows Spz × | Izch |, and FIG. 11G shows Δθc.

In the present embodiment, when a high frequency voltage is applied in the zc-axis direction, and the peak value of the generated high-frequency current is | Izchp | in the zc-axis positive direction and | Izchn | in the zc-axis negative direction, Is assumed to have the relationship of the expression (3) with respect to the axis error Δθ (FIG. 10B). Note that k1 in the equation is a proportionality coefficient and is determined by actual measurement of characteristics of the electric motor 1 to be controlled or electromagnetic field analysis.

The coordinate converter 500 calculates the φc axis current detection value Iφc and the zc axis current detection value Izc using the dc axis current detection value Idc, the qc axis current detection value Iqc, and the phase command value γ ^* of the high frequency voltage. To do. Note that the value of the detection current Izc that can be actually obtained by the control device is the value at the timing indicated by the white circle in FIG. From the obtained zc-axis current detection value Izc, a delay unit 501 and an adder 502 are used to calculate a first-order difference value ΔIzch of Izc for each control calculation cycle (FIG. 11 (d)), and an absolute value calculator In 503, the absolute value | ΔIzch | is calculated.

Further, the current polarity calculator 504 determines a current polarity signal Spz that determines which peak value in the zc-axis positive direction or the zc-axis negative direction is | ΔIzch | obtained by the absolute value calculator 503 (FIG. 11E). Is generated from Vh ^* . A current change amount calculator 505 serving as a high-frequency current extraction unit calculates a current change amount | Izchp | in the zc-axis positive direction and a current change amount | Izchn | in the zc-axis negative direction from | ΔIzch | and Spz. The axis error calculator 506 calculates the axis error estimated value Δθc from | Izchp | and | Izchn | according to the equation (3) (FIG. 11 (g)). Thus, the magnetic pole position can be estimated by obtaining the axis error estimated value Δθc.

但し、|Ｖｈ０^＊|は、キャリア周波数ｆｃの基準値ｆｃ０の高周波電圧指令Ｖｈ^＊の振幅であり、|Ｖｈ^＊|は、キャリア周波数ｆｃに対する高周波電圧指令Ｖｈ^＊の振幅である。すなわち、制御装置１００は、高周波電圧の振幅に基づいて、電動機電流の平均値の大きさを制御している。
この場合、キャリア周波数ｆｃの変化による高周波電流の波高値の増減を相殺できるので、（３）式の比例係数ｋ１を一定と扱うことができる。 The above description is based on the assumption that the carrier frequency fc of the pulse width modulator 12 is fixed. However, when the carrier frequency fc is variable, the high frequency voltage command generator 1301 sets the amplitude | Vh ^* | ).

However, | Vh0 ^* | is the amplitude of the high-frequency voltage command Vh ^* of the reference value fc0 of the carrier frequency fc, and | Vh ^* | is the amplitude of the high-frequency voltage command Vh ^{* with} respect to the carrier frequency fc. That is, the control device 100 controls the magnitude of the average value of the motor current based on the amplitude of the high frequency voltage.
In this case, since the increase / decrease in the peak value of the high frequency current due to the change in the carrier frequency fc can be offset, the proportionality coefficient k1 in the equation (3) can be treated as constant.

The method for estimating the magnetic pole position with the magnet magnetic flux Φm of the electric motor 1 being constant has been described above. However, since the magnetic flux Φm changes depending on the temperature of the permanent magnet, even if the current supplied to the electric motor 1 is the same, the interlinkage magnetic flux Φ1. Will change. For this reason, it is desirable to change the output of the command value generator 8 according to the change of the magnet magnetic flux Φm. Since the magnetic flux Φm decreases as the temperature of the permanent magnet increases, the interlinkage magnetic flux Φ1 also decreases, and the degree of magnetic saturation of the iron core is alleviated.
Therefore, the higher the temperature of the permanent magnet, the smaller the amplitude of the high frequency current generated by applying the high frequency voltage. The rise in magnet temperature also depends on the physique of the electric motor 1 to be controlled, but the time constant is in units of seconds or minutes. Compared to the magnetic pole position estimation and current control response of the control device, the time constant Is long enough.

