JP7707953B2 - Gear Fault Detection System - Google Patents

Description

The present invention relates to a gear failure detection system that detects gear failures.

Gears can experience various types of fatigue, and can also suffer spontaneous damage depending on the length of use. In general, even if there is no damage, gears can generate vibrations due to changes in meshing stiffness over time and other reasons. The presence of damage increases the amount of vibration.

In a gear transmission mechanism, vibrations of the gears are transmitted to a case of the gear transmission mechanism through a shaft and bearings that support the gears. Patent Document 1 discloses a device for detecting an abnormality in the gears of a gear transmission mechanism.
The device disclosed in Patent Document 1 has an acceleration sensor that detects vibrations that occur when two gears mesh with each other. The acceleration sensor is fixed to a casing of the transmission.

JP 2017-181282 A

A part of the vibration energy generated when the gears mesh is released from the case that houses the gears to the outside as audible noise. Along the transmission path of this audible noise, the vibration excited by the original meshing is affected by the vibration transmission path specific to the case as it travels from the vibration excitation point to the pick-up point (the installation position of the acceleration sensor). Therefore, the vibration characteristics measured at the pick-up point are changed compared to the vibration at the excitation point. This is due to the stiffness and mass properties of the objects in the transmission path, as well as damping, which changes the vibration time series in terms of altered amplitude and phase across the entire frequency band.

In the configuration disclosed in Patent Document 1, the acceleration sensor is attached to the casing of the speed-up gear, so the vibrations generated when two gears (gear pairs) mesh are transmitted to the acceleration sensor via the casing of the speed-up gear. The vibrations that reach the acceleration sensor are affected by the vibration transmission path specific to the casing, so the vibrations measured by the acceleration sensor may have different vibration characteristics compared to the vibrations at the vibration generating point (the meshing gear pairs). In addition, the configuration of Patent Document 1 cannot detect abnormalities in individual teeth of the gears.
In addition, it is expected that the higher the performance of the sensor, the higher the diagnostic accuracy will be, but high-performance sensors are expensive.

Therefore, the present invention aims to provide a technology that can accurately detect gear abnormalities even if the sensor performance is low.

According to one aspect of the present invention, a gear failure detection system is provided for detecting gear failure. The gear failure detection system includes an acceleration sensor that is provided in a bearing that rotatably supports a shaft on which a gear is provided, and that measures instantaneous values of acceleration caused by vibration of the bearing each time the shaft rotates for a predetermined time shorter than the time required for one rotation of the shaft; an acceleration profile generation unit that rearranges multiple instantaneous values of acceleration measured by the acceleration sensor while the shaft rotates multiple times, and generates an acceleration profile that represents the change in acceleration per one rotation of the shaft by connecting the rearranged instantaneous values of acceleration; and a failure determination unit that determines gear failure based on the acceleration profile generated by the acceleration profile generation unit.

In one aspect of the present invention, multiple instantaneous values of acceleration measured by an acceleration sensor are accumulated while the shaft rotates multiple times. A faulty location in the gear is detected as an acceleration anomaly at any angular position within a 360° angular range, and the acceleration anomaly corresponding to the faulty location in the gear is repeatedly measured during multiple rotations of the shaft. In one aspect of the present invention, multiple instantaneous values of acceleration accumulated during multiple rotations of the shaft are sorted, and the sorted instantaneous values of acceleration are connected together to generate an acceleration profile that represents the change in acceleration per rotation of the shaft, even if the performance of the acceleration sensor is low, so that gear failure can be accurately determined.

FIG. 1 is a diagram showing a gear failure detection system having a gear failure determination device and a gear speed change mechanism according to an embodiment of the present invention. FIG. 2 is a block diagram showing the configuration of the gear failure determination device of FIG. FIG. 3 is an exploded perspective view of the first bearing of FIG. FIG. 4 is an exploded perspective view of the first bearing of FIG. 1 as seen from a different direction. FIG. 5A is a plan view showing the cover, the coil board, the circuit board group, and the storage battery. FIG. 5B is a plan view showing the coil substrate in detail. FIG. 6 is a block diagram for explaining the mechanism of self-power generation. FIG. 7 is a graph for explaining the operation of the power management IC. FIG. 8 is a block diagram showing the configuration of a measuring device in the first bearing. FIG. 9 is a timing chart for explaining the processing of the measuring device. FIG. 10 is a diagram for explaining the sampling period of the measured acceleration value and the measured angle value. FIG. 11 is a flowchart for explaining the process of the measuring device. FIG. 12 is a graph showing the change over time in acceleration and angle measured by the measuring device. FIG. 13 is a block diagram showing elements of the gear fault detection system that are particularly relevant to gear fault determination. FIG. 14 is a graph of an acceleration profile and a graph of an angle profile generated by the gear failure determination device according to the embodiment. FIG. 15 shows graphs of acceleration profiles and angle profiles obtained using expensive acceleration and angle sensors. FIG. 16 is a diagram for explaining the failure determination process executed by the gear failure determination device. FIG. 17 is an angle vs. tooth map without the use of angle offset values. FIG. 18 is a map of angles and teeth using angle offset values. FIG. 19 is a block diagram showing elements particularly related to gear failure determination in a gear failure detection system according to a modified embodiment. FIG. 20 is a diagram for explaining another example of the sampling period of the measured acceleration values and the measured angle values according to the modified example. FIG. 21 is a diagram for explaining linear interpolation executed by the gear fault determination device in the modified example of FIG. FIG. 22 is a diagram showing an example of an acceleration profile obtained in the embodiment. FIG. 23 is a graph showing the relative signal intensity coefficients for each of the three axes calculated under certain conditions. FIG. 24 is a graph showing triaxial FFT results obtained from a triaxial acceleration profile obtained under certain conditions. FIG. 25 is a diagram for explaining the resolution of an angle sensor. FIG. 26 is a graph showing the acceleration profile used as the basis for the simulation. FIG. 27 is a graph showing an acceleration profile obtained by the simulation. FIG. 28 is a graph showing an acceleration profile obtained in another simulation. FIG. 29 is a block diagram showing elements particularly related to gear failure determination in a gear failure detection system according to yet another modified example of the embodiment. FIG. 30 is a diagram for explaining angle estimation performed by the gear failure determination device in the modified example of FIG. 19.

