JP2007124835A - Method of estimating rotating angle of synchronous machine having saliency - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce magnetic noise and vibration and reduce power loss while securing the accuracy in judgement of a magnetic pole and reducing errors in the estimation of a rotor position. <P>SOLUTION: A high frequency current controller 209 for judgement of magnetic poles calculates the difference between the maximum amplitude value of a high frequency current at the position of a magnetic pole in specified polarity and the maximum amplitude value of a high frequency current at the position of a magnetic pole in opposite polarity, and adjusts the amplitude of high frequency voltage so that this difference may be within a suitable range. Hereby, it can reduce magnetic noise and power loss while securing the accuracy in judgement of magnetic polarity. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a method for estimating a rotation angle of a synchronous machine having a saliency that detects a rotor position of a synchronous machine having a magnetic saliency by superimposing a high-frequency voltage on an armature coil.

  For example, in a brushless DC motor, in order to detect a rotor position (also referred to as a magnetic pole position or a rotation angle, or a rotor position or a rotor angle), sensorless rotation is used to estimate the rotor position using an armature coil current or voltage. An angle estimation method has been proposed.

  However, in this sensorless rotation angle estimation method, when the current or voltage of the armature coil is used, the rotor position detection accuracy decreases when the rotor is stopped or when the rotor is rotating at a low speed. For this reason, for example, the following Patent Documents 1 to 3 propose that the rotor position is estimated by superimposing a high-frequency voltage for estimating a rotation angle higher than the frequency (also referred to as drive frequency) on the drive voltage. Hereinafter, this method is referred to as a high frequency voltage superposition method. In the rotor position estimation by this high-frequency voltage superposition method, the rotor position detector can be omitted in a synchronous machine that needs to detect the rotor position, and the rotor position detection accuracy is ensured well even in a stationary state or a low-speed rotation state. be able to. This high-frequency voltage superposition method utilizes the phenomenon that the high-frequency current flowing in the armature coil due to the high-frequency voltage superimposed on the armature coil changes periodically due to the period change of the armature coil inductance of the salient pole type synchronous machine due to the rotor rotation. is doing.

  The conventional high frequency voltage superposition method will be described more specifically.

  In the salient pole type synchronous machine, the inductance of the armature coil is minimized in the d-axis direction on the rotation orthogonal coordinate (dq axis coordinate) system. Therefore, when a high-frequency voltage rotating at a predetermined frequency with a constant amplitude is applied to the armature coil, the locus of the high-frequency current on the stationary orthogonal coordinate (αβ-axis coordinate) system at that time becomes an elliptical shape. Hereinafter, this ellipse is also referred to as a current ellipse locus. The apex of the high frequency current rotates on the current ellipse locus at the frequency of the high frequency voltage (and high frequency current). Further, the current elliptical locus rotates on the stationary orthogonal coordinate system at the driving frequency. Since the major axis of the current elliptical locus is the d-axis direction, the rotor position θ with reference to the α axis can be estimated by calculating the phase between the α axis and the major axis of the stationary orthogonal coordinate system.

  FIG. 3 shows the locus of the high-frequency voltage Vh and the locus of the high-frequency current Ih on the stationary orthogonal coordinate system. The locus of the high frequency voltage Vh is circular, and the locus of the high frequency current Ih is elliptic. In FIG. 3, the rotor position is determined as the d-axis rotor position θ with respect to the α-axis. However, in the following description, the d-axis refers to the direction of the N-pole magnet center as viewed from the rotor rotation axis center. If the α-axis component of the maximum value Imax of the amplitude of the high-frequency current Ih in the current elliptical locus of one cycle of the high-frequency voltage is Imaxα and the β-axis component of the maximum value Imax of the high-frequency current is Imaxβ, the formula tanθ = Imaxβ / Imaxα is established. Therefore, the rotor position θ can be calculated from the detected Imaxα and Imaxβ.

  However, when the current elliptical locus has a symmetric shape as shown in FIG. 3, in one cycle of the high frequency voltage, the maximum amplitude value (simply the maximum value) of the two high frequency currents Ih in the major axis direction for one cycle of the current elliptical locus. Also called Imax and Imax ′. Imax is the maximum value on the positive side in the major axis direction, and Imax 'is the maximum value on the negative side. That is, two maximum values of the high-frequency current Ih are extracted from one cycle of the current elliptical locus.

  However, since the angle between the N-pole direction of the magnet and the α axis is normally adopted as the rotor position, only the value of the maximum value Imax of the high-frequency current Ih in the major axis direction N-pole direction of the current elliptical locus is Used for the rotor position θ calculation. In this case, if the maximum value Imax ′ of the high-frequency current in the major axis direction S pole of the current elliptical locus is always smaller than the maximum value Imax of the high-frequency current in the major axis direction N pole of the current elliptical locus, the major axis direction N It should be possible to discriminate the maximum value Imax of the pole-oriented high-frequency current and use it reliably in the calculation of the rotor position θ.

