JP2009060752A - Angle-error learning device and angle detection system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To learn a detection error on an angle detection means regarding the angle detecting means detecting the angle of rotation of a rotating machine. <P>SOLUTION: In a learning processing section 60, command currents Idc and Iqc are set so that a current flowing through a U phase reaches zero, and the Idc and Iqc are converted into command voltages Vdc and Vqc on a dq axis from a voltage equation and output to a three-phase conversion section 52 through selectors S1 and S2. In the three-phase conversion section 52, the command voltages Vdc and Vqc are converted into the command voltages vuc, vvc and vwc having three-phases. The angle of rotation θn for a learning used in the three-phase conversion section 52 is operated so that an actual current iu having the U phase at a time when an inverter 20a is operated reaches zero on the basis of these command voltages vuc, vvc and vwc, and a difference between the angle of rotation θn for the learning at the time when the actual current iu reaches zero and the angle of rotation θ detected by a resolver 36 is learnt as an angle error. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an angle error learning device for learning a detection error of an angle detection means for detecting a rotation angle of a rotating machine and an angle detection system equipped with the device.

As this type of learning device, for example, as seen in Patent Document 1 below, the detected value of the rotation angle when energizing a current that locks the output shaft of the electric motor mounted on the power steering at a predetermined rotation angle, and the above There has also been proposed one that learns a detection error of a detection value based on an error from a predetermined rotation angle. In other words, if the detected value deviates from this even though the output shaft of the motor is locked at a predetermined rotation angle by energization, it can be determined that a detection error has occurred and this can be learned. .
JP 2003-319680 A

  By the way, the learning device is premised on that the output shaft of the motor can be rotated by energization. For this reason, in an electric motor or the like constituting an in-vehicle power generation device, the output shaft of which is connected to a drive wheel, the output shaft may not be able to rotate to a predetermined rotation angle by energization. In this case, the detection error may be learned erroneously. Furthermore, even if the output shaft can be rotated, this means that the output shaft is intentionally rotated for learning, which may give the user a sense of incongruity.

  It should be noted that not only the electric motor that constitutes the in-vehicle power generation device, but also in general a rotating machine, such a situation in which inconvenience occurs due to the use of the rotation of the rotating shaft when learning in the above mode is generally common. It has become.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide an angle error learning device that can more appropriately learn the detection error of an angle detection unit that detects the rotation angle of a rotating machine. And providing an angle detection system on which the apparatus is mounted.

  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, in the angle error learning device for learning the detection error of the angle detection means for detecting the rotation angle of the rotating machine, the detected value of the current actually flowing through the rotating machine and the angle detecting means And learning means for learning a detection error of the angle detection means by controlling the current flowing through the rotating machine to a specified current based on the detected value.

  The difference between the detected value of the current actually flowing through the rotating machine and the specified current varies depending on the rotation angle used when the current flowing through the rotating machine is controlled to the specified current. That is, when the rotation angle used is separated from the actual rotation angle, the detected value of the actually flowing current is separated from the specified current. In the above invention, paying attention to this point, it is possible to learn the angle error included in the detected value of the angle detecting means based on the detected value of the actually flowing current.

  The learning means may perform the learning when the rotating machine is not energized in response to an operation request for the rotating machine.

  Further, the control of the current flowing through the rotating machine to the specified current may be a control that sets the current amount of at least one phase to a predetermined amount. In this case, if the rotation angle used when controlling to the predetermined amount has an angle error, the detected value of the current of the at least one phase is separated from the predetermined amount. Here, since the current of at least one phase is a current in a fixed coordinate system, it can be detected without using angle information. For this reason, the difference between the detected value and the predetermined amount can be suitably used as a parameter having a correlation with the angle error of the rotation angle used when controlling to the predetermined amount. Here, it is desirable to set the current amount of at least one phase to zero.

  Further, the rotating machine may be a rotating machine having a saliency (salient pole machine), and the specified current may be a current in a direction in which the inductance is minimized. That is, in the case of a salient pole machine, the inductance changes depending on the direction, and the current tends to flow in the direction where the inductance is minimized. For this reason, when the rotation angle to be used has an angle error, even if a current is passed in the direction where the inductance is estimated to be the minimum, the actual current will flow in a direction shifted from the direction where the inductance is the minimum. For this reason, the actual current is deflected in a direction in which the inductance is minimized. In other words, current flows in a direction deviating from the direction in which the inductance is estimated to be minimum based on the rotation angle. For this reason, the difference between the specified current and the detected current value can be used as a parameter having a correlation with the angle error.

