JP5459184B2 - Electric power steering device - Google Patents

Description

  The present invention relates to an electric power steering apparatus for assisting a steering operation of a steering wheel by a driver.

  2. Description of the Related Art Conventionally, there is known an electric power steering device that applies a steering assist torque to a driver's steering operation. The electric power steering device detects a steering torque acting on the steering shaft by a torque detection device, calculates a target assist torque that increases as the steering torque increases, and obtains the calculated target assist torque. Feedback control of the energization amount. Therefore, in the electric power steering device, the reliability of the torque detection device is particularly required.

The torque detection device calculates a steering torque proportional to the torsion angle by detecting the torsion angle of the torsion bar provided on the steering shaft. For example, the torque detection device proposed in Patent Document 1 employs a configuration that detects the twist angle of a torsion bar using two resolvers. In this torque detection device, a first resolver is provided on one end side of the torsion bar, and a second resolver is provided on the other end side, and the first rotation angle θ ₁ detected by the first resolver and the second resolver are detected. from the difference between the second rotational angle theta ₂ that detects the steering torque.

  Each resolver includes an excitation coil that is supplied with an excitation AC signal and energizes the rotor coil, and a pair of detection coils (sin phase detection coil, cos phase detection coil) that are fixed around the torsion bar. The pair of detection coils are assembled with an electrical angle shifted by 90 degrees (π / 2). The sin phase detection coil outputs an AC signal having an amplitude corresponding to the sin value of the rotation angle of the rotor, and the cos phase detection coil outputs an AC signal having an amplitude corresponding to the cos value of the rotation angle of the rotor. Therefore, the rotation angle of the rotor detected by the resolver can be calculated based on the arctangent value obtained by dividing the amplitude of the output signal of the sin phase detection coil by the amplitude of the output signal of the cos phase detection coil.

The torque detection device proposed in Patent Document 1 is disconnected when the rotation angle detected by a normal resolver is within a predetermined angle range even when one of the detection coils in one resolver is disconnected. The rotation angle is estimated using the output signal of the other detection coil that is paired with the detected coil. This is based on the condition that the mechanical angle difference between the first rotation angle θ ₁ detected by the first resolver and the second rotation angle θ ₂ detected by the second resolver is always a certain value or less. In a situation where the rotation angle detected by a normal resolver is within a predetermined angle range, the fact that the rotation angle can be uniquely determined from the output signal of the other detection coil paired with the disconnected coil is utilized. It is a thing. Thus, in a situation where the rotation angle detected in normal resolver is within a predetermined angular range, it is possible to detect the steering torque from the difference between the first rotation angle theta ₁ and the second rotation angle theta _2.

JP 2003-315182 A

However, based on the condition that the mechanical angle difference between the first rotation angle θ ₁ and the second rotation angle θ ₂ is equal to or less than a certain value, the other detection coil (a normal detection coil paired with the disconnected coil) Even if the rotation angle is obtained from the output signal of), there are many situations in which the steering torque cannot be detected because there is a wide range in which the rotation angle is not uniquely determined.

  The present invention has been made to cope with the above problem, and an object of the present invention is to reduce the situation in which the steering torque cannot be detected even when one of the detection coils is disconnected.

In order to achieve the above object, a feature of the present invention is that a first resolver (110 having a pair of detection coils for outputting a two-phase detection signal corresponding to a first rotation angle that is a rotation angle on one end side of a steering shaft. ) And a second resolver (120) having a pair of detection coils for outputting a two-phase detection signal corresponding to a second rotation angle that is a rotation angle on the other end side of the steering shaft. ) And the first rotation angle based on the two-phase detection signal output from the first resolver, and the second rotation angle based on the two-phase detection signal output from the second resolver. A rotation angle calculation means (S15) for calculating, a steering torque calculation means (S16) for calculating a steering torque acting on the steering shaft based on the difference between the calculated first rotation angle and the second rotation angle; An electric motor (21) for generating an assist torque for assisting the steering operation, an assist control means (31, 40) for driving and controlling the electric motor based on the steering torque calculated by the steering torque calculating means, In either the first resolver or the second resolver, an abnormality detection means (S13) for detecting that an abnormality has occurred on one side of the pair of detection coils, and an abnormality occurrence on one side of the detection coil are detected. When the rotation angle can be uniquely determined from the output signal of the other detection coil paired with the detection coil in which the abnormality occurs, from the output signal of the other detection coil, In the electric power steering apparatus including rotation angle estimation means (S30, S40) for estimating the rotation angle of the abnormal side resolver,
When there is an abnormality on one side of the detection coil, the rotation angle cannot be estimated uniquely from the output signal of the other detection coil paired with the detection coil in which the abnormality has occurred. A non-estimable motor electrical angle region storage means (33) storing the region as an unestimable motor electrical angle region corresponding to the electrical angle region of the electric motor, and a motor electrical angle detection means for detecting the electrical angle of the electric motor (61, 307) and when the occurrence of abnormality on one side of the detection coil is detected, the electrical angle of the electric motor detected by the motor electrical angle detection means and the non-estimable motor electrical angle area storage means are stored. Drive control of the electric motor so that the electric angle of the electric motor does not stay in the non-estimable motor electric angle region In that a putative non region passing controller (310) that.

  In the present invention, the first resolver is provided on one end side (for example, the input side) of the steering shaft, and the second resolver is provided on the other end side (for example, the output side). The first resolver receives an excitation AC signal, and from a pair of detection coils (sin phase detection coil, cos phase detection coil), has a two-phase according to the first rotation angle that is the rotation angle on one end side of the steering shaft. A detection signal is output. Similarly, the second resolver receives an excitation AC signal and responds to a second rotation angle, which is a rotation angle on the other end side of the steering shaft, from a pair of detection coils (sin phase detection coil, cos phase detection coil). A two-phase detection signal is output. The two-phase detection signal is an AC voltage signal generated by the input AC signal for excitation, and depends on the sin phase detection signal whose amplitude changes depending on the sin value of the rotation angle and the cos value of the rotation angle. And a cos phase detection signal whose amplitude changes. Therefore, in each resolver, a sin phase detection signal is output from one detection coil, and a cos phase detection signal is output from the other detection coil.

  The rotation angle calculation means calculates the first rotation angle based on the two-phase detection signal output from the first resolver, and calculates the second rotation angle based on the two-phase detection signal output from the second resolver. For example, the amplitude of the sin phase detection signal and the cos phase detection signal in each resolver is detected, and the rotation angle is calculated based on the arctangent value obtained by dividing the amplitude of the sin phase detection signal by the amplitude of the cos phase detection signal. can do. The steering torque calculation means calculates a steering torque that acts on the steering shaft based on the difference between the first rotation angle and the second rotation angle thus calculated. That is, by detecting the twist angle of the steering shaft, a steering torque proportional to the twist angle is calculated. In this case, a torsion bar is interposed on the steering shaft so that the torsion angle can be easily detected, and the first rotation angle on one end side and the second rotation angle on the other end side of the torsion bar are calculated by two resolvers. It is preferable to do.

  The assist control means drives and controls the electric motor based on the steering torque calculated by the steering torque calculation means. The electric motor is connected to the output side of the steering shaft via a reduction gear or the like, for example, and generates assist torque that assists the driver's steering operation. In this case, the electric motor may apply a rotational torque to the steering shaft, or may be connected to a rack bar connected to the steering shaft to apply an axial force to the rack bar. Good.

  When the signal output line of the detection coil of the resolver is disconnected, a normal detection signal is not output from the detection coil. At this time, when an abnormality has occurred in only one of the pair of detection coils, there is a situation in which the rotation angle can be uniquely determined from the output signal of the other detection coil. Therefore, when the abnormality detection means detects that an abnormality has occurred on one side of the pair of detection coils in either the first resolver or the second resolver, the rotation angle estimation means outputs the output signal of the other detection coil. From the output signal of the other detection coil, the rotation angle of the abnormal side resolver (the resolver having the detection coil in which an abnormality has occurred) is estimated. Here, “estimating the rotation angle” means calculating the rotation angle in a situation where an abnormality has occurred on one side of the pair of detection coils.

  Therefore, in a situation where the rotation angle can be uniquely determined from the output signal of the other detection coil, the steering torque can be calculated, and the assist control can be performed based on the steering torque. However, there are situations where the rotation angle cannot be uniquely determined from the output signal of the other detection coil. Therefore, the present invention includes a non-estimable motor electrical angle area storage means, a motor electrical angle detection means, and a non-estimable area passage control means.

  The non-estimable motor electrical angle region storage means associates a rotation angle non-estimable region in which the rotation angle cannot be determined uniquely from the output signal of the other detection coil with the electrical angle region of the electric motor. Remember as. Therefore, when the electrical angle of the electric motor falls within the motor electrical angle region where estimation is impossible, the rotation angle of the resolver in which an abnormality has occurred cannot be calculated.

  Therefore, the non-estimable area passage control means stores the electrical angle of the electric motor detected by the motor electrical angle detection means and the non-estimable motor electrical angle area storage means when the occurrence of an abnormality on one side of the detection coil is detected. Based on the estimated non-estimable motor electrical angle region, the electric motor is driven and controlled so that the electrical angle of the electric motor does not remain in the unestimable motor electrical angle region. Therefore, the rotational angle of the resolver in which an abnormality has occurred passes through the rotational angle estimation impossible region. As a result, according to the present invention, it is possible to reduce the situation in which the steering torque cannot be detected even when one of the detection coils is disconnected.

  Another feature of the present invention is that the non-estimable region passage control means is configured such that when an abnormality is detected on one side of the detection coil, the electric angle of the electric motor enters the non-estimable motor electric angle region. The target current of the electric motor set by the assist control means is increased and corrected.

  In the present invention, when the electric angle of the electric motor falls within the unestimable motor electric angle region, the target current of the electric motor set by the assist control means is increased and corrected. As a result, the electric angle of the electric motor can easily pass through the non-estimable motor electric angle region.

  Another feature of the present invention is that motor rotation speed detection means (312) for detecting the rotation speed of the electric motor is provided, and the non-estimable region passage control means is used when the rotation speed of the electric motor is small. In comparison, the increase correction amount of the target current of the electric motor is reduced (313, 314).

  In the present invention, the increase correction amount of the target current of the electric motor is smaller when the rotation speed of the electric motor is large than when the rotation speed is small. Therefore, in a situation where the rotational inertia of the electric motor is large and it is easy to pass through the motor electrical angle region where the rotational angle cannot be estimated, the increase correction amount of the target current is reduced. The motor is prevented from turning.

  Another feature of the present invention is that the rotation angle estimation impossibility region is the first rotation angle or the second rotation angle per calculation period in which the rotation angle calculation means calculates the first rotation angle and the second rotation angle. Is set based on the rotation angle change condition that the change amount of the rotation angle that changes is within a predetermined change amount.

  In this case, when the detection coil in which the abnormality is detected is a sin phase detection coil, the electrical angle of the abnormal side resolver is 0 ° or 180 ° during one calculation cycle. If the detection coil in which the abnormality is detected is a cos phase detection coil, the electrical angle of the abnormal resolver is 90 ° or 270 during one calculation cycle. It is good to set it in the area that may pass through.

  When the calculation cycle of the rotation angle calculation means is multiplied by the rotation speed of the steering shaft, the change amount of the rotation angle that has changed during one calculation cycle, that is, the rotation angle change amount is obtained. On the other hand, since the speed at which the steering wheel is rotated is limited, the rotation angle change amount is within a predetermined change amount. Therefore, it is possible to estimate the rotation angle that can uniquely calculate the rotation angle of the abnormal side resolver having the detection coil in which the abnormality is detected based on the electrical angle of the abnormal side resolver calculated before one calculation cycle. It is possible to discriminate between an area and a rotation angle estimation impossible area that cannot be calculated uniquely.

