JP2019191011A - Rotation angle detection device - Google Patents

Abstract

To provide a rotation angle detection device that can improve detection accuracy of a rotation angel.SOLUTION: A rotation angle detection device (1) comprises: a calculation unit (21) that calculates a rotation angle of a rotation shaft (14) on the basis of a detection result detected from a sensor (92) as a pre-correction rotation angle; a correction value learning unit (22) that detects and learns an error correction value to be used in an error correction of a pre-correction rotation angle on the basis of the pre-correction rotation angle detected during a detection period in which the rotation shaft rotates from a reference position to a prescribed angle, and learns the error correction value; and a second correction unit (23) that, when a rotation velocity of the rotation shaft is less than a threshold, estimate the error correction value of the pre-correction rotation angle with a reference to an error map, and when the rotation velocity is equal to or more than the threshold while the pre-correction rotation angle is corrected by the estimated error correction value, corrects the pre-correction rotation angle by the error correction value learned by the correction value learning unit.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a rotation angle detection device that detects a rotation angle of a rotation shaft in a rotating electrical machine.

  2. Description of the Related Art Conventionally, a rotation angle detection device that detects a rotation angle on a rotation shaft of a motor using a magnetic sensor is known (for example, Patent Documents 1 and 2). In these rotation angle detection devices, the rotation angle detection error is calculated based on the leakage magnetic flux or the like or the attachment error of the magnetic sensor or the like, and the detection error is corrected, thereby improving the rotation angle detection accuracy.

JP 2006-133376 A JP 2006-128772 A

  By the way, when calculating the detection error of the rotation angle, for example, when the rotation axis rotates one rotation from the reference position (the position where the north marker signal is output) and the rotation angle before correction detected by the magnetic sensor. There is a method of calculating and learning a rotation angle error based on the rotation time of the above. Specifically, the rotation angle error is calculated by comparing the ideal value calculated based on the rotation time on the assumption that the speed of the rotation axis is constant and the detected rotation angle before correction, and learning. To do. And it is the method of correct | amending the rotation angle before correction | amendment detected in the rotation after learning using the learned detection error.

  By the way, when performing such learning correction, it is premised that the speed of the rotating shaft is constant. Therefore, when the rotational speed fluctuates greatly during one rotation, the calculation accuracy decreases. In particular, when the rotation speed is low, the rotation speed is likely to fluctuate greatly during one rotation compared to the case where the rotation speed is high, and the calculation accuracy of the rotation angle error is likely to decrease.

  The present invention has been made in view of the above circumstances, and a main object thereof is to provide a rotation angle detection device capable of improving the detection accuracy of the rotation angle.

  In order to solve the above-described problem, the first means includes a signal output unit that is provided on a rotating shaft of a rotating electrical machine and rotates together with the rotating shaft, and a sensor that detects a signal from the signal output unit, In a rotation angle detection device that detects a rotation angle of the rotation shaft based on a detection result of the sensor, a calculation unit that calculates the rotation angle of the rotation shaft as a rotation angle before correction based on the detection result detected from the sensor; Correction value learning for calculating and learning an error correction value used for error correction of the rotation angle before correction based on a rotation angle before correction detected in a detection period in which the rotation shaft rotates from a reference position to a predetermined angle. And the rotation speed of the rotary shaft is less than a threshold value, an error correction value of the pre-correction rotation angle is estimated with reference to an error map, and the pre-correction is calculated by the estimated error correction value A first correction unit that corrects a rotation angle; and a second correction unit that corrects the pre-correction rotation angle based on an error correction value learned by the correction value learning unit when the rotation speed is equal to or greater than a threshold value; Is provided.

  With the above configuration, when the rotation speed is less than the threshold value, the error correction value is estimated with reference to the error map, and the rotation angle before correction is corrected by the estimated error correction value. In some cases, the rotation angle before correction is corrected by the error correction value learned by the correction value learning unit. As a result, if the rotation speed is slow and fluctuations in the rotation speed during the detection period are likely to increase, that is, if an error is likely to occur when the correction value learning unit calculates the error correction value, error correction of the error map is performed. By using the value, the detection accuracy of the rotation angle can be improved.

  On the other hand, if the rotation speed is fast and fluctuations in the rotation speed during the detection period tend to be small, i.e., if the correction value learning unit does not easily generate an error when calculating the error correction value, the learned error correction value is used. By using this, the detection accuracy of the rotation angle can be improved.

