JP5543388B2 - Control device for permanent magnet synchronous motor - Google Patents

Description

  Embodiments described herein relate generally to a permanent magnet synchronous motor control apparatus having a rotor position estimation function.

  Variable speed motor control devices using inverters are applied in various fields, and further improvements in performance and reliability such as improved torque and speed control accuracy, high efficiency and low noise are expected in the future. In particular, position / speed sensorless control without using a speed sensor or position sensor is useful in terms of improving reliability and improving constraints on the installation environment.

  Since about 1980, a technique for detecting the rotor magnetic flux direction of an electric motor having a magnetic flux in the rotor by a permanent magnet or the like without using a position sensor has been developed. However, since the induced voltage is 0 or very small when the motor is stopped or rotated at a low speed, it is easily affected by the output voltage error due to the dead time compensation error and the setting error of the primary resistance used in the estimation formula, and depends on the induced voltage. Sufficient control characteristics cannot be obtained only by means for estimating the rotor magnetic flux direction.

  Therefore, as one of the rotor magnetic flux position estimation techniques, there is means for superimposing a high-frequency alternating current or alternating voltage on the estimated rotor magnetic flux direction in the control device. The estimated direction of the rotor magnetic flux is defined as the D axis, the direction of the current vector contributing to the torque is defined as the Q axis, and a current control device that controls each current component is provided.

  In order to estimate the rotor magnetic flux direction, when a high-frequency alternating current is superimposed, a voltage value in the axial direction in which no high-frequency current is passed may be used. That is, the rotor magnetic flux direction can be estimated using the voltage value of the Q axis that superimposes the high frequency current on the D axis and does not superimpose the high frequency current. Further, as an equivalent means, when a high-frequency alternating voltage is superimposed, an axial current value where the high-frequency voltage is not superimposed may be used. That is, it is possible to estimate the rotor magnetic flux position by superimposing a high frequency voltage on the D axis and using a Q axis current value without superimposing the high frequency voltage.

US Patent No. 4763058 JP-A-62-138074 Japanese Patent No. 3454212

Schroedl, M., `` DETECTION OF THE ROTOR POSITION OF A PERMANENT MAGNET MACHINE AT STANDSTILL '', ICEM (International Conference on Electrical Machine) 1988 in Pisa (Italy), p.195-197

  In order to estimate the rotor magnetic flux direction using the high-frequency current that is superimposed on the high-frequency voltage and that flows through the armature current, the high-frequency current component that appears in the D-axis current or the Q-axis current is extracted, and the actual rotor is based on that. It is necessary to obtain information on the error Δθ between the magnetic flux direction and the estimated rotor magnetic flux direction. One of the simplest methods for extracting a high-frequency current component is a method of extracting a difference between when the high-frequency current value is maximum and when it is minimum based on the superimposed high-frequency voltage.

  However, according to this method, the high frequency component is extracted by the current detection value only twice in the high frequency voltage period, and therefore, it is easily affected by the switching noise of the inverter circuit. When an error occurs in one of the two current detection values, the position error information calculated from the high-frequency current component includes a detection error in the high-frequency voltage period, and therefore the estimation result of the rotor magnetic flux direction May have an effect.

  In particular, when the inductance is large, if the frequency of the superimposed high-frequency voltage is increased, the high-frequency current component is reduced, so that the frequency of the superimposed high-frequency voltage cannot be increased. In this case, the time until the next maximum or minimum detection value of the high-frequency current is obtained is increased. As a result, when the estimation error of the rotor position becomes large, there is a possibility that the torque cannot be maintained and a step-out occurs.

  A problem to be solved by the present invention is to provide a controller for a permanent magnet synchronous motor that is less susceptible to the detection error of a high frequency current in a configuration in which a high frequency current is applied by applying a high frequency voltage to estimate a magnetic pole position. That is.