  In this way, the specified current is set such that the amount of current in a direction orthogonal to the direction in which the inductance is minimized in the salient pole machine is zero, or the current flowing through at least one phase in the non-salient pole machine is zero. When it is assumed that it is possible to properly detect an error in the rotation angle used for control to the specified current based on the difference between the detected current value and the specified current.

  According to a second aspect of the present invention, in the first aspect of the present invention, the angle detection means includes a member that rotates integrally with the shaft of the rotating machine.

  The angle detecting means may have a detection error due to a displacement due to an arrangement accuracy or a secular change of a member that rotates integrally with the shaft of the rotating machine. Therefore, the invention according to claim 1 is applied. It is effective.

  According to a third aspect of the present invention, in the first or second aspect of the invention, the detection error is learned separately for each of a plurality of values for the detection value of the angle detection means.

  The detection error of the angle detection means can change according to the rotation angle of the rotating machine. In this case, if the learning is performed as a single detection error regardless of the rotation angle of the rotating machine, the rotation angle dependency of the detection error cannot be appropriately learned. In this regard, in the above invention, the detection error is learned appropriately for each of a plurality of detection values, so that the detection error can be appropriately learned even when the detection error has a rotation angle dependency. Can do.

  The invention according to claim 4 is the invention according to any one of claims 1 to 3, wherein the learning is performed when the rotational speed of the rotating machine is equal to or lower than a specified speed.

  When learning is performed when the rotation speed is high, the rotation angle may change during a period from when the control of flowing the specified current is performed based on the detection value of the angle detection means to when the actual current associated therewith is detected. For this reason, if learning is performed when the rotation speed is high, it may be difficult to specify at which rotation angle the detection error is learned. In this regard, in the above invention, the detection error in the rotation angle detected by the angle detection unit can be appropriately learned by learning when the rotation speed is equal to or less than the specified speed.

  When the invention according to claim 4 is applied to the invention according to claim 3, the specified speed is the rotating machine in a period from the acquisition of the detection value of the angle detection means to the acquisition of the detection value of the actually flowing current. It is desirable that the amount of rotation of be smaller than the difference between any two values of the plurality of values.

  Further, the learning means may perform the control when the rotating machine that is not requested to operate the rotating machine is stopped.

  Further, the rotating machine may be an on-vehicle rotating machine, and the learning unit may perform control to the specified current when a travel permission switch of the vehicle is turned on.

  According to these, learning can be performed when the rotation speed is zero.

  According to a fifth aspect of the present invention, in the invention according to any one of the first to fourth aspects, the learning unit actually operates the rotating machine by operating a rotation angle used for control to the specified current. And means for feedback-controlling the flowing current to the specified current, and means for performing the learning based on a deviation between the operated rotation angle and a detected value of the angle detecting means.

  By operating the rotation angle by the feedback control, the rotation angle used for controlling the specified current converges to the actual rotation angle of the rotating machine. When the detection value of the angle detection unit deviates from the converged rotation angle, it can be considered that an angle error has occurred in the angle detection unit.

  The invention according to claim 6 is the invention according to any one of claims 1 to 5, wherein the specified current is a value required for the torque generated by the rotating machine to rotate the shaft of the rotating machine. It is set to be less than the minimum value.

  In the said invention, it can avoid that a rotary machine shows the behavior which is not intended at the time of learning.

  The invention according to claim 7 is the invention according to any one of claims 1 to 6, wherein the rotating machine constitutes an in-vehicle power generation device, and an output shaft of the rotating machine is connected to a drive wheel. It is characterized by.

  In the above invention, since the shaft of the rotating machine is connected to the drive wheel, the shaft of the rotating machine is difficult to rotate even if a current is passed through the rotating machine during learning. For this reason, learning based on a change in the rotation angle is difficult. Further, even if the rotation angle can be changed, this means that the vehicle exhibits unintended behavior. On the other hand, in the inventions according to claims 1 to 7, since learning can be performed without changing the rotation angle, the invention according to claim 7 has the utility value of the inventions according to claims 1 to 6. To increase.

