JP2010054282A - Device for measuring physical quantity of rotary member - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a structure that starts to rotate again when a rotary member stops rotation, and to perform appropriate filtering processing. <P>SOLUTION: When rotation of a hub constituting a rolling bearing unit for wheel support as the rotary member stops, the relationship between the circumferential position of an encoder at that time and an error component of a primary rotation component such as swinging based on an assembling error of the encoder to the hub is stored by a storing means such as memory. Simultaneously with restart of rotation of the hub, the filtering processing by an adaptation filter is started based on the relationship stored in the storing means. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The physical quantity measuring device for a rotating member according to the present invention supports, for example, a wheel of a vehicle (automobile) so as to be rotatable with respect to a suspension device, and the rotational speed and angle of the wheel or the magnitude of a load applied to the wheel. Is used to ensure stable operation of the vehicle. Or, it is incorporated in a rolling bearing unit for supporting the spindle of various machine tools, and is used to measure the load applied to the spindle and the displacement due to thermal expansion, etc., and to adjust the feed rate of the tool appropriately. .

  For example, a rolling bearing unit is used to rotatably support a vehicle wheel with respect to a suspension device. In order to ensure the running stability of the vehicle, a running state stabilizing device for the vehicle such as an antilock brake system (ABS) or a traction control system (TCS) is widely used. According to these running state stabilization devices such as ABS and TCS, it is possible to stabilize the running state of the vehicle during braking or acceleration, but in order to ensure this stability even under more severe conditions It is necessary to control the brakes and the engine by incorporating more information that affects the running stability of the vehicle.

  That is, in the case of the conventional running state stabilizing device such as ABS or TCS, since so-called feedback control is performed in which slippage between the tire and the road surface is detected to control the brake and the engine, the brake and engine Control is delayed for a moment. In other words, in order to improve performance under severe conditions, the so-called feed-forward control prevents slippage between the tire and the road surface, or the so-called brake one-side effect where the left and right wheels have extremely different braking forces. Cannot be prevented. Furthermore, it is impossible to prevent the running stability of a truck or the like from being deteriorated based on the poor loading state.

  In order to cope with such a problem, in order to perform the feedforward control or the like, one of a radial load and an axial load applied to the wheel or a rolling bearing unit for supporting the wheel with respect to the suspension device or It is conceivable to incorporate a load measuring device for measuring both. As a wheel support rolling bearing unit with a load measuring device that can be used in such a case, Patent Document 1 discloses a structure as shown in FIGS. An invention relating to a wheel bearing rolling bearing unit 1 with a load measuring device capable of measuring a load and a moment applied to a hub 3 which is a ring and an inner ring is described.

  In this conventional structure, rolling elements 6, between the double row outer ring raceways 4, 4 formed on the inner peripheral surface of the outer ring 2 and the double row inner ring raceways 5, 5 provided on the outer peripheral surface of the hub 3, A plurality of 6 are provided so as to roll freely for each row. The hub 3 is rotatably supported on the inner diameter side of the outer ring 2. In the illustrated example, a ball is used as the rolling element, but in the case of a rolling bearing unit for supporting a wheel of a heavy vehicle, a tapered roller may be used as each rolling element.

  Such a load measuring device 7 (also having a function as a rotational speed and rotational angle detecting device) that is a physical quantity measuring device combined with the rolling bearing unit 1 is externally fitted and fixed to the intermediate portion of the hub 3. One encoder 8 made of a permanent magnet and a pair of sensors 9, 9 provided in a portion between the rolling elements 6, 6 arranged in a double row at the axially intermediate portion of the outer ring 2. And a calculator (not shown). The detected surface 10, which is the outer peripheral surface of the encoder 8, has a cylindrical surface shape, and the portion magnetized in the N pole and the portion magnetized in the S pole are alternately arranged in the circumferential direction on the detected surface 10. Are arranged at equal intervals. In addition, the boundary between the portion magnetized in the N pole and the portion magnetized in the S pole is inclined by the same angle with respect to the axial direction of the encoder 8, and the inclined direction with respect to the axial direction is inclined with respect to the encoder. 8 are opposite to each other with an axial middle portion of 8 as a boundary. Therefore, the portion magnetized in the N pole and the portion magnetized in the S pole have a “<” shape with the axially middle portion protruding (or recessed) most in the circumferential direction. Accordingly, the magnetic characteristics of the detected surface 10 change alternately and at equal intervals in the circumferential direction, but the changing phase gradually changes in the axial direction of the detected surface 10.