Sectional drawing of ISG. The block diagram of a control system. The figure which shows the rotation angle before correction | amendment for every interruption angle, and elapsed time. The flowchart which shows a correction value learning process. The flowchart which shows a correction process. The block diagram of the control system in another example.

  Hereinafter, embodiments will be described with reference to the drawings. In the following embodiments, parts that are the same or equivalent to each other are denoted by the same reference numerals in the drawings, and the description of the same reference numerals is used. The rotating electrical machine according to the present embodiment is mounted on a vehicle. FIG. 1 shows a rotating electrical machine according to the present embodiment, and FIG. 2 shows a control system for controlling the rotating electrical machine.

  The rotating electrical machine shown in FIG. 1 and FIG. 2 is a generator with a motor function having a motor 10 that is an electric part and a control device 20 that is a control part that controls the motor 10, and is an integrated electromechanical ISG (Integrated Type). Starter Generator). Hereinafter, it is simply indicated as ISG100. The motor 10 is of a wound field type, and is specifically a wound field type synchronous machine having a three-phase winding. The ISG 100 has a power generation function for generating power (regenerative power generation) by rotation of a crankshaft 200a and an axle of an engine 200 as an internal combustion engine, and a power running function for applying a driving force (rotational force) to the crankshaft 200a.

  The motor 10 includes a housing 11, a stator 12 fixed to the housing 11, a rotor 13 that rotates with respect to the stator 12, and a rotating shaft 14 to which the rotor 13 is fixed. Hereinafter, in the present embodiment, the axial direction indicates the axial direction of the rotating shaft 14 (indicated by an arrow Y1 in the figure), and the radial direction indicates the radial direction of the rotating shaft 14 (indicated by the arrow in the figure). Y2).

  The housing 11 is formed in a cylindrical shape, and its axis is coaxial with the rotating shaft 14. A control device 20 is fixed to the outside of the housing 11 in the axial direction (right side in FIG. 1). On the other hand, a stator 12 and a rotor 13 are accommodated in the housing 11.

  The stator 12 is provided in a cylindrical shape along the inner periphery of the housing 11 at the approximate center in the axial direction of the housing 11, and is fixed to the housing 11. The stator 12 constitutes a part of a magnetic circuit, and includes a stator core 12a and an armature winding 12b.

  The stator core 12 a is formed in an annular shape from a magnetic material, and its axis is coaxial with the rotary shaft 14. The stator core 12a holds the armature winding 12b. The stator core 12a includes a plurality of slots for accommodating the armature windings 12b, and the armature windings 12b are accommodated and held in the slots.

  The armature winding 12b is composed of two sets of Y-connected three-phase windings. The armature winding 12b generates magnetic flux when electric power (AC power) is supplied. The armature winding 12b generates electric power (alternating current power) by interlinking with the magnetic flux generated by the rotor 13.

  The rotor 13 constitutes a part of a magnetic circuit, and includes a rotor core 13a made of a magnetic material and a field winding 13b held by the rotor core 13a.

  The rotor core 13a is a so-called Landel-type pole core, and includes an annular hollow portion, and the field winding 13b is accommodated in the hollow portion. The rotor core 13a is arranged so as to face the outer peripheral surface thereof in a state of being separated from the inner peripheral surface of the stator core 12a. A rotating shaft 14 is inserted through the rotor core 13 a and is fixed to the rotating shaft 14 so as to rotate integrally with the rotating shaft 14.

  The field winding 13b generates magnetic flux when DC power is supplied, and forms magnetic poles on the outer peripheral surface of the rotor core 13a. As a result, when AC power is supplied to the stator 12, the rotor 13 rotates by interlinking with the magnetic flux generated by the stator 12. Further, when the rotor 13 rotates, AC power is generated in the armature winding 12b of the stator 12.

  The rotating shaft 14 is rotatably supported by the housing 11 via bearings 11 a and 11 b provided on the housing 11. A rotor 13 is fixed to the rotary shaft 14 at the central portion in the axial direction. Further, in the axial direction, both end portions of the rotary shaft 14 protrude from the housing 11, and the end portion on the side opposite to the control device 20 (left side in FIG. 1) is connected to the crankshaft 200 a of the engine 200, the axle, or the like. Is done.