  Conveniently, in the PM type synchronous machine (magnet rotor type synchronous machine), when the magnetic circuit has a magnetic saturation tendency with respect to the high frequency current, the direction of the magnetic flux due to the high frequency current (also referred to as current elliptical locus magnetic flux) and When the directions of the magnetic fluxes of the rotor coincide with each other, the armature coil inductance decreases due to the influence of magnetic saturation as compared with the case where these directions are reversed, and the major axis of the current elliptical locus increases accordingly. That is, as shown in FIG. 4, the current ellipse trajectory has a distorted shape extending in the major axis direction N-pole direction due to magnetic saturation. Therefore, the maximum value is obtained from each data of the high-frequency current Ih for one period of the current ellipse trajectory. If selected, the maximum value Imax ′ in the major axis direction S pole is smaller than the maximum value Imax in the major axis direction N pole, and therefore can be easily eliminated.

The estimation of the rotor position θ by the above-described high-frequency voltage superposition method and the discrimination on one side in the major axis direction (N pole side) using magnetic saturation are already well-known matters, so further explanation will be omitted. And
JP-A-8-205578 JP 2004-129430 A JP 2005-237172 A

  In the determination of the magnetic pole position indicated by the maximum value of the current elliptical locus using the magnetic saturation phenomenon of the conventional synchronous machine having the saliency described above, the difference between the two current amplitude maximum values in the major axis direction of the current elliptical locus is small. In some cases, the magnetic pole determination accuracy is lowered, and an undesirable erroneous determination may occur. In order to solve this problem, it is only necessary to increase the high-frequency voltage and supply a large high-frequency current to reinforce magnetic saturation that occurs at the magnetic pole position on one side. However, such an increase in high-frequency current causes an increase in magnetic noise and power loss, which is not preferable.

  The present invention has been made in view of the above problems, and provides a rotational angle estimation method for a synchronous machine having a saliency capable of improving rotor position estimation accuracy while suppressing an increase in power consumption and magnetic noise. Is a problem to be solved.

  The rotation angle estimation method for a synchronous machine having a saliency according to the first and second aspects of the present invention solves the above problem, and has a stationary coordinate system having a frequency higher than a driving voltage frequency of the synchronous machine having a saliency and a constant amplitude. A high-frequency voltage superimposing means for superimposing a high-frequency voltage rotating on the armature coil of the synchronous machine; a high-frequency current detecting means for detecting a high-frequency current flowing in the armature coil by the high-frequency voltage; and the detected high-frequency current A magnetic pole position by comparing the maximum amplitude value of the high-frequency current at the magnetic pole position of a predetermined polarity and the maximum amplitude value of the high-frequency current at the magnetic pole position of the opposite polarity. The present invention is applied to a method for estimating the rotation angle of a synchronous machine having saliency with a magnetic pole determining means for determining. Such devices are known.

  Particularly in the first invention, the high-frequency voltage superimposing means calculates a difference between the maximum amplitude value of the high-frequency current at the magnetic pole position of the predetermined polarity and the maximum amplitude value of the high-frequency current at the magnetic pole position of the opposite polarity, and the difference is predetermined. It is characterized in that the amplitude of the high frequency voltage is adjusted so as to be equal to or greater than the threshold value. In this way, the difference between the amplitude values can be reliably maintained at a required level, so that accurate magnetic pole determination can always be performed when magnetic pole determination is necessary.

  In a preferred aspect, the high-frequency voltage superimposing means adjusts the amplitude of the high-frequency voltage so that the difference is equal to or smaller than a second predetermined threshold value that is greater than the predetermined threshold value. In this way, since a high-frequency current does not flow more than necessary, magnetic noise and power consumption can be reduced.

  In a preferred aspect, the high frequency voltage superimposing means performs adjustment for a predetermined adjustment period when the synchronous machine is started. In this way, it is possible to reliably perform the magnetic pole determination at the start-up in which the past determination result cannot be used and the necessity of magnetic pole determination is the most important. After startup, the magnetic pole position having the maximum value of the high-frequency current may be determined by following the past magnetic pole determination history. Thereby, it is not necessary to increase the high-frequency current for determining the magnetic poles other than at the time of start-up, so that magnetic noise and power consumption can be reduced.

  In a preferred aspect, the high-frequency voltage superimposing means performs the adjustment for a predetermined adjustment period every predetermined period during operation of the synchronous machine. In this way, step-out due to disturbance can be prevented by performing the magnetic pole determination every predetermined period even during motor driving.

  In a preferred aspect, the predetermined period is shortened when the drive voltage frequency is high or when the change in the drive voltage frequency is large. In this way, the magnetic pole determination is performed with high frequency only when there is a high possibility of erroneous determination of the magnetic pole, so that magnetic pole determination accuracy can be ensured while suppressing an increase in magnetic noise and power loss.