  The pair of sensors 9 and 9 is an active type magnetic sensor, and a magnetic detecting element such as a Hall element or a magnetoresistive element is provided in the detection part of both the sensors 9 and 9. The characteristics of such a magnetic detection element change between a state facing the N pole and a state facing the S pole. In the two sensors 9 and 9, the detection portions of both the sensors 9 and 9 are closely opposed to the detection surface 10 of the encoder 8. Note that the positions where the detection portions of these sensors 9 and 9 face the detected surface 10 of the encoder 8 are the same with respect to the circumferential direction of the encoder 8. In the state where an axial load is not applied between the outer ring 2 and the hub 3, the axially intermediate portion between the portion magnetized in the N pole and the portion magnetized in the S pole is the most in the circumferential direction. The installation position of each member 8, 9, 9 is regulated so that the protruding part (the part where the inclination direction of the boundary changes) is just at the center position between the detection parts of both sensors 9, 9. ing.

  In the case of the conventional structure described in Patent Document 1 configured as described above, when an axial load is applied between the outer ring 2 and the hub 3, the phase in which the output signals of the sensors 9, 9 change is shifted. Although detailed description is described in the above-mentioned Patent Document 1, the phase shift of the output signal is shifted in a direction corresponding to the axial load applied between the outer ring 2 and the hub 3. Further, the degree of the phase shift of the output signals of the sensors 9 and 9 due to this axial load increases as the axial load increases. Therefore, the relative displacement amount in the axial direction between the outer ring 2 and the hub 3 based on the presence or absence of the phase shift of the output signals of the sensors 9, 9 and the direction and magnitude of the shift, if any. The direction and magnitude of the axial load acting between the outer ring 2 and the hub 3 can be obtained.

  By the way, in the case of the wheel support rolling bearing unit 1 as described in Patent Document 1, the encoder 8 is rotated along with the rotation of the hub 3 due to an assembly error when the encoder 8 is assembled to the hub 3. The detected surface 10 may be apparently displaced (vibrates with rotation) regardless of the load. When the geometric center axis of the detected surface 10 of the encoder 8 and the rotation center axis do not coincide with each other due to the assembling error or the like, that is, both the center axes deviate in the radial direction or are inclined with respect to each other. Regardless of the load, the position in the width direction of the detection surface 10 of the encoder 8 where the detection portions of the sensors 9 and 9 face each other is shifted. For example, when the load to be detected is an axial load as in the load measuring device 7, such a deviation is caused by the rotation of the hub 3 on the detected surface 10 of the encoder 8. It appears as a rotation primary axial displacement movement such as a whirl. As a result, even if the load does not change, the pattern in which the output signals of the sensors 9 and 9 change may change, and the measurement accuracy of the load may deteriorate.

Therefore, the load measuring device 7 combined with the rolling bearing unit 1 is configured as shown in FIG. 4 so that the output signals of the sensors 9 and 9 (the output signal of one sensor 9 and the output signal of the other sensor 9). The error component of the primary rotation component such as the above-mentioned swirl is eliminated from the phase difference). This operation will be briefly described below.
Of FIG 4, d _d represents the phase difference between the output signals of the accompanying these two sensors 9, 9 to the actual displacement with respect to the two sensors 9, 9 of the encoder 8, d _n is the rotation such as the whirling An error component of the primary component, x represents a reference signal correlated with the error component d _n, and y represents a cancel signal that is an output of the adaptive filter 11.
Further, d is a phase difference between output signals actually measured by the sensors 9 and 9 (hereinafter referred to as an output signal), and the actual phase difference d _d and the error component d _n are superimposed (d = d). _d + d _n ).
Further, e is a correction signal, which is obtained after the output signal d is filtered by the adaptive filter 11 (e = dy).