  When the detection coil in which the abnormality is detected is a sin phase detection coil, if the electrical angle of the abnormal side resolver can pass 0 ° or 180 ° during one calculation cycle, the cos phase detection coil The rotation angle cannot be uniquely determined from the amplitude of the output detection signal. This is because it is not known which side of the electrical angle of the abnormal side resolver is 0 ° or 180 °. Similarly, when the detection coil in which an abnormality is detected is a cos phase detection coil, if the electrical angle of the abnormal side resolver can pass 90 ° or 270 ° during one calculation cycle, The rotation angle cannot be uniquely determined from the amplitude of the detection signal output from the detection coil. This is because it is not known which side of the electrical angle of the abnormal side resolver is 90 ° or 270 ° across.

  Therefore, in the present invention, when the detection coil in which the abnormality is detected in the rotation angle estimation impossible region is a sin phase detection coil, the electrical angle of the abnormal side resolver is 0 ° or 180 ° during one calculation cycle. When the detection coil in which an abnormality is detected is a cos phase detection coil, the electrical angle of the abnormal resolver is set to 90 ° or 270 ° during one calculation cycle. Set to an area that may pass. Therefore, according to the present invention, the rotation angle estimation impossibility region can be narrowed. Therefore, even when the electric motor is driven and controlled by the estimation impossibility region passage control means, the influence on the steering feeling can be reduced.

  In the above description, in order to help the understanding of the invention, the reference numerals used in the embodiments are attached to the configuration of the invention corresponding to the embodiments in parentheses. It is not intended to be limited to the embodiment defined by.

It is a schematic structure figure of the electric power steering device as an embodiment. It is a functional block diagram of an assist calculating part. It is a circuit block diagram of a torque sensor unit. It is a figure showing the range from which a rotation angle is uniquely determined based on 1st conditions, Comprising: (a) represents the range from which a rotation angle is uniquely determined from the detection signal of only a cos phase, (b) is a sin phase. Represents a range in which the rotation angle is uniquely determined from only the detection signal. It is a figure showing the range in which a rotation angle is uniquely determined based on 2nd conditions, (a) represents the range from which a rotation angle is uniquely determined from the detection signal of only a cos phase, (b) is a sin phase. Represents a range in which the rotation angle is uniquely determined from only the detection signal. It is a flowchart showing a steering torque detection routine. It is a flowchart showing a 2nd rotation angle calculation routine (subroutine). It is a flowchart showing a 1st rotation angle calculation routine (subroutine). It is a figure showing the range from which a rotation angle is uniquely determined based on 1st conditions, Comprising: (a) represents the range from which a rotation angle is uniquely determined from the detection signal of only a cos phase, (b) is a sin phase. Represents a range in which the rotation angle is uniquely determined from only the detection signal. It is a figure showing the range in which a rotation angle is uniquely determined based on 2nd conditions, (a) represents the range from which a rotation angle is uniquely determined from the detection signal of only a cos phase, (b) is a sin phase. Represents a range in which the rotation angle is uniquely determined from only the detection signal. It is a flowchart showing a raising current calculation routine. FIG. 6 is a graph showing a raised current map, where (a) shows a raised current map when the sin phase is abnormal, and (b) shows a raised current map when the cos phase is abnormal. It is a graph showing a gain setting map. It is a figure showing the relationship between the amplitude of a sin phase, the amplitude of a cos phase, and an electrical angle.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic configuration diagram of an electric power steering apparatus as an embodiment.

  The electric power steering device for a vehicle includes a steering mechanism 10 that steers left and right front wheels FW1 and FW that are steered wheels by steering a steering handle 11, and a power assist unit 20 that is provided in the steering mechanism 10 and generates steering assist torque. And an assist control device 50 (hereinafter referred to as an assist ECU 50) for driving and controlling the electric motor 21 of the power assist unit 20, a vehicle speed sensor 60, and a resolver unit 100.

  The steering mechanism 10 includes a steering shaft 12 that is rotatable. A steering handle 11 is connected to the upper end of the steering shaft 12 so as to be integrally rotatable, and a pinion gear 13 is connected to the lower end of the steering shaft 12 so as to be integrally rotatable. The pinion gear 13 meshes with rack teeth formed on the rack bar 14 to constitute a rack and pinion mechanism. Left and right front wheels FW1, FW2 are connected to both ends of the rack bar 14 via a tie rod and a knuckle arm (not shown) so as to be steerable. The left and right front wheels FW1 and FW2 are steered to the left and right according to the axial displacement of the rack bar 14 as the steering shaft 12 rotates about the axis.

  A power assist unit 20 is assembled to the rack bar 14. The power assist unit 20 includes a steering assist electric motor 21 (for example, a three-phase DC brushless motor) and a ball screw mechanism 22. The rotating shaft of the electric motor 21 is connected to the rack bar 14 through the ball screw mechanism 22 so as to be able to transmit power, and assists the steering of the left and right front wheels FW1, FW2 by the rotation. The ball screw mechanism 22 functions as a speed reducer and a rotation-linear converter, and decelerates the rotation of the electric motor 21 and converts it into a linear motion and transmits it to the rack bar 14. Instead of assembling the electric motor 21 to the rack bar 14, the electric motor 21 is assembled to the steering shaft 12, and the rotation of the electric motor 21 is transmitted to the steering shaft 12 via the speed reducer so that the shaft 12 is axially connected. You may comprise so that it may drive around.

  The electric motor 21 is provided with a rotation angle sensor 61 for detecting the rotation angle of the rotation shaft. The rotation angle sensor 61 is incorporated in the electric motor 21 and outputs a detection signal corresponding to the rotation angle position of the rotor of the electric motor 21, and is constituted by a resolver. The detection signal of the rotation angle sensor 61 is used for calculating the rotation angle and rotation speed of the electric motor 21. On the other hand, the rotation angle of the electric motor 21 corresponds to the steering angle of the front wheels FW1 and FW2, which are steered wheels, and is used for detection of the steering angle.

  The steering shaft 12 is provided with a torsion bar 12a at an intermediate position in the axial direction. In the steering shaft 12, a portion connecting the upper end of the torsion bar 12a and the steering handle 11 is called an input shaft 12in, and a portion connecting the lower end of the torsion bar 12a and the pinion gear 13 is called an output shaft 12out.

A resolver unit 100 is provided on the steering shaft 12. The resolver unit 100 includes a torsion bar 12a, a first resolver 110 assembled to the input shaft 12in, and a second resolver 120 assembled to the output shaft 12out. The first resolver 110 outputs a signal according to the rotation angle of the input shaft 12in (the rotation angle at one end of the torsion bar 12a and corresponds to the first rotation angle of the present invention), and the second resolver 120 A signal corresponding to the rotation angle of the output shaft 12out (the rotation angle at the other end of the torsion bar 12a and corresponding to the second rotation angle of the present invention) is output. When the steering handle 11 is turned, torque acts on the steering shaft 12 and the torsion bar 12a is twisted. The torsion angle of the torsion bar 12a is proportional to the steering torque acting on the steering shaft 12. Therefore, the steering torque acting on the steering shaft 12 is detected by obtaining the difference between the first rotation angle θ ₁ detected by the first resolver 110 and the second rotation angle θ ₂ detected by the second resolver 120. Can do. The first resolver 110 and the second resolver 120 are connected to the assist ECU 50.

  The assist ECU 50 includes a calculation unit 30 including a microcomputer and a signal processing circuit, and a motor drive circuit 40 (for example, a three-phase inverter circuit) configured by a switching circuit. The calculation unit 30 includes an assist calculation unit 31, a torque calculation unit 32, and a storage unit 33. The torque calculator 32 is connected to the resolver unit 100 and detects the steering torque acting on the steering shaft 12 by calculation. Further, the torque calculation unit 32 is connected with a warning lamp 65 for notifying the driver of the abnormality, and turns on the warning lamp 65 when disconnection is detected, which will be described later. The assist calculation unit 31 calculates a control amount of the electric motor 21 based on the steering torque detected by the torque calculation unit 32 and outputs a control signal corresponding to the control amount to the motor drive circuit 40. The storage unit 33 includes a memory that stores various control programs and control data used by the assist calculation unit 31 and the torque calculation unit 32.

  The motor drive circuit 40 receives the PWM control signal from the assist calculation unit 31 and adjusts the energization amount to the electric motor 21 by controlling the duty ratio of the internal switching element. The motor drive circuit 40 is provided with a current sensor 41 that detects a current flowing through the electric motor 21.

  The assist calculation unit 31 connects a current sensor 41, a vehicle speed sensor 60, and a rotation angle sensor 61. The vehicle speed sensor 60 outputs a vehicle speed detection signal representing the vehicle speed vx. The assist calculation unit 31 inputs the calculation result of the steering torque calculated by the torque calculation unit 32.

  Next, the assist calculation unit 31 will be described. FIG. 2 is a functional block diagram of the assist calculation unit 31 processed by program control of the microcomputer of the assist ECU 50. The assist calculation unit 31 controls energization of the electric motor 21 by current vector control represented by the dq coordinate system. Focusing on the function, the assist torque setting unit 301, the assist current command unit 302, and feedback control are performed. Unit 303, three-phase / two-phase coordinate conversion unit 304, two-phase / three-phase coordinate conversion unit 305, PWM signal generation unit 306, electrical angle conversion unit 307, and a raising current setting unit which is a characteristic part in the present embodiment 310 is provided.

  The assist torque setting unit 301 acquires the vehicle speed vx detected by the vehicle speed sensor 60 and the steering torque Tr calculated by the torque calculation unit 32, and based on the acquired vehicle speed vx and the steering torque Tr, the target assist torque T * Is calculated. The target assist torque T * is set so as to increase as the steering torque Tr increases and decrease as the vehicle speed vx increases with reference to an assist map (not shown) stored in the storage unit 33. .

  The assist torque setting unit 301 outputs the target assist torque T * to the assist current command unit 302. The assist current command unit 302 calculates the q-axis command current Iq * in the dq coordinates by dividing the target assist torque T * by the torque constant. The assist current command unit 302 sets the d-axis command current Id * to zero (Id * = 0). The q-axis command current Iq * works to generate torque by the electric motor 21. The d-axis command current Id * is used for field-weakening control of the electric motor 21, but here, it will be described as zero.

  The assist current command unit 302 outputs the d-axis command current Id * to the feedback control unit 303, and outputs the q-axis command current Iq * to the raised current setting unit 310. The raising current setting unit 310 corrects the q-axis command current Iq * input from the assist current command unit 302 by executing a later-described raising current calculation routine, and finally outputs the corrected q-axis command current Iq *. Is output to the feedback control unit 303 as a q-axis command current Iq *. The raising current setting unit 310 will be described later.

  The feedback control unit 303 calculates a deviation ΔIq obtained by subtracting the q-axis actual current Iq from the q-axis command current Iq *, and the q-axis actual current Iq is changed to the q-axis command current Iq * by proportional-integral control using the deviation ΔIq. The q-axis command voltage Vq * is calculated so as to follow. Similarly, a deviation ΔId obtained by subtracting the d-axis actual current Id from the d-axis command current Id * is calculated, and the d-axis actual current Id follows the d-axis command current Id * by proportional-integral control using the deviation ΔId. D-axis command voltage Vd * is calculated.