In the control device for a permanent magnet synchronous motor of the embodiment, the estimated axis of the rotor magnetic flux of the permanent magnet synchronous motor is the D axis, the axis at a position advanced by π / 2 with respect to the D axis is the Q axis, and the permanent magnet synchronous motor A high frequency voltage is applied to the rotor to cause a high frequency current to flow to estimate the magnetic pole position of the rotor, and the permanent magnet synchronous motor is controlled based on the estimated magnetic pole position. The current detection means detects the current flowing through the permanent magnet synchronous motor at a predetermined current detection cycle. The high frequency voltage application means selects a half cycle of the applied high frequency voltage to be N times the current detection cycle (N = 1, 2,...), And sets the phase of the high frequency voltage to (2π) ÷ (2N). Treated as a discrete phase as a unit, the voltage component of the sinusoidal function of this discrete phase is superimposed on the D-axis voltage. The first position error information extraction means multiplies at least the Q-axis current of the detected D-axis current and Q-axis current of the polarity inversion signal of the cosine wave function having a discrete phase, and multiplies the multiplied value by 0 of the high-frequency voltage. To 2Mπ (M = 1, 2,...) To calculate a position error extraction amount including error information of the estimated magnetic pole position. The second position error information extraction unit adds the multiplication value over a phase interval whose phase is different by Mπ with respect to the phase interval, and calculates a position error extraction amount including error information of the estimated magnetic pole position. The convergence calculation means cancels the error of the estimated magnetic pole position based on the position error extraction amount obtained from the first position error information extraction means and the position error extraction amount obtained from the second position error information extraction means. Thus, the output frequency is converged. The first position error information extraction unit and the second position error information extraction unit sequentially add the multiplication values to the same memory area when adding the multiplication values over each phase interval of the high frequency voltage.

The block block diagram of the motor control apparatus which shows 1st Embodiment. Diagram showing superimposed sinusoidal function The figure which shows the relationship between MT axis coordinates and DQ axis coordinates The figure which shows the signal value and calculation value which are used in the process of extracting the high frequency component ΔIqh by superimposing the pulsed high frequency voltage vd (k) FIG. 4 equivalent view showing the second embodiment FIG. 4 equivalent view showing the third embodiment

(First embodiment)
The first embodiment will be described with reference to FIGS. 1 to 4. FIG. 1 shows a configuration of an electric motor control device having a function of estimating a rotor magnetic pole position. The motor control device 1 is a sensorless control device that drives the permanent magnet synchronous motor 2 without using a speed sensor and a position sensor. The inverter 3 is composed of six switching elements connected in a three-phase bridge. The inverter 3 converts a DC voltage between the DC power supply lines 4p and 4n into a three-phase AC voltage based on the command voltages uref, vvref, and vwref. Supply. The current detectors 5 arranged in the U phase and the W phase are current detection means for detecting the current flowing through the electric motor 2.

  The motor control device 1 uses the estimated axis of the rotor magnetic flux (estimated rotor magnetic flux direction) as the D axis, and sets the axis in the torque current direction at a position advanced by π / 2 with respect to the D axis as the Q axis. Then, the rotor magnetic pole position θest (estimated magnetic pole position) that is the direction of the rotor magnetic flux is estimated by applying a high frequency voltage vdhf to the motor 2 and causing a high frequency current to flow through the stator winding, and based on this estimated magnetic pole position. To control the motor 2.

  The coordinate conversion unit 6 converts the detected currents iu and iw on the three-phase stationary axis into the detected currents id and iq on the synchronous DQ axis by using the estimated rotor magnetic pole position θest. The current control unit 7 outputs the command voltages vdref and vqref, for example, by PI calculation with respect to the current deviation so that the command currents idref and iqref and the detected currents id and iq match.

  An adder 8 serving as a high-frequency voltage application means adds the high-frequency voltage vdhf to the command voltage vdref that is the current control result of the D axis in order to estimate the rotor magnetic pole position θest. The superposed high-frequency voltage vdhf is a sinusoidal function. Similar to the sine wave function, the sine wave function is a function having a positive value in the period from 0 to π and a negative value in the period from π to 2π, as in the case of the high frequency voltage vdhf. As an example of a sine wave function, as shown in FIG. 2, (a) a sine wave voltage (sine wave function), (b) a 50% duty pulse voltage having two different voltage levels (for example, positive and negative directions) Square wave voltage (square wave function)), (c) triangular wave voltage (triangular wave function), (d) sawtooth voltage (sawtooth wave function), and the like. In this embodiment, as will be described later, a sine wave voltage and a square wave voltage (pulse voltage) will be described as an example.

  The response of the current control is adjusted as a response sufficiently slower than the frequency of the superimposed high frequency voltage vdhf. That is, the high frequency mentioned here is a sufficiently high frequency with respect to the operating frequency of the electric motor 2 and a sufficiently high frequency with respect to the frequency band of current control.

  The coordinate conversion unit 9 converts the command voltages vdref and vqref on the synchronous DQ axis into command voltages vuref, vvref and vwref on the three-phase stationary axis using the estimated rotor magnetic pole position θest. The command voltages vuref, vvref, and vwref are compared with a carrier wave formed of a triangular wave and subjected to PWM modulation, and the PWM-modulated drive signal is applied to the switching element of the inverter 3.