The adaptive filter 11 uses the reference signal x and the correction signal e as information, and based on a synchronous LMS algorithm as described in Patent Document 1, a filter coefficient sequence constituting the adaptive filter 11 by self-learning W (consisting of the same number of filter coefficients as the number of pulses of one rotation of the encoder 8) is successively updated appropriately. When the self-learning is complete, the adaptive filter 11 forms a FIR having the same characteristics as the transmission characteristics of the actual signal flow "d _n → d '(finite impulse response) filter.
According to such a configuration, if the cancel signal y of the adaptive filter 11 after the self-learning is completed is subtracted from the output signal d, the error component d is output from the output signals d of the sensors 9 and 9. removal of the _n become (d-d _n) that is equivalent. When the cancel signal y includes a DC level (direct current component), the correction signal e or the cancel signal y is corrected for the DC level by the process described in Patent Document 1.

FIG. 5 shows a result of applying the filtering process to the output signal d by the adaptive filter 11 having the configuration shown in FIG. 4 as described above to eliminate the error component d _n of the rotation first-order component such as the swing. ing. (A) in FIG. 5 shows the output signals d of the sensors 9 and 9 that have not been subjected to the filtering process. Such in FIG. 5 (A), the rotation first order error component d _n based on the assembly error of the encoder 8 has appeared larger. On the other hand, FIG. 5B shows the modified signal e obtained as a result of performing the filtering process by the adaptive filter 11 on the output signal d. From such FIG. 5 (B), the by the filtering process, it is understood that it is possible to reduce the rotational first-order error component d _n which is a problem when obtaining the axial load.
However, as shown in (B) of FIG. 5, from the output signal d to sufficiently reduce the error component d _n requires a predetermined time (about 2 seconds in (B) of FIG. 5) . This is because the cancel signal y of the adaptive filter 11 has not converged because the self-learning of the adaptive filter 11 has not progressed sufficiently. Thus, the data during which the self-learning of the adaptive filter 11 is not sufficiently advanced cannot be used for control or the like because the reliability is low.
Further, in the case of the structure of Patent Document 1 as described above, in order to detect the phase difference, it is necessary to measure the time from edge to edge of a plurality of pulses. Therefore, infinitely long time cannot be measured. Therefore, if the rotation speed of the hub 3 becomes lower than a certain speed or the rotation stops, the calculation must be stopped. Therefore, when the rotation starts again and the calculation is restarted, it is necessary to relearn the adaptive filter 11, and as shown in FIG. 5B, until the cancel signal y of the adaptive filter 11 converges. Data cannot be used for control or the like for a predetermined time.

JP 2007-40954 A

  In the present invention, in view of the circumstances as described above, when the rotation member stops rotating, the rotation 1 starts rotating again, and at the same time, by performing an appropriate filtering process, the rotation 1 based on the assembly error from the output signal of the sensor. The invention was invented to realize a physical quantity measuring device for a rotating member that can effectively reduce the following error components.

An apparatus for measuring a physical quantity of a rotating member according to the present invention includes an encoder, a sensor, a filter circuit, and a calculator.
Of these, the encoder is supported by a part of the rotating member concentrically with the rotating member, and the characteristics of the detection surface are alternately changed in the circumferential direction.
The sensor is supported by a portion that does not rotate with the detection portion facing the detection surface, and changes its output signal in response to a change in the characteristics of the detection surface.
The filter circuit performs a filtering process on an output signal of the sensor (including a processing signal obtained based on the output signal by a waveform shaping circuit or the like). That is, the filter circuit eliminates an error component based on an error related to a characteristic change in the circumferential direction of the detected surface, among the fluctuations of the output signal. The filter circuit described in the present specification and claims includes not only an independent filter analog circuit but also a structure for executing a filter process (filter operation) in a program installed in an arithmetic unit. In this case, a part of the arithmetic unit corresponds to the filter circuit.
Further, the arithmetic unit calculates a physical quantity of the rotating member based on the output signal subjected to the filtering process by the filter circuit. That is, the arithmetic unit has a function of calculating the physical quantity based on a pattern in which the output signal changes after passing through the filter circuit.