  The invention according to claim 8 is the invention according to any one of claims 1 to 7, wherein the angle detection device is an electromagnetic induction type angle detector.

  The electromagnetic induction type detector may cause an angle error due to a connection error with the shaft of the rotating machine. Furthermore, due to variations in the internal structure of the electromagnetic induction type detector, the angle error may change depending on the rotation angle of the rotating machine. For this reason, the utility value of the inventions of claims 1 to 7 is particularly high. In particular, in the invention having the invention specific matter of claim 3, since the angle dependency of the angle error can be learned, its utility value is further increased.

  The invention according to claim 9 is an angle detection system comprising the angle error learning device according to any one of claims 1 to 8 and the angle detection means.

(First embodiment)
Hereinafter, a first embodiment in which an angle error learning device and an angle detection system according to the present invention are applied to an angle error learning device and an angle detection system mounted on a hybrid vehicle will be described with reference to the drawings.

  FIG. 1 shows an overall configuration of a hybrid system (series / parallel hybrid system) according to the present embodiment.

  As shown in the figure, the power of the internal combustion engine 10 is distributed to the first motor generator (generator 14) and the second motor generator (electric motor 16) via the power split mechanism 12. Specifically, the power split mechanism 12 includes a planetary gear mechanism. The planetary gear 12p is an output shaft of the internal combustion engine 10, the sun gear 12s is a rotating shaft of the generator 14, and the ring gear 12r is an electric motor 16. Are connected to each output shaft.

  The load torque of the generator 14 and the torque of the electric motor 16 are controlled by the power control unit 20. The power control unit 20 is connected to a high voltage battery that stores high voltage power of, for example, several hundred volts. Then, the generated energy of the generator 14 is charged into a high voltage battery as a DC power source via the power control unit 20, and the electric motor 16 is operated by the electric power of the high voltage battery. The torque of the electric motor 16 is transmitted to the drive wheels 18 of the vehicle.

  The electronic control device (ECU 30) is a control device for the in-vehicle power generation system, and is turned on by power supply from the low-voltage battery B when the start switch 34 of the ECU 30 serving as a permission switch for allowing the vehicle to travel is turned on. Become. The ECU 30 takes in the output of the electromagnetic induction type angle detector (resolver 36) that detects the rotation angle of the electric motor 16, the output of the resolver 37 that detects the rotation angle of the generator 14, and the like, and based on this, the power control unit 20, the power generation amount (torque) of the generator 14 is controlled, and the torque of the motor 16 is controlled. The ECU 30 controls the combustion state by operating various actuators of the internal combustion engine 10. The resolver 36 and the resolver 37 are configured to include a rotor directly connected to the output shaft of the electric motor 16 or the generator 14 and a wound stator, and an electromagnetic coupling rate between the rotor and the stator is high. The rotation angle is detected by utilizing the change depending on the rotation angle.

  The ECU 30 also includes a constant memory holding device 32. Here, the constant storage device 32 is a storage device that always holds data regardless of the main connection state (the state of the start switch 34) with the low voltage battery B as the power supply means of the ECU 30. As this constant memory holding device, for example, a backup memory which is always in a power supply state regardless of the state of main connection between the ECU 30 and the low voltage battery B, or a non-volatile memory (EEPROM or the like) which holds data regardless of the presence or absence of power supply and so on.

  FIG. 2 shows a process for controlling the electric motor 16 among the processes performed by the ECU 30.

  The illustrated inverter 20 a is provided in the power control unit 20 and applies the voltage of the above-described high voltage battery to each phase of the electric motor 16. The current sensors 20b and 20c detect currents flowing through the U phase and the W phase of the electric motor 16, respectively.

  The command current generation unit 40 generates a command current on the dq axis for the electric motor 16 based on the required torque Trqcom. The required torque Trqcom is set according to the accelerator operation amount or the like. On the other hand, the dq converter 42 converts the actual currents iu and iw detected by the current sensors 20b and 20c into actual currents Id and Iq on the dq axis.