  Therefore, in this embodiment, the temperature estimator 15 is provided for the purpose of estimating the temperature of the permanent magnet from the amplitude of the high frequency current generated by applying the high frequency voltage. The temperature estimator 15 shown in FIG. 12 includes delay units 1501 and 1503, adders 1502 and 1504, a high-frequency current amplitude calculator 1505, a low-pass filter 1506, and a temperature estimation calculator 1507, and a dc-axis current. An estimated temperature Temp is calculated from the detected value Idc and the qc-axis current detected value Iqc.

  Delay device 1501 and adder 1502 calculate first-order difference value ΔIdch of Idc, and delay device 1503 and adder 1504 calculate first-order difference value ΔIqch of qc-axis current detection value Iqc. The high frequency current amplitude calculator 1505 calculates the amplitude I1h of the high frequency current by calculating the sum of squares of ΔIdch and ΔIqch. Further, the low-pass filter 1506 calculates an average value of the amplitude of the high-frequency current with a time constant corresponding to the temperature increase of the permanent magnet. The temperature estimation calculator 1507 estimates the temperature of the permanent magnet (estimated temperature Temp) from the average value of the amplitude of the high-frequency current. Since the characteristics given to the temperature estimation calculator 1507 are different depending on the electric motor 1 to be controlled, for example, it may be calculated and given in advance by electromagnetic field analysis.

The command value generator 8 corrects the current command values Id ^* and Iq ^* and the phase command value γ ^{* of the} high frequency voltage from the estimated temperature Temp of the permanent magnet estimated by the temperature estimator 15. As a correction method, for example, a characteristic calculated in advance by electromagnetic field analysis may be given as table data or an approximate expression. Since the command value generator 8 functions as an average current control means and the command value generator 8 corrects the current command values Id ^* and Iq ^* , the motor current is determined based on the estimated temperature Temp of the permanent magnet. The average value is adjusted.

  Even if the magnet magnetic flux Φm is constant, when the sudden change in the load torque τL occurs and the axial error Δθ becomes transiently large, the phase of the current flowing through the motor 1 changes, so the current magnetic flux Φi also changes. At this time, in the direction in which the interlinkage magnetic flux Φ1 increases, the peak value of the high-frequency current increases and there is a risk of reaching a stop level due to an excessive current of the power converter 2 at a heavy load. If it is significantly reduced, the flux linkage Φ1 becomes smaller than the magnetic saturation point Φ10 even if the magnetic flux change due to the high-frequency current is superimposed, and the axis error Δθ may not be estimated.

  Therefore, the high-frequency current amplitude determiner 16 calculates the amplitude I1h of the high-frequency current similarly to the temperature estimator 15, determines whether the amplitude I1h is within a predetermined range, and outputs a determination signal Sig. The high-frequency current amplitude determiner 16 shown in FIG. 13 includes delay units 1601 and 1603, adders 1602 and 1604, a high-frequency current amplitude calculator 1605, and an upper and lower limit determiner 1606, and inputs Idc and Iqc. The determination signal Sig is output.

  Delay device 1601 and adder 1602 calculate first-order difference value ΔIdch of Idc, and delay device 1603 and adder 1604 calculate first-order difference value ΔIqch of Iqc. The high frequency current amplitude calculator 1605 is the same as the temperature estimator 15 until the square sum of ΔIdch and ΔIqch is calculated to obtain the amplitude I1h of the high frequency current. However, the high-frequency current amplitude determiner 16 directly inputs the obtained amplitude I1h to the upper / lower limit determiner 1606, and determines whether the amplitude I1h is within an appropriate range based on the characteristics calculated in advance by electromagnetic field analysis. The determination is made and the result is output as a determination signal Sig.