  The q-axis actual current Iq and the d-axis actual current Id are obtained by converting the detected values Iu, Iv, and Iw of the three-phase current actually flowing in the coil of the electric motor 21 into the two-phase current of the dq coordinate. Conversion from the three-phase currents Iu, Iv, and Iw to the two-phase currents Id and Iq in the dq coordinates is performed by the three-phase / 2-phase coordinate conversion unit 304. The three-phase / two-phase coordinate conversion unit 304 receives the motor electrical angle θme output from the electrical angle conversion unit 307, and based on the motor electrical angle θme, the three-phase currents Iu, Iv, Iw is converted into two-phase currents Id and Iq in dq coordinates.

  The electrical angle conversion unit 307 calculates the electrical angle θme of the electric motor 21 from the detection signal representing the rotation angle θm output from the rotation angle sensor 61, and the calculated motor electrical angle θme is a three-phase / 2-phase coordinate conversion unit 304. The data is output to the 2-phase / 3-phase coordinate conversion unit 305 and the raising current setting unit 310.

  The q-axis command voltage Vq * and the d-axis command voltage Vd * calculated by the feedback control unit 303 are output to the 2-phase / 3-phase coordinate conversion unit 305. The two-phase / 3-phase coordinate conversion unit 305 converts the q-axis command voltage Vq * and the d-axis command voltage Vd * into the three-phase command voltages Vu *, Vv *, Vw * based on the motor electrical angle θme, The converted three-phase command voltages Vu *, Vv *, Vw * are output to the PWM signal generator 306. The PWM signal generator 306 outputs a PWM control signal corresponding to the three-phase command voltages Vu *, Vv *, Vw * to a switching element (not shown) of the motor drive circuit 40. Thus, a target current in a direction rotating in the same direction as the driver's steering direction flows through the electric motor 21 by current feedback control. As a result, the driver's steering operation is appropriately assisted by the torque generated by the electric motor 21.

  Next, the resolver unit 100 will be described. FIG. 3 shows a schematic circuit configuration of the resolver unit 100. The first resolver 110 includes an input shaft 12in as a rotor. A first excitation coil 111 wound along the circumferential direction of the rotor is fixedly provided on the stator on the outer peripheral side of the input shaft 12in. A first rotor coil 114 is fixedly provided on the input shaft 12in serving as a rotor. The first rotor coil 114 rotates as the rotor rotates. The first rotor coil 114 is electrically connected to the first excitation coil 111 in a non-contact manner via a transformer (not shown) provided in the rotor, and is energized by an AC voltage applied to the first excitation coil 111. Is done. Although not shown, a plurality of first rotor coils 114 are arranged at equiangular intervals so that the electrical angle is N times the mechanical rotation angle of the rotor in order to increase the resolution of the rotation angle. .

  The first resolver 110 includes a first sin phase detection coil 112 and a first cos phase detection coil 113 on a stator on the outer peripheral side of the input shaft 12in. The first sin phase detection coil 112 and the first cos phase detection coil 113 are arranged at positions where the electrical angles are shifted from each other by π / 2.

  The first sin phase detection coil 112 and the first cos phase detection coil 113 are arranged on the rotation plane of the first rotor coil 114 and output an alternating voltage signal by the magnetic flux generated by the first rotor coil 114. The amplitude value of the AC voltage signal generated in the first sin phase detection coil 112 and the first cos phase detection coil 113 varies depending on the rotational position of the first sin phase detection coil 112 and the first cos phase detection coil 113 with respect to the first rotor coil 114. To do. That is, the first sin phase detection coil 112 outputs an AC voltage signal having an amplitude corresponding to the sin value of the rotation angle of the input shaft 12in, and the first cos phase detection coil 113 sets the cos value of the rotation angle of the input shaft 12in. An AC voltage signal having a corresponding amplitude is output.

  One end of the first excitation coil 111 is connected to the first excitation signal output port 50pe1 of the assist ECU 50 via the first excitation line 210. Further, one end of the first sin phase detection coil 112 is connected to the first sin phase signal input port 50ps1 of the assist ECU 50 via the first sin phase detection line 212. One end of the first cos phase detection coil 113 is connected to the first cos phase signal input port 50pc1 of the assist ECU 50 via the first cos phase detection line 213. In the figure, reference numeral 100pe1 is a first excitation signal input port of the resolver unit 100, reference numeral 100ps1 is a first sin phase signal output port, and reference numeral 100pc1 is a first cos phase signal output port. Therefore, the portion that becomes the wire harness includes the first excitation line 210 between the first excitation signal output port 50pe1 and the first excitation signal input port 100pe1, the first sin phase signal output port 100ps1, and the first sin phase signal input port 50ps1. And a first cos phase detection line 212 between the first cos phase signal output port 100pc1 and the first cos phase signal input port 50pc1.

  The other end of the first excitation coil 111, the other end of the first sin phase detection coil 112, and the other end of the first cos phase detection coil 113 are connected to the ground port 50pg of the assist ECU 50 via the common ground line 240. In the figure, reference numeral 100pg denotes a ground port of the resolver unit 100. Therefore, the common ground line 240 between the ground port 50pg and the ground port 100pg is a wire harness portion.

  The second resolver 120 includes an output shaft 12out as a rotor. A second excitation coil 121 wound along the circumferential direction of the rotor is fixedly provided on the stator on the outer peripheral side of the output shaft 12out. A second rotor coil 124 is fixedly provided on the output shaft 12out serving as a rotor. The second rotor coil 124 rotates as the rotor rotates. The second rotor coil 124 is electrically connected to the second excitation coil 121 in a non-contact manner via a transformer (not shown) provided in the rotor, and is energized by an AC voltage applied to the second excitation coil 121. Is done. Although not shown, a plurality of second rotor coils 124 are arranged at equal angular intervals so that the electrical angle is N times the mechanical rotation angle of the rotor in order to increase the resolution of the rotation angle. .

  The second resolver 120 includes a second sin phase detection coil 122 and a second cos phase detection coil 123 on the outer peripheral side of the output shaft 12out. The second sin phase detection coil 122 and the second cos phase detection coil 123 are arranged at positions where the electrical angles are shifted from each other by π / 2.

  The second sin phase detection coil 122 and the second cos phase detection coil 123 are arranged on the rotation plane of the second rotor coil 124, and output an AC voltage signal by the magnetic flux generated by the second rotor coil 124. The amplitude value of the AC voltage signal generated in the second sin phase detection coil 122 and the second cos phase detection coil 123 varies depending on the rotational position of the second sin phase detection coil 122 and the second cos phase detection coil 123 with respect to the second rotor coil 124. To do. That is, the second sin phase detection coil 122 outputs an AC voltage signal having an amplitude corresponding to the sin value of the rotation angle of the output shaft 12out, and the second cos phase detection coil 123 sets the cos value of the rotation angle of the output shaft 12out. An AC voltage signal having a corresponding amplitude is output.

  One end of the second excitation coil 121 is connected to the second excitation signal output port 50pe2 of the assist ECU 50 via the second excitation line 220. Also, one end of the second sin phase detection coil 122 is connected to the second sin phase signal input port 50ps2 of the assist ECU 50 via the second sin phase detection line 222. One end of the second cos phase detection coil 123 is connected to the second cos phase signal input port 50pc2 of the assist ECU 50 via the second cos phase detection line 223. In the figure, reference numeral 100pe2 is a second excitation signal input port of the resolver unit 100, reference numeral 100ps2 is a second sin phase signal output port, and reference numeral 100pc2 is a second cos phase signal output port. Accordingly, the portion that becomes the wire harness includes the second excitation line 220 between the second excitation signal output port 50pe2 and the second excitation signal input port 100pe2, the second sin phase signal output port 100ps2, and the second sin phase signal input port 50ps2. And a second cos phase detection line 222 between the second cos phase signal output port 100pc2 and the second cos phase signal input port 50pc2.

  The other end of the second excitation coil 121, the other end of the second sin phase detection coil 122, and the other end of the second cos phase detection coil 123 are connected to the ground port 50pg of the assist ECU 50 via the common ground line 240.

The assist ECU 50 includes a coil drive circuit 52. The coil drive circuit 52 includes a first excitation coil drive circuit 521 and a second excitation coil drive circuit 522. The first excitation coil drive circuit 521 outputs an excitation AC voltage having a constant period and amplitude from the first excitation signal output port 50pe1. Hereinafter, the excitation AC voltage outputted from the first excitation signal output port 50pe1 referred to as a first excitation signal, referred to the voltage value of the first excitation signal and the first excitation voltage V _1. The first excitation voltages V _1, when the amplitude and E _1, represented by the following equation.
V ₁ = E ₁ · sin (ωt)

Further, the second excitation coil drive circuit 522 applies the excitation AC voltage set to have the same frequency and the same phase as the excitation AC voltage output from the first excitation coil drive circuit 521 to the second excitation. Output from the signal output port 50pe2. Hereinafter, the excitation AC voltage outputted from the second excitation signal output port 50pe2 called the second excitation signal, referred to the voltage value of the second excitation signal and the second excitation voltage V _2. Second excitation voltage V _2, when the amplitude and E _2, is expressed by the following equation.
V ₂ = E ₂ · sin (ωt) (2)

The amplitudes E ₁ and E ₂ of the first excitation voltage V ₁ and the second excitation voltage V ₂ are set according to the characteristics of the first resolver 110 and the second resolver 120. In addition, the first excitation coil drive circuit 521 and the second excitation coil drive circuit 522 are shared, and an excitation signal is supplied to the resolver unit 100 from one excitation coil drive circuit (for example, the first excitation coil drive circuit 521). You may do it. In this case, the first excitation line 210 and the second excitation line 220 can be made one.

  When the first excitation coil 111 of the first resolver 110 is excited by the first excitation signal, an alternating voltage is generated in the first sin phase detection coil 112 and the first cos phase detection coil 113. Further, when the second excitation coil 121 of the second resolver 120 is excited by the second excitation signal, an AC voltage is generated in the second sin phase detection coil 122 and the second cos phase detection coil 123.

The AC voltage signal output from the first sin phase detection coil 112 is referred to as a first sin phase detection signal, and the voltage value is referred to as a first sin phase detection voltage Es1. The AC voltage signal output from the first cos phase detection coil 113 is referred to as a first cos phase detection signal, and the voltage value is referred to as a first cos phase detection voltage Ec1. The first sin phase detection voltage Es1 and the first cos phase detection voltage Ec1 are expressed by the following equations.
_{Es1 = K 1 · E 1 ·} sin (N · θ 1) · sin (ωt + α) ··· (3)
_{Ec1 = K 1 · E 1 ·} cos (N · θ 1) · sin (ωt + α) ··· (4)

The AC voltage signal output from the second sin phase detection coil 122 is referred to as a second sin phase detection signal, and the voltage value is referred to as a second sin phase detection voltage Es2. The AC voltage signal output from the second cos phase detection coil 123 is referred to as a second cos phase detection signal, and the voltage value thereof is referred to as a second cos phase detection voltage Ec2. The second sin phase detection voltage Es2 and the second cos phase detection voltage Ec2 are expressed by the following equations.
Es2 = K ₂ · E ₂ · sin (N · θ ₂ ) · sin (ωt + α) (5)
Ec2 = K ₂ · E ₂ · cos (N · θ ₂ ) · sin (ωt + α) (6)

Here, θ ₁ is the angle of the rotor of the first resolver 110 directly connected to the input shaft 12 in, θ ₂ is the angle of the rotor of the second resolver 120 directly connected to the output shaft 12 out, K ₁ is the transformation ratio of the first resolver 110, K ₂ is the transformation ratio of the second resolver 120, N is the shaft multiple angle of the first resolver 110 and the second resolver 120, α is the phase delay amount (input / output phase difference) with respect to the excitation signal, ω is the excitation frequency, and t is the time. Represent.