  The current deviation calculation unit 44 calculates a difference ΔId of the actual current Id with respect to the command current Idc on the d axis. Further, the current deviation calculation unit 46 calculates a difference ΔIq of the actual current Iq with respect to the command current Iqc on the q axis. The PI control unit 48 calculates a command voltage Vdc on the d axis based on the difference ΔId. Further, the PI control unit 50 calculates a command voltage Vqc on the q axis based on the difference ΔIq.

  The three-phase converter 52 converts the command voltages Vdc and Vqc on the dq axis into three-phase command voltages vuc, vvc, and vwc. The PWM control unit 54 generates an operation signal for the inverter 20a by modulating the command voltages vuc, vvc, vwc with a carrier so as to apply the three-phase command voltages vuc, vvc, vwc to the electric motor 16. Here, a triangular wave or a saw wave may be employed as the carrier. When the operation signals for the inverter 20a are generated in this way, the PWM control unit 54 outputs these operation signals to the inverter 20a. Thereby, the command voltages vuc, vvc, vwc are applied to the electric motor 16.

  The rotation angle of the electric motor 16 detected by the resolver 36 is used for the processing of the dq conversion unit 42 and the three-phase conversion unit 52. Below, the process which learns the detection error of the resolver 36 performed in the learning process part 60 is demonstrated.

  FIG. 3 shows the procedure of the detection error learning process of the resolver 36 according to this embodiment. This process is executed by the ECU 30 as a trigger when the start switch 34 is turned on.

  In this series of processes, when the activation switch 34 is turned on, first, in step S10, the rotation angle θ detected by the resolver 36 is set as the initial value of the learning rotation angle θn. This learning rotation angle θn is used in the three-phase converter 52. That is, during the learning process, the learning rotation angle θn is output from the learning processing unit 60 to the three-phase conversion unit 52 by switching the selector S3 shown in FIG. In subsequent step S12, the learning processing unit 60 calculates command currents Idc and Iqc at which the actual current iu flowing through the U phase becomes zero based on the learning rotation angle θn. Here, there is a relationship of the following equation (c1) between the three-phase actual currents iu, iv, iw and the two-phase actual currents Id, Iq.

ここで、「ｉｕ＝０」とすることで、指令電流Ｉｄｃ,Ｉｑｃは下記の式（ｃ２）となる。

Here, by setting “iu = 0”, the command currents Idc and Iqc are expressed by the following equation (c2).

ただし、上記指令電流Ｉｄｃ，Ｉｑｃのベクトルの長さは、電動機１６の生成するトルクが電動機１６の出力軸を回転させるのに要する値の最小値未満となるようにする。続くステップＳ１４では、上記の式（ｃ２）の指令電流を周知の電圧方程式に代入することで得られるｄｑ軸上での指令電圧Ｖｄｃ、Ｖｑｃを、３相変換部５２にて３相の指令電圧Ｖｄｃ、ｖｖｃ、ｖｗｃに変換する処理を行う。すなわち、学習制御時においては、電流偏差算出部４４，４６やＰＩ制御部４８，５０の処理を行うことなく、指令電流Ｉｄｃ，Ｉｑｃから指令電圧Ｖｄｃ、Ｖｑｃを電圧方程式にて算出し、これら算出した指令電圧Ｖｄｃ、Ｖｑｃを、先の図２に示すセレクタＳ１、Ｓ２を介して３相変換部５２に出力する。

However, the vector lengths of the command currents Idc and Iqc are set so that the torque generated by the electric motor 16 is less than the minimum value required to rotate the output shaft of the electric motor 16. In the subsequent step S14, the command voltages Vdc and Vqc on the dq axis obtained by substituting the command current of the above formula (c2) into a known voltage equation are converted into three-phase command voltages by the three-phase conversion unit 52. A process of converting to Vdc, vvc, and vwc is performed. That is, at the time of learning control, the command voltages Vdc and Vqc are calculated from the command currents Idc and Iqc by the voltage equation without performing the processes of the current deviation calculation units 44 and 46 and the PI control units 48 and 50, and these calculations are performed. The command voltages Vdc and Vqc are output to the three-phase conversion unit 52 via the selectors S1 and S2 shown in FIG.