The determination signal Sig is input to the command value generator 8 and the high frequency voltage command generator 13. The command value generator 8 changes the current command values Id ^* and Iq ^* so that the linkage flux Φ1 is decreased when the amplitude I1h is greater than or equal to the upper limit value, and is increased when the amplitude I1h is less than or equal to the lower limit value. To do. The high frequency voltage command generator 13 decreases the amplitude | Vh ^* | of the high frequency voltage command when the amplitude I1h is equal to or higher than the upper limit value, and increases | Vh ^* |
By combining the command value generator 8 and the high-frequency voltage command generator 13, the electric motor 1 can be driven without causing an excessive current stop or step-out.

(Second Embodiment)
Next, a second embodiment will be described. This embodiment differs in the configuration of the axis error estimator 5 and the axis error estimation method. As described above, the peak value of the high-frequency current generated by applying the high-frequency voltage increases as the phase to which the high-frequency voltage is applied is closer to the positive direction of the φ axis, and the peak value decreases as the phase is closer to the negative direction of the φ axis. Compared with the case where the phase for applying the high-frequency voltage is close to the z-axis direction, the winding inductance is larger in the z-axis direction than in the φ-axis direction when the average value of the amplitude close to the φ-axis is large. That is, it can be considered that the electric motor 1 has saliency.

  FIG. 14 is a diagram showing the configuration of the axis error estimator 5b using saliency, and FIG. 15 is a diagram showing the principle of axis error estimation. FIG. 16 is a diagram showing an actual operation of the axis error estimator 5 when the frequency of the high-frequency voltage is half of the PWM triangular wave carrier frequency. Note that the calculation cycle of the control device is half of the PWM triangular wave carrier cycle, and the current detection value is obtained at the upper and lower timings of the PWM triangular wave carrier.

The application direction in which the high frequency voltage is applied is the zc-axis direction as in the first embodiment. The present embodiment is characterized by using a φc-axis high-frequency current component Iφch orthogonal to the application direction in which a high-frequency voltage is applied. It is assumed that the high-frequency current component Iφch changes with the equation (5) with respect to the axis error Δθ. Here, Lφ is a winding inductance in the φ-axis direction, and Lz is a winding inductance in the z-axis direction.

  The difference from the first embodiment is the difference in the high-frequency current used. That is, the change of the high-frequency current with respect to the axial error Δθ of one electrical angle cycle is one cycle in the first embodiment, but two cycles in the present embodiment. The coordinate converter 500 (FIG. 14) is the same as in the first embodiment (FIG. 9), and the value of Iφc actually obtained by the control device is the value at the timing indicated by the white circles in FIG. 16 (c). . In the present embodiment, the delay unit 507 and the adder 508 obtain the first-order difference value ΔIφch of Iφc for each control calculation cycle.

The current polarity calculator 509 generates a current polarity signal Spφ that determines which peak value of the first-order difference value ΔIφch is in the positive direction of the φc axis or the negative direction of the φc axis from Vh ^* . When ΔIφch is multiplied by Spφ by the multiplier 510, an amount corresponding to the right side of the equation (5) is obtained. Therefore, the axis error calculator 511 calculates the axis error estimated value Δθc based on the equation (5).
The axis error estimator 5 may use either the configuration shown in FIG. 9 or the configuration shown in FIG. 15 or may be used in combination.

(Third embodiment)
FIG. 17 is an overall configuration diagram of an electric motor drive system according to the third embodiment.
The control device 110 of the motor drive system 210 is different from that shown in FIG. 1 in that the second axis error estimator 18 estimates the axis error based on the induced voltage of the motor 1 and the movement for calculating the fundamental wave component from the detected current Idqc. An average processor 17 and a switch 19 for selecting one of the outputs of the axis error estimator 5 and the second axis error estimator 18 are added as estimated values of the axis error. The second axis error calculator 18 uses the PM motor constants (R, Lq) to estimate the axis error using equation (6).