  Further, in the axial direction, the end of the rotating shaft 14 (the end on the control device 20 side) is inserted into the control device 20. A slip ring 14 a that contacts the brush 80 and supplies DC power supplied from the battery via the brush 80 to the field winding 13 b is provided at the end of the rotating shaft 14. The slip ring 14a is connected to the field winding 13b via a wiring (not shown).

  In addition, a magnet 91 as a signal output unit is disposed at the tip of the rotating shaft 14 (tip on the control device 20 side). Further, a magnetic sensor 92 is fixed on the control device 20 side as a sensor facing the magnet 91 disposed at the tip.

  The magnet 91 is formed in a disk shape and generates a magnetic flux (signal). Of the circular surface of the magnet 91, an N pole is formed in one semicircular portion and an S pole is formed in the other semicircular portion. The magnet 91 is fixed to the rotating shaft 14 with the center of the circle being coincident with the axis of the rotating shaft 14.

  The magnetic sensor 92 is an element that detects a magnetic flux density in a predetermined direction. The magnetic sensor 92 is fixed to the control device 20 and is arranged at a predetermined distance from the magnet 91. The thickness direction is the axial direction of the rotating shaft 14, and the center is provided to coincide with the axis of the rotating shaft 14. The output terminal of the magnetic sensor 92 is connected to the control device 20.

  Next, the control device 20 will be described. As described above, the control device 20 is fixed to the outside in the axial direction of the housing 11 of the motor 10. The control device 20 is mainly configured by a microcomputer including a CPU, a ROM, a RAM, an I / O, and the like, and various functions are realized by the CPU executing programs stored in the ROM. Note that the various functions may be realized by an electronic circuit that is hardware, or may be realized at least in part by software, that is, processing executed on a computer.

  For example, the control device 20 has a function of converting electric power from the outside (for example, a battery) and supplying the motor 10 to generate a driving force, and converting electric power from the motor 10 to supply electric power to the outside. It has a function to supply. Further, for example, the control device 20 has a function of controlling the motor 10 by using information on the rotation angle input from the magnetic sensor 92. The rotation angle detection device 1 of the present embodiment includes a control device 20, a magnetic sensor 92, and a magnet 91.

  Incidentally, the rotation angle calculated based on the detection result detected by the magnetic sensor 92 usually includes an angle error. As a factor of the angle error, for example, there is one based on an installation error of the magnetic sensor 92, the rotary shaft 14, the magnet 91, or the like. In addition, for example, a leakage magnetic flux generated based on a current flowing through the field winding 13b may affect the magnetic sensor 92 and cause an angle error. Further, the magnetic flux density generated by the magnet 91 may change depending on the temperature of the magnet 91, and an angle error may occur. Therefore, the control device 20 has a function of correcting the angle error. More specifically, the control device 20 includes a calculation unit 21, a correction value learning unit 22, and a correction unit 23.

  The calculation unit 21 calculates the rotation angle of the rotating shaft 14 as the rotation angle before correction based on the detection result detected from the magnetic sensor 92. That is, when the magnet 91 rotates with respect to the magnetic sensor 92 as the rotating shaft 14 rotates, the magnetic flux density detected by the magnetic sensor 92 changes. Therefore, the calculation unit 21 calculates the pre-correction rotation angle of the rotary shaft 14 based on the change in magnetic flux density detected by the magnetic sensor 92. The rotation angle before correction may include an angle error as described above.

  The correction value learning unit 22 calculates an error correction value used for error correction of the rotation angle before correction based on the rotation angle before correction detected in the detection period in which the rotation shaft 14 rotates from the reference position to a predetermined angle. To learn. The reference position is a position where a reference position signal (north marker signal) is input. The north marker signal is a signal that is output when the rotor 13 (and the rotation shaft 14) is positioned at a reference angle. The predetermined angle is 360 ° (one rotation). That is, the period from when the reference position signal is input to when the next reference position signal is input corresponds to the detection period.