Various embodiments of the present invention will now be described with reference to the accompanying drawings. The drawings are not necessarily drawn to scale, and some features may be exaggerated or omitted.

The gear failure detection system of this embodiment monitors the condition of the gears and detects gear failures. The gear failure detection system of this embodiment is configured to monitor the gears using a self-powered sensor-equipped bearing (bearing device) and to evaluate the vibration signal of each gear tooth. The bearing device used in this embodiment has the functions of self-power generation, sensing, and wireless transmission in addition to the original function of a bearing, which is to rotatably support the shaft. In other words, the bearing device is a device that supports the shaft, and also functions as a sensor device (measurement device). A sensor device can also be called a measurement device, so the sensor-equipped rolling bearing can also be called a measurement device-equipped bearing.

In the following description, the gear transmission mechanism may be referred to as a gear box. In addition, in this embodiment, for ease of explanation, a gear transmission mechanism having a simple one-stage gear mechanism is exemplified as the gear transmission mechanism.
The scope of the present invention is not limited to the following embodiments, and changes and modifications can be made within the scope of the technical idea of the present invention.

Figure 1 shows a gear failure determination device 10 and a gear transmission mechanism 20 according to an embodiment of the present invention. The gear transmission mechanism 20 has a housing 21, and a first gear 22 and a second gear 23 provided within the housing 21. The housing 21 has a first wall 21a and a second wall 21b that face each other at a distance. The first gear 22 and the second gear 23 mesh with each other. The gear transmission mechanism 20 also has a first shaft 24a to which the first gear 22 is attached, and a first bearing 25 and a second bearing 26 that support both ends of the first shaft 24a. The first bearing 25 is provided in the first wall 21a of the housing 21. The second bearing 26 is provided in the second wall 21b of the housing 21.

The gear transmission mechanism 20 further includes a second shaft 24b to which the second gear 23 is attached, and a third bearing 27 and a fourth bearing 28 that support both ends of the second shaft 24b. The third bearing 27 is provided on the first wall 21a of the housing 21 below the first bearing 25. The fourth bearing 28 is provided on the second wall 21b of the housing 21 below the second bearing 26. In this embodiment, the first bearing 25 to the fourth bearing 28 may be rolling bearings, for example, tapered roller bearings or deep groove ball bearings.

In this embodiment, only one of the four bearings 25 to 28 (for example, the first bearing 25) is a bearing device with a built-in measuring device 100 (FIG. 3). The measuring device 100 measures the vibration generated by the meshing of the first gear 22 and the second gear 23. The measuring device 100 also detects the rotational angle position of the first bearing 25 (the rotational angle position of the first shaft 24a). The measuring device 100 has a communication function and can wirelessly communicate with the gear failure determination device 10. Furthermore, the measuring device 100 also has a power generation unit 149 (FIG. 8). The gear failure determination device 10 detects a failure of the first gear 22 based on a signal received from the measuring device 100 of the first bearing 25. In this embodiment, the gear failure detection system 30 is configured by the gear failure determination device 10 and the measuring device 100. The first bearing 25 is equipped with the measuring device 100, so it may be called a bearing device.

Fig. 2 is a block diagram showing the configuration of the gear failure determination device 10. As shown in Fig. 2, the gear failure determination device 10 has an input unit 11, a control unit 12, a storage unit 13, a display unit 14, and a communication unit 15. The gear failure determination device 10 may be, for example, a personal computer.
The input unit 11 includes a button, a switch, a mouse, a touch panel, etc., and a user performs various inputs, etc., via the input unit 11 (for example, inputs the number of teeth of the first gear 22). The user uses the input unit 11 when operating the gear failure determination device 10. The input unit 11 can also be referred to as an operation unit.

The control unit 12 is composed of one or more CPUs (Central Processing Units) or MPUs (Microprocessor Units), and controls the operation of the input unit 11, memory unit 13, display unit 14, and communication unit 15 of the gear failure judgment device 10. The control unit 12 processes signals received from the measurement device 10 and judges whether a gear has a failure (performs the failure judgment process). The control unit 12 controls the gear failure judgment device 10 and executes the gear failure judgment process by executing a control program stored in the memory unit 13.

The storage unit 13 is configured with a hard disk drive (HDD), a read only memory (ROM), a random access memory (RAM), and/or an integrated circuit (IC) memory card, and stores various information such as a control program executed by the control unit 12 and an algorithm for generating a map (described later) showing the relationship between the angle and teeth of the first gear 22. The map is generated by the control unit 12 executing the control program stored in the storage unit 13. The storage unit 13 can store data received from the measuring device 100.
The display unit 14 is configured with a liquid crystal display or the like, and displays various data, numerical values, characters, images, etc. The display unit 14 can display data received from the measuring device 100 on the display unit. Furthermore, under the control of the control unit 12, the display unit 14 can display the processing results of the gear failure determination processing.
The communication unit 15 performs wireless communication with the measuring device 100 .

