JP2008125207A - Controller of multi-phase rotary machine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem in which a rotation speed is not be correctly calculated on the basis of electric state amount of a multi-phase rotary machine in controlling the output of the multi-phase rotary machine by operating a switching element of an inverter in a transition or the like. <P>SOLUTION: A high-frequency voltage setting section 40 superposes a high-frequency signal "vhdc" on an output signal of an inverter 34. An error calculating section 42 extracts a high-frequency signal to be actually propagated through an electric motor 10 at this time from actual currents iα, iβ. Then, a phase difference (error quantity Δ) between two signals, namely, the extracted signals and a high-frequency signal "vhdc". A PI control section 44 calculates a rotation angle θ on the basis of the error quantity. A speed calculating section 46 corrects the change quantity of the rotation angle θ with the error quantity Δ, thereby calculating a rotation speed ω. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention acquires information on the rotation angle of the multiphase rotating machine based on the electrical state quantity of the multiphase rotating machine when controlling the output of the multiphase rotating machine by operating the switching element of the inverter. The present invention relates to a control device for a multi-phase rotating machine.

  As this type of control device, for example, as seen in Patent Document 1 below, a high-frequency signal is superimposed in the estimated d-axis direction of the output signal of the inverter for a three-phase motor having saliency, and the frequency actually propagated at this time There has also been proposed a method for calculating a rotation angle based on a signal. That is, since the inductance of the three-phase motor is the smallest in the d-axis direction, current tends to flow in the d-axis direction compared to the q-axis direction. For this reason, regardless of the phase angle of the superimposed high-frequency signal, the actually propagated current signal is a signal having a large value in the d-axis direction. Therefore, if the frequency signal that actually propagates when the high-frequency signal is superimposed in the estimated d-axis direction is deviated from the estimated d-axis, it can be seen that the estimated d-axis is different from the actual d-axis. In the control device, paying attention to this point, it is possible to acquire information about the rotation angle by correcting the estimated d-axis so as to reduce the difference between the actually propagated frequency signal and the estimated d-axis.

  Further, in this control device, the rotation speed is calculated as a time differential value of the rotation angle.

As a control device for calculating the rotation angle based on the actually propagated frequency signal, there is, for example, the following Patent Document 2 in addition to the above Patent Document 1.
JP 2004-254423 A Japanese Patent No. 3312472

  By the way, during transient operation in which the rotation speed of the three-phase motor changes, there is a delay in the calculation of the rotation angle calculated in the above manner, and the calculation accuracy of the rotation angle may temporarily decrease. At this time, the calculation accuracy of the rotation speed as the time differential value of the rotation angle also decreases. When the rotational speed calculation accuracy decreases, if the output control of the three-phase motor is performed based on the rotational speed calculated in such a manner, the controllability decreases.

  In addition, when controlling the output of the multiphase rotating machine by operating the switching element of the inverter, not only the control device, the rotation angle of the multiphase rotating machine is based on the electrical state quantity of the multiphase rotating machine. In the case of acquiring information related to the above, such a situation that the calculation accuracy of the rotational speed may be lowered is generally common.

  The present invention has been made in order to solve the above-mentioned problems, and its purpose is to control the output of the multi-phase rotating machine by operating the switching element of the inverter. An object of the present invention is to provide a control device capable of calculating the rotational speed with higher accuracy based on the state quantity.

  Hereinafter, means for solving the above-described problems and the operation and effects thereof will be described.

  According to the first aspect of the present invention, based on the electrical state quantity, an error calculation means for calculating an error amount for the currently acquired rotation angle, and a change amount of the acquired rotation angle are calculated. And a speed calculation means for calculating a rotation speed of the multiphase rotating machine based on the change amount and the error amount.

  When the calculated rotation angle is different from the actual rotation angle, such as when the actual rotation speed of the multiphase rotating machine changes, the amount of change in the calculated rotation speed is the actual rotation speed. It has become out of place. At this time, the magnitude relationship between the calculated rotation angle and the actual rotation angle is a parameter expressing the magnitude relationship between the actual rotation speed and the calculated rotation speed. In the above configuration, paying attention to this point, the rotational speed is calculated based on the amount of change in the rotational angle while grasping the magnitude relationship between the actual rotational speed and the calculated rotational speed based on the error amount. As a result, the rotation speed can be calculated with higher accuracy.

  The invention according to claim 2 is the invention according to claim 1, further comprising means for removing a high frequency component from the change amount calculated by the change amount calculating means and outputting the same to the speed calculating means. .

  When high-frequency noise is mixed in the electrical state quantity of the multiphase rotating machine, the high-frequency noise is also mixed in information related to the rotation speed acquired based on the electrical state quantity. The influence of this noise becomes particularly significant when calculating the amount of change in the rotation angle. In this regard, in the above configuration, the influence of the high frequency noise can be suitably removed from the calculated value of the rotation speed by removing the high frequency noise from the amount of change in the rotation angle used for calculating the rotation speed.

  The invention according to claim 3 is the invention according to claim 1 or 2, further comprising means for removing a high-frequency component from the error amount calculated by the error calculation means and outputting it to the speed calculation means. To do.

  When high-frequency noise is mixed in the electrical state quantity of the multi-phase rotating machine, the high-frequency noise is also mixed in the error amount for the rotation speed calculated based on the electrical state quantity. In this regard, in the above configuration, the influence of the high frequency noise can be suitably removed from the calculated value of the rotation speed by removing the high frequency noise from the error amount of the rotation angle used for calculating the rotation speed.