The second axis error estimator 18 does not require application of a high-frequency voltage, and when the motor 1 is a PM motor, the speed is approximately 10% or more of the rated speed, and can be applied even at a lower speed depending on the characteristics of the PM motor. However, it cannot be applied when stopped or when the speed is very low. On the other hand, the control methods of the first embodiment and the second embodiment can be applied even when stopped or when the speed is very low, but the high-frequency current generated by the application of the high-frequency voltage increases the peak value of the motor current. The output torque can be limited.
Therefore, according to the speed of the electric motor 1, the output torque is more preferably switched to the axis error calculator 5 when the speed is low from the stop and switched to the second axis error estimation calculator 18 when the rated speed is approximately 10% or more. It is advantageous to use it to the limit.

A switching sequence of axis error estimation during acceleration / deceleration in this embodiment will be described with reference to FIG. FIG. 18 shows an example of switching based on the estimated electrical angular velocity value ω1c. 18A shows the transition of the estimated electrical angular velocity value ω1c due to acceleration / deceleration, FIG. 18B shows the output of the switch 19 that switches the axis error estimated value output, and FIG. 18C shows the high-frequency voltage command Vh. ^* the amplitude | Vh * ^| of showing changes, FIG 18 (d) shows the operation timing of the moving average processor 17.

First, at the time of starting acceleration from the stop (t0 to t2), the amplitude | Vh ^* | of the high-frequency voltage command Vh ^* (FIG. 8) is made non-zero and a high-frequency voltage is applied. A shaft error estimated value Δθc is output. When the electrical angular velocity estimated value ω1c reaches the first angular velocity threshold value ω1s1 (t2), the output of the switch 19 is switched to the output Δθc2 of the second axis error estimator 18, and at the same time, | Vh ^* | The acceleration is continued by the magnetic pole position estimation using the error estimator 18. On the other hand, when performing deceleration from high speed, the output of the switch 19 is switched to the output Δθc2 of the second axis error estimator 18, and at the same time, the amplitude | Vh ^* |

When the electrical angular velocity estimated value ω1c falls below the first angular velocity threshold ω1s1 (t3), the output of the switch 19 remains Δθc2, and the high-frequency voltage command generator 13 | Vh as shown in FIG. ^* Gradually increase | to start applying high-frequency voltage. Further, as shown in FIG. 18D, the moving average processor 17 operates to remove the influence of the generated high frequency current from the estimation result of the second axis error estimator 18. The moving average processor 17 averages the detected current value Idqc over one period of the high-frequency voltage, and gives the average current value Idqc0 to the second axis error estimator. When the electrical angular velocity estimated value ω1c further decreases and falls below the second angular velocity threshold ω1s2 (t4), the output of the switch 19 is switched to the axis error estimated value Δθc by the axis error estimator 5. By performing this switching operation, it is possible to operate with higher torque in a speed region where no high-frequency voltage is applied. Further, there is no need to keep the flux linkage Φ1 at or above the magnetic saturation point Φ10, and the current value especially at light load can be reduced. That is, the motor drive system 210 controls the motor current.

In the present embodiment, during the period when no high frequency voltage is applied, the temperature estimator 15 estimates the temperature of the permanent magnet by the following method. When the set value of the induced voltage constant Ke is Ke ^* , the set value of the winding resistance R1 is R1 ^* , and the set values of the winding inductance Ld and Lq are Ld ^* and Lq ^* , respectively, the set values of the constants are If it matches with the actual constant of the PM motor, the voltage equation calculated by the vector controller 10 is the equation (7).

Here, it is assumed that the steady state is assumed, and the electrical angular velocity estimation value ω1c is equal to the actual angular velocity ω1. At this time, Id ^** = Id ^* , Iq ^** = Iq ^* is established. From this state, assuming that only the induced voltage constant changes to Ke ′ due to temperature fluctuations of the permanent magnet, if the actual angular velocity ω1 of the electric motor 1 is sufficiently large, the influence of the winding resistance can be ignored. Is represented by equation (8).