  The assist ECU 50 converts the first sin phase detection signal, the first cos phase detection signal, the second sin phase detection signal, and the second cos phase detection signal into the first sin phase detection line 212, the first cos phase detection line 213, and the second sin phase detection line 222, respectively. , Input via the second cos phase detection line 223. The assist ECU 50 inputs the first sin phase detection signal, the first cos phase detection signal, the second sin phase detection signal, and the second cos phase detection signal to the amplifiers 51s1, 51c1, 51s2, and 51c2, and amplifies the voltage of each detection signal with respect to the ground potential. Then, the amplified voltage signal is converted into a digital value by an A / D converter (not shown), and the digital value is input to a microcomputer to perform torque calculation processing.

  The torque calculation unit 32 in the assist ECU 50 includes a circuit that amplifies the first sin phase detection signal, the first cos phase detection signal, the second sin phase detection signal, and the second cos phase detection signal, converts them into digital signals, and inputs them to the microcomputer, and coil drive The circuit 52 and a functional unit that performs torque calculation processing by a microcomputer are included.

Next, a method for calculating the steering torque will be described. First, the method of detecting the rotor angles θ ₁ and θ ₂ of the first resolver 110 and the second resolver 120 will be described. Since both are calculated by the same method, here, a method of detecting the first rotation angle θ ₁ of the first resolver 110 will be described as an example.

ここで、ｆ_i，t_iは、それぞれサンプリングされた値と時間（サンプリング周期）の離散値である。 The amplitude of the _first sin phase detection signal is represented by K ₁ · E ₁ · sin (N · θ ₁ ) from the equation (3). Further, the amplitude of the _first cos phase detection signal is expressed by K ₁ · E ₁ · cos (N · θ ₁ ) from the equation (4). Here, the amplitude of the first sin phase detection signal is assumed to be As1 (As1 = K ₁ · E ₁ · sin (N · θ ₁ )). Further, the amplitude of the first cos phase detection signal is Ac1 (Ac1 = K ₁ · E ₁ · sin (N · θ ₁ )). The torque calculation unit 32 of the assist ECU 50 samples the voltage values of the first sin phase detection signal and the first cos phase detection signal at a cycle shorter than the cycle of the first excitation signal. For example, the voltage value of each detection signal is sampled four times at equal time intervals for one cycle of the first excitation signal. Since the sampled voltage value is a discrete value, the amplitudes As1 and Ac1 can be calculated by the following equations (7) and (8) using the least square method.

Here, f _i and t _i are discrete values of the sampled value and time (sampling period), respectively.

こうして、第１回転角θ_１は、次式（１０）により計算することができる。

Therefore, the following equation (9) can be calculated from the calculated values of these two equations.

Thus, the first rotation angle θ ₁ can be calculated by the following equation (10).

ここで、Ａs2は第２ｓｉｎ相検出信号の振幅（Ａs2＝Ｋ_２・Ｅ_２・ｓｉｎ（Ｎ・θ_２））であり、Ａc2は第２ｃｏｓ相検出信号の振幅（Ａc2＝Ｋ_２・Ｅ_２・ｃｏｓ（Ｎ・θ_２））である。 Similarly, the second rotation angle θ ₂ of the second resolver 120 can be calculated by the following equation (11).

Here, As2 is the amplitude of the _second sin phase detection signal (As2 = K ₂ · E ₂ · sin (N · θ ₂ )), and Ac2 is the amplitude of the second cos phase detection signal (Ac2 = K ₂ · E ₂ · cos (N · θ ₂ )).

When the first rotation angle θ ₁ and the second rotation angle θ ₂ are thus obtained, the steering torque Tr can be calculated by the following equation (12).
Tr = Kb · (θ ₁ −θ ₂ ) (12)
Here, Kb is a proportionality constant (spring constant) determined according to the torsional characteristics of the torsion bar 12a, and is stored in the storage unit 33 in advance.

Next, a method of detecting the rotation angle θ ₁ (θ ₂ ) when an abnormality occurs in either the sin phase or the cos phase coil in each of the resolvers 110 and 120 will be described. First, the case where an abnormality occurs in the sin phase of the first resolver 110 (when the first sin phase detection signal cannot be detected normally) will be described as an example. The abnormality of the coil is not only the abnormality of the coils 112, 113, 122, and 123 itself, but also the first sin phase detection line 212, the first cos phase detection line 213, the second sin phase detection line 222, and the second cos phase detection line 223. This means a state in which the sin phase detection signal or the cos phase detection signal is not normally input to the assist ECU 50, such as disconnection.

As described above, the first rotation angle θ ₁ is obtained by calculation using the above equation (10) from the amplitude As1 of the first sin phase detection signal and the amplitude Ac1 of the first cos phase detection signal. For this reason, when the amplitude As1 cannot be detected, it cannot be calculated using the above equation (10). In this case, the rotation angle θ ₁ is calculated from the amplitude of the detection signal output from the phase coil in which no abnormality is detected (here, the first cos phase detection coil 113).

As described above, the amplitude As1 of the _first sin phase detection signal is represented by As1 = K ₁ · E ₁ · sin (N · θ ₁ ), and the amplitude Ac1 of the first cos phase detection signal is Ac1 = K ₁ · E. ₁ · cos (N · θ ₁ ). Therefore, the following relational expressions (13) and (14) are obtained.

Here, the value of K ₁ · E ₁ can be measured and stored in advance when the resolver unit 100 is normal. The relationship between K ₁ · E ₁ and the amplitudes As 1 and Ac 1 is as shown in FIG. Therefore, the following formula (15) is obtained from these two formulas, and the calculation formula for the first rotation angle θ ₁ shown in the following formula (16) is obtained.

Similarly, when an abnormality occurs in the cos phase of the first resolver 110, the calculation formula of the rotation angle theta ₁ showing the amplitude As1 of the 1sin phase detection signal to the following equation (17) is obtained.

尚、Ｋ_２・Ｅ_２と振幅Ａs2，Ａc2の関係は、図１４（ｂ）に示すものとなる。 Further, the same can be considered when one side phase coil abnormality occurs in the second resolver 120. When an abnormality occurs in the sin phase, the calculation formula of the rotation angle θ ₂ shown in the following equation (18) is obtained from the amplitude Ac2 of the second cos phase detection signal, and when abnormality occurs in the cos phase, from the amplitude As2 of 2sin phase detection signal calculation formula for the rotation angle theta ₂ in the following equation (19) is obtained.

The relationship between K ₂ · E ₂ and the amplitudes As ₂ and Ac ₂ is as shown in FIG.

  Since these calculation formulas (16), (17), (18), and (19) are for obtaining the rotation angle from the amplitude of the detection signal of the one-sided phase, two solutions are obtained, and the rotation is uniquely performed. I can't derive the corner. Therefore, in this embodiment, focusing on the fact that a situation in which one of the two solutions can be specified occurs based on the following first condition and second condition, such a situation has occurred. Sometimes the rotation angle is uniquely determined from the amplitude of the detection signal of one side phase.

  First, the first condition and a situation where the rotation angle can be uniquely calculated from the amplitude of the detection signal of the one-sided phase determined based on the first condition (hereinafter referred to as a first condition estimable situation) will be described. To do.

The torsion bar 12a provided on the steering shaft 12 is twisted by a steering operation, but the twisting angle has a limit. In a normal steering operation, for example, it is less than 6 [deg] (mechanical angle). Yes. That is, the difference between the first rotation angle θ ₁ that is the angle of the rotor of the first resolver 110 and the second rotation angle θ ₂ that is the angle of the rotor of the second resolver 120 is always less than 6 [deg]. When this is expressed in terms of the electrical angle of the resolver, when the axial multiple angle N of the resolver is 8 (N = 8), the electrical angle θe ₁ (= N · The difference between θ ₁ ) and the electrical angle θe ₂ (= N · θ ₂ ) of the second resolver 120 is less than 48 [deg].
| Θe ₁ −θe ₂ | <48 ° (20)
Thus, the first condition is that the first rotation angle θ ₁ and the second rotation angle θ ₂ are always less than the predetermined angle difference. In the formula, [deg] is represented by [°].

The reason why the electrical angle θe ₁ is uniquely determined only from the amplitude Ac1 of the cos phase is that the electrical angle θe ₁ is in a range from 0 ° to 180 ° as shown in the following equations (21) and (22). Or it is a case where it is known beforehand that it is the range to 180 degrees-360 degrees.
0 ° <θe ₁ <180 ° (21)
180 <θe ₁ <360 ° (22)

Therefore, from the first condition (formula (20)), if the electrical angle θe ₂ of the second resolver 120 on the normal side is in the range of the following formulas (23) and (24), the electrical angle can be obtained only from the cos phase amplitude Ac1. θe ₁ is uniquely determined.
0 ° + 48 ° <θe ₂ <180 ° -48 ° (23)
180 + 48 ° <θe ₂ <360-48 ° (24)

If this is rearranged, the following equations (25) and (26) are obtained.
48 ° <θe ₂ <132 ° (25)
228 ° <θe ₂ <312 ° (26)
Therefore, when an abnormality in the sin phase of the first resolver 110 has occurred, the electrical angle .theta.e ₂ of the second resolver 120 of the normal side is equation (25), a situation that has entered the range represented by (26) The first condition can be estimated.

Similarly, even when an abnormality occurs in the cos phase of the first resolver 110 and only the amplitude As1 of the sine phase is known, in the following situation, the first rotation angle θ _{1 is determined} only from the amplitude As1. Can be calculated.

The reason why the electrical angle θe ₁ is uniquely determined only from the amplitude As1 of the sin phase is that the electrical angle θe ₁ is in a range from −90 ° to 90 ° as shown in the following equations (27) and (28). Or it is a case where it is known beforehand that it is the range from 90 degrees to 270 degrees.
−90 ° <θe ₁ <90 ° (27)
90 <θe ₁ <270 ° (28)

Therefore, the first condition (equation (20)) from the electrical angle .theta.e ₂ of the second resolver 120 of the normal side is equation (29), an electrical angle of only the amplitude As1 of long range, sin phase (30) θe ₁ is uniquely determined.
−90 ° + 48 ° <θe ₂ <90 ° -48 ° (29)
90 + 48 ° <θe ₂ <270-48 ° (30)

If this is rearranged, the following equations (31) and (32) are obtained.
−42 ° <θe ₂ <42 ° (31)
138 ° <θe ₂ <222 ° (32)
In addition, when the electrical angle is expressed by a value between 0 ° and 360 °, it can be expressed by the following equations (33), (34), and (35).
0 ° <θe ₂ <42 ° (33)
138 ° <θe ₂ <222 ° (34)
318 ° <θe ₂ <360 ° (35)
Accordingly, when an abnormality occurs in the cos phase of the first resolver 110, the electrical angle θe ₂ of the normal second resolver 120 is also in the range represented by the equations (33) to (35). The first condition can be estimated.

  Hereinafter, the ranges represented by the formulas (25) and (26) and the ranges represented by the formulas (33) to (35) are referred to as ranges that can be estimated under the first condition.

Similarly, the case where abnormality occurs in any of the coils of the sin phase or cos phase of the second resolver 120 is also based on the electrical angle .theta.e ₁ of the first resolver 110 of the normal side, the electrical angle .theta.e ₁ in advance When entering the set angle range (a range in which θe ₂ is replaced by θe ₁ in the formulas (25), (26), (33) to (35)), the situation can be estimated under the first condition. As2 or can be calculated second rotation angle theta ₂ from only the amplitude Ac2.