  The invention according to claim 4 is the invention according to claim 2 or 3, further comprising means for variably setting the high frequency to be removed according to information on the rotational speed.

  The electrical state quantity of the polyphase rotating machine has periodicity due to the mechanical rotation of the output shaft. The high frequency noise is caused by a change in the relative positional relationship between the materials constituting these members due to a change in the relative positional relationship between the stator and the rotor accompanying the mechanical rotation of the output shaft. Many things are caused by doing. The high-frequency noise generated in this way depends on the rotational speed of the multiphase rotating machine. In this regard, in the above configuration, by setting the high frequency to be removed variably based on the information about the rotational speed, the target is set to the high frequency noise assumed to be generated due to the above factors, and this is suitably removed. Can do.

  Note that the information on the rotation speed is a rotation speed calculated by the speed calculation means or a change amount calculated by the change amount calculation means.

  The invention according to claim 5 is the invention according to any one of claims 1 to 4, further comprising angle calculation means for calculating the rotation angle based on the error amount.

  The actual rotation angle is the sum of the difference between the currently acquired rotation angle with respect to the actual rotation speed and the currently acquired rotation speed. For this reason, the actual rotation angle can be calculated by using the error amount.

  According to a sixth aspect of the present invention, in the invention according to any one of the first to fourth aspects, the angle calculation means for calculating the rotation angle based on a time integral value of the rotation speed calculated by the speed calculation means. It is characterized by providing.

  When the rotation angle is calculated by, for example, integral control based on the error amount related to the rotation angle, the calculated rotation angle is deviated from the actual rotation angle in a transient state where the rotation speed of the multiphase rotating machine changes. It is easy to become a thing. In this regard, in the above configuration, the tracking delay of the calculated rotation angle can be suitably suppressed or avoided by using the time integral value of the rotation speed calculated in the above aspect.

  The invention according to claim 7 is the invention according to claim 6, wherein the angle calculation means calculates the rotation angle by correcting the time integral value based on the error amount.

  In the above configuration, by correcting the time integration value based on the error amount, feedback correction of the rotation angle based on the error amount can be performed, and as a result, the rotation angle can be calculated with higher accuracy.

  The invention according to claim 8 is the invention according to claim 7, wherein the speed calculation means corrects the change amount based on the error amount, and the correction amount of the change amount based on the error amount and the correction amount The correction amount of the time integral value based on the error amount is set separately.

  The correction amount for calculating the rotation speed with high accuracy by correcting the change amount based on the error amount is the correction amount for calculating the rotation angle with high accuracy by correcting the time integral value based on the error amount. It can be different. In this regard, in the above configuration, the rotation speed and the rotation angle can be calculated with higher accuracy by setting the two correction amounts separately.

  The invention according to claim 9 is the invention according to any one of claims 1 to 8, wherein the cycle has a period different from the period of the electrical angle of the multiphase rotating machine based on the information about the rotation angle and is arbitrary. Superimposing means for superimposing a frequency signal oscillating in the phase angle direction on the output signal of the inverter, and the error calculating means calculates the error amount based on the frequency signal actually propagated by the superposition. Features.

  In the above configuration, the frequency signal is superimposed on the output signal of the inverter. Thereby, it is assumed that a predetermined frequency signal is propagated through the multiphase rotating machine. However, when the acquired rotation angle and the actual rotation angle are shifted, it is considered that a shift occurs between the assumed frequency signal and the actually propagated frequency signal. In the above configuration, paying attention to this point, the error amount is calculated based on the deviation between the assumed frequency signal and the actually propagated frequency signal.

  Note that the frequency signal actually propagated by superimposition is a state quantity that is not the same as the state quantity that is directly operated as the output of the inverter in order to superimpose the frequency signal among the detected values of the electrical state quantity of the rotating machine. The signal is calculated based on the detected value. For example, when the output voltage of the inverter (phase voltage of the rotating machine) is manipulated to the command voltage to superimpose the frequency signal, the actually propagated frequency signal is a detection of a state quantity (eg current) other than the voltage. The signal is calculated based on the value.

  The invention described in claim 10 is the invention described in claim 9, wherein the multi-phase rotating machine has a saliency, and the superimposing means is configured based on information on the rotation angle. The frequency signal is superimposed in a direction in which the inductance is assumed to be minimum, and the error calculation means, as the error amount, a frequency signal actually propagated by the superposition and a frequency signal superimposed by the superposition means. The outer product value or inner product value of two vector signals consisting of

  When there is a deviation in the information about the rotation angle, a phase difference occurs between two vector signals consisting of the frequency signal actually propagated by superimposition and the frequency signal superimposed by the superimposing means, which is an error in the rotation angle. It has become. However, in order to directly calculate the phase difference, the processing becomes complicated, for example, an inverse trigonometric function is calculated. In this regard, in the above configuration, by using an outer product value or an inner product value between vector signals, a value having a phase difference and a correlation can be easily calculated as the error amount.

(First embodiment)
Hereinafter, a first embodiment in which a control device for a multiphase rotating machine according to the present invention is applied to a control device for a three-phase motor mounted in a hybrid vehicle will be described with reference to the drawings.

  FIG. 1 shows an overall configuration of a control system according to the present embodiment.

  The illustrated electric motor 10 is an embedded magnet synchronous motor (IPMSM). That is, as shown in FIG. 2, the rotor 10a of the electric motor 10 is configured by embedding a permanent magnet in an iron body.