  In a preferred embodiment, this adjustment period is set to approximately 2 to 10 cycles of the high frequency voltage. In this way, it is possible to reliably determine the magnetic pole and not to use a large high-frequency current for a long time, so that magnetic noise and power consumption can be reduced.

  In a preferred aspect, the increase rate of the high-frequency voltage at the start of the adjustment period is set to be larger than the decrease rate of the high-frequency voltage at the stop of the adjustment period. Data for one cycle after the start-up can be discarded, but cannot be discarded after the start-up. Therefore, by gradually changing the data, the current stability can be improved and the detection error can be reduced.

  In a preferred embodiment, when the difference between the maximum amplitude value of the high-frequency current at the magnetic pole position of the predetermined polarity and the maximum amplitude value of the high-frequency current at the magnetic pole position of the opposite polarity is less than a predetermined threshold, the magnetic pole position determination is stopped. The magnetic pole position determination is permitted when the difference between the maximum amplitude values exceeds a predetermined threshold value. When the difference is equal to or smaller than the predetermined value regardless of the adjustment, the magnetic pole determination is stopped, so that a magnetic pole determination error can be prevented.

  In the second invention, in particular, only in a predetermined magnetic pole determination period after the start of the synchronous machine, the high frequency voltage superimposing means is instructed to superimpose the high frequency voltage, and the magnetic pole determining means is instructed to determine the magnetic pole position. It is characterized by having magnetic pole determination high-frequency current control means for increasing the amplitude value of the high-frequency voltage during the period more than the amplitude value of the high-frequency voltage during the period other than the magnetic pole determination period.

  It should be noted that the magnetic pole determination result determined in the immediately preceding magnetic pole determination period may be inverted every time the maximum value of the high-frequency current is detected and used as the current value of the magnetic pole determination result during the period when the magnetic pole determination is not performed. As the magnetic pole determination period, at the time of start-up is preferable, and thereafter, it is preferable to carry out at a predetermined interval. In this way, it is possible to achieve both improvement in magnetic pole determination accuracy and reduction in magnetic noise and power loss.

  Hereinafter, preferred embodiments of a high-frequency superposition type rotation angle detection device for a synchronous machine according to the present invention will be described with reference to the drawings. However, the present invention should not be construed as being limited to the following embodiments, and it goes without saying that the technical idea of the present invention may be implemented by a combination of other techniques.

(Embodiment 1)
A motor control apparatus provided with the rotation angle estimation method for a synchronous machine having saliency according to the first embodiment will be described with reference to a block circuit diagram shown in FIG.

(Circuit configuration)
In FIG. 1, 1 is a current command unit, 2 and 3 are subtractors, 4 and 5 are PI amplifiers, 6 is a dq / 3-phase coordinate conversion unit, 7 to 9 are adders, 10 is a PWM voltage generation circuit, 11 is Three-phase inverters, 12 and 13 are current sensors that detect phase currents, M is a synchronous machine having saliency, 14, 15 are low-pass filters, 16 are three-phase / dq coordinate converters, 17 and 18 are band-pass filters, Reference numeral 19 denotes a three-phase / two-phase coordinate conversion unit, and reference numeral 20 denotes a rotor position estimation unit which forms a main part of this embodiment. Current command unit 1, subtracters 2 and 3, PI amplifiers 4 and 5, dq / 3 phase coordinate conversion unit 6, PWM voltage generation circuit 10, three phase inverter 11, current sensors 12 and 13, 3 phase / dq coordinate conversion unit The current control unit 16 is the same as the conventional configuration and performs the same motor control operation as the conventional one.

  The band-pass filters 17 and 18, the three-phase / two-phase coordinate conversion unit 19, and the rotor position estimation unit 20 constitute a high-frequency voltage superimposition type rotor position estimation system that characterizes the present invention. This rotor position estimation system functions to detect the position of the rotor of the synchronous machine 13 having saliency. In short, this rotor position estimation system uses three-phase high-frequency voltage commands Vuh, Vvh, Vwh to be superimposed on armature coils of the synchronous machine M having saliency as drive voltage commands by adders 7-9. Are added to the three-phase voltage commands Vu ′, Vv ′, Vw ′, and the rotor position θ is estimated from the high-frequency currents iuh, ivh generated thereby.

  In this embodiment, the detection currents iu and iv of the current sensors 12 and 13 include the high-frequency currents iuh and ivh by the high-frequency voltage commands Vuh, Vvh, and Vwh superimposed on the three-phase voltage commands Vu ′, Vv ′, and Vw ′. Therefore, in order to prevent this from flowing to the current control system, low-pass filters 14 and 15 are provided in front of the three-phase / dq coordinate conversion unit 16. The low-pass filters 14 and 15 may be provided on the output side of the three-phase / dq coordinate conversion unit 16 or may be omitted in some cases.