  In the subsequent step S16, it is determined whether or not the absolute value of the actual current iu is greater than or equal to a predetermined value α (> 0). This determination is for determining whether or not the learning rotation angle θn is deviated from the actual rotation angle of the electric motor 16. Here, the relational expression for converting the actual currents Id and Iq on the dq axis into the actual currents iu and iw is the following expression (c3).

このため、学習用回転角度θｎが電動機１６の実際の回転角度に対して角度誤差Δθを有している場合、上記の式（ｃ２）にて定義される指令電流によって実際に電動機１６を流れる実電流ｉｕ，ｉｗは、以下の式（ｃ４）となる。

Therefore, when the learning rotation angle θn has an angle error Δθ with respect to the actual rotation angle of the electric motor 16, the actual flow through the electric motor 16 by the command current defined by the above equation (c2). The currents iu and iw are expressed by the following equation (c4).

上記において、実電流ｉｕは、「ｉｕ＝ｉｗ・２・ｓｉｎΔθ／√３」となり、角度誤差Δθを有すると、実電流ｉｕはゼロとはならない。このため、実電流ｉｕの値を、角度誤差Δθと相関を有するパラメータとして利用することができる。

In the above, the actual current iu is “iu = iw · 2 · sin Δθ / √3”, and the actual current iu does not become zero if the angle error Δθ is present. Therefore, the value of the actual current iu can be used as a parameter having a correlation with the angle error Δθ.

  The predetermined value α is set to a value at which it can be determined that the learning rotation angle θn is deviated from the actual rotation angle of the electric motor 16. If it is determined in step S16 that the value is greater than or equal to the predetermined value α, the learning rotation angle θn is corrected in step S18. That is, for example, if the actual current iu is positive, the angle error Δθ is positive, so that the learning rotation angle θn is corrected to be decreased. If the actual current iu is negative, the angle error Δθ is negative, so the learning rotation angle θn. Increase the correction.

  When the correction of the learning rotation angle θn in step S18 is completed, the process returns to step S12. The processes in steps S12 to S18 are continued until it is determined in step S16 that the value is less than the predetermined value α. When it is determined in step S16 that it is less than the predetermined value α, it is determined that the current learning rotation angle θn matches the actual rotation angle of the electric motor 16 with high accuracy, and the process proceeds to step S20. In step S20, the error Δθ of the rotation angle θ detected by the resolver 36 is learned as “θn−θ”.

  In the subsequent step S22, the error Δθ is stored in the constant memory holding device 32 in association with the rotation angle θ detected by the resolver 36. Specifically, as shown in FIG. 3, a region where the rotation angle θ can be taken is previously divided into a plurality of regions, and among the divided regions, the rotation angle θ detected this time by the resolver 36 is included. The error Δθ is stored in the area. Here, the reason why the error Δθ is separately learned for each value of the rotation angle θ is that the detection error of the resolver 36 can change according to the rotation angle.

  In addition, when the process of step S22 is completed, this series of processes is once complete | finished.

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

  (1) The command currents Idc and Iqc are set based on the learning rotation angle θn so that the current flowing through the U phase of the electric motor 16 is zero, and the learning rotation angle θn is corrected so that the actual current iu becomes zero. However, the detection error of the resolver 36 was learned. Thus, the detection error of the resolver 36 can be learned while using the current flowing through the U phase of the current actually flowing through the electric motor 16 as a parameter having a correlation with the error of the learning rotation angle θn.

  (2) The detection error was learned for each of a plurality of values for the detection value of the resolver 36. Thereby, even if the detection error of the resolver 36 changes according to the rotation angle of the electric motor 16, the detection error can be learned appropriately.

  (3) Learning was performed when the rotation speed of the electric motor 16 was zero by using the start switch 34 to be turned on as a trigger. As a result, it is possible to accurately identify at which rotation angle the detection error has been learned. As a result, it is possible to appropriately learn the detection error at the rotation angle detected by the resolver 36.

  (4) The learning rotation angle θn was manipulated to feedback control the actual current iu to zero. Thereby, the learning rotation angle θn can be controlled to the actual rotation angle of the electric motor 16.

  (5) The torque generated by the electric motor 16 is set to be less than the minimum value required to rotate the output shaft of the electric motor 16. Thereby, it can avoid that the electric motor 16 shows the behavior which is not intended at the time of learning.