  The αβ converter 20 shown in FIG. 1 converts the current flowing through the motor 10 into the current in the stationary coordinate system based on the U-phase actual current iu and the V-phase actual current iv out of the current actually flowing through the motor 10. That is, it is a part for converting into current vectors of α axis and β axis. Here, for example, the U phase is set to the same phase as the α axis, and the β axis is determined so as to be orthogonal thereto.

  The dq converter 22 is a part that converts the actual current iα on the α axis and the actual current iβ on the β axis into the current of the rotating coordinate system, that is, the actual current id on the d axis and the actual current iq on the q axis. is there. In this conversion, the rotation angle θ of the output shaft of the electric motor 10 is used. More precisely, the rotation angle θ is an electrical angle and is a rotation angle in the positive direction of the d axis with respect to the α axis. At this time, the low-pass filter also performs a process of removing a high-frequency signal described later from the output of the αβ converter 20. For this reason, the dq converter 22 extracts the d-axis component and the q-axis component of the current used for output control from the current that actually flows through the electric motor 10.

  The command current setting unit 24 is a part for setting the command current idc on the d axis and the command current iqc on the q axis based on the required torque Td for the electric motor 10.

  The command voltage setting unit 26 is a part that calculates the command voltage vdc on the d axis and the command voltage vqc on the q axis based on the command current idc, the command current iqc, the actual current id, and the actual current iq. This conversion is basically performed by feedback control of the actual current id on the d-axis to the command current idc and feedback control of the actual current iq on the q-axis to the command current iqc. This feedback control may be proportional integral control, for example. In this embodiment, well-known non-interacting control is used in combination. That is, in this embodiment, the command voltages vdc and vqc are set by feedback control and non-interacting control. In the non-interference control, the rotational speed ω of the electric motor 10 is used.

  The three-phase converter 28 converts the command voltage vqc on the q-axis and the output of the adder 30 according to the command voltage vdc on the d-axis into the u-phase command voltage vuc, the v-phase command voltage vvc, and w This is the part that converts to the phase command voltage vwc. In this conversion, the rotation angle θ is used.

  The PWM signal generation unit 32 is a part that generates an operation signal of the inverter 34 for applying the command voltages vuc, vvc, and vwc to the electric motor 10. As a result, the switching element of the inverter 34 is operated, and the voltage of the high voltage battery 36 is applied to the electric motor 10.

  Next, processing related to acquisition of the rotation angle θ of the electric motor 10 according to the present embodiment will be described.

  In the present embodiment, when controlling the output of the electric motor 10, a high-frequency signal having a cycle shorter than the cycle of the electrical angle of the electric motor 10 is superimposed on the output of the inverter 34. In other words, a high-frequency signal having a cycle shorter than the cycle of the current that actually flows through the electric motor 10 is superimposed according to the command currents idc and iqc. Then, the rotation angle θ of the electric motor 10 is calculated based on the high-frequency signal that actually propagates through the electric motor 10. This is done in view of the electric motor 10 having saliency.

  That is, since the electric motor 10 has saliency due to its structure, the inductance in the d-axis direction is minimum and the inductance in the q-axis direction is maximum. For this reason, current flows more easily in the d-axis direction than in the q-axis direction. Therefore, when the high-frequency signal is superimposed, the high-frequency signal actually propagated through the electric motor 10 is deflected in the d-axis direction. Specifically, as shown in FIG. 3A, when the estimated d-axis (estimated d-axis) is advanced with respect to the actual d-axis (real d-axis), the estimated d-axis When a high-frequency signal (broken line in the figure) is superimposed on the direction, the direction of the high-frequency signal actually propagated (solid line in the figure) is retarded with respect to the estimated d-axis in order to be deflected to the real d-axis side. Sneak away. Further, as shown in FIG. 3B, when the estimated d-axis and the actual d-axis coincide with each other, the high-frequency signal that is actually propagated when a high-frequency signal (broken line in the figure) is superimposed in the estimated d-axis direction. The direction of the signal (solid line in the figure) coincides with the estimated d-axis. Further, as shown in FIG. 3C, when the estimated d-axis is retarded with respect to the actual d-axis, when the high-frequency signal (broken line in the figure) is superimposed in the estimated d-axis direction, The direction of the high-frequency signal propagating to (in the figure, the solid line) is shifted to the advance side with respect to the estimated d-axis in order to be deflected to the actual d-axis.

  If the above property is used, the d-axis can be estimated and calculated, and thus the rotation angle θ can be calculated. That is, by repeating the superposition of the high frequency signal while making the direction in which the high frequency signal actually propagates be the estimated d-axis direction, the phase angle of the superposed high frequency signal can be matched with the phase angle of the actually propagated high frequency signal, As a result, the estimated d-axis can be matched with the actual d-axis.

  Specifically, as shown in FIG. 1, the high frequency voltage setting unit 40 applies the voltage signal vhdc to the adder 30 in order to apply the voltage signal vhdc as a high frequency signal in the direction assumed to be the d axis. Output. For this reason, a signal in which the voltage signal vhdc is superimposed on the command voltage vdc is input to the three-phase conversion unit 28.

  On the other hand, the error calculation unit 42 calculates an error amount Δ between the rotation angle θ and the actual rotation angle based on the actual current iα on the α axis, the actual current iβ on the β axis, and the voltage signal vhdc. FIG. 4 shows the configuration of the error calculation unit 42.