3 and 4 are exploded perspective views showing the configuration of the first bearing 25. FIG. 3 is an exploded perspective view of the first bearing 25 as viewed from the left side. FIG. 4 is an exploded perspective view of the first bearing 25 as viewed from the right side. As shown in FIG. 3, the first bearing 25 has a measuring device 100 and a bearing unit 250. The measuring device 100 is attached to the left side of the bearing unit 250. The measuring device 100 includes a cover 110, a coil board 120 (FIG. 4), a rotating unit 130, a circuit board group 140 (FIG. 4), a storage battery (capacitor) 150, a retainer (not shown), and a Z-phase magnet unit 160. The retainer is a thin ring-shaped member and is located between the cover 110 and the Z-phase magnet unit 160. The circuit board group 140 is composed of a plurality of boards, as described below.

The rotating part 130 is attached to the first shaft 24a and rotates together with the first shaft 24a. When the cover 110 is attached to the bearing part 250, the cover 110 is attached to the outer ring 250A of the bearing part 250. Therefore, the cover 110 does not rotate. The retainer is attached to the inner ring 250B of the bearing part 250. The Z-phase magnet unit 160 is fixed to the inner ring 250B of the bearing part 250 by the retainer. Therefore, the Z-phase magnet unit 160 rotates together with the inner ring 250B of the bearing part 250 (and thus the first shaft 24a).

The cover 110 is a circular flat plate member, and is made of a magnetic material such as silicon steel plate, carbon steel (JIS SS400 or S45C), martensitic stainless steel (JIS SUS420), or ferritic stainless steel (JIS SUS430). Since a sensor is attached to the cover 110, the cover 110 may be referred to as a sensor case.

As shown in FIG. 4, the coil board 120, the circuit board group 140, and the storage battery 150 are attached to the surface of the cover 110 facing the bearing portion 250. The circuit board group 140 includes a power supply control board 141, an angle sensor board 142, and a control board 143. The power supply control board 141, the angle sensor board 142, and the control board 143 are fixed to the cover 110, for example, by fastening bolts made of a non-magnetic material such as brass into female threaded holes drilled in the cover 110. The bolts have a length that does not protrude from the cover 110 when attached to the cover 110.

The power supply control board 141 has a rectifier circuit 261, a smoothing circuit 262, a power supply management IC 263, and a FET 264 (field effect transistor) as shown in FIG. 6. The circuits 261 to 263 and the FET 264 of the power supply control board 141 will be described later with reference to FIG. 6.

A through hole 111 is formed in the cover 110, and the through hole 111 is sealed with a lid 117 made of a non-magnetic material such as resin. As described below, an antenna 147 ( FIG. 5A ) is mounted on the control board 143. The cover 110 is magnetic, and therefore acts to shield electromagnetic waves from the antenna 147. However, the antenna 147 is disposed in a position opposite the lid 117, and thus the electromagnetic waves from the antenna 147 can reach the external gear fault determination device 10 via the non-magnetic lid 117.
The coil substrate 120 is fixed to the cover 110 by, for example, an adhesive.

5A is a plan view showing an example of the configuration of the cover 110, the coil board 120, the circuit board group 140, and the storage battery 150. FIG. 5B is a view showing only the cover 110 and the coil board 120. As shown in FIG. 5B, the coil board 120 has a flexible board 121, a coil pattern 123 provided on the flexible board 121, and a plurality of yokes 125 provided on the flexible board 121. The yokes 125 may be installed arbitrarily. The shape of the flexible board 121 in a planar view is a perfect circular ring shape centered on the rotation center axis Ax. The coil pattern 123 has a plurality of planar coils stacked in the thickness direction of the flexible board 121. The planar coil is a pattern of a conductor patterned and provided on a predetermined surface of an insulator. In this embodiment, the conductor pattern is formed on multiple surfaces of the insulator. However, this is not limited to this, and the conductor pattern may be formed on one surface of the insulator. The number of turns of the coil pattern 123 is proportional to the number of layers of the planar coil. The amount of power generated can be adjusted by changing the number of layers in the planar coil.

In addition, the coil pattern 123 is arranged so that, in a plan view, the concaves and convexes are arranged alternately in the circumferential direction of a circle centered on the central axis of rotation Ax. The yokes 125 are arranged in the concaves of the concaves and convexes. The coil pattern 123 may be shaped so that a portion of the circle is missing in order to place an angle sensor at a position where it can detect magnetic changes in the encoder magnet, which will be described later.

As shown in FIG. 5A, the power supply control board 141, angle sensor board 142, control board 143, and storage battery 150 are attached radially outward from the coil board 120 in a plan view. The power supply control board 141 (more specifically, the power supply management IC 263 described later) has two DC-DC converters (one for step-down and one for step-up), and steps down the DC voltage supplied from the storage battery (capacitor) 150 and supplies the DC voltage to the angle sensor board 142 and control board 143.