(Basic control explanation)
First, motor drive control based on a torque command will be described. However, since this control itself is the same as the conventional one, the description will be given simply. The low frequency components are extracted by the low-pass filters 14 and 15 from the U-phase current iu detected by the current sensors 12 and 13 and the V-phase current iv and sent to the three-phase / dq coordinate conversion unit 16. The three-phase / dq coordinate conversion unit 16 converts the input signal from the stationary three-phase coordinate system to the dq-axis rotation orthogonal coordinate system, and outputs currents iq and id.

  The current command unit 1 forms current command values id ′ and iq ′ based on the torque command input from the outside and the rotor position θ input from the rotor position estimation unit 20. The detected d-axis current id and q-axis current iq are sent to the subtracters 2 and 3, and the subtracters 2 and 3 obtain the d-axis current id from the d-axis current command value id 'and the q-axis current command value iq'. The q-axis current iq is subtracted.

  The current deviations Δid and Δiq output from the subtracters 2 and 3 are PI-converted by the PI amplifiers 4 and 5 into voltage commands Vd and Vq that make these current deviations Δid and Δiq 0, and the voltage commands Vd and Vq are dq / 3. After the coordinate conversion to the three-phase voltage commands Vu ′, Vv ′, Vw ′ by the phase coordinate conversion unit 6, the final voltage command is added by high frequency voltage commands Vuh, Vvh, Vwh (to be described later) by the adders 7 to 9. Vu, Vv, and Vw. The final voltage commands Vu, Vv, and Vw are converted into PWM voltages by the PWM voltage generation circuit 10 to drive the three-phase inverter 11, whereby the three-phase current iu, iv and iw are energized.

(Explanation of basic operation for rotor position estimation)
The band-pass filters 17 and 18 extract narrow-band current components including high-frequency voltage commands Vuh, Vvh, and Vwh, that is, high-frequency currents iuh and ivh, from the phase currents iu and iv, and a three-phase / two-phase coordinate conversion unit. Send to 19. The three-phase / two-phase coordinate conversion unit 19 converts the input high-frequency currents iuh and ivh into currents iα and iβ on the two-phase stationary orthogonal coordinate system, and sends the currents iα and iβ to the rotor position estimation unit 20. . The rotor position estimation unit 20 estimates the rotor position θ from the input currents iα and iβ, and forms high-frequency voltage commands Vα and Vβ on the two-phase stationary orthogonal coordinate system, which are converted into two-phase / 3-phase. The coordinates are converted into three-phase high-frequency voltage commands Vuh, Vvh, Vwh, and then sent to adders 7-9. The rotor position estimating unit 20 may directly form the three-phase high-frequency voltage commands Vuh, Vvh, Vwh.

(Description of Rotor Position Estimation Unit 20)
Next, the configuration and operation of the rotor position estimation unit 20 will be described in more detail with reference to FIG.

  The rotor position estimation unit 20 includes a drive frequency detection unit 201 that detects a drive voltage frequency f, a high frequency voltage generation unit 202 that generates high frequency voltage commands Vuh, Vvh, and Vwh, and a carrier voltage that generates a PWM carrier voltage Vc that is a triangular wave. The generator 203, the high frequency currents iuh and ivh output from the bandpass filters 17 and 18 are sampled and held for a predetermined current detection period and a predetermined sampling period, and the sample hold is performed only during the immediately preceding current detection period. A maximum amplitude calculator 205 that detects the maximum amplitude value from the predetermined number of current value data, and a rotor position calculator 206 that estimates the rotor position θ based on the maximum amplitude value detected by the maximum amplitude calculator 205. The magnetic pole judgment for determining whether or not the maximum amplitude value detected by the maximum amplitude calculator 205 is on the N pole side. Part 207 has a rotor position correcting unit 208 for correcting the rotor position θ on the basis of the rotor position θ and the magnetic pole determination result.

(Description of operation)
(Explanation of rotor position estimation operation)
First, the estimation operation of the rotor position θ by the rotor position estimation unit including the current sample hold unit 204, the maximum amplitude calculation unit 205, the rotor position calculation unit 206, the magnetic pole determination unit 207, and the rotor position correction unit 208 will be described below. explain.

  First, the high frequency current Ih (= iα + jiβ) is sampled and held at a predetermined sampling timing by the current sample hold unit 204. As the sampling timing, the maximum point and the minimum point of the triangular wave voltage which is the PWM carrier voltage output to the PWM voltage generation circuit 10 are employed. Of the data of the sampled and held high-frequency current Ih, the data for the half cycle of the immediately preceding high-frequency voltage Vh is output to the maximum amplitude calculation unit 205 for each half cycle of the high-frequency voltage Vh. The maximum amplitude calculation unit 205 selects the maximum value as the maximum value Imax from the input data for the current elliptical locus half cycle. Each data is vector data including time information. The extracted maximum value Imax is input to the rotor position calculation unit 206. The rotor position calculation unit 206 divides the maximum value Imax into an α-axis component Imaxα and a β-axis component Imaxβ, and these α-axis component Imaxα and β-axis The rotor position θ is estimated from the component Imaxβ by the above-described formula.