  FIG. 4 (a) shows the estimable range (electrical angle of the normal resolver) under the first condition when an abnormality occurs in the sin phase, and FIG. 4 (b) shows the state when the abnormality occurs in the cos phase. It represents the range that can be estimated under the first condition (the electrical angle of the normal resolver). In the figure, a range indicated by an arrow is a range that can be estimated under the first condition, and a filled range is a range in which the rotation angle is not uniquely determined. As can be seen from this figure, the estimable range of the first condition set from the first condition is about half of the whole and is not so wide.

  Therefore, in the present embodiment, a situation in which the rotation angle can be uniquely calculated from the amplitude of the detection signal of the one-sided phase based on the second condition (hereinafter referred to as a situation that can be estimated under the second condition). Set. Hereinafter, the second condition and the situation that can be estimated under the second condition will be described.

The torque calculation unit 32 of the assist ECU 50 repeatedly executes a steering torque detection routine (FIG. 6) to be described later at a constant short cycle. In this steering torque detection routine, the first rotation angle θ ₁ and the second rotation angle θ ₂ are repeatedly calculated in the calculation cycle. On the other hand, since there is a limit to the speed at which the steering handle 11 is rotated, the amount of change in the rotation angle at which the first rotation angle θ ₁ and the second rotation angle θ ₂ change during one calculation cycle, that is, the rotation angle change. The amount is within a predetermined amount of change. Hereinafter, the assumed maximum value of the rotation angle change amount, which is the predetermined change amount, is referred to as a maximum change amount. When the first rotation angle θ ₁ and the second rotation angle θ ₂ are not distinguished, the rotation angles are referred to as rotation angles θ, and the angle obtained by converting the rotation angle θ into an electrical angle is referred to as an electrical angle θe. Call. Further, when the first resolver 110 and the second resolver 120 are not distinguished, they are simply referred to as a resolver.

When the value representing the maximum change amount in electrical angle is Δθemax, the maximum change amount Δθemax can be expressed by the following equation (36).
Δθemax = N · (360 / 2π) · ωmax · Tc (36)
Here, N shaft angle multiplier of the resolver, .omega.max the rotational speed of the steering wheel 11 (hereinafter, referred to as the steering speed) the maximum value envisaged for, Tc is torque calculating section 32 first rotation angle theta ₁ and the second rotary it is a calculation cycle to calculate the angle theta _2. For example, when N = 8, Tc = 200 [μs], and ωmax = 10 [rad / s], the maximum change amount Δθemax is about 0.92 [deg].

Thus, the second condition is that the rotation angle change amount Δθe at which the electrical angle θe changes during one calculation cycle is less than the maximum change amount Δθemax. With this second condition, in a situation where abnormal side electrical angle .theta.e calculated to one calculation period before the resolver _(n-1) is in the range shown below, the electrical angle .theta.e _n in the next calculation cycle ( this time, it is possible to uniquely determine the electrical angle .theta.e _n) to be calculated from the amplitude of one side phases.

Here, an example when abnormality occurs in the sin phase will be described. When the maximum change amount Δθemax is set to 0.92 [deg], as shown in the following equation (37), the electrical angle θe _(n−1) calculated this time before one calculation cycle of the abnormal-side resolver is calculated this time. the difference between the electrical angle .theta.e _n is less than 0.92 [deg].
| Θe _n −θe _(n−1) | <0.92 ° (37)

The only amplitudes of cos phase electrical angle .theta.e _n uniquely determined, it electrical angle .theta.e _n is in the range of up to 0 ° to 180 °, or in advance in the range of up to 180 ° to 360 ° This is the case. Therefore, from the second condition (Expression (37)), if the electrical angle θe _(n−1) calculated one operation period before the abnormal side resolver is within the range shown in the following Expressions (38) and (39), electrical angle .theta.e _n from only the amplitude of the cos phase uniquely determined.
0 ° + 0.92 ° <θe _(n−1) <180 ° −0.92 ° (38)
180 + 0.92 ° <θe _(n−1) <360−0.92 ° (39)

If the electrical angle of the abnormal side resolver can pass 0 ° or 180 ° during one calculation cycle, the rotation angle cannot be uniquely determined from the amplitude of the cos phase. This is because the electrical angle of the abnormal-side resolver cannot be determined on either side of 0 ° or 180 ° only by the amplitude of the cos phase. On the other hand, if the situation is not likely to electrical angle of abnormal side resolver passes 0 ° or 180 ° in between one operation cycle, the electrical angle .theta.e _n from only the amplitude of the cos phase uniquely determined. Thus, the above equations (38) and (39) are derived.

By rearranging the equations (38) and (39), the following equations (40) and (41) are obtained. The equation (40), when satisfied (41), the electrical angle .theta.e _n is determined uniquely only from the amplitude of the cos phase.
0.92 ° <θe _(n−1) <179.08 ° (40)
180.92 ° <θe _(n−1) <359.08 ° (41)
Therefore, when an abnormality occurs in the sin phase of the resolver, the electrical angle θe _(n−1) calculated before one calculation cycle of the abnormal side resolver falls within the range represented by the equations (40) and (41). This is a situation that can be estimated under the second condition.

  Similarly, even if an abnormality occurs in the cos phase of the resolver and only the amplitude of the sin phase is known, in the following situation, the rotation angle of the abnormal side resolver is calculated only from the amplitude of the sin phase. can do.

The only amplitude sin phase electrical angle .theta.e _n uniquely determined, it electrical angle .theta.e _n is in the range of up to -90 ° to 90 °, or in the range of up to 90 ° to 270 ° This is a case where it is known in advance. Therefore, from the second condition (Expression (37)), if the electrical angle θe _(n−1) calculated one operation period before the abnormal side resolver is within the range shown in the following Expressions (42) and (43), electrical angle .theta.e _n from only the amplitude of sin phase uniquely determined.
−90 ° + 0.92 ° <θe _(n−1) <90 ° −0.92 ° (42)
90 + 0.92 ° <θe _(n−1) <270−0.92 ° (43)

If the electrical angle of the abnormal side resolver can pass 90 ° or 270 ° during one calculation cycle, the rotation angle cannot be uniquely determined from the amplitude of the sin phase. This is because the electrical angle of the abnormal-side resolver cannot be determined on either side of 90 ° or 270 ° by the amplitude of the sin phase alone. On the other hand, if the situation is not likely to electrical angle of abnormal side resolver passes 90 ° or 270 ° in between one operation cycle, the electrical angle .theta.e _n from only the amplitude of sin phase uniquely determined. Thus, the above equations (42) and (43) are derived.

By arranging the equations (42) and (43), the following equations (44) and (45) are obtained. The equation (44), when satisfied (45), the electrical angle .theta.e _n is determined uniquely only from the amplitude of the sin phase.
−89.08 ° <θe _(n−1) <89.08 ° (44)
90.92 ° <θe _(n−1) <269.08 ° (45)
In addition, when the electrical angle is represented by a value between 0 ° and 360 °, it can be expressed by the following equations (46), (47), (48).
0 ° <θe _(n−1) <89.08 ° (46)
90.92 ° <θe _(n−1) <269.08 ° (47)
270.92 ° <θe _(n−1) <360 ° (48)
Therefore, when an abnormality occurs in the cos phase of the resolver, the electrical angle θe _(n−1) calculated before one calculation cycle of the abnormal side resolver falls within the range represented by the equations (46) to (48). The situation of being able to be estimated under the second condition is also possible.

  Hereinafter, the ranges represented by the equations (40) and (41) and the ranges represented by the equations (46) to (48) are referred to as ranges that can be estimated under the second condition.

FIG. 5A shows an estimable range (range of θe _(n−1)) when abnormality occurs in the sin phase, and FIG. 5B shows that abnormality occurs in the cos phase. Represents the estimable range (range of θe _(n-1)) under the second condition. In the figure, a range indicated by an arrow is a range that can be estimated under the second condition, and a filled range is a range in which the rotation angle is not uniquely determined. As can be seen from the figure, the estimable range under the second condition set from the second condition is very wide.

  Next, the steering torque detection process executed by the torque calculator 32 will be described. FIG. 6 is a flowchart showing a steering torque detection routine. The steering torque detection routine is stored in the storage unit 33 as a control program. The steering torque detection routine is repeatedly executed at a predetermined short period during a period in which the ignition key is in the on state. The torque calculator 32 operates the coil drive circuit 52 together with the start of the steering torque detection routine, and the first excitation signal and the second excitation signal from the first excitation signal output port 50pe1 and the second excitation signal output port 50pe2. Starts output.

  In step S11, the torque calculation unit 32 determines the first sin phase detection voltage Es1, the first cos phase detection voltage, which are voltage values of the first sin phase detection signal, the first cos phase detection signal, the second sin phase detection signal, and the second cos phase detection signal. Ec1, the second sin phase detection voltage Es2, and the second cos phase detection voltage Ec2 are read. The torque calculator 32 is a sampling routine different from the steering torque detection routine, and samples the instantaneous values of the detected voltages Es1, Es1, Es2, and Ec2 at, for example, four sampling periods per period of the excitation signal. In step S11, the detection voltages Es1, Es1, Es2, and Ec2 sampled by the sampling routine are read. Subsequently, in step S12, the first sin phase detection signal, the first cos phase detection signal, the second sin phase detection signal, and the second cos phase are detected using the least square method based on the read detection voltages Es1, Ec1, Es2, and Ec2. The detection signal amplitudes As1, Ac1, As2, and Ac2 are calculated.

  Subsequently, in step S13, the torque calculation unit 32 checks the coil abnormality based on the detected voltages Es1, Ec1, Es2, and Ec2. The coil abnormality is caused not only by the coil abnormality in the resolver but also by disconnection of the first sin phase detection line 212, the first cos phase detection line 213, the second sin phase detection line 222, the second cos phase detection line 223, and the like. This is mainly due to disconnection of the wire harness and disconnection of the connector. When such a coil abnormality occurs, the detection voltages Es1, Ec1, Es2, and Ec2 deviate from the normal range. For example, the detection voltages Es1, Ec1, Es2, and Ec2 may be offset, or the AC voltage signal may not be detected. Step S13 identifies the presence or absence of a coil abnormality, the resolver in which the coil abnormality has occurred, and its abnormal phase, based on whether or not these detection voltages Es1, Ec1, Es2, and Ec2 are out of the normal range. To do.

Subsequently, in step S14, the torque calculation unit 32 determines whether or not a coil abnormality has occurred based on the coil abnormality check result. If no coil abnormality has occurred (S14: No), in step S15, the second rotation angle theta ₂ of the above formula of the first rotation angle theta ₁ and the second resolver 120 of the first resolver 110 (10), is calculated using (11).

  Subsequently, the torque calculation unit 32 calculates the steering torque Tr using the above formula (12) in step S16, and then outputs the calculated steering torque Tr to the assist calculation unit 31 in step S17. The assist calculation unit 31 calculates a target assist torque using the steering torque Tr, and outputs a PWM control signal to the motor drive circuit 40 so that a target current corresponding to the target assist torque flows to the electric motor 21. Thereby, an appropriate steering assist torque is generated from the electric motor 21.

  When the torque calculation unit 32 performs the process of step S17, the steering torque detection routine is temporarily ended. Then, the steering torque detection routine is repeated at a predetermined short cycle. If such a process is repeated and a coil abnormality is detected in step S13, the determination in step S14 is “Yes”. In this case, the torque calculation unit 32 advances the process to step S18 and turns on the warning lamp 65 of the vehicle. Thereby, it can be made to recognize that abnormality has arisen with respect to a driver. The means for notifying the driver of the abnormality is not limited to the warning lamp 65, and an abnormality message may be notified using a speaker, a display screen, or the like.