Hereinafter, the learning method will be described in detail with reference to FIG. The correction value learning unit 22 acquires the elapsed time t (n) from the reference position every time the pre-correction rotation angle matches the predetermined interrupt angle A (n) during the detection period. The elapsed time t (n) from the reference position is the elapsed time from the input of the reference position signal. The elapsed time is measured by the control device 20. Then, the elapsed time t (n) is stored in the storage device of the control device 20 in association with the interrupt angle A (n). Further, the interrupt angle A (n) is obtained by dividing one rotation (360 °) by a constant N as shown in Equation (1). The constant N is a natural number equal to or greater than 2, and the variable n is a natural number equal to or less than the constant N.
Interrupt angle A (n) = 360 ° / constant N × variable n (1)

  In this embodiment, N = 16, and the interrupt angle A (n) is 22.5 ° (n = 1), 45 ° (n = 2), 67.5 ° (n =) as shown in FIG. 3),..., 337.5 ° (n = 15), 360 ° (n = 16). Further, the time elapsed from the start to the end of the detection period, that is, the time from the input of the reference position signal to the input of the next reference position signal is acquired and stored as the rotation time T. FIG. 3 shows the relationship between the interrupt angle A (n) and the elapsed time t (n). In FIG. 3, the solid line indicates the rotation angle before correction based on the detection result of the magnetic sensor 92, and the broken line indicates that the rotation time T elapses on the assumption that the rotating shaft 14 rotates at a constant speed. Sometimes an ideal angle (ideal value) when the rotating shaft 14 makes one rotation is shown.

Next, the correction value learning unit 22 assumes that the rotating shaft 14 is rotating at a constant speed using the rotation time T and the elapsed time t (n) based on Equation (2). The interrupt actual angle I (n), which is an ideal angle when the elapsed time t (n) has elapsed from the reference position, is calculated. That is, assuming that the rotation angle (ideal value) and the elapsed time t (n) are in a proportional relationship, the actual interrupt angle I (n) is calculated by proportionally distributing the rotation angle.
Interruption angle I (n) = 360 ° × elapsed time t (n) / rotation time T (2)

  Then, the correction value learning unit 22 compares the detected interrupt angle A (n) with the ideal interrupt actual angle I (n) to obtain an angle error. Specifically, the correction value learning unit 22 calculates the fluctuation error E (n) by subtracting the interrupt actual angle I (n) from the interrupt angle A (n) based on Equation (3). The elapsed time t (N) is the rotation time T, and the fluctuation error E (N) is “0”.

Fluctuation error E (n) = interrupt angle A (n) −interrupt actual angle I (n) (3)
If the speed of the rotary shaft 14 is not constant, there may be a difference between when the pre-correction rotation angle is 360 ° from the reference position and when the reference position is reached again from the reference position. In this case, the fluctuation error E (n) calculated by the above equation (3) includes an error based on this deviation.

Therefore, the correction value learning unit 22 uses Equations (4) to (6) to remove an error based on this deviation from the fluctuation error E (n), and calculates an error correction value C (n). That is, the correction value learning unit 22 calculates an integrated value of the fluctuation error E (n) as shown in the mathematical formula (4).
Integrated value X = variation error E (1) + variation error E (2) +... + Variation error E (N) (4)

Here, it is assumed that the rotation shaft 14 returns to the original reference position during one rotation (the error becomes zero) and that the rotation speed is constant. In other words, assuming that the deviation during one rotation occurs evenly at each interrupt angle, the correction value learning unit 22 divides the integrated value X by a constant N as shown in Equation (5), thereby changing each variation. A reference position deviation G included in the error E (n) is calculated.
Reference position deviation G = integrated value X / constant N (5)

Then, the correction value learning unit 22 subtracts the reference position deviation G from each variation error E (n), as shown in Equation (6), thereby obtaining an error correction value C (() at each interrupt angle A (n). n) is calculated.
Error correction value C (n) = variation error E (n) −reference position deviation G (6)

  Then, the correction value learning unit 22 learns the calculated error correction value C (n) for each interrupt angle A (n). That is, the correction value learning unit 22 stores the error correction value C (n) in the storage device of the control device 20 in association with each interrupt angle A (n).

  The correcting unit 23 corrects the rotation angle before correction based on the learned error correction value C (n). An arbitrary method may be used to correct the pre-correction rotation angle based on the learned error correction value C (n). For example, an approximate curve of the error correction value is calculated from the error correction value C (n), and an error correction value for each pre-correction rotation angle is specified using the approximate curve of the error correction value, and the pre-correction rotation angle is corrected. May be. Further, for example, every time the rotation angle before correction coincides with the interrupt angle A (n), the error correction value C (n) corresponding to the interrupt angle A (n) is added or subtracted. It may be calculated.