(Description of magnetic pole judgment operation)
Next, the magnetic pole determination unit 207 performs magnetic pole determination to determine whether the maximum value Imax extracted from each data for the current half cycle of the current elliptical locus is in the major axis direction N-pole direction of the current elliptical locus. There are various methods for this magnetic pole determination. In this embodiment, the magnetic pole determination operation described later is performed at start-up or at a predetermined timing to determine the magnetic pole, and this determination result can be rephrased every half cycle of the current elliptical locus. Inverted and stored every high-frequency voltage half cycle. A magnetic pole determination operation performed by the magnetic pole determination unit 207 at startup or at a predetermined timing will be described below. In this operation, the maximum value Imax obtained from the previous half cycle of the current elliptical locus is compared with the maximum value Imax obtained from the half cycle of the current elliptical locus that is one half cycle before, and the larger value is compared with the N pole side, that is, the d axis. Position direction. The magnetic pole determination result is output to the rotor position correction unit 208.

  Based on the rotor position θ input from the rotor position correction unit 208 and the magnetic pole determination result input from the magnetic pole determination unit 207, the rotor position correction unit 208 determines that the rotor position θ is in the normal d-axis direction (N pole If the rotor position θ is the rotor position θ in the opposite d-axis direction (S pole), add π to the rotor position θ. And output as a typical rotor position θ.

  For the magnetic pole determination, various other methods can be considered. For example, current data for one cycle of a current elliptical locus is acquired at a constant interval during start-up or motor operation, and the maximum value Imax is extracted therefrom, thereby obtaining the normal rotor position θ. In this case, a high-frequency current Ih having a magnitude that causes magnetic saturation in the normal d-axis direction is supplied. Thereby, the rotor position θ with respect to the correct magnetic pole direction can be obtained. After that, π is added to the rotor position θ calculated from the maximum value Imax obtained in the odd-numbered half cycle of the current elliptical locus, and calculation is performed from the maximum value obtained in the even-numbered half cycle of the current elliptical locus. Π is not added to the rotor position θ. As a result, the position can be estimated accurately despite the rotor position being estimated every half cycle of the current elliptical locus. The above rotor position estimation operation is the same as the conventional one except that it is performed using current data for one half of the current elliptical locus instead of current data for one period of the current elliptical locus. Is omitted. The circuit for performing the rotor position estimation operation may be configured by a hardware circuit or may be performed by software calculation.

(Description of high-frequency voltage generation operation)
Next, generation control of the high frequency voltage Vh and the PWM carrier voltage performed by the drive frequency detection unit 201, the high frequency voltage generation unit 202, and the carrier voltage generation unit 203 will be described below.

  The drive frequency detector 201 detects the frequency, that is, the drive frequency f from the drive voltage. The driving frequency f can be detected from various signals. For example, the phase current output from the low-pass filter 14 or 15 can be extracted by the zero crossing of the low-frequency component iul or ivl. In addition, it can be extracted from the rotor position θ output by the rotor position estimation unit 20. The high frequency voltage generation unit 202 generates a three-phase high frequency voltage command Vh having a frequency fh that is positively correlated with the drive frequency f detected by the drive frequency detection unit 201. The high-frequency voltage command Vh includes a U-phase high-frequency voltage command Vuh, a V-phase high-frequency voltage command Vvh, and a W-phase high-frequency voltage command Vwh. The amplitudes of the high-frequency voltage commands Vuh, Vvh, Vwh are constant unless there is a special command. Therefore, the frequency of the high-frequency current Ih is fh. The carrier voltage generation unit 203 also generates a triangular wave voltage having a frequency fc positively correlated with the drive frequency f detected by the drive frequency detection unit 201 as the carrier voltage of the PWM voltage generation circuit 10. Here, the drive frequency f, the frequency fh of the high-frequency voltage Vh, and the frequency fc of the carrier voltage are such that the error of the rotor position θ output from the rotor position estimation unit 20 is less than a predetermined allowable level, and more preferably a minimum value. Is set to be

(Description of high-frequency current control operation for magnetic pole determination)
Next, the magnetic pole determination high-frequency current control operation performed by the magnetic pole determination high-frequency current control unit 209 will be described.

  This high-frequency current control operation for magnetic pole determination has been performed by paying attention to the problem that magnetic noise increases if the magnetic circuit of the motor M is sufficiently magnetically saturated in order to accurately perform the magnetic pole determination described above. is there. It has been found that this magnetic noise is caused in particular by periodic vibrations of magnetic force acting in the radial direction on the teeth of the stator core. However, when the amplitude of the high-frequency current Ih is reduced, the difference ΔI between the maximum amplitude values Imax and Imax ′ of the high-frequency current Ih on both sides in the major axis direction of the current elliptical locus shown in FIG. Resulting in.