  Subsequently, in Step S19, the torque calculation unit 32 determines whether or not there is one coil abnormality (one phase). If there are a plurality of coil abnormalities, the steering torque Tr cannot be detected, and therefore a torque detection impossible signal is output to the assist calculation unit 31 in step S20. Thereby, the assist calculation unit 31 stops the steering assist control.

On the other hand, if there is only one coil abnormality (S19: Yes), it is determined in step S21 whether or not it is the first resolver 110 in which the coil abnormality has occurred. If it is the first resolver 110, at step S22, the second rotation angle theta ₂ of the second resolver 120 is calculated using the above equation (11).

Subsequently, the torque calculator 32 calculates the first rotation angle θ ₁ of the first resolver 110 in step S30. This process will be described later using the flowchart of FIG. When the first rotation angle θ _{1 of} the first resolver 110 is calculated in step S30, the steering torque Tr is calculated based on the first rotation angle θ ₁ and the second rotation angle θ ₂ calculated in step S22. (S16), the steering torque Tr is output to the assist calculator 31 (S17).

If it is determined in step S21 that the coil abnormality has not occurred in the first resolver 110, that is, if the second resolver 120 has detected a coil abnormality, in step S23, the first resolver 110 A first rotation angle θ ₁ of 110 is calculated using the above equation (10).

Subsequently, the torque calculating section 32, at step S40, calculates a second rotation angle theta ₂ of the second resolver 110. This process will be described later using the flowchart of FIG. When the second rotation angle θ _{2 of} the second resolver 120 is calculated in step S40, the steering torque Tr is calculated based on the second rotation angle θ ₂ and the first rotation angle θ ₁ calculated in step S23. (S16), the steering torque Tr is output to the assist calculator 31 (S17).

The torque calculator 32 calculates the electrical angles θe ₁ , θe ₁ , θ2 corresponding to the rotation angles θ ₁ , θ ₂ every time the first rotation angle θ ₁ and the second rotation angle θ ₂ are calculated in steps S15, S30, S40. θe ₂ is stored and updated in the storage unit 33 as electrical angles θe _{1 (n−1)} and θe _{2 (n−1)} one calculation cycle before. The electrical angles θe _{1 (n−1)} and θe _{2 (n−1)} stored in the storage unit 33 are used for determination of an estimable state under a second condition described later.

Next, the calculation process of the second rotation angle θ ₂ of the second resolver 120 in step S40 will be described. FIG. 7 is a flowchart showing a second rotation angle calculation routine (subroutine) incorporated as step S40 in the steering torque detection routine.

  When the second rotation angle calculation routine is started, the torque calculator 32 first determines in step S41 whether or not the coil abnormality of the second resolver 120 has occurred in the sin phase. If the coil abnormality has occurred in the sin phase, the sin phase fail signal Fails is output to the assist calculation unit 31 in step S42. On the other hand, if the coil abnormality has occurred in the cos phase, the cos phase fail signal Failc is output to the assist calculation unit 31 in step S47.

First, processing when a coil abnormality occurs in the sin phase will be described. Torque computing unit 32, in step S42, and outputs the sin phase fail signal Fails, subsequently, in step S43, the electrical angle θe corresponding to the first rotation angle theta ₁ of the first resolver 110 calculated in the previous step S23 It is determined whether ₁ is within the estimable range under the first condition. That is, it is determined whether or not the first condition can be estimated. The presumable range under the first condition is a range indicated by an arrow in FIG.

In this step S43, it is determined whether or not the electrical angle θe ₁ is within the range represented by the following equations (49) and (50).
θe _s1 <θe ₁ <θe _s2 (49)
θe _s3 <θe ₁ <θe _s4 (50)
As described above, the estimable range under the first condition is set based on the first condition that the first rotation angle θ ₁ and the second rotation angle θ ₂ are less than a predetermined angle difference. Using this example, θe _s1 = 48 °, θe _s2 = 132 °, θe _s3 = 228 °, and θe _s4 = 312 °. The predetermined angle difference is the angle represented by the illustrated electrical angle _{(θe s1 -0 °) [deg} ].

When the torque calculation unit 32 determines that the electrical angle θe ₁ is within the estimable range under the first condition (S43: Yes), the second rotation angle is calculated using the above equation (18) in step S44. Calculate θ ₂ . In this case, the second rotation angle theta ₂ is the solution of two types are present, selecting the first condition is satisfied solution described above. That is, the second rotation angle θ ₂ is selected such that the difference from the first rotation angle θ ₁ is less than the predetermined angle difference (6 [deg] in the above example).

When the torque calculator 32 calculates the second rotation angle θ ₂ of the second resolver 120 in this way, the process proceeds to step S16 of the steering torque detection routine that is the main routine.

Torque calculating unit 32, at step S43, when it is judged to be "No", i.e., when the electrical angle .theta.e ₁ is judged not in the first condition estimation range advances the process to step S45 . In this step S45, whether or not the electrical angle θe _{2 (n−1)} corresponding to the second rotation angle θ _{2 (n−1)} calculated before one calculation cycle is within the estimable range under the second condition. Judging. That is, it is determined whether the second condition can be estimated. The estimable range under the second condition is a range indicated by an arrow in FIG.

In step S45, it is determined whether or not the electrical angle θe _{2 (n−1)} is within a range represented by the following equations (51) and (52).
θe _sa <θe _{2 (n−1)} <θe _sb (51)
θe _sc <θe _{2 (n−1)} <θe _sd (52)
As described above, the estimable range under the second condition is set based on the second condition that the rotation angle change amount Δθe that can change the electrical angle θe during one calculation cycle is less than the maximum change amount Δθemax. , for example, Using the above _{example, θe sa = 0.92 °, θe} sb = 179.08 °, θe sc = 180.92 °, which is θe _sd = 359.08 °. This maximum change amount Δθemax is an angle represented by (θe _sa −0 °) [deg] in terms of electrical angle.

When the torque calculation unit 32 determines that the electrical angle θe _{2 (n−1)} is within the estimable range under the second condition (S45: Yes), the formula (18) is used in step S44. To calculate the second rotation angle θ ₂ . In this case, although there are two solutions at the second rotation angle θ ₂ , a solution that satisfies the second condition is selected. That is, the electrical angle .theta.e 2 _(n-1) second rotation angle theta ₂ which the difference is less than the maximum change amount Δθemax with corresponding to one calculation cycle before the second rotation angle theta ₂ is calculated in _(n-1) select.

When the torque calculator 32 calculates the second rotation angle θ ₂ of the second resolver 120 in this way, the process proceeds to step S16 of the steering torque detection routine that is the main routine.

  On the other hand, if it is determined as “No” in step S45, that is, if it is determined that the situation is not estimable under the second condition, the torque calculator 32 advances the process to step S46. In step S46, the steering torque Tr (n-1) calculated in the steering torque detection routine one calculation cycle before is set as the current steering torque Tr, and the steering torque Tr is output to the assist calculation unit 31 ( S17).

If it is determined in step S41 that the coil resolver of the second resolver 120 is abnormal in the cos phase in step S41, the cos phase fail signal Failc is output in step S47, and then in step S48. in, it is determined whether the electrical angle .theta.e ₁ corresponding to the first rotation angle theta ₁ of the first resolver 110 calculated in the previous step S23 is in the first condition estimation range. That is, it is determined whether or not the first condition can be estimated. The range that can be estimated under the first condition is a range indicated by an arrow in FIG.

In step S48, it is determined whether or not the electrical angle θe ₁ is within the range represented by the following equations (53), (54), and (55).
0 ° <θe ₁ <θe _c1 (53)
θe _c2 <θe ₁ <θe _c3 (54)
θe _c4 <θe ₁ <360 ° (55)
As described above, the estimable range under the first condition is set based on the first condition that the first rotation angle θ ₁ and the second rotation angle θ ₂ are less than a predetermined angle difference. Using this example, θe _c1 = 42 °, θe _c2 = 138 °, θe _s4 = 222 °, and θe _c4 = 318 °. This predetermined angle difference is, for example, an angle represented by (90 ° −θe _c1 ) [deg] in terms of electrical angle.

When the torque calculation unit 32 determines that the electrical angle θe ₁ is within the estimable range under the first condition (S48: Yes), the second rotation angle is calculated using the above equation (19) in step S49. Calculate θ ₂ . In this case, the second rotation angle theta ₂ is the solution of two types are present, selecting the first condition is satisfied solution described above. That is, the second rotation angle θ ₂ is selected such that the difference from the first rotation angle θ ₁ is less than the predetermined angle difference (6 [deg] in the above example).

When the torque calculator 32 calculates the second rotation angle θ ₂ of the second resolver 120 in this way, the process proceeds to step S16 of the steering torque detection routine that is the main routine.

Torque calculating unit 32, at step S48, the case of a "No" determination, i.e., when the electrical angle .theta.e ₁ is judged not in the first condition estimation range advances the process to step S50 . In this step S50, whether or not the electrical angle θe _{2 (n−1)} corresponding to the second rotation angle θ _{2 (n−1)} calculated before one calculation cycle is within the estimable range under the second condition. Judging. That is, it is determined whether the second condition can be estimated. The estimable range under the second condition is a range indicated by an arrow in FIG.

In step S50, it is determined whether or not the electrical angle θe _{2 (n−1)} is within the range represented by the following equations (56), (57), and (58).
0 ° <θe _{2 (n−1)} <θe _ca (56)
θe _cb <θe _{2 (n−1)} <θe _cc (57)
θe _cd <θe _{2 (n−1)} <360 ° (58)
As described above, the estimable range under the second condition is set based on the second condition that the rotation angle change amount Δθe that can change the electrical angle θe during one calculation cycle is less than the maximum change amount Δθemax. For example, using the example described above, θe _ca = 89.08 °, θe _cb = 90.92 °, θe _cc = 269.08 °, and θe _cd = 270.92 °. The maximum change amount Δθemax is an angle represented by, for example, (90 ° −θe _ca ) [deg] in terms of electrical angle.

When the torque calculation unit 32 determines that the electrical angle θe _{2 (n−1)} is within the estimable range under the second condition (S50: Yes), the equation (19) is used in step S49. To calculate the second rotation angle θ ₂ . In this case, there are two solutions for the second rotation angle θ ₂ , but a solution that satisfies the second condition is selected. That is, the electrical angle .theta.e 2 _(n-1) second rotation angle theta ₂ which the difference is less than the maximum change amount Δθemax with corresponding to one calculation cycle before the second rotation angle theta ₂ is calculated in _(n-1) select.

When the torque calculator 32 calculates the second rotation angle θ ₂ of the second resolver 120 in this way, the process proceeds to step S16 of the steering torque detection routine that is the main routine.

  On the other hand, when it is determined as “No” in step S50, that is, when it is determined that the situation is not estimable under the second condition, the torque calculator 32 advances the process to step S46 described above.

Next, the calculation process of the first rotation angle θ ₁ of the first resolver 110 in step S30 will be described. FIG. 8 is a flowchart showing a first rotation angle calculation routine (subroutine) incorporated as step S30 in the steering torque detection routine. The first rotation angle calculation routine may be considered in the same manner as the second rotation angle calculation routine described above, but the processing corresponding to steps S42 and S47 described above is not performed.

When the first rotation angle calculation routine is activated, the torque calculation unit 32 first determines in step S31 whether the coil abnormality of the first resolver 110 has occurred in the sin phase. When the coil abnormality has occurred in the sin phase, in step S32, the preceding step S22 the second resolver 120 second rotation angle θ electrical angle .theta.e ₂ corresponding to _two first condition estimation possible calculated in Determine if it is in range. That is, it is determined whether or not the first condition can be estimated. The presumable range under the first condition is a range indicated by an arrow in FIG.