An angle sensor 443 and a Z-phase detector (pulse generating unit) 444 are mounted on the angle sensor board 142. As described later, the angle sensor 443 used in this embodiment is a low-performance acceleration sensor with low resolution and low sampling frequency. The Z-phase detector 444 is, for example, a Hall IC. The Z-phase detector 444 generates one pulse every time the Z-phase magnet 162, which is stationary with respect to the inner ring 250B of the first bearing 25 and rotates together with the inner ring 250B, passes near the Z-phase detector 444. In other words, the Z-phase detector 444 generates one pulse every time the inner ring 250B of the first bearing 25 rotates once (every time the first shaft 24a rotates once).
The control circuit 143 is equipped with a control circuit 145, an antenna 147, an acceleration sensor 441, and a temperature sensor 442. The acceleration sensor 441, the temperature sensor 442, the angle sensor 443, the control circuit 145, and the antenna 147 may be configured as separate IC chips, or a part or all of them may be configured as one IC chip. The temperature sensor 442 is provided to detect the temperature when the acceleration and angle are detected (or before detection). The acceleration sensor 441 used in this embodiment is, for example, a MEMS (Micro Electro Mechanical Systems) acceleration sensor. Furthermore, the acceleration sensor 441 used in this embodiment is configured to measure acceleration values (obtain measured acceleration values) in each of three axes (X/Y/Z axes) as shown in FIG. 5A (i.e., it is a three-axis acceleration sensor). As will be described later, the acceleration sensor 441 used in this embodiment is a low-performance acceleration sensor with low resolution and low sampling frequency.

Both ends of the coil pattern 123 are connected to the power supply control board 141 via extensions 116 provided at predetermined positions on the outer periphery of the coil board 120. Note that instead of the extensions 116, the coil board 120 and the power supply control board 141 may be connected by lead wires. Alternatively, instead of the extensions 116, the coil board 120 and the power supply control board 141 may be connected by an FPC (Flexible Printed Circuit) connector. Connections using FPC connectors do not require soldering, which can increase the productivity of the measuring device 100.

3 and 4, the rotating part 130 has an annular magnetic track 131, an annular base material 133, and an annular mounting jig 135. The base material 133 and the mounting jig 135 are preferably made of a magnetic metal material. The magnetic track 131 is provided on the left surface of the base material 133. The base material 133 has an opening. The mounting jig 135 is fixed to the right surface of the base material 133. The mounting jig 135 protrudes from the right surface of the base material 133 through the opening of the base material 133 to the left surface side of the base material 133. The left surface of the base material 133 is the surface facing the cover 110. The magnetic track 131 may be configured to be detachable from the base material 133.

In this embodiment, the magnetic track 131 and the substrate 133 are collectively referred to as the encoder magnet. For example, the encoder magnet is formed by forming a plastic magnet on one side of the metal substrate 133, and magnetizing the surface of the formed plastic magnet with alternating north and south poles. The mounting jig 135 is a jig for mounting the encoder magnet (rotating portion 130) to the first shaft portion 24a.

The magnetic track 131 has a plurality of magnetic pole pairs 311 each consisting of an N pole 131N and an S pole 131S. The plurality of magnetic pole pairs 311 are arranged in the circumferential direction of the magnetic track 131. The N poles 131N and S poles 131S are arranged alternately.
The distance between the centers of adjacent north poles 131N and south poles 131S on the magnetic track 131 is the same as the distance between the centers of adjacent yokes 125 on the coil substrate 120 (see FIG. 5B).

In this embodiment, when the magnetic track 131 rotates relative to the coil substrate 120 around the central axis of rotation Ax (Figure 5B), when one of the adjacent yokes 125 faces the N pole, the other yoke 125 faces the S pole. Also, when one yoke 125 faces the S pole, the other yoke 125 faces the N pole. In other words, adjacent yokes 125 of the coil substrate 120 do not face the same magnetic pole of the magnetic track 131. As a result, the phase of the change in magnetic flux density passing through one yoke 125 and the phase of the change in magnetic flux density passing through the other yoke 125 are shifted by 180°.

In this way, when the magnetic track 131 rotates relative to the coil substrate 120, the magnetic poles facing the yoke 125 alternate. This causes the magnetic flux density passing through the yoke 125 to change periodically. In response to this periodic change in magnetic flux density, a voltage change (e.g., a sinusoidal AC voltage) is generated in the coil pattern 123 located around the yoke 125. That is, in this embodiment, when the inner ring 250B of the first bearing 25 rotates together with the first shaft 24a, power is generated by electromagnetic induction in the first bearing 25. This power generation is self-generated. The combination of the encoder magnet (magnetic track 131) and the coil substrate 120 is sometimes referred to as the power generation unit 149. The power generation unit 149 generates power based on the relative rotation of the outer ring 250A and the inner ring 250B of the first bearing 25. The DC voltage generated by the power generation unit 149 is stored in the storage battery 150.

The Z-phase magnet unit 160 has a Z-phase magnet holder 161 and a Z-phase magnet 162. The Z-phase magnet holder 161 is an annular member, and the Z-phase magnet 162 is embedded in a hole formed in the Z-phase magnet holder 161. The Z-phase magnet holder 161 is made of metal (e.g., aluminum). When the Z-phase magnet unit 160 rotates once together with the first shaft 24a, the Z-phase magnet 162 crosses the Z-phase detector 444 once. At this time, the Z-phase detector 444 generates one pulse.

6 is a schematic diagram of the self-power generation mechanism described above. In this figure, the coil C corresponds to the coil pattern 123 attached to the cover (metal case) 110, and the magnet M corresponds to the encoder magnet provided on the inner ring 250B of the first bearing 25.
As shown in Fig. 6, the coil C and the magnet M rotate relative to each other, causing electromagnetic induction and generating power. The generated current is AC, so it is converted to DC through a rectifier circuit 261. The rectifier circuit 261 is, for example, a diode bridge. Since the current converted by the rectifier circuit 261 may contain pulsating current, in this embodiment, the output of the rectifier circuit 261 is connected to a smoothing circuit 262 to make the current closer to DC.