  The high-frequency current extraction unit 42a extracts the high-frequency components of the real current iα and the real current iβ, so that the current signal ihα on the α-axis and the current signal ihβ on the β-axis are high-frequency signals that are actually propagated to the motor 10. And output. The αβ conversion unit 42b calculates the α-axis and β-axis components of the voltage signal vhdc output from the high-frequency voltage setting unit 40 using the rotation angle θ. The outer product value calculation unit 42c calculates the outer product value of the vector signal as the voltage signal vhdc and the vector signal as the current signals ihα and ihβ based on the components on the α axis and the β axis. This outer product value is a parameter having a correlation with the phase angle difference between the two vector signals of the voltage signal vhdc and the current signals ihα and ihβ. In other words, the parameter has a correlation with the angle difference between the rotation angle θ and the actual rotation angle. Therefore, in the present embodiment, this outer product value is set as the error amount Δ.

  The PI controller 44 shown in FIG. 1 is a part that calculates the rotation angle θ so as to reduce the difference in phase angle between the two vector signals of the current signals ihα, ihβ and the voltage signal vhdc. This is because it is considered that the voltage signal vhdc output from the high-frequency voltage setting unit 40 can be set in the direction with the smallest inductance if the difference in phase angle can be made zero. In the three-phase converter 28, command voltages vuc, vvc, vwc are set based on the rotation angle θ. For this reason, the voltage signal vhdc from the high-frequency voltage setting unit 40 is superimposed on the voltage signal vhdc in the direction in which the inductance is assumed to be minimum when viewed from the rotation angle θ, that is, in the direction assumed to be the d axis. If the difference can be set to zero, it is considered that the rotation angle θ matches the actual rotation angle. In this embodiment, the PI control unit 44 calculates the sum of the proportional term and the integral term based on the error amount Δ, and outputs this as the rotation angle θ.

  The speed calculation unit 46 calculates the rotation speed of the electric motor 10 using the rotation angle θ and the error amount Δ. FIG. 5 shows the configuration of the speed calculation unit 46.

  The change amount calculating unit 46a calculates a time differential value (change amount ω1) of the rotation angle θ. The correction amount calculation unit 46b calculates a correction amount Δω of the change amount ω1 based on the error amount Δ. This correction amount Δω is for compensating for the deviation between the change amount ω1 and the actual rotational speed. Here, this will be described with reference to FIG.

  For example, during the transient operation of the electric motor 10 in which the rotation speed changes, the calculated rotation angle θ indicated by the alternate long and short dash line is deviated from the actual rotation angle indicated by the solid line in FIG. . At this time, the change amount ω1 also deviates from the actual rotational speed. At this time, the magnitude relationship between the actual rotational speed and the change amount ω1 corresponds to the magnitude relationship between the actual rotational angle and the calculated rotational angle θ. That is, in the example shown in FIG. 6, the actual rotational speed is faster than the amount of change ω1, and in this case, the actual rotational angle is advanced by a shift amount δ from the calculated rotational angle θ. In this way, it is possible to grasp whether the change amount ω1 is small or large with respect to the actual rotational speed from the deviation amount δ. Therefore, it is considered that this can be approximated to the actual rotational speed by correcting the change amount ω1 based on the deviation amount δ.

  Therefore, the correction amount calculation unit 46b shown in FIG. 5 calculates the correction amount Δω using the error amount Δ that is a parameter correlated with the angle difference between the actual rotation angle and the calculated rotation angle θ. The correction amount Δω may be a proportional term based on the error amount Δ, for example. Further, in view of the fact that the separation speed between the actual rotation angle and the calculated rotation angle θ indicates the difference between the actual rotation speed and the change amount ω1, it may be a differential term based on the error amount Δ. Further, for example, a proportional term + derivative term, a proportional term + integral term, or the like may be used.

  The adder 46c calculates and outputs the sum of the change amount ω1 and the correction amount Δω calculated by the correction amount calculation unit 46b as the final rotational speed ω. As a result, the rotational speed ω can be calculated with higher accuracy even when the electric motor 10 is in transition.

  FIG. 7 shows a procedure for calculating the rotational speed ω according to the present embodiment. This process is repeatedly executed at a predetermined cycle by, for example, a microcomputer.

  In this series of processing, first, in step S10, high-frequency signals (current signals ihα and ihβ) that are actually propagated through the electric motor 10 are detected. In subsequent step S12, based on the superimposed high-frequency signal (voltage signal vhdc) and the detected high-frequency signal (current signals ihα, ihβ), an error amount Δ is calculated by quantifying the phase difference of these vector signals by the outer product value. To do. Further, in step S14, the rotation angle θ is calculated based on the error amount Δ. In the subsequent step S16, the change amount ω1 is calculated as a time differential value of the rotation angle θ. In step S18, a correction amount Δω is calculated based on the error amount Δ. In step S20, the rotational speed ω is calculated by correcting the change amount ω1 with the correction amount Δω.

  According to the embodiment described in detail above, the following effects can be obtained.

  (1) The rotational speed ω of the electric motor 10 is calculated based on the change amount ω1 and the error amount Δ of the rotation angle θ. As a result, the rotational speed ω can be calculated with higher accuracy.

  (2) The rotation angle θ was calculated based on the error amount Δ. Thereby, the rotation angle θ can be appropriately calculated.

  (3) When a high-frequency signal (voltage signal vhdc) that vibrates in the d-axis direction is superimposed on the output signal of the inverter 34 based on the rotation angle θ, based on the high-frequency signals (current signals ihα and ihβ) that are actually propagated by the superposition. The error amount Δ was calculated. Thus, the error amount Δ can be calculated based on the difference between the high-frequency signal that is actually propagated and the high-frequency signal that is actually propagated.