  By the way, when the rotation speed of the rotating shaft 14 is low, the rotation speed is likely to fluctuate during the detection period as compared with the case where the rotation speed is high. That is, when the rotation speed is low, the rotation speed from the reference position to one rotation is not constant and unevenness is likely to occur as compared with the case where the rotation speed is high. For example, a state in which the speed when rotating from 0 to 180 degrees is different from the speed when rotating from 180 to 360 degrees tends to occur.

  In particular, in the present embodiment, the rotating shaft 14 of the ISG 100 is connected to the crankshaft 200a of the engine 200 and is interlocked with the crankshaft 200a, so that it is easily affected by the rotation. That is, during driving of the engine 200, the rotating shaft 14 is accelerated during the combustion stroke, and is easily decelerated during the compression stroke. This is particularly noticeable when the rotational speed is low compared to when it is fast. That is, the fluctuation range of the rotation speed tends to increase.

  As described above, when the error correction value is calculated on the assumption that the rotation speed in the detection period is constant, as described above, the error correction value is calculated even though the fluctuation range of the rotation speed in the detection period is large. Accuracy is reduced. Therefore, in the present embodiment, when it is predicted that the rotation speed is slow and the error correction value cannot be calculated with high accuracy, the error correction value is estimated from the map. In addition, learning is prohibited when the rotation speed is slow and it is predicted that the error correction value cannot be calculated with high accuracy.

  Specifically, when the rotation speed is less than a predetermined threshold, the correction unit 23 as the first correction unit refers to an error map corresponding to the rotation angle before correction and an error correction value of the rotation angle before correction. And the pre-correction rotation angle is corrected by the estimated error correction value. The error map is, for example, a map in which an error correction value is specified using the elapsed time from the reference position as a parameter, and is prepared and stored for each rotation angle before correction. The rotation speed is calculated by the control device 20 from the rotation time, for example.

  In the present embodiment, the error map varies depending on whether the engine 200 is driven. That is, when the engine 200 is driven, the angle error of the rotating shaft 14 tends to increase due to the influence from the crankshaft 200a. Therefore, the error map is changed depending on whether or not the engine 200 is driven to improve the accuracy of angle error estimation.

  On the other hand, if the error correction value is learned in the previous detection period when the rotation speed is equal to or higher than a predetermined threshold, the correction unit 23 as the second correction unit learns the learned error as described above. Based on the correction value, the rotation angle before correction is corrected.

  The threshold value is set to a larger value when the engine 200 is being driven than when the engine 200 is stopped. In other words, when the rotation speed is low, it is more susceptible to the influence of driving the engine 200 than when it is high. That is, the fluctuation range of the rotation speed during the detection period tends to increase. On the other hand, when engine 200 is stopped, the fluctuation range of the rotational speed during the detection period tends to be small. Therefore, the threshold value is set larger when the engine 200 is being driven than when it is stopped.

  Further, the correction value learning unit 22 prohibits learning when the fluctuation range of the rotation speed in the current detection period is larger than the predetermined speed width compared to the rotation speed in the previous detection period. For example, when the ratio of the current rotation time to the previous rotation time is outside a predetermined allowable range, learning is prohibited.

  The correction value learning unit 22 prohibits learning when the fluctuation range of the current flowing through the ISG 100 is larger than a predetermined current width. The current flowing through the ISG 100 is, for example, the field current of the field winding 13b and the armature current flowing through the armature winding 12b. If the difference between the current detected in the previous detection period and the current detected in the current detection period is outside the predetermined range, learning is prohibited. What is necessary is just to measure an electric current with a current sensor etc., for example.

  As described above, the correction value learning unit 22 has a fluctuation width of the rotation speed in the current detection period within a predetermined speed width compared to the rotation speed in the previous detection period, and the fluctuation width of the current flowing in the ISG 100 is smaller. The error correction value is calculated and learned on condition that the current is within a predetermined current width.

  Next, the flow of the correction value learning process will be described with reference to FIG. The correction value learning process is executed by the control device 20 at predetermined intervals while the rotary shaft 14 is rotating.

  The control device 20 determines whether or not the rotation shaft 14 is at the reference position, that is, whether or not a reference position signal has been input (step S101). If this determination result is negative, the correction value learning process is terminated. On the other hand, when the determination result of step S101 is affirmative, the control device 20 calculates the pre-correction rotation angle based on the detection result of the magnetic sensor 92 (step S102). Then, the control device 20 determines whether or not the calculated rotation angle before correction matches the interrupt angle A (n) (step S103).