In particular, the physical quantity measuring apparatus for a rotating member according to the present invention comprises an error component (a filter circuit constituting a characteristic change over the circumferential position and the circumferential direction of the encoder at the time when the rotation of the rotating member is stopped. The information is stored in the storage means. Then, at the same time as the rotation of the rotating member is resumed, the filtering process is started based on the relationship stored in the storage means.
In the case of carrying out the present invention as described above, for example, the physical quantity is at least one of the rotation speed and the rotation angle of the rotating member.
Alternatively, the physical quantity is at least one of a displacement amount of the rotating member and an external force acting on the rotating member.
Further, in the case of obtaining the displacement amount or the external force in this way, preferably, a pair of sensors are installed in a state where the respective detection units are opposed to the positions separated in the width direction of the detected surface of the encoder. Moreover, the boundary where the characteristic changes in the circumferential direction of the portion of the detected surface where the detection portion of at least one sensor faces is inclined with respect to the width direction. Then, the phase of the change in the output signal of the at least one sensor is changed corresponding to the position in the width direction of the detected surface of the encoder, which the detection unit of the sensor faces.
When the present invention is carried out, for example, the rotating member is a rotating bearing ring of a rolling bearing unit or a member that is coupled and fixed to the rotating bearing ring and rotates together with the rotating bearing ring. Further, the rolling bearing unit exists on the circumferential surfaces of the rotating side bearing ring that rotates in the used state, the stationary side bearing ring that does not rotate in the used state, and the rotating side bearing ring and the stationary side bearing ring that face each other. And a plurality of rolling elements provided between the stationary-side track and the rotating-side track.
Further, when the present invention is implemented, the filter circuit is preferably an adaptive filter.

  The physical quantity measuring device for a rotating member of the present invention configured as described above has the relationship between the circumferential position of the encoder at that time and the error component related to the characteristic change in the circumferential direction when the rotating member stops rotating. By storing in the storage means, at the same time as the rotation member resumes rotation, an appropriate filtering process is applied to the output signal of the sensor to eliminate the error component included in the output signal. I can do things. For this reason, the reliability of the data immediately after resuming the rotation can be improved, and this data can be used for various controls such as TCS control at the time of starting.

  Hereinafter, embodiments of the present invention will be described. The feature of this example is that the configuration of the filter circuit is such that when the rotation of the hub, which is a rotating member, is stopped and the rotation is started again, an appropriate filtering process can be performed simultaneously with the start of the rotation. Is in the point which devised. Since the structure and operation other than the characteristic part of this example are the same as those of the conventional structure shown in FIGS. 2 to 4, overlapping illustration and description are omitted. Further, when it is necessary to explain the structure of this example, reference is made to FIGS.

  In the case of this example, when the rotation of the hub 3 constituting the wheel support rolling bearing unit 1 (see FIG. 2), which is a rotating member, is stopped, the circumferential position (rotation start position) of the encoder 8 at that time is stopped. As the reference position, the number of pulses from the reference position, the rotation angle, and the like, and the error component {rotation primary component such as run-out based on the assembly error of the encoder 8 with respect to the hub 3 {the adaptive filter 11 (FIG. And a filter coefficient sequence W (referred to as a filter coefficient w equal to the number of pulses of one rotation of the encoder 8)} is stored in a storage means such as a memory (not shown). At the same time when the hub 3 resumes rotation, filtering processing by the adaptive filter 11 is started based on the relationship between the circumferential position and the error component stored in the storage means.

  In the case of this example, when the rotation of the hub 3 that is a rotating member is stopped and the rotation is started again, the circle stored at the time when the rotation of the encoder 8 is stopped is stored in the storage means. The filtering process can be started in a state where an appropriate value corresponding to the circumferential position is set in the filter coefficient sequence W constituting the adaptive filter 11. Therefore, the encoder 8 is assembled from the output signals of both the sensors 9 and 9 immediately after the hub 3 starts rotating again (the phase difference between the output signal of one sensor 9 and the output signal of the other sensor 9). It is possible to effectively reduce an error component of a primary rotation component such as a run-out based on an error, and it is possible to improve the reliability of data used for control. The reason will be described below. Note that the learning algorithm of the adaptive filter 11 is described in detail in the above-mentioned Patent Document 1, and therefore omitted.