  As will be described in detail later, the position error information extraction unit 10 (position error information extraction means) multiplies the polarity inversion signal of the cosine wave function by the D axis detection current id and the Q axis detection current iq, and respectively multiplies the multiplied values by high frequency. A position error extraction amount (corresponding to high-frequency components ΔIdh and ΔIqh described later) including the error Δθ of the estimated rotor magnetic pole position θest is calculated by averaging over M periods (M = 1, 2,...) Of the voltage vdhf. .

  The convergence calculation unit 11 (convergence calculation means) obtains an error Δθ of the rotor magnetic pole position θest based on the position error extraction amount, and performs a convergence calculation on the rotor speed ωstat (output frequency) so as to cancel this error Δθ. The integrator 12 integrates the rotor speed ωstat to obtain the rotor magnetic pole position θest. The above-described processing such as coordinate conversion, current control, high-frequency voltage superposition, PWM modulation, position error extraction amount calculation, convergence calculation, and the like is executed by a microcomputer in accordance with a control program stored in advance in a nonvolatile memory. It has become.

Hereinafter, the estimation calculation of the rotor magnetic pole position θest by the position error information extraction unit 10 and the convergence calculation unit 11 will be described in detail. In a mathematical model of a rotating machine having a magnetic flux in the rotor, when the magnetic flux direction of the rotor is defined as the M axis and the position advanced by π / 2 from the M axis is defined as the T axis, generally, Indicated. iM and iT are M-axis current, T-axis current, R is the resistance of the stator winding, Ld and Lq are D-axis and Q-axis inductance of the stator winding, φ is the stator flux linkage of the rotor, and ωme is The rotational speed, p is a differential operator.

The M-axis direction, which is the actual magnetic flux direction of the electric motor 2, is defined as θ, and the error angle Δθ between the M-axis and the D-axis (the estimated magnetic pole position error Δθ) is determined as shown in Equation (2). The relationship between the MT axis and the DQ axis is as shown in FIG.

Since the position advanced by π / 2 from the D axis, which is an estimated axis for control, is defined as the Q axis, the voltage equation on the MT axis shown in the equation (1) is a mathematical model on the DQ axis. (3) is expressed as follows.

When the differential term is arranged on the left side, the following equation (4) is obtained.

Here, L0 and L1 are defined as in equation (5).

Since the frequency of the superimposed high-frequency voltage vdhf is sufficiently higher than the operating frequency of the electric motor 2, the voltage drop of the winding resistance is sufficiently smaller than the term related to the inductance. Therefore, ignoring the term related to the resistance R of the stator winding and further ignoring the term including the rotational speed ωme that becomes a small value at the time of stopping or in the low speed region, the equation (4) is an approximation of the equation (6). Can be represented.

If the high-frequency voltage signal is superimposed only in the D-axis direction, vd = vdhf and vq = 0, so that equation (6) becomes equation (7).

When this is expressed for each of the D-axis component and the Q-axis component, the following equations (8) and (9) are obtained.

From this, it can be seen that the D-axis current id and the Q-axis current iq include position error information corresponding to twice the estimated magnetic pole position error Δθ. Therefore, when the high-frequency voltage vdhf superimposed on the D-axis is a sine wave voltage that is a sine wave function, it can be expressed as in equation (10). However, ΔVh represents an amplitude and is a positive value.

At this time, equation (6) becomes equation (11).

In order to consider what kind of current flows when the high-frequency voltage vdhf, which is a sinusoidal function, is superimposed, equation (12) is obtained by integrating both sides of equation (11). id0 and iq0 are initial values at time t = 0 when the integration is started. If only the high-frequency component of equation (12) is written, equation (13) is obtained.

  According to the equation (13), the high-frequency current is a high-frequency current whose phase is delayed by π / 2 with respect to the applied sine wave high-frequency voltage vdhf, in other words, the polarity reversal component of the cosine wave (polarity reversal of the cosine wave function) Signal). According to the equation (5), L0−L1cos (2Δθ) ≧ 0, so that the amplitude of the D axis changes according to the estimated magnetic pole position error Δθ, and not only the amplitude but also the polarity of the Q axis changes. I understand.

If the portion (position error extraction amount) representing the amplitude and polarity of the high-frequency current that changes in accordance with the estimated magnetic pole position error Δθ is expressed as high-frequency components ΔIdh and ΔIqh on the D axis and Q axis, equation (13) becomes (14) It becomes like the formula.