The direct current that has passed through the smoothing circuit 262 is input to a power management IC (storage battery control unit, charging completion detection unit) 263 and stored in the storage battery (capacitor) 150. The power management IC 263 has two DC-DC converters (for step-down and step-up), and the minute electric power generated by the generator (coil C and magnet M in FIG. 6 ) is stepped up by the step-up DC-DC converter of the power management IC 263 and stored in the storage battery 150. Charging the storage battery 150, discharging during overcharging, and supplying power at a voltage Vcc to the load L (acceleration sensor 441, angle sensor 443, Z-phase detector 444, and control board 143) are managed by the power management IC 263.
In order to prevent runaway of the control board 143 due to the inflow of an unstable voltage, an FET 264 is provided between the power management IC 263 and the load L. In addition, the power management IC 263 can switch between supplying and cutting off power to the load L by controlling the FET 264.

Since the power obtained by the rotation of the first shaft 24a (relative rotation between the coil substrate 120 and the magnetic track 131) is extremely small, the power management IC 263 is, for example, a power management IC used for energy harvesting that can store even minute amounts of power (for example, the ultra-low power consumption harvester power management IC-BQ25570 manufactured by Texas Instruments).

The storage battery 150 is capable of storing an amount of charge sufficient to perform at least one operation of the measurement device 100. One operation of the measurement device 100 is an operation from measuring the angular position and acceleration of the gear teeth for a predetermined time (e.g., 25 ms) to transmitting the measurement results to the gear fault determination device 10.
When the charging of the storage battery 150 is completed, the power management IC 263 detects the completion of charging of the storage battery 150 and sends a charging completion signal notifying the FET 264 of the completion of charging. Then, the storage battery 150 resumes supplying power to the load L.
Since the lower the ESR (Equivalent Series Resistance) of the storage battery 150, the more current can be instantly extracted, it is preferable to use a storage battery with a low ESR. If a storage battery with a low ESR is not used, a capacitor or an all-solid-state battery that can be connected in parallel to the storage battery 150 may be provided. In this case, it is preferable to connect them in parallel.
The rectifier circuit 261, the smoothing circuit 262, the power management IC 263, and the FET 264 may be collectively referred to as a circuit group 431 (FIG. 8).

Figure 7 shows two graphs that explain the operation of the power management IC 263. The vertical axis of the upper graph in Figure 7 indicates the voltage of the storage battery 150, and the horizontal axis indicates time. The vertical axis of the lower graph indicates the charging completion signal. The vertical axis of this graph indicates the signal strength (High or Low), and the horizontal axis indicates time. When the charging completion signal is Low, the charging of the storage battery 150 is not completed (charging). When the charging completion signal is High, the storage battery 150 is supplying power. When the charging completion signal transitions from High to Low, it indicates that the storage battery 150 has stopped (ended) the power supply and is in a charging state. The charging completion signal is a signal generated by the power management IC 263, and by combining the two graphs in Figure 7, it can be seen how the power management IC 263 manages (controls) the voltage of the storage battery 150.

7, when the charge completion signal is Low, the charge amount of the storage battery 150 increases over time, and the charge completion signal becomes High at time t1. The voltage of the storage battery 150 at time t1 is V2. The voltage V2 is a threshold value for resuming power supply to the load L.
After time t1 (after power supply to the load L starts), if the power consumption of the load L is greater than the power generation amount, the voltage drops from V2 (solid line J1). When the voltage drops and reaches V1, the power management IC 263 stops power supply to the load L. In other words, V1 is a threshold for stopping power supply. During the power supply stop, the power management IC 263 charges the storage battery 150.
V1 has a value equal to or slightly larger than Vcc, because if V1 is smaller than Vcc, the step-up DC-DC converter of the power management IC 263 will not function.

After time t1, when the power consumption of the load L is equal to the power generation, the voltage remains at V2 (dashed double-dashed line J2).
After time t1, when the power consumption of the load L is smaller than the power generation, the voltage rises from V2 (dotted line J3). When the power rises and reaches V3, the power management IC 263 stops charging the storage battery 150. This is to prevent overcharging.

In this embodiment, the storage battery 150 can store the amount of charge required for one operation of the measurement device 100. Under normal conditions, the voltage of the storage battery 150 changes as shown by the solid line J1 in FIG. 7. The measurement device 100 can perform one operation between time t1 and time t2, when the voltage drops from V2 to V1. The period from time t1 to time t2 is the period during which the storage battery 150 can supply power to the load L (the angle sensor board 142, the control board 143, and the acceleration sensor 441, the angle sensor 443, the Z-phase detector 444, the antenna 147, etc. thereon), and the measurement of acceleration and angle and the transfer of the measured data are performed between time t1 and time t2.

8 is a block diagram showing the configuration of the measurement device 100. As described above, the measurement device 100 includes the power supply control board 141, the angle sensor board 142, and the control board 143. The angle sensor board 142 includes a sensor board 170 and a Z-phase detector board 171.
The power supply control board 141 includes a circuit group 431. The circuit group 431 includes the rectifier circuit 261, the smoothing circuit 262, the power supply management IC 263, and the FET 264 shown in FIG.

The power generating unit 149 includes a magnetic track 131 (FIG. 4) and a coil substrate 120 (FIG. 5B). The power generating unit 149 generates electricity based on the relative rotation of the outer ring 250A and the inner ring 250B of the bearing unit 250, and supplies the electricity to the angle sensor substrate 142 and the like.
The power generation unit 149 generates single-phase AC power and outputs it to the circuit group 431. The rectifier circuit 261 of the circuit group 431 full-wave rectifies the single-phase AC power generated by the power generation unit 149 and converts it into DC power. The DC power output from the rectifier circuit 261 is smoothed by the smoothing circuit 262 to be stable power. This power is then stored in the storage battery 150 via the power management IC 263 (boosted by the step-up DC-DC converter of the power management IC 263). The DC power stored in the storage battery 150 is supplied to the angle sensor board 142 and the control board 143 via the power management IC 263 (stepped down by the step-down DC-DC converter of the power management IC 263) at the timing of measurement.