上記（１）式中、ｋは時系列データのデータ番号、Ｎは上記適応フィルタ１１としているＦＩＲフィルタのタップ数である。又、ｗはＦＩＲフィルタのフィルタ係数列Ｗを構成するフィルタ係数を表し、ｗ_kはｋ番目のデータ処理をする場合に使用するフィルタ係数を表している。 As described in Patent Document 1, the cancel signal y (see FIG. 4) of the adaptive filter 11 is expressed by the following equation (1) using the reference signal x and the filter coefficient sequence W constituting the adaptive filter 11. ) Is uniquely determined.

In the above equation (1), k is the data number of the time series data, and N is the number of taps of the FIR filter as the adaptive filter 11. Further, w represents a filter coefficient constituting the filter coefficient string W of the FIR filter, and w _k represents a filter coefficient used when k-th data processing is performed.

この（２）式で、上記参照信号ｘが値１のインパルスとなる位置ｊは、時系列ｋが進んでいくのに従って右側に１個ずつずれて行き、一番右側の「Ｎ−１」番目までずれると、次の時系列では、新たなインパルス値が一番左の０番目に表れる事になる。即ち、上記参照信号ｘは、値１のインパルスの位置を０番目からＮ−１番目まで巡回させただけのデータ列となる。この（２）式を、前述の（１）式に当て嵌めると、次の（３）式を得られる。

Further, the reference signal x input to the adaptive filter 11 is the rotation n-th order (n is n A signal having a correlation with a (positive integer) component may be used, so that one impulse signal per one rotation of the encoder 8 may be used. Therefore, also in this example, the reference signal x is one impulse signal, and at the same time, the reference signal x having the same number of taps N of the adaptive filter as the number of pulses per one rotation of the encoder is used. In this case, the reference signal x used for calculation at the instant of time series k is expressed by the following equation (2).

In this equation (2), the position j where the reference signal x becomes an impulse of value 1 is shifted to the right by one as the time series k advances, and the rightmost “N−1” th In the next time series, a new impulse value appears at the leftmost 0th position. That is, the reference signal x is a data string obtained by circulating the position of the impulse having the value 1 from the 0th to the (N-1) th. When this equation (2) is applied to the aforementioned equation (1), the following equation (3) is obtained.

As can be seen from the above equation (3), when the reference signal x as in the above equation (2) is used, the adaptive filter 11 causes the position in the circumferential direction of the encoder 8 at the instant of time series k {from the reference position. As a cancel signal y (k) for erasing the error component at a position corresponding to the jth pulse (an angle corresponding to this), the jth filter coefficient w _k (j) of the filter coefficient sequence W Is output. That is, the circumferential position of the encoder 8 and the filter coefficient w constituting the filter coefficient string W correspond one-to-one. For this reason, it is stored how many pulses (an angle corresponding to this position) of the circumferential position of the encoder 8 from the reference position when the rotation of the hub 3 is stopped, When the hub 3 resumes rotation, the filter coefficient string W is reconstructed (filter coefficient rearrangement) so that the filter coefficient w of the filter coefficient string W corresponding to the stored circumferential position can be output. If the degree of processing) is performed, appropriate filtering processing by the adaptive filter 11 can be started from the moment when rotation is resumed.

  FIG. 1 shows the result of filtering the signal shown in FIG. 5A by the adaptive filter 11 having the configuration of this example. From the result of FIG. 1, it can be seen that the first-order error component of rotation can be reduced by the filtering process of the adaptive filter 11 from the moment when the rotation of the hub 3 is resumed. Although some high-frequency noise remains in the signal after filtering processing, this high-frequency noise can be easily reduced by a low-pass filter and without causing a response delay that causes a practical problem. It doesn't matter. When the rotation is resumed, the reference position is updated to the circumferential position where the rotation is resumed.