In step S32, it is determined whether or not the electrical angle θe ₂ is within the range represented by the following equations (59) and (60).
θe _s1 <θe ₂ <θe _s2 (59)
θe _s3 <θe ₂ <θe _s4 (60)

When the torque calculation unit 32 determines that the electrical angle θe ₂ is within the estimable range under the first condition (S32: Yes), in step S33, the first rotation angle is calculated using the above equation (16). Calculate θ ₁ . In this case, the first rotation angle theta ₁ is a solution of 2 types exist, selects the first condition is satisfied solution described above.

When the torque calculation unit 32 determines that the electrical angle θe ₂ is not within the estimable range under the first condition (S32: No), the first rotation angle θ calculated one calculation cycle before in step S34. ₁ electric angle .theta.e 1 _(n-1) corresponding to the _(n-1) to determine whether the entered second condition estimation range. That is, it is determined whether the second condition can be estimated. The estimable range under the second condition is a range indicated by an arrow in FIG.

In step S34, it is determined whether or not the electrical angle θe _{1 (n−1)} is within the range represented by the following equations (61) and (62).
θe _sa <θe _{1 (n−1)} <θe _sb (61)
θe _sc <θe _{1 (n−1)} <θe _sd (62)

When the torque calculation unit 32 determines that the electrical angle θe _{1 (n−1)} is within the estimable range under the second condition (S34: Yes), the formula (16) is used in step S33. To calculate the first rotation angle θ ₁ . In this case, one operation cycle first rotation angle theta ₁ is calculated before _(n-1) the electrical angle .theta.e 1 corresponding to _(n-1) first rotation angle difference is less than the maximum change amount Δθemax the theta ₁ Select.

When the torque calculator 32 calculates the first rotation angle θ ₁ of the first resolver 110 in this way, the process proceeds to step S16 of the steering torque detection routine that is the main routine.

  On the other hand, if it is determined as “No” in step S34, that is, if it is determined that the condition cannot be estimated under the second condition, the steering torque calculated in the steering torque detection routine before one calculation cycle in step S35. Tr (n-1) is set to the current steering torque Tr, and the steering torque Tr is output to the assist calculation unit 31 (S17).

In Step S31, when it is determined that “No”, that is, the coil abnormality of the first resolver 110 is occurring in the cos phase, in Step S36, the second resolver 120 calculated in the previous Step S22 is determined. It is determined whether or not the electrical angle θe ₂ corresponding to the second rotation angle θ ₂ is within an estimable range under the first condition. That is, it is determined whether or not the first condition can be estimated. The range that can be estimated under the first condition is a range indicated by an arrow in FIG.

In step S36, it is determined whether or not the electrical angle θe ₁ is within the range represented by the following equations (63), (64), and (65).
0 ° <θe ₂ <θe _c1 (63)
θe _c2 <θe ₂ <θe _c3 (64)
θe _c4 <θe ₂ <360 ° (65)

When the torque calculation unit 32 determines that the electrical angle θe ₂ is within the estimable range under the first condition (S36: Yes), in step S37, the first rotation angle is calculated using the above equation (17). Calculate θ ₁ . In this case, the difference between the first rotation angle theta ₁ selects the first rotation angle theta ₁ which is less than a predetermined angle difference.

When the torque calculator 32 calculates the first rotation angle θ ₁ of the first resolver 110 in this way, the process proceeds to step S16 of the steering torque detection routine that is the main routine.

When it is determined “No” in step S36, that is, when it is determined that the electrical angle θe ₂ is not within the estimable range under the first condition, the torque calculation unit 32 determines one calculation cycle in step S38. It is determined whether or not the electrical angle θe _{1 (n−1)} corresponding to the previously calculated first rotation angle θ _{1 (n−1)} is within an estimable range under the second condition. That is, it is determined whether the second condition can be estimated. The estimable range under the second condition is a range indicated by an arrow in FIG.

In step S38, it is determined whether or not the electrical angle θe _{1 (n−1)} is within the range represented by the following equations (66), (67), (68).
0 ° <θe _{1 (n−1)} <θe _ca (66)
θe _cb <θe _{1 (n−1)} <θe _cc (67)
θe _cd <θe _{1 (n−1)} <360 ° (68)

When the torque calculation unit 32 determines that the electrical angle θe _{1 (n−1)} is within the estimable range under the second condition (S38: Yes), the equation (17) is used in step S37. To calculate the first rotation angle θ ₁ . In this case, there are two solutions for the first rotation angle θ ₁ , but the electrical angle θe _{1 (n−1)} corresponding to the second rotation angle θ _{1 (n−1)} calculated before one calculation cycle. difference between selects the first rotation angle theta ₁ which is less than the maximum change amount Derutashitaemax.

When the torque calculator 32 calculates the first rotation angle θ ₁ of the first resolver 110 in this way, the process proceeds to step S16 of the steering torque detection routine that is the main routine.

  On the other hand, if it is determined as “No” in step S38, that is, if it is determined that the situation is not estimable under the second condition, the torque calculator 32 advances the process to step S35 described above.

  In this way, in the torque detection routine, not only the situation that can be estimated under the first condition derived from the first condition (twist angle upper limit) but also the second condition estimated from the second condition (rotation angle change amount upper limit). Even in a possible situation, the rotation angle is calculated based on the amplitude of the detection signal of the normal side phase. Therefore, even when a coil abnormality occurs in one side phase, the range in which the rotation angle can be uniquely calculated from the amplitude of the detection signal in the normal side phase becomes very wide, and the period during which the steering torque can be detected becomes long.

By the way, although there is a slight range, there is a range in which the rotation angle cannot be calculated. That is, if the sine phase is abnormal, the rotation angle θ ₁ falls within the painted range of FIG. 10A, and if the cosine phase is abnormal, the rotation angle θ ₁ falls within the painted range of FIG. , Θ ₂ cannot be calculated. This filled area corresponds to the rotation angle estimation impossible area of the present invention.

Therefore, in the present embodiment, the electric motor 21 is driven so that the electrical angle θe ₂ of the second resolver 120 does not remain in the range that cannot be estimated under the second condition (the filled range in FIG. 10: the rotation angle cannot be estimated range). This reduces the situation where the rotation angle cannot be detected when an abnormality occurs in the second resolver 120. For this purpose, the assist calculation unit 31 includes a raising current setting unit 310, and in the situation where the sin phase fail signal Fails or the cos phase fail signal Failc is output from the torque calculation unit 32, the electrical angle θe of the second resolver 120. _{When 2} enters the incapable estimation range under the second condition, the q-axis command current Iq * of the electric motor 21 is increased.

  Focusing on the function of the raising current setting unit 310, as shown in FIG. 2, the raising basic current setting unit 311, the motor rotational speed calculation unit 312, the gain setting unit 313, and the raising target current calculation unit 314 and a q-axis command current correction unit 315. These functional units are realized by program control of the microcomputer of the assist ECU 50.

  The raising basic current setting unit 311 sets the raising basic current Iqupo with reference to the raising current map when the sin phase is abnormal or the raising current map when the cos phase is abnormal. The sin phase abnormality raising current map and the cos phase abnormality raising current map are stored in the storage unit 33 in advance. As shown in FIG. 12A, the sin phase abnormality raising current map is associated with the electrical angle θme of the electric motor 21 in the case where the sine phase abnormality of the second resolver 120 has occurred. This is a basic current raising Iqupo for increasing * by a predetermined amount.

In the sin phase abnormality or to raise the current map, the second condition estimation impossible range electrical angle .theta.e ₂ of the second resolver 120 is shown with filled in FIG. _{10 (a) (θe sd ~θe} sa, θe sb ~θe sc ) In the range (θme _{sd to} θme _sa , θme _{sb to} θme _sc ) of the electric angle θme of the electric motor 21 corresponding to)), the raised basic current Iqupo is set to a preset Iqo. The second condition estimation range of the electric angle .theta.e ₂ of the second resolver 120 is indicated by arrows in FIG. _{10 (a) (θe sa ~θe} sb, θe sc ~θe sd) to electric motor 21 corresponding In the range of the angle θme (0 to θme _sd , θme _{sa to} θme _sb , θme _sc to...), The raising basic current Iqupo is set to zero (Iqupo = 0).

Similarly, as shown in FIG. 12B, the cos phase abnormality raising current map is associated with the electrical angle θme of the electric motor 21 when the second resolver 120 has a cos phase abnormality. The raising basic current Iqupo for increasing the command current Iq * by a predetermined amount is set. In this cos phase abnormality raising current map, the electrical angle θe ₂ of the second resolver 120 is in an unestimable range (θe _{ca to} θe _cb , θe _{cc to} θe _cd ) shown in FIG. ) In the range (θme _{ca to} θme _cb , θme _{cc to} θme _cd ) of the electric angle θme of the electric motor 21 corresponding to ()), the raised basic current Iqupo is set to a preset Iqo. The electric angle θe ₂ of the second resolver 120 corresponds to the electric range of the electric motor 21 corresponding to the second condition estimable range (θe _{cd to} θe _ca , θe _{cb to} θe _cc ) indicated by an arrow in FIG. range of angular _{θme (0~θme ca, θme cb ~θme} cc, θme cd ~ ···) in, raising the basic current Iqupo is set to zero (Iqupo = 0).

  The range of the electric angle θme of the electric motor 21 in which the raising basic current Iqupo is set to Iqo corresponds to the unestimable motor electric angle region in the present invention. Therefore, the storage unit 33 that stores the sin phase abnormality raising current map and the cos phase abnormality raising current map corresponds to the non-estimable motor electrical angle region storage means of the present invention.

  Since the second resolver 120 and the electric motor 21 are mechanically coupled via the output shaft 12out, the pinion gear 13, the rack bar 14, and the ball screw mechanism 22 so as not to be twisted, the electrical angle θe of the second resolver 120 is There is a one-to-one correspondence with the electrical angle θme of the electric motor 21. That is, the electrical angle θme of the electric motor 21 can be uniquely derived from the electrical angle θe of the second resolver 120. Therefore, it is possible to store and set the range that cannot be estimated under the second condition in the second resolver 120 in association with the range of the electrical angle θme of the electric motor 21 (the motor angle range that cannot be estimated).

  In this example, the size of the non-estimable range under one second condition is 1.84 deg (0.92 deg × 2). Therefore, for example, if the motor speed reduction ratio is 18.5 and the shaft angle multiplier of the rotation angle sensor 61 (resolver) of the electric motor 21 is 7 (= the number of pole pairs of the electric motor 21), the estimation range of one second condition is not possible. When the magnitude is expressed by the motor electrical angle θme, it is about 30 deg (1.84 ÷ 8 × 18.5 × 7). Since the rotation range of the second resolver 120 exceeds the range of 360 ° in the motor electrical angle θme, the sinusoidal abnormality raising current map and the cos phase abnormality raising current map are the reference positions of the motor electrical angle θme. It is expressed using the integrated value from. In setting the raising basic current Iqupo, an integrated value of the motor electrical angle θme from the reference position is calculated, and the raising basic current Iqupo corresponding to the integrated value is used.

  The raising basic current setting unit 311 receives the fail signal output from the torque calculation unit 32 and the motor electrical angle θme output from the electrical angle conversion unit 307. When the fail signal is the sin phase fail signal Fails, the raising current map is selected when the sin phase is abnormal. When the fail signal is the cos phase fail signal Failc, the raising current map is selected when the cos phase is abnormal. Then, the raising basic current Iqup is set based on the selected raising current map and the motor electrical angle θme.