An angle sensor 443 is mounted on the sensor board 170. A Z-phase detector 444 is mounted on the Z-phase detector board 171. An acceleration sensor 441, a temperature sensor 442, a microcomputer (hereinafter referred to as "microcomputer") 451, an external memory 452, and a wireless module (transmitter) 453 are mounted on the control board 143. The microcomputer 451, the external memory 452, and the wireless module 453 are included in the control circuit 145 (Figure 5).

The acceleration sensor 441, temperature sensor 442, and angle sensor 443 use DC power supplied from the power supply control board 141 to detect acceleration, temperature, and rotation angle (position), respectively.
For example, the angle sensor 443 is mounted on the sensor board 170 so as to be located to the side of the magnetic track 131. The rotating part 130 having the magnetic track 131 is fixed to the inner ring 250B of the bearing part 250, and the angle sensor 443 detects the rotation angle of the inner ring 250B relative to the outer ring 250A of the bearing part 250 by detecting the magnetic flux density that changes due to the rotation of the magnetic track 131 together with the inner ring 250B of the bearing part 250. The angle sensor 443 is a magnetic rotary encoder. The angle sensor 443 may have one reader that reads the rotation of the magnetic track 131, or may have multiple readers that each read the rotation of the magnetic track 131. In this embodiment, the angle sensor 443 has 25 readers arranged at equal angular intervals around the axis of the shaft 24a, and each reader reads the rotation of the magnetic track 131. Thus, the reader outputs a similar reading value to the reading value of the adjacent reader at a distance of 14.4°.
The angle sensor 443 may be of an absolute type, but in this embodiment, an incremental type and a pseudo-absolute system using a Z-phase detector 444 are used.

The microcomputer 451 includes a CPU 455, a DMA (Direct Memory Access) controller 456, and an internal memory 457. The microcomputer 451 writes the measured values obtained from the acceleration sensor 441 and the angle sensor 443 into the internal memory 457 and the external memory 452. The DMA controller is abbreviated as DMAC in the drawings.
The external memory 452 is used as an auxiliary memory when the storage capacity of the internal memory 457 is insufficient for the data of the measurement values obtained from the acceleration sensor 441 and the angle sensor 443. Therefore, the external memory 452 is not absolutely necessary. Hereinafter, the internal memory 457 and the external memory 452 may be referred to as the storage unit 230.

In addition, if the acceleration sensor 441 has a large-capacity FIFO (first in, first out) buffer and the angle sensor 443 has a large-capacity FIFO buffer, the DMA controller 456, the internal memory 457, and the external memory 452 may be omitted. In this case, the measurement value data of the acceleration sensor 441 can be stored in the FIFO buffer of the acceleration sensor 441, and the measurement value data of the angle sensor 443 can be stored in the FIFO buffer of the angle sensor 443, and a clock and DMA controller 456 for transferring data to the internal memory 457 and the external memory 452 are not required, and power consumption can be reduced.

The CPU 455 initializes the storage unit 230, the acceleration sensor 441, the temperature sensor 442, and the angle sensor 443, and initializes the DMA controller 456. When a DMA trigger (an INT signal from the acceleration sensor 441) is input, the DMA controller 456 starts DMA transfer. Specifically, when a DMA trigger is input, the DMA controller 456 transfers the latest measurement values (a plurality of discrete instantaneous values described later) held by the acceleration sensor 441 and the angle sensor 443 as unconverted data (raw data) to the storage unit 230 (the internal memory 457 and the external memory 452 as necessary) without going through the CPU 455.
The acceleration sensor 441 and the angle sensor 443 update and hold the measurement values (multiple discrete instantaneous values) at a predetermined update period. When the acceleration sensor 441 is a MEMS acceleration sensor, the update period is the ODR (Output Data Rate) of the MEMS acceleration sensor. The acceleration sensor 441 generates an INT signal (interrupt signal) every time data output becomes possible, and the INT signal is input to the DMA controller 456 as a trigger. The update period (sampling period) will be described later with reference to FIG. 10.

The wireless module 453 transmits the data stored in the storage unit 230 to the gear failure determination device 10 under the control of the CPU 455. The wireless module 453 includes an antenna 147 (FIG. 5A). For example, the wireless module 453 transmits the data to the gear failure determination device 10 by wireless communication such as BLE. BLE is an abbreviation for Bluetooth (registered trademark) Low Energy. The transmitted data is received by the communication unit 15 of the gear failure determination device 10 and processed by the control unit 12 of the gear failure determination device 10.
When BLE is used, data is transmitted, for example, one packet at a time. If it is desired to increase the throughput, More data or Data Length Extension may be used. Alternatively, the throughput can be increased by adopting PHY 2 Mbps defined in BLE 5.x.
When the first bearing 25 and the gear failure determination device 10 perform wireless communication, the wireless communication may be in accordance with a communication standard other than BLE. For example, Zigbee (registered trademark) or Thread may be adopted. Alternatively, a frequency band other than BLE (for example, a 920 MHz band specific low power radio) may be used.