Equation (14) means that the high frequency components ΔIdh and ΔIqh are uniquely determined according to the estimated magnetic pole position error Δθ. If the high-frequency components ΔIdh and ΔIqh can be extracted from the detected current value of the electric motor 2, the relational expression represented by the equation (15) can be obtained.

In equation (15), position estimation may be performed using only the high-frequency component ΔIqh corresponding to the numerator portion that is sinusoidal information with respect to the double angle of the estimated magnetic pole position error Δθ. Further, if approximation of cos2Δθ≈1 and sinΔθ≈2Δθ is performed, Expression (15) can be transformed into Expression (16).

By using the high-frequency component ΔIdh of the D-axis together with the high-frequency component ΔIqh of the Q-axis, information on the estimated magnetic pole position error Δθ can be expressed as shown in Equation (17).

According to the equation (17), when obtaining information on the estimated magnetic pole position error Δθ from information on the high-frequency components ΔIdh and ΔIqh, the convergence calculation for position estimation is performed by using the inductance values Ld and Lq obtained for the motor 2. It can be seen that the gain can be kept optimal. That is, in the PI convergence calculation of the rotor speed ωstat, the gain of the proportional term can be set to 1 (optimum value), and there is an effect that it is not necessary to perform gain adjustment. If only the Q-axis high-frequency component ΔIqh corresponding to the molecular part is used in the equation (17), the high-frequency component ΔIdh is made constant by, for example, ΔIdh = ΔVh / (ωh · Ld). , (17), a multiplication operation can be used instead of the division operation. The relational expression of ΔIdh = ΔVh / (ωh · Ld) described above can be derived by setting Δθ = 0 in the expression of ΔIdh described in the proviso of Expression (14).

  Next, means for extracting the high-frequency components ΔIdh and ΔIqh included in equation (17), which is the equation of the estimated magnetic pole position error Δθ, will be described. The high-frequency current flowing by the high-frequency voltage vdhf applied to the D-axis is a current delayed in phase by π / 2 with respect to the high-frequency voltage signal given by the sine wave function (sinωht component), as shown in equation (14). Become. In other words, it can be seen that the cosine wave function is expressed as a polarity inversion signal (-cosωht component).

  For example, in order to extract the value of the high-frequency component ΔIqh representing the amplitude and polarity of the Q-axis high-frequency current, the detection current iq is multiplied by −cosωht, and is one cycle of the high-frequency voltage vdhf (when M = 1). By integrating only for Th, it can be obtained as in equation (18).

  Although the calculation formula is omitted, the high-frequency component ΔIdh representing the amplitude of the D-axis high-frequency current can be calculated in the same manner by multiplying the detection current id by −cosωht and integrating only for Th.

  The case of superimposing the high-frequency voltage vdhf that is a sine wave function shown in the equation (10) has been described above, but in the following, the high-frequency component ΔIdh, The fact that ΔIqh can be extracted will be described.

As shown in FIG. 2B, the high-frequency voltage vdhf composed of the superimposed sinusoidal function is set to + Δvh when the sinusoidal function is positive (0 to π), and −Δvh () during the negative period (π to 2π). However, the pulse voltage is such that Δvh> 0). That is, the period from time t = 0 to t = Th / 2 is expressed by equation (19), and the period from time t = Th / 2 to t = Th can be expressed by equation (20). After time t = Th, + Δvh is output again. In this way, a case is considered where positive and negative alternating voltages are switched and applied every Th / 2 period.

First, consider how the flowing high-frequency current is represented. The current in the period from time t = 0 to t = Th / 2 is described by equation (21).

When both sides of the equation (21) are integrated, the current in the period from time t = 0 to t = Th / 2 can be expressed as the equation (22). id0 and iq0 are initial values at time t = 0 when the integration is started.

From equation (22), the current value at time t = Th / 2 can be expressed by equation (23).

Subsequently, the current in the period from time t = Th / 2 to t = Th is described by equation (24).

When both sides of the equation (24) are integrated, the current in the period from the time t = Th / 2 to t = Th can be expressed as the equation (25).

  In order to extract the value of the high-frequency component ΔIqh representing the amplitude and polarity of the Q-axis high-frequency current, the Q-axis current iq represented by the equations (22) and (25) is multiplied by −cosωht to obtain the high-frequency voltage vdhf. (26) is obtained by integrating only during Th, which is one period (when M = 1).