  When this determination result is affirmative, the control device 20 acquires an elapsed time t (n) from the input of the reference position signal, and stores it in association with the interrupt angle A (n) (step S104). When the determination result of step S103 is negative, or after the process of step S104, the control device 20 determines whether or not the detection period has elapsed (step S105). That is, the control device 20 determines whether or not the reference position signal is input again. When the determination result is negative, the control device 20 executes the processes after step S102 again after a predetermined time has elapsed.

  On the other hand, when the determination result of step S105 is affirmative, the control device 20 acquires the rotation time T (step S106). That is, the time from the input of the reference position signal in step S101 to the input of the reference position signal in step S105 is acquired as the rotation time T.

  Next, the control device 20 determines whether or not the error correction value C (n) can be calculated (step S107). In step S107, the control device 20 compares the rotation speed in the current detection period with the fluctuation width of the rotation speed in the current detection period within a predetermined speed width, and the fluctuation width of the current flowing in the motor 10 is less than the rotation speed in the previous detection period. It is determined whether it is within a predetermined current width. It is determined whether or not an error correction value with a predetermined accuracy or higher can be calculated.

  When the determination result in step S107 is affirmative, as described above, the control device 20 uses the equations (2) to (6) to calculate the error correction value C (n) for each interrupt angle A (n). Calculate (step S108). Then, the control device 20 stores (learns) the error correction value C (n) in the storage device of the control device 20 in association with each interrupt angle A (n) (step S109). Then, the correction value learning process ends. On the other hand, when the determination result of step S107 is negative, the control device 20 ends the correction value learning process.

  Next, the correction process will be described with reference to FIG. The correction process is executed by the control device 20 at predetermined intervals. Even when the correction value learning process is being executed, the correction value learning process is interrupted at the execution cycle, the correction process is executed, and the correction value learning is interrupted after the correction process is completed. Resume processing.

  The control device 20 determines whether or not the rotation speed is greater than or equal to a threshold value (step S201). The threshold value is changed depending on whether or not engine 200 is being driven. When the determination result is affirmative, the control device 20 determines whether or not the error correction value has already been learned (stored) in the correction value learning process (step S202).

  When the determination result of step S202 is positive, the control device 20 corrects the pre-correction rotation angle calculated based on the detection result of the magnetic sensor 92 based on the learned error correction value (step S203). Then, the correction process ends.

  On the other hand, when the determination result of step S201 is negative, that is, when it is less than the threshold value, or when the determination result of step S202 is negative, that is, when learning is not performed, the control device 20 specifies an error map ( Step S204). The error map is changed depending on whether or not the engine 200 is being driven.

  Then, the control device 20 refers to the specified error map and estimates an error correction value according to the elapsed time from the reference position (step S205). The control device 20 corrects the pre-correction rotation angle calculated based on the detection result of the magnetic sensor 92 based on the error correction value estimated in step S205 (step S206). Then, the correction process ends.

  After completing the correction process, the control device 20 performs control of the motor 10 based on the corrected rotation angle.

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

  When the rotational speed of the rotating shaft 14 is less than the threshold, the control device 20 estimates an error correction value with reference to the error map, and corrects the pre-correction rotation angle with the estimated error correction value. On the other hand, when the rotation speed is equal to or higher than the threshold value, the control device 20 corrects the pre-correction rotation angle with the error correction value C (n) learned by the correction value learning unit 22. Thereby, when the rotational speed of the rotating shaft 14 is slow and the fluctuation of the rotational speed during the detection period tends to be large, that is, when the correction value learning unit 22 calculates the error correction value C (n), an error is likely to occur. In this case, the accuracy of rotation angle detection can be improved by using the error correction value of the error map.

  On the other hand, when the rotation speed of the rotating shaft 14 is fast and fluctuations in the rotation speed during the detection period are likely to be small, that is, when the correction value learning unit 22 does not easily generate an error when calculating the error correction value C (n). The accuracy of rotation angle detection can be improved by using the learned error correction value C (n).