The power supply management IC 263 in the circuit group 431 cuts off the power supply to the load L every time the acceleration sensor 441 finishes measuring the acceleration for a predetermined time and the angle sensor 443 finishes measuring the angle for a predetermined time. Preferably, the power supply management IC 263 cuts off the power supply to the load L (the acceleration sensor 441, the angle sensor 443, the Z-phase detector 444, and the control board 143) after the wireless module 453 has communicated via BLE wireless communication the instantaneous value measured by the acceleration sensor 441 and the instantaneous value measured by the angle sensor 443.
Furthermore, when the power management IC 263 detects that charging to the storage battery 150 is completed (when the voltage of the storage battery 150 exceeds the threshold V2), it resumes the power supply to the load L.

Fig. 9 is a timing chart for explaining the operation of the measurement device 100. Fig. 9(A) shows a period during which power is supplied by the power management IC 263, and Fig. 9(B) shows a period during which the measurement device 100 is connected to the gear failure determination device 10 via a wireless network (BLE). Fig. 9(C) shows a pulse output by the Z-phase detector 444, and Fig. 9(D) shows a measured acceleration value output by the acceleration sensor 441 and stored in the memory unit 230 from the acceleration sensor 441. Fig. 9(E) shows a measured angle value output by the angle sensor 443 and stored in the memory unit 230 from the angle sensor 443.
When the power is turned on, the power management IC 263 starts supplying power from the storage battery 150 to the microcomputer 451 and the antenna 147 of the wireless module 453. Then, the CPU 455 of the measurement device 100 controls the wireless module 453 to exchange information required for wireless communication with the communication unit 15 of the gear failure determination device 10 and establish a connection. That is, the measurement device 100 pairs with the gear failure determination device 10. The "pairing time" is the time required for the measurement device 100 to establish communication (connection) with the gear failure determination device 10. The pairing time is, for example, 5 seconds. The "pairing completion" indicates the time when the BLE connection is established. After the pairing is completed, the power management IC 263 also starts supplying power from the storage battery 150 to the acceleration sensor 441, the angle sensor 443, and the Z-phase detector 444.

In this embodiment, the output of the Z-phase detector 444 is active low (Low), and when the Z-phase detector 444 detects the Z-phase magnet 162, it outputs a Low signal. However, in FIG. 9 (C), the pulse output when the Z-phase detector 444 detects the Z-phase magnet 162 is shown at a High level.
9, the "sensor preparation time" is the time to wait until the Z-phase detector 444, the acceleration sensor 441, and the angle sensor 443 can measure accurately after the start of power supply to the acceleration sensor 441, the angle sensor 443, and the Z-phase detector 444, and is, for example, 50 ms. The "processing waiting time" is set to a time equal to or longer than the time required for the CPU 455 to finish a predetermined monitoring process. The processing waiting time is, for example, 30 seconds.
After pairing is completed, when the Z-phase detector 444 detects the Z-phase magnet 162 for the first time after the sensor preparation period and the processing wait time are ended, the CPU 455 supplies a measurement start signal to the acceleration sensor 441 and the angle sensor 443. After the measurement start signal is supplied, the acceleration sensor 441 and the angle sensor 443 are initialized (acceleration sensor/angle sensor initialization time). The initialization of the acceleration sensor 441 and the angle sensor 443 means that the acceleration sensor 441 and the angle sensor 443 are supplied with power so that the angle sensor 443 can operate correctly. When the initialization of the acceleration sensor 441 and the angle sensor 443 is completed, the acceleration sensor 441 and the angle sensor 443 start measurement after a measurement start delay (for example, about 480 ms). The acceleration sensor/angle sensor initialization time is, for example, 100 ms.

As shown in FIG. 9D, after a delay in starting measurement, the acceleration sensor 441 measures the acceleration for each of the three axes (X/Y/Z axes) for a predetermined measurement time. The measurement value of the acceleration sensor 441 represents the vibration generated in the first gear 22 and the bearing part 250 due to the meshing of the first gear 22 and the second gear 23. Note that FIG. 9D shows the acceleration value measured by the acceleration sensor 441 for any one of the three axes. The predetermined measurement time is, for example, 25 ms, which is a time equivalent to less than one rotation of the shaft 24a. During the predetermined measurement time, the acceleration sensor 441 repeats measurement at a sampling period, so the measurement result is a plurality of discrete instantaneous values. In this way, the acceleration sensor 441 measures the instantaneous value of the acceleration caused by the vibration of the bearing part 250 for a certain measurement time from the measurement start point after the Z-phase detector 444 generates a pulse.
During the measurement start delay, the CPU 455 performs three processes (1) to (3).
In process (1), the CPU 455 acquires temperature data from the temperature sensor 442 before the start of measurement. The CPU 455 can determine whether abnormal heat generation is occurring inside the first bearing 25 based on the temperature measured by the temperature sensor 442. When the temperature measured by the temperature sensor 442 is equal to or higher than a predetermined value, the CPU 455 transmits a signal indicating abnormal heat generation to the gear failure determination device 10. The gear failure determination device 10, which receives the signal, displays a message indicating abnormal heat generation on the display unit 14. The CPU 455 can also correct the angle measured by the angle sensor 443 based on the temperature measured by the temperature sensor 442. Since the angle sensor 443 is a magnetic sensor, the influence of the magnetism due to the temperature change may appear as a measurement error of the angle sensor 443. Regarding this angle error, it is confirmed in advance what kind of influence/error appears in the measurement value of the angle sensor 443 for each temperature, and a correction table is created for performing correction to cancel the influence/error. Then, the CPU 455 corrects the angle data based on the correction table.
In process (2), the CPU 455 programs the DMA controller 456. The programming of the DMA controller 456 is to program the DMA controller 456 so as to acquire measurement data of the acceleration sensor 441 and the angle sensor 443 using the INT signal from the acceleration sensor 441 as a trigger, and to store the measurement data in the storage unit 230.
In process (3), the CPU 455 starts a wake-up timer for determining the end of measurement (timer out).