であるので、

となる。 here,

So

It becomes.

Therefore, the calculation is performed as follows for each term of the equation (26).

As a result, Equation (26) becomes Equation (26 ′).

  Thus, it can be seen that even when a pulse voltage that is a square wave function is applied as the high-frequency voltage vdhf, the high-frequency components ΔIdh and ΔIqh can be obtained by the same calculation method as the sine wave voltage. Further, since the amplitude of the pulse voltage is made equal to the amplitude of the sine wave, it can be seen that a high-frequency current component that is 4 / π times higher than the effective value of the sine wave of the same amplitude is included.

  As described above, a sine wave voltage or a square wave voltage is applied to the D axis, which is an estimated magnetic flux axis, as the high frequency voltage vdhf of the sine wave function, and the rotor magnetic pole position θest is estimated based on the high frequency current flowing at that time. To do. In this case, the detected D-axis current id and Q-axis current iq are multiplied by the polarity inversion signal −cosωht of the voltage of the cosine wave function (cosine wave voltage), and the addition average value during the M cycles of the high frequency voltage of the multiplied value is obtained. By calculating, it is possible to obtain high frequency components ΔIdh and ΔIqh which are position error extraction amounts. Then, the estimated magnetic pole position error Δθ (the estimated magnetic pole position error information) can be obtained from the high frequency components ΔIdh and ΔIqh. Based on this error information, the rotor speed ωstat can be obtained by performing a convergence operation such as PI compensation on the rotor speed ωstat so as to cancel the estimated magnetic pole position error Δθ. It can be converged to be zero.

When the estimation control of the rotor magnetic pole position described above is actually implemented, it is more convenient to convert it to a sample control system that is a discrete value system and a discrete time system because a digital processor equipped with a microcomputer is used. Is good. Therefore, the control method described above is converted into a sample control system. First, a case where a high-frequency voltage that is a sine wave function is superimposed will be described. The phase of the high-frequency voltage vdhf expressed by the equation (10) is expressed as follows in the discrete time system.

The superposed high-frequency voltage vd (k) can be expressed as in equation (27).

  k is an integer value from 1 to n, and n represents the number of samples (however, even number) during one period of the high frequency component. Here, a case where n = 8 is considered as an example. Specifically, the carrier frequency is 4 kHz, and current is detected for each peak or valley of each carrier cycle. If 500 Hz is selected as the high frequency voltage frequency to be superimposed, one cycle of the high frequency voltage vd (k) to be superimposed is 8 carrier cycles.

The detected D-axis and Q-axis high-frequency currents id (k) and iq (k) are expressed as a sample control system represented by equation (28).

In order to extract the value of the high-frequency component ΔIqh representing the amplitude and polarity of the Q-axis high-frequency current iq (k), the detected Q-axis current iq (k) is expressed as −cos (k ( 2π / n)) is multiplied to calculate the addition average in a section of 1 ≦ k ≦ n, which is one cycle of the high-frequency voltage vd (k) (when M = 1).

  By considering in the same manner as the equation (18) calculated as the continuous time system described above, the high-frequency component ΔIqh of the Q axis can be obtained as the equation (30). Although the result is omitted, the high-frequency component ΔIdh representing the amplitude of the D-axis high-frequency current id (k) can be obtained in the same manner.

Next, a case where a pulsed high-frequency voltage that is a square wave function is superimposed will be described. When superimposing the pulse voltage, it can be expressed by the equation (31) in the sample control system.

The detected current values id (k) and iq (k) in the range of 1 ≦ k ≦ n / 2 can be expressed by the equation (32).

The current value when k = n / 2 + 1, that is, t = Th / 2 is expressed by equation (33) in the same manner as equation (23).

The detected current values id (k) and iq (k) in the range of n / 2 + 1 ≦ k ≦ n can be expressed by the equation (34).

  In order to extract the high-frequency component ΔIqh of the Q axis, the current values id (k) and iq (k) represented by the equations (32) and (34) indicating the detected current value are set to −cos (k (2π / N)) and calculating the addition average in the section of 1 ≦ k ≦ n, which is one cycle of the high-frequency voltage vd (k), the equation (35) is obtained for the Q axis. Although the result is omitted, the D-axis high-frequency component ΔIdh can be obtained in the same manner.

である。 Here, assuming that the error with respect to the continuous operation result is small with respect to the addition operation in equation (35),

It is.

Therefore, equation (35) becomes equation (35 ').