  The motor rotation speed calculation unit 312 receives the motor electrical angle θme output from the electrical angle conversion unit 307, and calculates the motor rotation speed ωm by differentiating the motor electrical angle θme with time. The motor rotation speed calculation unit 312 outputs the calculated motor rotation speed ωm to the gain setting unit 313.

  The gain setting unit 313 obtains the gain K from the magnitude | ωm | of the motor rotation speed ωm output from the motor rotation speed calculation unit 312 with reference to the gain setting map shown in FIG. The gain setting map is stored in the storage unit 33. In this gain setting map, in a range where the motor rotation speed | ωm | is smaller than ωmo, a constant gain K is set (for example, K = 1), and the motor rotation speed | ωm | becomes larger than ωmo. In the range, a gain K that decreases as the motor rotation speed | ωm | increases is set. The gain setting map may set a gain K that decreases stepwise as the motor rotation speed | ωm | increases.

  The raising target current calculation unit 314 inputs the raising basic current Iqupo set by the raising basic current setting unit 311 and the gain K set by the gain setting unit 313, and gain K is added to the raising basic current Iqupo. Multiply Then, the multiplication result is output to the q-axis command current correction unit 315 as the raised target current Iqup * (= K · Iqupo).

  The q-axis command current correction unit 315 inputs the q-axis command current Iq * output from the assist current command unit 302 and the raising target current Iqup * output from the raising target current calculation unit 314, and receives the q-axis The target current Iqup * is added to the command current Iq *. Then, the addition result (Iq * + Iqup *) is output to the feedback control unit 303 as the final q-axis command current Iq *. That is, the q-axis command current correction unit 315 increases the q-axis command current Iq * and increases it by the target current Iqup *. The direction of the raising target current Iqup * is set in the same direction as the q-axis command current Iq *, that is, the direction in which the torque in the same direction as the target assist torque T * is generated.

  The overall processing of the raising current setting unit 310 will be described using a flowchart. FIG. 11 shows a raising current calculation routine executed by the raising current setting unit 310. This raising current calculation routine is repeatedly executed at a predetermined short period in a period in which the ignition key is in the ON state.

  When the raising current calculation routine is activated, the raising current setting unit 310 first determines whether or not the fail signal output from the torque calculation unit 32 is input in step S101. This fail signal is a signal output in step S42 or step S47 of the second rotation angle calculation routine described above. If no fail signal is input, in step S102, the q-axis command current Iq * output from the assist current command unit 302 is output without correction to the feedback control unit 303, and this calculation routine is temporarily executed. finish.

  The raising current setting unit 310 repeats such processing, and when a fail signal is input in step S101, it determines in step S103 whether or not the fail signal is a sin-phase fail signal Fails. If the fail signal is the sin phase fail signal Fails (S103: Yes), in step S104, the sin phase abnormality raising current map is selected, and if the fail signal is the cos phase fail signal Failc (S103). : No), in step S105, the cos phase abnormality time raising current map is selected.

Subsequently, in step S106, the motor electrical angle θme output from the electrical angle conversion unit 307 is read. Next, in step S107, the raising basic current Iqupo corresponding to the motor electrical angle θme is referred to with reference to the raising current map. Set. In this case, raising the basic current Iqupo, if sin phase abnormality, the motor electric angle θme corresponding to the second condition estimation impossible range fills in FIG. 10 of the electrical angle .theta.e ₂ of the second resolver 120 (a) It is set to the set value Iqo (> 0) in the range (non-estimable motor electrical angle region), and is set to zero in other ranges. Also, if cos phase abnormality, the second resolver 120 range (estimated impossible motor electric angle range of the motor electric angle θme corresponding to the second condition estimation impossible range fills in FIG. 10 of the electric angle .theta.e ₂ (b) of ) Is set to the set value Iqo, and is set to zero in other ranges.

  Subsequently, in step S108, the raising current setting unit 310 calculates the motor rotation speed ωm by differentiating the motor electrical angle θme read in the previous step S106 with respect to time. Subsequently, in step S109, the gain K corresponding to the motor rotation speed | ωm | is set with reference to the gain setting map. Subsequently, in step S110, the raising basic current Iqupo is multiplied by the gain K to calculate the raising target current Iqup * (= K · Iqupo).

  Subsequently, in step S111, the raising current setting unit 310 adds a value obtained by adding the raising target current Iqup * to the q-axis command current Iq * output from the assist current command unit 302 as a final q-axis command current Iq. * Is output to the feedback control unit 303 to end the present calculation routine once.

According to the electric power steering apparatus of the present embodiment described above, the motor electrical angle θme cannot calculate the rotation angle of the second resolver 120 in the situation where the one-side phase abnormality of the second resolver 120 has occurred. When entering the corner region, the q-axis command current Iq * is increased and corrected. That is, the assist torque is corrected to increase. Thus, the electric motor 21 is rotated, the motor electric angle θme is now not remain in the region, conditions are reduced in which it becomes impossible to detect the second rotation angle theta _2. Therefore, the situation in which the steering torque cannot be detected is reduced, and the period during which proper steering assist can be continued can be lengthened.

  Further, the increase target current Iqup * is used to increase and correct the q-axis command current Iq *. The increase target current Iqup * is set smaller as the motor rotational speed ωm increases. Accordingly, in a situation where the rotational inertia of the electric motor 21 is large and it is easy to pass through the range where the rotation angle cannot be estimated, the increase correction amount of the q-axis command current Iq * is reduced, so that an appropriate q-axis command current Iq * is calculated. Thus, the electric motor 21 is prevented from being rotated more than necessary.

  Further, since the angle range in which the q-axis command current Iq * is corrected to be increased is a very narrow range as represented by the filled range in FIG. 10, the influence on the steering feeling can be reduced. .

  When a coil abnormality occurs, the warning lamp 65 is lit, so that the driver can be made aware of the abnormality while continuing the steering assist. Thereby, since repair is performed at an early stage, the probability of falling into a situation where the steering assist is stopped due to a double failure can be extremely reduced.

  Although the present embodiment has been described above, the present invention is not limited to the above-described embodiment, and various modifications can be employed within the scope of the present invention.

  For example, in this embodiment, in calculating the raising target current Iqup *, the gain K corresponding to the motor rotation speed ωm is multiplied by the raising basic current Iqupo, but it is not always necessary to multiply the gain K. The raising target current Iqup * may be a constant value.

  In the present embodiment, it is determined whether the rotation angle can be uniquely calculated from the detection signal of one side phase based on the two conditions. However, the two conditions are not necessarily used. The determination may be made based on only two conditions or only on the first condition.

DESCRIPTION OF SYMBOLS 11 ... Steering handle, 12 ... Steering shaft, 12in ... Input shaft, 12out ... Output shaft, 12a ... Torsion bar, 20 ... Power assist part, 21 ... Electric motor, 31 ... Assist calculation part, 32 ... Torque calculation part, 33 ... Storage unit 40 ... motor drive circuit 50 ... assist ECU 51s1, 51c1, 51s2, 51c2 ... amplifier 52 ... coil drive circuit 61 ... rotation angle sensor 100 ... resolver unit 110 ... first resolver 111 ... first DESCRIPTION OF SYMBOLS 1 excitation coil, 112 ... 1st sin phase detection coil, 113 ... 1st cos phase detection coil, 114 ... 1st rotor coil, 120 ... 2nd resolver, 121 ... 2nd excitation coil, 122 ... 2nd sin phase detection coil, 123 ... Second cos phase detection coil, 124 ... second rotor coil, 212 ... first sin phase detection line, 213 ... first cos phase detection line, 222 ... second sin phase detection line, 223 ... second cos phase detection line, 307 ... electric angle converter, 310 ... raising current setting unit, 311 ... raising basic current setting unit, 312 ... Motor rotation speed calculation unit, 313 ... gain setting unit, 314 ... uplift target current calculation unit, 315 ... q-axis command current correction unit, As1, Ac1, As2, Ac2 ... amplitude, Es1, Ec1, Es2, Ec2 ... detection voltage , Tr: steering torque, θ ₁ : first rotation angle, θ ₂ ... second rotation angle, Failc ... cos phase fail signal, Fails ... sin phase fail signal, Iq * ... q-axis command current, Iqo ... set value, Iqupo ... Raising basic current, Iqup * ... Raising target current, K ... Gain, θme ... Motor electrical angle, ωm ... Motor rotation speed.

Claims (5)

A first resolver having a pair of detection coils for outputting a two-phase detection signal corresponding to a first rotation angle that is a rotation angle on one end side of the steering shaft, and a second angle that is a rotation angle on the other end side of the steering shaft. Sensor means comprising a second resolver having a pair of detection coils for outputting a two-phase detection signal corresponding to the rotation angle;
Rotation that calculates the first rotation angle based on a two-phase detection signal output from the first resolver and calculates the second rotation angle based on a two-phase detection signal output from the second resolver Angle calculation means;
Steering torque calculation means for calculating a steering torque acting on the steering shaft based on a difference between the calculated first rotation angle and second rotation angle;
An electric motor that generates assist torque for assisting a driver's steering operation;
Assist control means for driving and controlling the electric motor based on the steering torque calculated by the steering torque calculation means;
An abnormality detection means for detecting that an abnormality has occurred on one side of the pair of detection coils in either the first resolver or the second resolver;
When the occurrence of an abnormality on one side of the detection coil is detected, the rotation angle can be uniquely determined from the output signal of the other detection coil paired with the detection coil in which the abnormality has occurred. In addition, in the electric power steering apparatus comprising: a rotation angle estimating means for estimating the rotation angle of the abnormal resolver from the output signal of the other detection coil;
When there is an abnormality on one side of the detection coil, the rotation angle cannot be estimated uniquely from the output signal of the other detection coil paired with the detection coil in which the abnormality has occurred. A non-estimable motor electrical angle region storage means that stores the region as an unestimable motor electrical angle region corresponding to the electrical angle region of the electric motor;
Motor electrical angle detection means for detecting an electrical angle of the electric motor;
When the occurrence of an abnormality on one side of the detection coil is detected, the electrical angle of the electric motor detected by the motor electrical angle detection means and the non-estimable motor electrical angle stored in the non-estimable motor electrical angle area storage means And a non-estimable region passage control means for driving and controlling the electric motor so that the electric angle of the electric motor does not stay in the non-estimable motor electric angle region based on the region. Steering device. The non-estimable region passage control means is
When an abnormality is detected on one side of the detection coil and the electrical angle of the electric motor falls within the non-estimable motor electrical angle region, the target current of the electric motor set by the assist control means is increased and corrected. The electric power steering apparatus according to claim 1, wherein Motor rotation speed detecting means for detecting the rotation speed of the electric motor;
3. The electric power according to claim 2, wherein the non-estimable region passage control means reduces the increase correction amount of the target current of the electric motor when the rotational speed of the electric motor is large compared to when the rotational speed is small. Steering device.   In the rotation angle non-estimable region, the rotation angle change amount by which the first rotation angle or the second rotation angle changes per calculation cycle in which the rotation angle calculation means calculates the first rotation angle and the second rotation angle. The electric power steering apparatus according to any one of claims 1 to 3, wherein the electric power steering apparatus is set based on a rotation angle change condition of being within a predetermined change amount.   When the detection coil in which the abnormality is detected is a sin-phase detection coil, the electrical angle of the abnormal-side resolver passes 0 ° or 180 ° during one calculation cycle. When the detection coil in which the abnormality is detected is a cos-phase detection coil, the electrical angle of the abnormal resolver passes 90 ° or 270 ° during one calculation period. The electric power steering apparatus according to claim 4, wherein the electric power steering apparatus is set in a region where there is a possibility of occurrence.

Images