  When the rotary shaft 14 is connected so as to be linked to the crankshaft 200a of the engine 200, the rotation speed of the rotary shaft 14 is affected by the rotation of the crankshaft 200a when the engine 200 is driven. When the engine 200 is driven, the crankshaft 200a is decelerated during the compression stroke and accelerated during the combustion stroke. For this reason, when the engine 200 is driven at a low speed, the rotation speed is particularly likely to be uneven in the detection period in which the rotation shaft 14 rotates from the reference position to a predetermined angle (360 ° in the present embodiment). That is, the fluctuation range of the rotation speed during the detection period tends to increase. On the other hand, when engine 200 is stopped, the fluctuation range of the rotational speed during the detection period tends to be small. Therefore, the threshold value is set larger when the engine 200 is being driven than when it is stopped. Thereby, the detection accuracy of a rotation angle can be improved.

  Moreover, when the rotating shaft 14 of the motor 10 is interlocked with the crankshaft 200a of the engine 200, generally, the range from the lower limit to the upper limit of the rotational speed tends to be widened. In this case, it is desirable to reduce the trouble of creating an error map by using the learned error correction value C (n) for correction.

  When the engine 200 is driven, unlike the case where the engine 200 is stopped, the rotational speed is likely to fluctuate and the angle error tends to increase. Accordingly, the control device 20 improves the estimation accuracy of the error correction value by differentiating the error map that is referred to when the engine 200 is being driven and the error map that is referred to when the engine 200 is being stopped. Thereby, the detection accuracy of a rotation angle can be improved.

  The control device 20 compares the rotational speed in the current detection period with the fluctuation width of the rotational speed within the predetermined speed width and the fluctuation width of the current flowing through the ISG 100 within the predetermined current width compared to the rotational speed in the previous detection period. As a condition, an error correction value C (n) is calculated and learned. That is, when the fluctuation width of the rotational speed is larger than the predetermined speed width, or when the fluctuation width of the current flowing through the rotating electrical machine is larger than the predetermined current width, the error correction value C (n) is calculated and learned. Banned. As a result, the error correction value C (n) can be calculated based on the detection result detected in the period in which the current fluctuation range and the rotation speed fluctuation range are small and the error based on the fluctuation is likely to be small. . Thereby, the calculation accuracy of the error correction value C (n) can be improved, and the detection accuracy of the rotation angle can be improved.

(Other embodiments)
The present invention is not limited to the above embodiment, and may be implemented as follows, for example. In the following, parts that are the same or equivalent to each other in the respective embodiments are denoted by the same reference numerals, and the description of the same reference numerals is used.

  In the above-described embodiment, the rotation angle detection device 1 may include the disturbance magnetic flux density calculation unit 24 that calculates the magnetic flux density due to the disturbance among the magnetic flux densities detected by the magnetic sensor 92. Then, the correction value learning unit 22 may calculate the error correction value in consideration of the magnetic flux density calculated by the disturbance magnetic flux density calculation unit 24.

  A rotation angle detection device 1 in another example will be specifically described with reference to FIG. As shown in FIG. 6, the control device 20 as the disturbance magnetic flux density calculation unit 24 detects the field current flowing in the field winding 13 b with the current sensor 301 during the detection period. Then, the control device 20 calculates the leakage magnetic flux based on the field current, and among the magnetic flux densities detected by the magnetic sensor 92 based on the leakage magnetic flux, the magnetic flux density based on the leakage magnetic flux (magnetic flux density due to disturbance). Is calculated by map calculation or model calculation.

  Then, the control device 20 removes the magnetic flux density based on the disturbance from the magnetic flux density detected by the magnetic sensor 92 and calculates the pre-correction rotation angle based on the removed magnetic flux density. Note that the control device 20 may calculate an angle error based on the magnetic flux density based on the disturbance and reflect the angle error in the error correction value. Further, the current that causes the leakage magnetic flux is not limited to the field current, but the current flowing through the armature winding 12b and the current flowing through the control device 20, and these currents are detected to calculate the leakage magnetic flux. May be.

  Further, the disturbance is not limited to the leakage magnetic flux. For example, the control device 20 may detect the temperature of the magnet 91 with a temperature sensor in the detection period. And the control apparatus 20 calculates the magnetic flux density (magnetic flux density by disturbance) based on the temperature change of the said magnet 91 among the magnetic flux densities detected by the magnetic sensor 92 by map calculation or model calculation. As described above, the control device 20 may remove the magnetic flux density based on the disturbance from the magnetic flux density detected by the magnetic sensor 92 and calculate the pre-correction rotation angle based on the removed magnetic flux density.