  (4) A high-frequency signal (voltage signal vhdc) is superimposed in a direction (d-axis direction) in which the inductance of the electric motor 10 is assumed to be the minimum, and is composed of a high-frequency signal that is actually propagated by superposition and a high-frequency signal that is superimposed 2 The error amount Δ was calculated based on the outer product value of two vector signals. Thereby, the error amount Δ can be easily calculated.

(Second Embodiment)
Hereinafter, the second embodiment will be described with reference to the drawings with a focus on differences from the first embodiment.

  FIG. 8 shows a configuration of the speed calculation unit 46 according to the present embodiment. In FIG. 8, the same members as those shown in FIG. 5 are given the same reference numerals for the sake of convenience.

  As illustrated, in the present embodiment, the output of the change amount calculation unit 46a is input to the low-pass filter 46d. The low-pass filter 46d removes high-frequency components from the change amount ω1 output from the change amount calculation unit 46a. Here, the high frequency component is basically generated by a change in the relative positional relationship between the stator of the electric motor 10 and the magnet and iron of the rotor 10a. That is, in the portion of the rotor 10 a that is close to an arbitrary portion of the stator, a portion in which the magnet is embedded and a portion in which the rotor is not embedded are alternately generated as the electric motor 10 rotates. For this reason, when the adjacent portion changes from a portion where iron is not embedded to a portion where iron is embedded or when changing from a portion where iron is embedded to a portion where iron is not embedded, The physical properties of adjacent parts change discontinuously. And this discontinuous change becomes high frequency noise and mixes in the electrical state quantity of the electric motor 10. For this reason, the high frequency noise resulting from this will mix in real current iu and iv. For this reason, high-frequency noise is mixed into the error amount Δ calculated based on the actual currents iu and iv, and consequently also into the rotation angle θ and the change amount ω1. In particular, since the change amount ω1 is obtained by differentiating the rotation angle θ with respect to time, the influence of high-frequency noise tends to be significant.

  Therefore, in this embodiment, prior to the calculation of the rotational speed ω by the adder 46c, the high-frequency component is removed from the change amount ω1 by the low-pass filter 46d. Thereby, the influence of the high frequency component can be suitably removed from the rotational speed ω.

  According to this embodiment described above, the following effects can be obtained in addition to the effects (1) to (4) of the first embodiment.

  (5) After removing the high frequency component from the change amount ω1, the rotational speed ω was calculated. Thereby, the influence of high frequency noise can be suitably removed from the calculated value of the rotational speed ω.

(Third embodiment)
Hereinafter, the third embodiment will be described with reference to the drawings with a focus on differences from the first embodiment.

  FIG. 9 shows a configuration of the speed calculation unit 46 according to the present embodiment. In FIG. 8, the same members as those shown in FIG. 5 are given the same reference numerals for the sake of convenience.

  As illustrated, in the present embodiment, the error amount Δ is input to the low-pass filter 46e. The low pass filter 46e removes a high frequency component from the error amount Δ. As the cause of the generation of the high frequency component, the same one as assumed in the second embodiment is assumed. Thus, in the present embodiment, prior to calculating the correction amount Δω, the high-frequency component is removed from the error amount Δ by the low-pass filter 46e. Thereby, the influence of the high frequency component can be suitably removed from the rotational speed ω.

  According to this embodiment described above, the following effects can be obtained in addition to the effects (1) to (4) of the first embodiment.

  (6) After removing the high frequency component from the error amount Δ, the correction amount Δω was calculated. Thereby, the influence of the high frequency noise can be preferably removed from the correction amount Δω, and the influence of the high frequency noise can be suitably removed from the calculated value of the rotational speed ω.

(Fourth embodiment)
Hereinafter, the fourth embodiment will be described with reference to the drawings with a focus on differences from the first embodiment.

  FIG. 10 shows a configuration of the speed calculation unit 46 according to the present embodiment. In FIG. 10, the same members as those shown in FIG. 5 are given the same reference numerals for the sake of convenience.

  As shown in the figure, the present embodiment includes both a low-pass filter 46d that removes a high-frequency component from the change amount ω1 output from the change amount calculation unit 46a, and a low-pass filter 46e that removes a high-frequency component from the error amount Δ. . Thereby, the influence of the high frequency component can be more suitably removed from the rotational speed ω.

(Fifth embodiment)
Hereinafter, a fifth embodiment will be described with reference to the drawings, focusing on differences from the first embodiment.

  FIG. 11 shows a configuration of the speed calculation unit 46 according to the present embodiment. In FIG. 9, the same members as those shown in FIG. 5 are given the same reference numerals for the sake of convenience.

  In the present embodiment, the high frequency component removed by the low pass filter 46d is variably set according to the rotational speed ω. This is because high-frequency noise mixed in the electrical state quantity of the motor 10 (high-frequency noise mixed in the actual currents iu and iv) depends on the rotational speed ω. That is, as described above, the cause of the high-frequency noise is that, as the motor 10 rotates, the portion of the rotor 10a that is close to an arbitrary portion of the stator is alternately embedded with magnets and portions that are not embedded. This is because a phenomenon in which the physical properties of adjacent parts change discontinuously occurs periodically. Since the period of this discontinuous change is determined by the rotational speed ω, the frequency of the high-frequency noise changes according to the rotational speed ω. Therefore, in the present embodiment, the high frequency component removed by the low pass filter 46d is variably set to a higher frequency as the rotational speed ω increases, so that the high frequency noise caused by the above factors can be more appropriately removed.