Therefore, if Id ^* and Idc are equal and Iq ^* and Iqc are equal by the current controller 9, the amount of change in the induced voltage constant can be extracted from the equation (9).

That is, the setting error ΔKe ^* of the induced voltage constant is obtained from the difference between the current command value Id ^* and the second current command value Id ^** for the d axis. If the change of the induced voltage constant with respect to the temperature of the permanent magnet is calculated in advance by electromagnetic field analysis, the temperature estimator 15 calculates the temperature estimation value Temp of the permanent magnet using the equation (7) simultaneously with ΔKe ^*. it can. As shown in FIG. 18E, the operation of the temperature estimator 15 in the present embodiment is during a period in which the switch 19 outputs the axis error estimated value Δθc2 from the second axis error estimator 18 (t2 to t4). Is always operated, and is decelerated and stopped while maintaining the set error ΔKe ^* and the estimated temperature value Temp immediately before the output of the switching unit 19 is switched to the estimated axis error value Δθc by the estimated axis error estimator 5. Since the time constant of the temperature fluctuation of the permanent magnet is long as described above, if the deceleration time is short, the temperature fluctuation of the permanent magnet can be ignored from the stop to the low speed range, and the calculation load of the control device can be reduced. When the deceleration time is longer than the time constant of the temperature fluctuation of the permanent magnet, a temperature estimator having the configuration shown in FIG.

  That is, according to this embodiment, in the low speed range from the stop, when the speed of the PM motor further increases while minimizing the limit of torque output due to the high frequency current generated by applying the high frequency voltage, the axis error estimation method By switching the torque, it is possible to increase the torque to a current that can be allowed by the power converter 2, and it is also possible to estimate the temperature fluctuation of the permanent magnet.

(Fourth embodiment)
Up to the third embodiment, the method for estimating and controlling the magnetic pole position while operating the PM motor has been described. However, when starting the PM motor, it is necessary to give the initial value of the dc axis position θdc as the initial magnetic pole position. In this embodiment, a method for estimating the initial magnetic pole position will be described. The configuration of the drive system of the synchronous motor may be any of FIG. 1 and FIG. When estimating the initial magnetic pole position, it is necessary to change the operations of the command value generator 8, the high-frequency voltage command generator 13, and the axis error estimator 5. Hereinafter, the operation of each component will be described with reference to FIG.

The command value generator 8 performs (a) Id ^* , (b) Iq ^* , (c) γ ^* in accordance with FIG. 19 regardless of the load torque command value τ ^* during the initial magnetic pole position estimation period. In order. As the number of combinations increases, the estimation accuracy of the initial magnetic pole position improves, but an estimation time proportional to the number of combinations is required. In the present embodiment, there are four combinations, and the periods for holding the respective states are period 1, period 2, period 3, and period 4. As shown in FIG. 19 (d), the high-frequency voltage command generator 13 outputs the high-frequency voltage by a certain number of cycles within each period.

Specifically, in a state in which the current controller is moved, in period 1, DC current I0 flows in the initial dc axis positive direction, and in period 2, DC current I0 flows in the dc axis negative direction. This corresponds to the application of the high-frequency voltage command Vh ^* for four cycles in the axial direction. That is, the control device 100 estimates the magnetic pole position in a state in which the high-frequency voltage is applied with a phase aligned with the phase in which the motor current flows. Also, in period 3, DC current I0 is passed in the positive direction of qc axis, and in period 4, DC current I0 is passed in the negative direction of qc axis, and high-frequency voltage command Vh ^* is applied for four cycles with a phase in the direction of qc axis. It corresponds to applying. The zc-axis at this time is equal to the initial dc-axis positive direction in period 1, equal to the negative dc-axis direction in period 2, equal to the positive qc-axis direction in period 3, and equal to the same qc-axis in period 4. Equal to the negative direction.