  By adopting the above configuration, it is possible to calculate and learn an error correction value by excluding a change in magnetic flux density due to disturbance, for example, an error due to a change in temperature of the magnet 91 or a change in leakage magnetic flux. Thereby, the accuracy of the error correction value calculated by the correction value learning unit 22 can be improved, and the detection accuracy of the rotation angle can be improved.

  In the above-described embodiment, the control unit 20 of the ISG 100 includes the calculation unit 21, the correction value learning unit 22, and the correction unit 23. However, the calculation unit 21 and the correction value are different from the control unit 20 of the ISG 100. One or all of the functions of the learning unit 22 and the correction unit 23 may be provided.

  In the above embodiment, a plurality of reference positions may be provided. That is, the reference position signal may be output at a predetermined angle (for example, 180 °) interval. In this case, a period from the first reference position (0 °) to the second reference position (180 °) may be set as the detection period.

  In the above embodiment, the threshold value may not be changed depending on whether or not the engine 200 is being driven. Further, the error map referred to when the engine 200 is being driven may be the same as the error map referred to when the engine 200 is being stopped. Further, an error map may be provided for each rotation speed, and an error map corresponding to the detected rotation speed may be referred to.

  In the above embodiment, the correction value learning unit 22 may calculate and learn an error correction value even if the fluctuation range of the rotation speed is not within the predetermined speed range. Similarly, the error correction value may be calculated and learned even if the fluctuation range of the current flowing through the ISG 100 is not within the predetermined current width.

  In the above embodiment, a resolver may be employed instead of the magnet 91 and the magnetic sensor 92. In this case, for example, the exciting coil of the resolver provided in the rotor corresponds to the signal output unit, and the detection coil corresponds to the sensor.

  DESCRIPTION OF SYMBOLS 1 ... Rotation angle detection apparatus, 14 ... Rotary axis, 21 ... Calculation part, 22 ... Correction value learning part, 23 ... Correction part, 91 ... Magnet, 92 ... Magnetic sensor, 100 ... ISG.

Claims (5)

A signal output unit (91) that is provided on a rotating shaft of the rotating electrical machine (100) and rotates together with the rotating shaft (14); and a sensor (92) that detects a signal from the signal output unit. In the rotation angle detection device (1) for detecting the rotation angle of the rotation shaft based on the detection result of
A calculation unit (21) that calculates a rotation angle of the rotation shaft as a rotation angle before correction based on a detection result detected from the sensor;
A correction value learning unit that calculates and learns an error correction value used for error correction of the rotation angle before correction based on a rotation angle before correction detected in a detection period in which the rotation axis rotates from a reference position to a predetermined angle. (22)
When the rotation speed of the rotation axis is less than a threshold value, an error correction value of the rotation angle before correction is estimated with reference to an error map, and the rotation angle before correction is corrected by the estimated error correction value. 1 correction unit (23);
A rotation angle detection device comprising: a second correction unit (23) that corrects the rotation angle before correction based on the error correction value learned by the correction value learning unit when the rotation speed is equal to or greater than a threshold value. The rotating shaft of the rotating electric machine is connected to the crankshaft (200a) of the internal combustion engine (200),
The rotation angle detection device according to claim 1, wherein the threshold value is set to a larger value when the internal combustion engine is being driven than when the internal combustion engine is being stopped. The rotating shaft of the rotating electrical machine is connected to interlock with the crankshaft of the internal combustion engine,
3. The first correction unit according to claim 1, wherein the error map that is referred to when the internal combustion engine is being driven differs from the error map that is referred to when the internal combustion engine is stopped. Rotation angle detection device.   The correction value learning unit has a fluctuation range of the rotation speed in the current detection period within a predetermined speed range as compared to the rotation speed in the previous detection period, and the fluctuation range of the current flowing in the rotating electrical machine is a predetermined current. The rotation angle detection device according to claim 1, wherein the error correction value is calculated and learned on condition that the width is within a range. In the detection period, a disturbance magnetic flux density calculation unit (24) for calculating a magnetic flux density due to a disturbance among magnetic flux densities detected by the sensor,
The said correction value learning part calculates the said error correction value in consideration of the magnetic flux density by the said disturbance calculated by the said disturbance magnetic flux density calculation part, and learns any one of Claims 1-4. Rotation angle detection device.

Images