  (6) The output shaft of the electric motor 16 is connected to the drive wheel 18. For this reason, even if a current is passed through the motor 16 during learning, the output shaft of the motor 16 is difficult to rotate. For this reason, learning based on a change in the rotation angle is difficult. Further, even if the rotation angle can be changed, this means that the vehicle exhibits unintended behavior. On the other hand, in this embodiment, learning can be performed without changing the rotation angle.

  (7) (Resolver 36) was used as the angle detection device. Since the resolver 36 can change the angle error depending on the rotation angle of the electric motor 16, the effects (2) and (3) can be achieved particularly suitably.

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

  In the present embodiment, a motor having saliency is used as the electric motor 16. Since the salient pole machine has the minimum d-axis inductance, even if the command currents Idc and Iqc are vectors in an arbitrary direction, the vector of the current that actually flows through the motor 16 is deflected in the d-axis direction. In the present embodiment, the detection error of the resolver 36 is learned using this property.

  FIG. 4 shows a procedure of detection error learning processing of the resolver 36 according to the present embodiment. This process is repeatedly executed by the ECU 30, for example, at a predetermined cycle. In the process shown in FIG. 4, the same process as the process shown in FIG. 3 is given the same step number for convenience.

  In this series of processes, first, in step S30, it is determined whether or not the rotational speed of the electric motor 16 is equal to or lower than a predetermined speed β. This process can be performed depending on whether or not the time differential value of the rotation angle θ detected by the resolver 36 is equal to or less than a predetermined speed β. Here, the predetermined speed β is different from one of rotation angle regions into which the rotation angle θ detected by the resolver 36 is divided in order to store the error Δθ in a period during which the error Δθ is learned. It is set to a value that does not change into the area.

  When an affirmative determination is made in step S30, the same processing as step S10 of FIG. 3 is performed, and then the process proceeds to step S12a. In step S12a, energization control is performed in the d-axis direction based on the learning rotation angle θn. In this process, the learning processing unit 60 sets the absolute value of the command current Idc in the d-axis direction to be larger than zero and sets the command current Iqc in the q-axis direction to zero, and then converts them into a command on the dq axis using a voltage equation. This can be done by converting the voltages to Vdc and Vqc and outputting them to the three-phase converter 52. The absolute value of the command current Idc in the d-axis direction is set to a value that does not cause the torque generated in the motor 16 to rotate the drive wheels 18 even if a current flows in the q-axis direction due to the detection error of the resolver 36. Is done.

  In the subsequent step S16a, it is determined whether or not the absolute value of the actual current Iq on the q axis is equal to or greater than a predetermined value α (> 0). This determination is for determining whether or not the learning rotation angle θn is deviated from the actual rotation angle of the electric motor 16. That is, as described above, since the electric motor 16 is a salient pole machine, the inductance component in the d-axis direction is minimized. For this reason, even if the command currents Idc and Iqc are vectors in an arbitrary direction, the vector of the current that actually flows through the electric motor 16 is deflected in the d-axis direction. For this reason, especially when the current that actually flows when the current flows in the d-axis direction estimated at the learning rotation angle θn has an estimated q-axis component, the learning rotation angle θn is determined from the actual rotation angle. It is thought that it has shifted. If it is determined in step S16a that the value is greater than or equal to the predetermined value α, the learning rotation angle θn is corrected in step S18.

  That is, in this case, as indicated by a one-dot chain line in FIG. 5A, the d (θn) axis and the q (θn) axis derived from the learning rotation angle θn are shifted from the true dq axis. Therefore, as shown by the thick line in the figure, when current flows in the d (θn) axis direction, the current that actually flows through the motor 16 is deflected in the true d-axis direction of the motor 16. For this reason, the current flowing through the electric motor 16 has a component in the q (θn) axis direction derived from the learning rotation angle θn. Therefore, the learning rotation angle θn is corrected so that the q (θn) axis direction component derived from the learning rotation angle θn becomes zero.

  When the correction of the learning rotation angle θn in step S18 of FIG. 4 is completed, the process returns to step S12a. The processes in steps S12a to S18 are continued until it is determined in step S16a that the value is less than the predetermined value α. If it is determined in step S16a that it is less than the predetermined value α, it is determined that the current learning rotation angle θn matches the actual rotation angle of the motor 16 with high accuracy, as shown in FIG. The processes of steps S20 and S22 of FIG. 3 are performed.