  In order to improve this problem, the magnetic pole determination high-frequency current control unit 209 mainly uses two solutions. One of them is feedback control of the difference ΔI to the minimum level that can ensure the magnetic pole determination accuracy. The other is that the magnetic pole determination operation is not always performed, but only during a specific period, and the high-frequency current Ih is increased during this specific period in order to improve the magnetic pole determination accuracy. It is to reduce Ih.

  Hereinafter, an example of a specific control operation of the magnetic pole determination high-frequency current control unit 209 will be described with reference to the flowcharts shown in FIGS. This control is performed, for example, by interrupt processing at a predetermined short interval.

  First, it is checked whether or not an activation command for the motor M has been input, and if activated, the magnetic pole determination should be performed and the process proceeds to step S204 (step S200). Otherwise, it is determined whether or not the magnetic pole determination period has arrived (step S200). If it is determined that it has arrived (S202), the process proceeds to step S104, and if not, the process proceeds to step S214. The magnetic pole determination period here refers to a period for magnetic pole determination that is set to a predetermined length of time at a predetermined magnetic pole determination interval.

  In step S204, the high frequency voltage generator 202 is instructed to superimpose the high frequency voltage Vh, the magnetic pole determination unit 207 is instructed to determine the magnetic pole, and the maximum amplitude calculation unit 205 uses the two maximum amplitude value data immediately before the high frequency current Ih. A certain maximum amplitude value Imax, Imax ′ (see FIG. 4) is read. In this embodiment, since the maximum amplitude value Imax or Imax ′ is detected once in the half cycle of the current elliptical locus, the maximum amplitude values Imax and Imax ′ to be read are the maximum amplitude value obtained in the immediately preceding half cycle, and further And a maximum amplitude value before a half cycle.

  Next, a difference ΔI between the read maximum amplitude values Imax and Imax ′ is calculated (step S206), and whether or not the difference ΔI exists between a predetermined low threshold value Ith1 and a predetermined high threshold value Ith2. If the difference ΔI is less than or equal to the low threshold value Ith1, the high frequency voltage generator 202 is commanded to increase the high frequency voltage Vh by a predetermined value ΔV (step S208), and the difference ΔI increases. If it is equal to or greater than the threshold value Ith2, the high-frequency voltage generator 202 is commanded to reduce the high-frequency voltage Vh by a predetermined value ΔV. The low threshold value Ith1 is the minimum value of the high-frequency current Ih necessary to ensure the necessary magnetic pole determination accuracy, and the high threshold value Ith2 is sufficient high-frequency current Ih to ensure the necessary magnetic pole determination accuracy. The value of

  Next, it is determined whether or not the magnetic pole determination period has ended. If not, the process returns to step S204, and if it has ended, the process proceeds to step S214. Next, it is determined whether or not the drive voltage frequency f is higher than a predetermined threshold value, that is, whether or not the motor is in a high rotation state. If not, the process proceeds to step S216, otherwise proceeds to step S220. In step S216, it is checked whether or not the change in the drive voltage frequency f is greater than or equal to a predetermined threshold value. If not, the process proceeds to step S218, and if so, the process proceeds to step S220.

  In step S218, the magnetic pole determination interval is set to a predetermined long interval, and in step S220, the magnetic pole determination interval is set to a predetermined short interval. An example of the magnetic pole determination interval is shown in FIG. In FIG. 8, the high-frequency voltage Vh is suppressed to an amplitude that allows the maximum value of the current elliptical locus to be determined outside the magnetic pole determination period.

  FIG. 9 shows a suitable example of the rising waveform of the high-frequency voltage Vh at the start of the magnetic pole determination period and the falling waveform of the high-frequency voltage Vh at the end of the magnetic pole determination period. In FIG. 9, at the start of the magnetic pole determination period, the amplitude of the high-frequency voltage Vh is increased during one cycle of the high-frequency voltage Vh from the level at which the maximum amplitude value of the high-frequency current Ih can be determined to the preferred level for determining the magnetic pole. . In addition, at the end of the magnetic pole determination period, the amplitude of the high frequency voltage Vh is decreased during the three periods of the high frequency voltage Vh from the current magnetic pole determination suitable level to the maximum amplitude value determinable level of the high frequency current Ih. That is, the increase of the high frequency voltage Vh is performed quickly, and the decrease is performed slowly. Data for one cycle after the start-up can be discarded, but cannot be discarded after the start-up. Therefore, by gradually changing the data, the current stability can be improved and the detection error can be reduced.

  In addition, after the end of the magnetic pole determination period, until the magnetic pole determination is performed in the next magnetic pole determination period, the magnetic pole determination data obtained in the immediately preceding magnetic pole determination period may be inverted and used every half cycle of the high-frequency voltage Vh. .