  According to this embodiment described above, in addition to the effects (1) to (4) of the previous first embodiment and the effect (5) of the previous second embodiment, The effect will be obtained.

  (7) The high frequency to be removed is variably set according to the rotational speed ω. Thereby, a target can be set to the high frequency noise assumed to be generated in the above-described manner, and this can be suitably removed.

(Sixth embodiment)
Hereinafter, the sixth embodiment will be described with reference to the drawings with a focus on differences from the first embodiment.

  FIG. 12 shows the calculation process of the rotation angle θ and the calculation process of the rotation speed ω according to the present embodiment.

  The correction amount calculation unit 50 is a part that calculates a correction amount Δθ for a rotation angle (more precisely, a time integration value θ1 described later) based on the error amount Δ. This correction amount Δθ may be calculated as the sum of a proportional term and an integral term based on the error amount Δ. Alternatively, for example, it may be calculated as a proportional term, or may be calculated as the sum of a proportional term, a differential term, and an integral term.

  The integral calculation unit 52 calculates a time integral value θ1 of the rotation speed ω. The adder 54 calculates the rotation angle θ as the sum of the time integration value θ1 and the correction amount Δθ. The change amount calculation unit 56 is a part that calculates the change amount ω1 as a time differential value of the rotation angle θ. The low-pass filter 58 removes the high-frequency noise described in the second to fifth embodiments from the change amount ω1.

  On the other hand, the correction amount calculation unit 60 calculates a correction amount Δω for correcting the change amount ω1 (more precisely, a value obtained by removing the high frequency component from the change amount ω1) based on the error amount Δ. Here, the correction amount Δω may be calculated as the sum of the proportional term and the integral term based on the error amount Δ. Alternatively, for example, it may be calculated as a proportional term or a differential term, and may be further calculated as a proportional term and a differential term. Then, the adding unit 62 calculates the rotational speed ω by correcting the change amount ω1 (more precisely, a value obtained by removing the high-frequency component from the change amount ω1) with the correction amount Δω.

  Thus, also in this embodiment, the rotational speed ω can be calculated with high accuracy by correcting the change amount ω1 with the correction amount Δω. Furthermore, in this embodiment, since the rotation angle θ is calculated based on the rotation speed ω, if the rotation speed ω deviates from the actual rotation speed, this is feedback-corrected as an error amount Δ. As a result, the rotational speed ω can be calculated with higher accuracy. Further, since the time integration value θ1 is corrected based on the correction amount Δθ, the rotation angle θ can be calculated with high accuracy.

  In particular, when the rotational speed ω is calculated, processing for removing high-frequency components from the amount of change ω1 is performed, and thus the value after this removal may be delayed by filtering. In the present embodiment, this delay is compensated by correction with the correction amount Δω with respect to the change amount ω1 (more precisely, a value obtained by removing the high-frequency component from the change amount ω1). As a result, the effects of delay and high frequency noise can be suitably removed from the rotational speed ω by filtering, and as a result, the rotational speed ω and the rotational angle θ can be calculated with higher accuracy.

  FIG. 13 shows the experimental results of the relationship between the rotational speed ω calculated by the above processing at the time of transition and the actual rotational speed. The solid line in FIG. 13A indicates the actual rotational speed (detected value by the sensor), and the two-dot chain line indicates the rotational speed ω. Specifically, in the figure, the two-dot chain line indicates the case where the correction amounts Δθ and Δω are calculated as the sum of the proportional and integral terms of the error amount Δ in both of the correction amount calculation units 50 and 60 of FIG. ing. As illustrated, the calculated rotational speed ω represents the actual rotational speed with high accuracy. On the other hand, when the rotation speed is calculated as a time differential value of the rotation angle θ calculated in the process shown in FIG. 1 (change amount ω1 in FIG. 5 above), 1 in FIG. Shown with a dotted line. Thus, the change amount ω1 is deviated from the actual rotational speed.

  According to the present embodiment described above, in addition to the effects (1), (3), (4) of the previous first embodiment and the effect (5) of the previous second embodiment. In addition, the following effects can be obtained.

  (8) The rotation angle θ was calculated based on the rotation speed ω. Thereby, the tracking delay of the calculated rotation angle θ can be suitably suppressed or avoided.

  (9) The rotation angle θ was calculated by correcting the time integration value θ1 based on the error amount Δ. Thereby, feedback correction of the rotation angle θ based on the error amount Δ can be performed, and as a result, the rotation angle θ can be calculated with higher accuracy.

  (10) The correction amount Δω of the change amount ω1 based on the error amount Δ and the correction amount Δθ of the time integration value θ1 based on the error amount Δ are set separately. Thereby, the rotational speed ω and the rotational angle θ can be calculated with higher accuracy.

(Seventh embodiment)
Hereinafter, the seventh embodiment will be described with reference to the drawings with a focus on differences from the first embodiment.

  FIG. 14 shows the overall configuration of the control system according to the present embodiment. In FIG. 14, members corresponding to those shown in FIG. 1 are given the same reference numerals for convenience.

  As shown in the figure, in the present embodiment, a high-frequency current setting unit 70 is provided to superimpose a high-frequency signal on the output signal of the inverter 34. The high frequency current setting unit 70 adds the current signal ihdc as an adder so as to superimpose the current signal ihdc as a high frequency signal in the direction estimated as the d-axis direction by the rotation angle θ, that is, the direction in which the inductance is assumed to be minimum. 72. The adder 72 outputs to the command voltage setting unit 26 the current signal ihdc superimposed on the command current idc on the d axis set by the command current setting unit 24.