  Also according to the present embodiment described above, an effect according to the first embodiment can be obtained.

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

  In each of the above embodiments, the initial value of the learning rotation angle θn is the rotation angle θ detected by the resolver 36, but is not limited thereto. For example, the learning rotation angle θn may be corrected by feedback-controlling an actually flowing current to a specified current with an arbitrary rotation angle as an initial value. Even in this case, when the current that actually flows becomes the specified current, the learning rotation angle θn used for setting the specified current matches the actual rotation angle of the electric motor 16. Therefore, the error Δθ of the rotation angle θ can be learned based on this and the rotation angle θ detected by the resolver 36.

  In each of the above embodiments, the current flowing through the electric motor 16 is feedback-controlled to the specified current (the current at which the U phase becomes zero, the current in the d-axis direction), but is not limited thereto. For example, the error Δθ is set according to the absolute value and sign of the U-phase actual current iu and the q-axis actual current Iq when performing open-loop control to a specified current (current at which the U-phase becomes zero, current in the d-axis direction). Presumed learning may be performed.

  In the first embodiment, learning may be performed when the rotational speed is equal to or lower than the predetermined speed β as in the second embodiment.

  In the second embodiment, as in the first embodiment, learning may be performed with the start switch 34 being turned on as a trigger.

  The angle detection means configured to include a member that rotates integrally with the shaft of the rotating machine is not limited to the resolver 36. For example, it may be a means for detecting a rotation angle provided with a magnetic sensing element such as a Hall element, an encoder, or the like.

  The motor 16 is not limited to a synchronous motor. For example, an induction motor or a step motor may be used.

  The rotating machine is not limited to an electric motor, and may be a generator 14 or the like, for example. Further, the rotating machine is not limited to that mounted on a hybrid vehicle, and may be mounted on an electric vehicle, for example. At this time, even if the rotating machine is an electric motor, the rotating machine is not necessarily configured as a vehicle power generation device, and may be mounted on a power steering, for example.

The figure which shows the whole structure of the hybrid system concerning 1st Embodiment. The block diagram which shows the process regarding control of the electric motor concerning the embodiment. The flowchart which shows the procedure of the error learning process of the rotation angle concerning the embodiment. 10 is a flowchart showing a procedure of rotation angle error learning processing according to the second embodiment. The figure which shows the said error learning process aspect.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Internal combustion engine, 14 ... Generator, 16 ... Electric motor, 30 ... ECU (one embodiment of an angle error learning device), 36 ... Resolver.

Claims (9)

In the angle error learning device for learning the detection error of the angle detection means for detecting the rotation angle of the rotating machine,
Learning means for learning a detection error of the angle detecting means by controlling the current flowing through the rotating machine to a specified current based on the detected value of the current actually flowing through the rotating machine and the detected value of the angle detecting means. An angle error learning device comprising:   The angle error learning device according to claim 1, wherein the angle detection unit includes a member that rotates integrally with a shaft of the rotating machine.   The angle error learning device according to claim 1, wherein the learning unit learns the detection error separately for each of a plurality of values of detection values of the angle detection unit.   The angle error learning device according to claim 1, wherein the learning is performed when a rotation speed of the rotating machine is equal to or less than a specified speed.   The learning means is a means for feedback-controlling the current actually flowing through the rotating machine to the specified current by operating a rotation angle used for control to the specified current, and the operated rotation angle and the angle detecting means. The angle error learning apparatus according to claim 1, further comprising: a unit that performs the learning based on a deviation from the detected value.   6. The specified current is set such that a torque generated by the rotating machine is less than a minimum value required for rotating a shaft of the rotating machine. The angle error learning device according to item.   The angle error learning device according to any one of claims 1 to 6, wherein the rotating machine constitutes an in-vehicle power generation device, and an output shaft of the rotating machine is connected to driving wheels. .   The angle error learning device according to claim 1, wherein the angle detection device is an electromagnetic induction type angle detector. The angle error learning device according to any one of claims 1 to 8,
An angle detection system comprising the angle detection means.

Images