  That is, in the sample control system, the high frequency components ΔIdh and ΔIqh can be extracted in the same manner as in the continuous calculation. In the subsequent convergence calculation, an estimated magnetic pole position error Δθ is calculated based on the equations (15) to (17), and a convergence calculation such as PI compensation is performed on the rotor speed ωstat so as to cancel the estimated magnetic pole position error Δθ. Just do it.

  FIG. 4 shows various signal values and calculated values used in the process of extracting the high frequency component ΔIqh of the Q axis by superimposing the pulsed high frequency voltage vd (k) in the sample control system. As described above, the high-frequency voltage to be superimposed can be handled in the same manner as when it is a sine wave function. FIG. 4A shows a carrier wave. The current detector 5 detects the current of the electric motor 2 in synchronization with this carrier wave. Here, an example is shown in which current is detected at the timing of a carrier wave valley.

  As shown in FIG. 5B, the cycle of the applied high frequency voltage vd (k) is 2N times (N = 1, 2,...) The current detection cycle Tsamp, that is, the high frequency voltage vd (k) The half cycle is selected to be N times the current detection cycle Tsamp. This is for the same number of multiplication values to be added in the positive half cycle and the negative half cycle of the high-frequency voltage vd (k). Accordingly, the high-frequency voltage phase θh (k) shown in FIG. 5C is handled as a discrete phase with (2π) ÷ (2N) as a unit. Note that the number of samples n and N described above have a relationship of n = 2N.

  In this sample control system, the coordinate conversion unit 6 converts the detected current value into a D-axis current id (k) and a Q-axis current iq (k). The position error information extraction unit 10 generates -cos (θh (k)) =-cos (k (2π / n)), which is a polarity inversion signal of a cosine wave having a discrete phase shown in FIG. Then, it is multiplied by the D-axis current id (k) and the Q-axis current iq (k), respectively. Then, the multiplication value (only the Q axis is shown) shown in FIG. 6 (f), which is the multiplication result, is added and averaged over one period (when M = 1) of the high-frequency voltage vd (k), so that the rotor magnetic pole position High-frequency components ΔIdh and ΔIqh including an error Δθ of θest are obtained ((g) and (h) in the figure; only the Q axis is shown).

  In this arithmetic processing, when performing averaging, the current values id (k), iq (k) and the polarity inversion signal of the cosine function are obtained for each discrete phase θh (k) (k = 1, 2,...). The multiplication value with −cos (θh (k)) is not stored in different memories, but is sequentially added to the same memory area every sampling period. Thereby, the amount of volatile memory such as RAM can be saved, and the cost of the digital processing apparatus can be reduced.

  As described above, the motor control apparatus 1 superimposes the high-frequency voltage vd (k) of the sine wave function having the discrete phase θh (k) on the D-axis voltage, and reverses the polarity of the cosine wave function having the discrete phase. The signal -cos (θh (k)) is multiplied by the detected D-axis current id (k) and Q-axis current iq (k), and the multiplied value is M cycles (M = 1, 2) of the high-frequency voltage vd (k). ,...), The high frequency components ΔIdh and ΔIqh including the error information Δθ of the rotor magnetic pole position θest are calculated. Based on the high frequency components ΔIdh and ΔIqh, the rotor speed ωstat is converged so as to cancel the estimated magnetic pole position error Δθ.

  According to this configuration, by extracting the high-frequency components ΔIdh and ΔIqh based on the current detection result for M periods of the superimposed high-frequency voltage vd (k), the amplitude (size) of the estimated magnetic pole position error Δθ and Polarity can be obtained simultaneously. As a result, the estimated magnetic pole position error Δθ can be more reliably and stably converged. Moreover, if the estimated magnetic pole position error Δθ is obtained by the calculation shown in the equation (17), the gain for convergence calculation can be kept optimal, and the gain design for the convergence calculation becomes unnecessary.

  Since the high-frequency components ΔIdh and ΔIqh are extracted using a plurality of current detection results detected during M periods of the high-frequency voltage vd (k), a detection error is included in part of the current detection value due to the influence of noise. Even if this is the case, the influence of the detection error is averaged, and the effect of reducing the influence of noise can be obtained. In the above description, the derivation process for one cycle (M = 1) of the high-frequency voltage vd (k) is shown. In general, the arithmetic mean of the high-frequency voltage vd (k) over M cycles (M = 1, 2,...) Is calculated. May be. By increasing the value of M, the number of samples of the current detection value used for extracting the high frequency components ΔIdh and ΔIqh increases, and the influence of switching noise and the like of the inverter 3 can be further reduced.