  As described above, the command voltage setting unit 26 sets the command voltage vdc on the d axis based on the difference between the actual current id on the d axis and the command current idc, and determines the actual current iq and the command current iqc on the q axis. A command voltage vqc on the q axis is set based on the difference. Here, since the low-pass filter processing is not performed in the dq conversion unit 22a according to the present embodiment, high-frequency signals actually propagated through the motor 10 are mixed in the actual currents id and iq. For this reason, when the current signal ihdc is superimposed on these command voltages vdc and vqc (more precisely, when the command voltage is set to superimpose the current signal ihdc), the frequency signal that actually propagates through the motor 10 It is mixed.

  The command voltages vdc and vqc are converted into a command voltage vαc on the α axis and a command voltage vβc on the β axis in the αβ conversion unit 74. Then, the error calculation unit 76 extracts the voltage signal vhα on the α axis and the voltage signal vhβ on the β axis corresponding to the high frequency signal actually propagated through the electric motor 10 from the command voltage vαc and the command voltage vβc. Then, the outer product value of the vector signal of the current signal ihα and the vector signals of the voltage signals vhα and vhβ is calculated as the error amount Δ.

  As described above, in this embodiment, the current signal ihdc is used as the high-frequency signal to be superimposed on the output signal of the inverter 34, and at this time, the high-frequency components of the command voltages vαc and vβc into which the high-frequency signal actually propagated through the motor 10 is mixed. The error amount Δ can be calculated from the outer product value of two vector signals of the current signal ihdc.

  Also according to the present embodiment described above, it is possible to obtain effects according to the above-described effects of the first embodiment.

(Eighth embodiment)
Hereinafter, the eighth embodiment will be described with reference to the drawings with a focus on differences from the first embodiment.

  FIG. 15 shows the overall configuration of the control system according to the present embodiment. In FIG. 15, members corresponding to those shown in FIG. 1 are given the same reference numerals for convenience.

  In the present embodiment, the error calculation unit 78 calculates the error amount Δ based on the actual currents id and iq from which the high-frequency components have been removed and the high-frequency components extracted from the actual currents iα and iβ. FIG. 16 shows relation information used for calculating the error amount Δ by the error calculation unit 78.

  FIG. 16 shows the phase angle of the actual current vector (actual current id, iq) for output control, the output torque of the electric motor 10, and the amplitude value (target amplitude value) of the frequency signal that actually propagates through the electric motor 10. Show the relationship. Specifically, this relationship indicates a relationship when a high-frequency signal in a direction in which the inductance of the electric motor 10 is minimized is superimposed on the output signal of the inverter 34. As shown in the drawing, the target amplitude value varies depending on the phase angle of the current vector for output control and the output torque. For this reason, when the actual amplitude value deviates from the target amplitude value, it is considered that the value deviated from the actual phase angle is recognized as the phase angle of the output control current. In other words, it is considered that the rotation angle deviated from the actual rotation angle is recognized as the rotation angle (more precisely, the electrical angle) of the electric motor 10.

  The error calculator 78 estimates the phase angle and output torque of the current vector from the actual currents id and iq. And based on these and the said relationship information, a target amplitude value is calculated. Then, an error amount Δ is calculated based on the difference between the amplitude of the high frequency component extracted from the actual currents iα and iβ and the target amplitude value.

  According to the present embodiment described above, the same effects as the effects (1), (3), and (4) of the first embodiment can be obtained.

(Other embodiments)
Each of the above embodiments may be modified as follows.

  In the third, fourth, and sixth embodiments, the high frequency that is to be removed by the low-pass filters 46d, 46e, and 58 may be variably set according to the rotational speed ω. Further, the information regarding the rotational speed referred to when the high frequency to be removed is made variable is not limited to the rotational speed ω but may be a change amount ω1.

  In the first to seventh embodiments, the outer product value is calculated using the fixed coordinate system component, but may be calculated using the rotating coordinate system component (for example, the d-axis and q-axis components).

  The vector signal for calculating the outer product value is not limited to those exemplified in the above embodiments. For example, in FIG. 1, the high-frequency components of the actual currents iα and iβ may be converted into voltage vectors vα and vβ, and the outer product value of this and the voltage signal vhdc may be calculated. In short, among the electrical state quantities of the motor 10, a high-frequency signal calculated from a detection value other than the electrical state quantity that is a direct operation target to superimpose the high-frequency signal on the output signal of the inverter 34 is converted into the motor. 10 may be a high-frequency signal that actually propagates. That is, for example, when a high frequency signal is superimposed by manipulating the output voltage of the inverter 34 (phase voltage of the electric motor 10) to a command voltage, the high frequency calculated based on the detected value of an electrical state quantity (for example, current) other than the voltage. The signal may be a high-frequency signal that actually propagates through the electric motor 10.

  The calculation method of the error amount Δ using the property that the current is deflected when the high-frequency signal superimposed on the output signal of the inverter 34 is actually propagated by the electric motor 10 is not limited to those exemplified in the above embodiments. For example, based on the inner product of the vector signal A as the superimposed high-frequency signal and the vector signal B as the actually propagated high-frequency signal, (1-A · B / | A || B |) may be used as the error amount Δ. Good. This also makes it possible to calculate the error amount Δ as a parameter for quantifying the phase difference between the vector signal A as the superposed high-frequency signal and the vector signal B as the actually propagating high-frequency signal without performing the inverse trigonometric function. Can be calculated. However, it is not limited to calculating the outer product or inner product value of these vector signals, and this may be used as the error amount Δ by calculating the phase difference between these vector signals based on the inverse trigonometric function.