(Modification)
In the flowchart shown in FIG. 6, the feedback control of the high-frequency current Ih performed in steps S208 and S210 and the magnetic pole determination performed by the magnetic pole determination unit 207 are performed in parallel, but the value of the high-frequency current Ih is controlled by the feedback control. After setting to an appropriate value, the magnetic pole determination unit 207 may perform magnetic pole determination or allow the use of the magnetic pole determination result.

(Modification)
In the flowchart shown in FIG. 7, the magnetic pole determination frequency is changed in two stages while the motor is in a high rotation state or a large change in the number of rotations. However, it is continuously changed according to the number of rotations or the magnitude of the change. Also good.

(Modification)
The superposition of the high-frequency voltage Vh may be stopped if the motor rotation speed is equal to or higher than a predetermined threshold value. In this case, the estimation of the rotor position can also be performed using the phase of the detected phase current, for example.

(Modification)
In the above embodiment, the high-frequency current control unit 21 is performed by the software processing shown in FIGS. 5 to 7, but it is natural that the high-frequency current control unit 21 may be configured by a hardware circuit.

(Modification)
Instead of using the magnetic pole determination result determined in the immediately previous magnetic pole determination period as the current value of the magnetic pole determination result by inverting it every time the maximum value of the high frequency current is detected during the period when the magnetic pole determination is not performed The entire magnetic pole determination result of a plurality of times is examined and classified into a first determination result group and a second determination result group opposite to the first determination result group, and the magnetic pole direction until the next magnetic pole determination is determined according to the determination result group having a large number of data. It may be estimated.

(Modification)
As shown in FIG. 10, by adding steps S222 to S226 to the flowchart shown in FIG. 6, when the difference ΔI between the maximum amplitude values Imax and Imax ′ is smaller than the low threshold value Ith1, the magnetic pole determination unit 207 The magnetic pole determination based on the comparison between the maximum amplitude values Imax and Imax ′ is stopped, and instead, the memory-type magnetic pole determination is performed to determine the magnetic pole direction opposite to the magnetic pole direction during the immediately preceding current elliptical locus half cycle. Good.

(Modification)
In the above embodiment, the current sampling period is twice the PWM carrier period, but the current sampling period and the PWM carrier period may be equal. In this case, current sampling may be performed at either the maximum value point or the minimum value point of the PWM carrier voltage Vc.

(Modification)
In the above embodiment, the period of the high-frequency voltage Vh is an even multiple of the current sampling period, that is, the period of the PWM carrier voltage, but it is also possible to shift both. In this case, the position where the maximum current sampling value exists on each current elliptical locus is shifted every time, so the value in the major axis direction of the current elliptical locus is sampled as the maximum current sampling value once every several times. be able to.

(Modification)
In the above embodiment, the rotor position is estimated from each current data corresponding to the current elliptical locus half cycle, that is, the maximum amplitude value Imax is extracted and the rotor position θ is calculated using the rotor position θ. However, the current data is sampled and held over the immediately preceding cycle, 3/2 cycles, 2 cycles or more, and the rotor position θ is calculated by extracting the maximum amplitude value Imax from the obtained many current data. May be.

  Alternatively, the current rotor position θ may be calculated from a plurality of rotor positions θ individually obtained for each of the immediately preceding current elliptical locus half cycles by a predetermined arithmetic expression. For example, each rotor position θ obtained for each half cycle of each current elliptical locus should be shifted by the rotor angular velocity, but each rotor position θ is a value related to each other. For example, in constant speed operation, the angle difference between the rotor positions θ (rotor angle advance amount during a half cycle of the current elliptical locus) is constant, the angle difference gradually increases in the acceleration state, and the angle difference in the deceleration state It gradually decreases. Therefore, the current value of the rotor position θ can be corrected using the tendency of the angle difference between the plurality of rotor positions θ obtained immediately before.

(effect)
According to the above-described embodiment, magnetic noise and power loss can be reduced while preventing a decrease in magnetic pole determination accuracy.

1 is a block diagram illustrating a motor control device according to a first embodiment. It is a block diagram which shows the rotor position estimation part shown in FIG. It is a figure which shows the vector locus | trajectory of the high frequency voltage and high frequency current on a stationary orthogonal coordinate system. It is a figure which shows the vector locus | trajectory of the high frequency voltage and high frequency current on a stationary orthogonal coordinate system in a magnetic saturation state. It is a flowchart which shows the control operation | movement of the high frequency current for magnetic pole determination. It is a flowchart which shows the control operation | movement of the high frequency current for magnetic pole determination. It is a flowchart which shows the control operation | movement of the high frequency current for magnetic pole determination. It is a timing chart which shows an example of a magnetic pole determination interval. It is a timing chart which shows the example of a waveform of the high frequency voltage in a magnetic pole determination period. It is a flowchart which shows the deformation | transformation aspect of the control operation | movement of the high frequency current for magnetic pole determination.