  The D-axis and Q-axis high-frequency components ΔIdh and ΔIqh are extracted as position error extraction amounts including error information of the estimated magnetic pole position. The Q-axis of which both the amplitude and polarity of the high-frequency current change according to the estimated magnetic pole position error Δθ. Only the high frequency component ΔIqh may be extracted. In this case, if the estimated magnetic pole position error Δθ is calculated with the D-axis high-frequency component ΔIdh as a constant value, the division operation of equation (17) can be omitted.

(Second Embodiment)
FIG. 5 is a view corresponding to FIG. 4 in the second embodiment. With respect to the first embodiment, a modification is added to the arithmetic processing method of the addition average by the position error information extraction unit 10. When the arithmetic processing method of the first embodiment is used, the high frequency components ΔIdh and ΔIqh, which are position error extraction amounts, are updated every cycle of the high frequency voltage vd (k). For this reason, for example, when the frequency of the high-frequency voltage vd (k) cannot be increased because the inductance of the motor 2 is large and the impedance to the high frequency is large, the position error extraction amount update period becomes long, and the position estimation response is increased. There are concerns that there will be constraints that must be lowered. As a result, the speed response and the response of the position control system may be restricted.

  Therefore, in the present embodiment, as shown in FIGS. 5G and 5H, as an interval for calculating the high frequency component ΔIqh, an addition average is calculated in the interval of 0 to 2π of the high frequency voltage phase θh (k), and The addition average is calculated even in a section shifted by π from a section of 0 to 2π. The same applies to the calculation of the high frequency component ΔIdh. Thereby, the update cycle of the high frequency components ΔIdh and ΔIqh can be set to a half cycle of the high frequency voltage vd (k).

  In general, even when the multiplication values are averaged over M periods (M = 1, 2,...) Of the high-frequency voltage vd (k), the phase interval from 0 to 2Mπ of the high-frequency voltage vd (k) If the high frequency components ΔIdh and ΔIqh are calculated by adding multiplication values for the phase sections whose phases are different by Mπ, the update cycle can be shortened from 2Mπ to Mπ. According to this arithmetic processing method, although the amount of use of the volatile memory holding the addition average value is doubled, the position error extraction amount update cycle can be halved, so that the speed response and The position control response can be improved.

(Third embodiment)
FIG. 6 is a diagram corresponding to FIG. 4 in the third embodiment. Compared to the first and second embodiments, the arithmetic processing method of the addition average by the position error information extraction unit 10 is different. That is, as shown in FIG. 6G, the position error information extraction unit 10 calculates the detected current values id (k), iq (k) and the cosine function polarity inversion signal -cos (θh (k)). Each multiplication value is stored in a separate area of the volatile memory over one period (when M = 1) of the high-frequency voltage vd (k) to be superimposed, and an average of the data group 1 is calculated.

  When the next detected current value was obtained, it was newly obtained instead of the multiplication value (indicated by the broken line in the figure) one cycle before the high frequency voltage vd (k) as shown in FIG. The multiplication value (indicated by hatching in the figure) of the current values id (k), iq (k) and the polarity inversion signal −cos (θh (k)) of the cosine function was stored in the previous cycle. It is stored (overwritten) in the memory area, and the average of the data group 2 is calculated.

  In general, even when the multiplication values are averaged over M periods (M = 1, 2,...) Of the high-frequency voltage vd (k), 2MN multiplications are performed for the phase interval from 0 to 2Mπ of the high-frequency voltage vd (k). The value is memorized and added to calculate the high frequency components ΔIdh and ΔIqh. Thereafter, every time a new multiplication value is calculated, the newly calculated multiplication value is stored and added instead of the previous multiplication value by 2Mπ. Thus, the high frequency components ΔIdh and ΔIqh may be calculated.

  According to this calculation processing method, the memory usage requires an area for holding the multiplication value data for M cycles for which the averaging is performed, but the update cycle of the position error extraction amount is defined as the current detection cycle Tsamp. Since they can be made equal, the speed control response and the position control response can be made high.

(Other embodiments)
In addition to the embodiment described above, the following configuration may be adopted.
A similar extraction method can also be used in the method of calculating the position error extraction amount using the voltage value of the axis on which the high frequency current is not superimposed on the motor 2 and the high frequency current is not superimposed.