  In each of the embodiments described above, the rotational speed ω is used for the non-interacting control.

  ・ As a method to obtain information on the rotation angle of the multi-phase rotating machine based on the electrical state quantity of the multi-phase rotating machine, the frequency signal that actually propagates the motor when the frequency signal is superimposed on the output signal of the inverter is used. It is not restricted to what is performed based on. For example, it may be detected based on an induced voltage (see, for example, “Japanese Patent Laid-Open No. 2006-230120” or “Electronic Engineering Handbook 6th Edition, Electrical Society of Japan, Volume 21, 3.6”).

  -Due to the structure, the electric motor having saliency is not limited to the electric motor 10 described above. For example, an embedded magnet synchronous motor (IPMSM) as shown in FIG. 17A or a synchronous reluctance motor (SynRM) as shown in FIG. 17B may be used.

  -As a multiphase rotating machine, not only an electric motor but a generator may be sufficient.

  In each of the above embodiments, the control device according to the present invention is applied to a hybrid vehicle. Furthermore, the control device of the present invention may be applied to an electric motor as power transmission means such as power steering in a vehicle using an internal combustion engine as a power source.

The figure which shows the whole structure of the control system of the electric motor concerning 1st Embodiment. The figure which shows the rotor of the electric motor concerning the embodiment. The figure which shows the problem regarding the detection of a rotation angle. The figure which shows the process of the error calculation part concerning the said embodiment. The figure which shows the process of the speed calculation part concerning the said embodiment. The figure explaining the calculation principle of the rotational speed by a speed calculation part. The flowchart which shows the procedure of the calculation process of the rotational speed concerning the said embodiment. The figure which shows the process of the speed calculation part concerning 2nd Embodiment. The figure which shows the process of the speed calculation part concerning 3rd Embodiment. The figure which shows the process of the speed calculation part concerning 4th Embodiment. The figure which shows the process of the speed calculation part concerning 5th Embodiment. The figure which shows the calculation process of the rotation angle and rotation speed concerning 6th Embodiment. The time chart which shows the calculation accuracy of the rotational speed concerning the embodiment. The figure which shows the whole structure of the control system of the electric motor concerning 7th Embodiment. The figure which shows the whole structure of the control system of the electric motor concerning 8th Embodiment. The figure which shows the relationship between the phase of the electric current for output control, output torque, and a target amplitude value. The figure which shows the rotor of the electric motor in the modification of each said embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Electric motor, 42 ... Error calculation part (one embodiment of error calculation means), 46 ... Speed calculation part (one embodiment of speed calculation means).

Claims (10)

When controlling the output of the multiphase rotating machine by operating the switching element of the inverter, the multiphase rotating machine acquires information on the rotation angle of the multiphase rotating machine based on the electrical state quantity of the multiphase rotating machine In the control device of
Based on the electrical state quantity, an error calculating means for calculating an error quantity for the currently acquired rotation angle;
A change amount calculating means for calculating a change amount of the acquired rotation angle;
A control device for a multi-phase rotating machine, comprising: a speed calculating unit that calculates a rotation speed of the multi-phase rotating machine based on the change amount and the error amount.   2. The control device for a multi-phase rotating machine according to claim 1, further comprising means for removing a high frequency component from the amount of change calculated by the amount of change calculating means and outputting the same to the speed calculating means.   3. The control device for a multi-phase rotating machine according to claim 1, further comprising means for removing a high frequency component from the error amount calculated by the error calculating means and outputting the same to the speed calculating means.   4. The control device for a multi-phase rotating machine according to claim 2, further comprising means for variably setting the high frequency to be removed according to information on the rotational speed.   The control device for a multi-phase rotating machine according to any one of claims 1 to 4, further comprising angle calculation means for calculating the rotation angle based on the error amount.   The multiphase rotating machine according to any one of claims 1 to 4, further comprising an angle calculation unit that calculates the rotation angle based on a time integral value of the rotation speed calculated by the speed calculation unit. Control device.   The control device for a multi-phase rotating machine according to claim 6, wherein the angle calculation means calculates the rotation angle by correcting the time integral value based on the error amount. The speed calculation means corrects the change amount based on the error amount,
8. The control device for a multi-phase rotating machine according to claim 7, wherein a correction amount of the change amount based on the error amount and a correction amount of the time integration value based on the error amount are set separately. Superimposing means for superimposing on the output signal of the inverter a frequency signal having a period different from the period of the electrical angle of the multiphase rotating machine and oscillating in an arbitrary phase angle direction based on the information on the rotation angle Prepared,
The control device for a multi-phase rotating machine according to claim 1, wherein the error calculation unit calculates the error amount based on a frequency signal actually propagated by the superposition. The multi-phase rotating machine has saliency,
The superimposing means superimposes the frequency signal in a direction in which the inductance of the multi-phase rotating machine is assumed to be minimum based on information on the rotation angle,
The error calculation means uses, as the error amount, an outer product value or an inner product value of two vector signals composed of a frequency signal actually propagated by the superposition and a frequency signal superposed by the superposition means. The control device for a multi-phase rotating machine according to claim 9.

Images