At this time, paying attention to the zc-axis current Izc shown in FIG. 19 (e), the average current for each period is equal to I0, but there is a difference in the amplitude of the high-frequency current shown in FIG. 19 (f). This is because the direct current I0 has affected the magnetic saturation. When the direction (phase) of the direct current I0 is closest to the positive direction of the d-axis, the amplitude of the high-frequency current becomes maximum. In FIG. 19, since period 1 corresponds, the closest to the d-axis positive direction is the dc-axis positive direction. The next largest current amplitude is period 3, so it can be estimated that there is a d-axis between the dc-axis positive direction and the qc-axis positive direction.
In addition, when estimating the initial magnetic pole position by the method of this embodiment, it is desirable that the rotor of the motor 1 to be controlled is mechanically fixed. Further, the number of Vh ^* periods in each period is not limited to four periods, and may be determined as necessary.
Although each embodiment has been described on the assumption that the electric motor 1 is a PM motor, it is obvious that the present invention can be applied to a synchronous reluctance motor if the magnetic flux Φm is zero.

1 is an overall configuration diagram of an electric motor drive system according to a first embodiment of the present invention. It is explanatory drawing which shows the definition of axial error (DELTA) (theta). It is a block diagram of a current controller. It is a schematic diagram which shows the estimation principle using the magnetic saturation in a conventional system. It is a schematic diagram which shows the estimation principle using the magnetic saturation in this invention. It is explanatory drawing which shows the definition of the coordinate system inside a control apparatus. It is a schematic diagram which shows the determination method of the electric current value with respect to load. It is a block diagram of the high frequency voltage command generator in 1st Embodiment. It is a block diagram of the axis error estimator in 1st Embodiment. It is explanatory drawing of the operation principle of the axis error estimator in 1st Embodiment. It is a schematic diagram of the operation waveform of the axis error estimator in the first embodiment. It is a block diagram of the temperature estimator in 1st Embodiment. It is a block diagram of the high frequency current amplitude determination device in 1st Embodiment. It is a block diagram of the axis error estimator in 2nd Embodiment. It is explanatory drawing of the operation principle of the axis error estimator in 2nd Embodiment. It is operation | movement explanatory drawing of the axis error estimator in 2nd Embodiment. It is a whole block diagram of the electric motor drive system in 3rd Embodiment. It is operation | movement explanatory drawing of the control operation switching method in 3rd Embodiment. It is operation | movement explanatory drawing of the initial magnetic pole position estimation in 4th Embodiment.

Explanation of symbols

1 Electric motor (synchronous motor)
2 Power converter 3 Current detector (current detection means)
4 dq converter 5, 5a, 5b Axis error estimator (magnetic pole position estimation means)
6 Speed estimator 7 Integrator 8 Command value generator (mean current control means)
9 current controller 10 vector controller 11 dq inverse converter 12 pulse width modulator 13 high frequency voltage command generator (high frequency signal generating means)
14, 502, 508, 1502, 1504 Adder 15 Temperature estimator 16 High frequency current amplitude determiner 17 Moving average processor 18 Second axis error estimator 19 Switch 21 Magnetic pole position estimating means 100, 110 Controller 200, 210 Motor Drive system 500 Coordinate converter 501, 507 Delay unit 503 Absolute value calculator 504, 509 Current polarity calculator 505 Current change calculator (high-frequency current extraction means)
506, 511 Axis error calculator 510 Multiplier 901, 903 Differentiator 902, 904 PI controller 1301 High frequency voltage command generator 1302 Coordinate converter 1501, 1503 Delay device 1505 High frequency current amplitude calculator 1506 Low pass filter 1507 Temperature estimation Calculator 1605 High-frequency current amplitude calculator 1606 Upper / lower limit value determiner

Claims (7)