  The half cycle of the high frequency voltage to be applied is selected to be N times the current detection cycle (N = 1, 2,...). When N = 1, only two current detection values are included in one cycle of the high frequency voltage. Will be used. Therefore, it is preferable to select N as an integer value of 2 or more.

  According to at least one embodiment described above, the magnitude and polarity of the estimated magnetic pole position error can be obtained simultaneously by extracting a high frequency component based on a current detection result for at least one cycle of the superimposed high frequency voltage. Can do. Further, even if a detection error is included in a part of the current detection value due to the influence of noise, a high frequency component is obtained by using a plurality of current detection results corresponding to the number of times of current detection performed during at least one cycle of the high frequency voltage. Since the extraction is performed, the influence of the detection error is averaged, and the effect of reducing the influence of noise is obtained.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

  In the drawings, 1 is a motor control device (control device for a permanent magnet synchronous motor), 2 is a permanent magnet synchronous motor, 5 is a current detector (current detection means), 8 is an adder (high frequency voltage application means), and 10 is a position. An error information extraction unit (position error information extraction unit) 11 is a convergence calculation unit (convergence calculation unit).

Claims (4)

The estimated axis of the rotor magnetic flux of the permanent magnet synchronous motor is D axis, the axis that is π / 2 advanced from the D axis is Q axis, and a high frequency current is applied by applying a high frequency voltage to the permanent magnet synchronous motor. In the control device of the permanent magnet synchronous motor that estimates the magnetic pole position of the rotor and controls the permanent magnet synchronous motor based on the estimated magnetic pole position,
Current detection means for detecting the current flowing in the permanent magnet synchronous motor at a predetermined current detection period;
The half cycle of the high frequency voltage to be applied is selected to be N times the current detection cycle (N = 1, 2,...), And the phase of the high frequency voltage is discrete in units of (2π) ÷ (2N). High-frequency voltage applying means for treating the voltage component of the discrete phase sinusoidal function on the D-axis voltage,
The polarity inversion signal of the discrete phase cosine wave function is multiplied by at least the Q-axis current of the detected D-axis current and Q-axis current, and the multiplication value is 0 to 2Mπ (M = 1, 2,..., And first position error information extraction means for calculating a position error extraction amount including error information of the estimated magnetic pole position;
Second position error information extraction means for calculating the position error extraction amount including error information of the estimated magnetic pole position by adding the multiplication value over a phase period having a phase different by Mπ with respect to the phase interval;
Based on the position error extraction amount obtained from the first position error information extraction means and the position error extraction amount obtained from the second position error information extraction means, the error of the estimated magnetic pole position is canceled out. And a convergence calculation means for calculating a convergence of the output frequency ,
The first position error information extraction unit and the second position error information extraction unit sequentially add the multiplication values to the same memory area when adding the multiplication values over the phase intervals of the high-frequency voltage. A control device for a permanent magnet synchronous motor. The first position error information extraction means and the second position error information extraction means multiply the detected D-axis current and Q-axis current by the polarity inversion signals of the discrete phase cosine wave functions, respectively. A position error extraction amount ΔIdh is calculated by adding the multiplication value of the D-axis current over each phase interval of the high-frequency voltage, and the position is determined by adding the multiplication value of the Q-axis current over each phase interval of the high-frequency voltage. Calculate the error extraction amount ΔIqh,
The convergence calculation means includes:
Lq / (Ld−Lq) × (ΔIqh / ΔIdh)
2. The control apparatus for a permanent magnet synchronous motor according to claim 1, wherein Ld is D-axis inductance and Lq is Q-axis inductance to calculate the error [Delta] [theta] of the estimated magnetic pole position. The first position error information extraction unit and the second position error information extraction unit multiply the detected Q-axis current by a polarity inversion signal of the cosine wave function having the discrete phase, respectively, and the Q-axis current To calculate the position error extraction amount ΔIqh by adding the multiplication value of the high frequency voltage over each phase interval of the high frequency voltage,
The convergence calculation means includes:
Lq / (Ld−Lq) × (ΔIqh / ΔIdh)
However, Ld is a D-axis inductance, Lq is a Q-axis inductance, ΔIdh = ΔVh / (ωh · Ld), ΔVh is an amplitude of the high-frequency voltage, and ωh is an error Δθ of the estimated magnetic pole position based on an angular frequency of the high-frequency voltage. The control device for a permanent magnet synchronous motor according to claim 1. The control apparatus for a permanent magnet synchronous motor according to any one of claims 1 to 3, wherein the sine wave function is a sine wave function, a square wave function, a triangular wave function, or a sawtooth function .

Images