Explanation of symbols

  1 is a current command unit, 2 and 3 are subtractors, 4 and 5 are PI amplifiers, 6 is a dq / 3-phase coordinate conversion unit, 7 to 9 are adders, 10 is a PWM voltage generation circuit, 11 is a three-phase inverter, Reference numerals 12 and 13 are current sensors for detecting phase currents, M is a synchronous machine having saliency, 14 and 15 are low-pass filters, 16 is a three-phase / dq coordinate converter, 17 and 18 are band-pass filters, and 201 is a drive frequency. Detecting unit, 202 is a high frequency voltage generating unit, 203 is a carrier voltage generating unit, 204 is a current sample and hold unit, 205 is a maximum amplitude calculating unit, 206 is a rotor position calculating unit, 207 is a magnetic pole determining unit, and 208 is a rotor position. A correction unit 209 is a magnetic pole determination high-frequency current control unit.

Claims (9)

High-frequency voltage superimposing means for superimposing a high-frequency voltage rotating on a stationary coordinate system with a frequency higher than a driving voltage frequency of a synchronous machine having saliency and a constant amplitude on an armature coil of the synchronous machine;
High-frequency current detecting means for detecting a high-frequency current flowing in the armature coil by the high-frequency voltage;
Rotor position estimating means for estimating a rotor position based on the detected high-frequency current;
Magnetic pole determination means for determining a magnetic pole position by comparing the maximum amplitude value of the high-frequency current at a magnetic pole position of a predetermined polarity and the maximum amplitude value of the high-frequency current at a magnetic pole position of the opposite polarity;
With
The high frequency voltage superimposing means includes:
The difference between the maximum amplitude value of the high-frequency current at the magnetic pole position of the predetermined polarity and the maximum amplitude value of the high-frequency current at the magnetic pole position of the opposite polarity is calculated, and the high-frequency voltage is set so that the difference is not less than a predetermined threshold value. A method for estimating the rotation angle of a synchronous machine having a saliency, characterized by adjusting the amplitude of the motor. In claim 1,
The high frequency voltage superimposing means includes:
A method for estimating a rotation angle of a synchronous machine having a saliency, wherein the amplitude of the high-frequency voltage is adjusted so that the difference is equal to or smaller than a second predetermined threshold value that is greater than the predetermined threshold value. In claim 1 or 2,
The high frequency voltage superimposing means includes:
A method for estimating a rotation angle of a synchronous machine having a saliency that performs the adjustment for a predetermined adjustment period when the synchronous machine is started. In any one of Claims 1 thru | or 3,
The high frequency voltage superimposing means includes:
A method for estimating a rotation angle of a synchronous machine having a saliency that performs the adjustment for a predetermined adjustment period every predetermined period during operation of the synchronous machine. In claim 4,
The method for estimating a rotation angle of a synchronous machine having a saliency that is shortened when the driving voltage frequency is high or when the driving voltage frequency changes greatly during the predetermined period. In claim 3 or 4,
The adjustment period is a method of estimating a rotation angle of a synchronous machine having a saliency set to approximately 2 to 10 cycles of the high-frequency voltage. In any one of Claims 1 thru | or 6.
A method for estimating a rotation angle of a synchronous machine having a saliency in which an increase rate of a high-frequency voltage at the start of the adjustment period is set larger than a decrease rate of the high-frequency voltage at the stop of the adjustment period. In any one of Claims 1 thru | or 7,
The rotor position estimating means includes
When the difference between the maximum amplitude value of the high-frequency current at the magnetic pole position of the predetermined polarity and the maximum amplitude value of the high-frequency current at the magnetic pole position of the opposite polarity is less than a predetermined threshold, the magnetic pole position determination is stopped, and the maximum A method of estimating a rotation angle of a synchronous machine having a saliency that permits the magnetic pole position determination when a difference in amplitude value exceeds the predetermined threshold value. High-frequency voltage superimposing means for superimposing a high-frequency voltage rotating on a stationary coordinate system with a frequency higher than a driving voltage frequency of a synchronous machine having saliency and a constant amplitude on an armature coil of the synchronous machine;
High-frequency current detecting means for detecting a high-frequency current flowing in the armature coil by the high-frequency voltage;
Rotor position estimating means for estimating a rotor position based on the detected high-frequency current;
Magnetic pole determination means for determining a magnetic pole position by comparing the maximum amplitude value of the high-frequency current at a magnetic pole position of a predetermined polarity and the maximum amplitude value of the high-frequency current at a magnetic pole position of the opposite polarity;
With
Only during a predetermined magnetic pole determination period after the start of the synchronous machine, the high-frequency voltage superimposing means is instructed to superimpose the high-frequency voltage, and the magnetic pole determining means is instructed to determine the magnetic pole position, and the magnetic pole determination A synchronous machine having a saliency of a synchronous machine, characterized by having high-frequency current control means for magnetic pole determination that increases an amplitude value of the high-frequency voltage in a period more than an amplitude value of the high-frequency voltage in a period other than the magnetic pole determination period Rotation angle estimation method.

Images