A non-saliency synchronous motor, a power converter that supplies AC power to the non-salience synchronous motor, and a PWM control signal transmitted to the power converter to rotate the non-saliency synchronous motor to a target value a control device that controls dq vector, in the drive system of the synchronous motor and a current detecting means for detecting a motor current flowing through the non-salient pole synchronous motor as a motor detection current,
The controller is
a high frequency signal generating means for generating a high-frequency signal to match the phase d c-axis,
The target value, and based on the value of the motor detection current, to generate the PWM control signal whose pulse width proportional to the sum voltage value obtained by adding the voltage value of the high frequency signal and the voltage command value vector operations Control signal generating means;
High-frequency current extraction means for extracting a high-frequency current from the motor detection current;
Magnetic pole position estimation means for estimating the magnetic pole position of the non-salient-pole synchronous motor with reference to the dc axis using the high-frequency extraction current extracted by the high-frequency current extraction means;
The electric motor detection current is decomposed into a qc-axis current and a dc-axis current,
The dc-axis current is set to a value at which the magnetic flux of the non-salient-pole synchronous motor is magnetically saturated based on a command value of the qc-axis current vector-calculated based on the target value. Synchronous motor drive system. The magnetic values of saturation, the drive system of the synchronous motor according to claim 1, characterized in Nedea Rukoto determined by the magnetic properties of the core to be used before Symbol non-salient synchronous motor. The magnetic pole position estimating means includes a φ-axis component that is a direction of magnetic flux interlinking the high-frequency extracted current extracted by the high-frequency current extracting means with the stator winding of the non-salient-pole synchronous motor, and the φ axis 2. The drive system for a synchronous motor according to claim 1, wherein the magnetic pole position is estimated based on at least one of components separated into a z-axis component that is an axis orthogonal to the axis. The non-salient pole type synchronous motor is a permanent magnet type synchronous motor,
The control device further includes temperature estimation means for estimating a temperature of a permanent magnet included in the non-salient-pole synchronous motor based on the detected motor detection current,
2. The synchronous motor drive system according to claim 1, wherein the dc-axis current is adjusted based on the temperature of the permanent magnet estimated by the temperature estimation means.   2. The synchronous motor drive system according to claim 1, wherein the control device adjusts the magnitude of the amplitude of the high-frequency signal based on the magnitude of the high-frequency component of the electric motor detection current. The high-frequency signal generating means includes
Before starting the synchronous motor from a stopped state, the motor current is made to flow in at least four phases in order while keeping the magnitude of the average value constant, and the phase is aligned with the phase in which the motor current flows. Apply a high frequency signal ,
2. The synchronous motor drive system according to claim 1, wherein the magnetic pole position estimating means estimates the magnetic pole position of the non-salient-pole synchronous motor by using the high-frequency extracted current. A control apparatus for a synchronous motor controlled dq vector to the target value of the rotation of said transmit a PWM control signal non-salient synchronous motor to a power converter for supplying AC power to the non-salient pole type synchronous motor,
The controller is
a high frequency signal generating means for generating a high-frequency signal to match the phase d c-axis,
Based on the target value and the value of the motor current flowing through the non-salient-pole synchronous motor , the pulse width is proportional to the added voltage value obtained by adding the voltage command value vector-calculated and the voltage value of the high-frequency signal. Control signal generating means for generating a PWM control signal;
High-frequency current extraction means for extracting a high-frequency current from the value of the motor current ;
Magnetic pole position estimation means for estimating the magnetic pole position of the non-salient-pole synchronous motor with reference to the dc axis using the high-frequency extraction current extracted by the high-frequency current extraction means;
The value of the motor current is decomposed into a qc-axis current and a dc-axis current,
The dc-axis current is set to a value at which the magnetic flux of the non-salient-pole synchronous motor is magnetically saturated based on a command value of the qc-axis current vector-calculated based on the target value. Control device for synchronous motor.

Images