In the case of the configuration of this example as described above, when the power source is switched from OFF to ON and the hub 3 is rotated for the first time, an appropriate value is set in the filter system sequence W of the adaptive filter 11. I can't. For this reason, similarly to the structure described in Patent Document 1, the data until the cancellation signal y of the adaptive filter 11 converges (until self-learning of the adaptive filter 11 sufficiently proceeds) is reliable. It cannot be used for control. Note that it is possible to store the relationship in the storage means even after the power is turned off, but it is also possible that the vehicle moves in the OFF state. In this case, the reliability of the control after the ON cannot be ensured. It is not preferable.
In addition, in order to detect the phase difference, it is necessary to measure the time from edge to edge of a plurality of pulses in the same manner as the structure described in Patent Document 1 above. Therefore, infinitely long time cannot be measured. Therefore, if the rotation speed of the hub 3 becomes lower than a certain speed or the rotation stops, the calculation must be stopped. However, since only the pulse edge recognition of the encoder 8 can be continued as long as the power is not turned off, the circumferential position of the encoder 8 can be recognized. As described above, even when the rotational speed is extremely low, the circumferential position of the encoder 8 is continuously recognized, and the circumferential position at the moment when the rotation of the encoder 8 is completely stopped is stored. . Further, after the encoder 8 stops rotating, the circumferential position is recognized even when the encoder 8 starts rotating at such a low speed that the above calculation cannot be resumed.
If the encoder 8 can be rotated in both forward and reverse directions, not only the circumferential position is recognized, but also the rotational direction is recognized by, for example, a rotary encoder using an AB two-phase incremental method. It can respond by doing. Also in this case, the same effect as this example can be obtained by setting an appropriate value in the filter coefficient sequence W of the adaptive filter 11 based on the information on the circumferential position and the rotation direction. Further, the method for recognizing the circumferential position of the encoder 8 is not limited to the relative position recognizing method from the reference value as in this example, and an absolute position recognizing method can also be used. Such a position recognition method may be determined by design in consideration of use, accuracy, cost, and the like.

In the embodiment described above, the equation (2) to have been a reference signal x data such as shown, this reference signal x is, as described in the above patent document 1, the correlation between the error component d _n Any signal may be used. For example, if the influence of the swing of the encoder 8 is a waveform whose primary component is the rotation primary, and the encoder 8 has N pulses per rotation, N data may be one cycle. Sine waves, triangular waves, sawtooth waves, rectangular waves, pulse waves, etc. can be used. Even when such a reference signal x is used, it is possible to obtain the same effect as that of the above embodiment by reconstructing the filter coefficient sequence W of the adaptive filter 11 as described above.
In addition, the present invention is not limited to the conventional structure described in Patent Document 1 and the structure for obtaining an axial load as described in the embodiment, but also rolling as described in Patent Document 1. The present invention can be applied to various structures for measuring physical quantities (rotational speed, rotational angle, displacement, external force, etc.) of a rotating member, such as a structure for obtaining a radial load applied to the bearing unit.

The diagram which shows the output signal after a filtering process in one example of embodiment of this invention. The fragmentary sectional view of the rolling bearing unit for wheel support which shows an example of the conventional structure. Similarly, the perspective view of an encoder. Similarly, the block diagram of the adaptive filter which filters the data based on the output signal of a sensor. Similarly, a diagram (A) showing an output signal of a sensor including an error component, and a diagram (B) showing a signal obtained by filtering the output signal of the sensor with an adaptive filter.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Rolling bearing unit 2 Outer ring 3 Hub 4 Outer ring raceway 5 Inner ring raceway 6 Rolling element 7 Load measuring device 8 Encoder 9 Sensor 10 Detected surface 11 Adaptive filter

Claims (1)

  An encoder that is supported and fixed to a part of the rotating member concentrically with the rotating member, and whose characteristics of the surface to be detected are alternately changed with respect to the circumferential direction, and which rotates in a state where the detecting portion faces the surface to be detected. A sensor that is supported by a portion that is not supported and changes its output signal in response to a change in the characteristics of the detected surface, a filter circuit that performs a filtering process on the output signal of the sensor, and a filtering process that is performed by the filter circuit An arithmetic unit that calculates a physical quantity related to the rotating member based on the output signal, and the filter circuit is based on an error related to a characteristic change in the circumferential direction of the detected surface among the fluctuations of the output signal. An error component is eliminated, and the arithmetic unit performs a filtering process by the filter circuit in a pattern in which the output signal changes. Accordingly, in the physical quantity measuring apparatus for a rotating member having a function of calculating the physical quantity, when the rotation of the rotating member is stopped, the encoder is moved in the circumferential position and the circumferential direction at that time. The relationship with the error component relating to the characteristic change is stored in the storage means, and at the same time as the rotation of the rotating member is resumed, the filtering process is started based on the relationship stored in the storage means. An apparatus for measuring a physical quantity of a rotating member.

Images