The measurement start delay is, for example, about 480 ms. After the above three processes are performed, the CPU 455 goes into a sleep state. Acceleration and angle measurements are performed while the CPU 455 is in a sleep state by activating only the minimum necessary functional blocks such as the DMA controller 456, acceleration sensor 441, and acceleration sensor 443. Acceleration and angle measurement ends when the wake-up timer expires. When acceleration and angle measurement ends, the CPU 455 is released from sleep and returns to an operating state. In this way, only the minimum necessary functional blocks operate, and a measurement system with low power consumption can be realized.

Also, as shown in FIG. 9E, after the measurement start delay, the angle sensor 443 measures the angular position of the first shaft 24a for a predetermined measurement time. The predetermined measurement time is, for example, 25 ms, which is a time equivalent to less than one rotation of the shaft 24a. During the predetermined measurement time, the angle sensor 443 repeats measurement at a sampling period, so the measurement result is a plurality of discrete instantaneous values. In this way, the angle sensor 443 measures the instantaneous value of the angle of the shaft 24a for a certain measurement time from the measurement start point after the Z-phase detector 444 generates a pulse. In this embodiment, the 25 readers of the angle sensor 443 simultaneously measure the angular position of the shaft 24a for a predetermined measurement time and at a sampling period.
The waveform of the angle sensor 443 is stored in the memory unit 230 with a delay of "DMA delay" compared to the waveform of the acceleration sensor 441. Theoretically, the measurement values of the acceleration sensor 441 and the angle sensor 443 are DMA transferred (sampled) using the INT signal from the acceleration sensor 441 as a common timing source (trigger). However, in reality, there is a delay of nanoseconds between the transfer of the measurement values of the acceleration sensor 441 and the measurement values of the angle sensor 443, and this delay is called "DMA delay."
After completing the measurement of the acceleration and angle, data (a pair of instantaneous values of acceleration measured by the acceleration sensor and instantaneous values of angle measured by the angle sensor) is transferred from the measuring device 100 to the gear fault determination device 10 during a "data transfer time". The data transfer time is, for example, one second. The period from when the power is turned on to the end of the data transfer time constitutes one operation of the measuring device 100.

When the data transfer is complete, the power management IC 263 stops supplying power to the load L and starts charging the storage battery 150. When charging of the storage battery 150 is complete, the power management IC 263 sends a charge completion signal to the FET 264 to notify the completion of charging, and it becomes possible to supply power to the load L again (the angle sensor board 142, the control board 143, and the acceleration sensor 441, the angle sensor 443, the Z-phase detector 444, the antenna 147, etc. on them).

After that, when the Z-phase detector 444 generates a pulse for detecting the Z-phase magnet 162 for the first time after the pairing time, the sensor preparation time, and the processing wait time are over, the CPU 455 supplies a measurement start signal to the acceleration sensor 441 and the angle sensor 443. However, the time when the previous measurement start signal was supplied was immediately after the generation of the Z-phase detection pulse (detection of the Z-phase magnet 162 by the Z-phase detector 444), whereas the time when the current measurement start signal is supplied is shifted by a certain time after the generation of the Z-phase detection pulse. This is because the measurement start time is changed every time the acceleration sensor 441 measures the acceleration for a predetermined time and the angle sensor 443 measures the angle. Therefore, the acceleration sensor 441 and the angle sensor 443 measure instantaneous values corresponding to different mechanical angles (actual angles around the shaft 24a) every time a measurement is performed for a predetermined time. In this way, it is possible to prevent the measurement start time from overlapping.
The CPU 455 (measurement start time changing unit) determines the time point at which the next measurement start signal is supplied for the Z-phase detection pulse. The supply time point of the measurement start signal for the Z-phase detection pulse can be determined to be delayed from the previous supply time point, for example.

After the measurement start signal, when the initialization of the acceleration sensor 441 and the angle sensor 443 is completed, the acceleration sensor 441 and the angle sensor 443 start measuring again after a measurement start delay. After the measurement of the acceleration and angle is completed, the data is transferred from the measurement device 100 to the gear failure determination device 10. When the data transfer is completed, the power management IC 263 stops the supply of power to the load L and starts charging the storage battery 150. When the charging of the storage battery 150 is completed, the power management IC 263 sends a charge completion signal to the FET 264 to notify the completion of charging, and the supply of power to the load L becomes possible again. After this, the pairing time, the sensor preparation period, and the processing wait time are completed, and after the Z-phase detection pulse is generated, the CPU 455 supplies a measurement start signal to the acceleration sensor 441 and the angle sensor 443. This operation is repeated thereafter. Therefore, each time the acceleration sensor 441 finishes measuring the instantaneous value of acceleration for a predetermined time (a time equivalent to less than one rotation of the shaft 24a) and the angle sensor 443 finishes measuring the instantaneous value of angle for a predetermined time (a time equivalent to less than one rotation of the shaft 24a), the power management IC 263 stops the power supply to the acceleration sensor 441, angle sensor 443, etc. from the storage battery 150, and charges the electricity generated by the power generation unit 149 to the storage battery 150.

Figure 10 shows the update period of the measurement value of the acceleration sensor 441 and the measurement value of the angle sensor 443. In this embodiment, as shown in Figure 10, the update completion timing of the measurement value of the acceleration sensor 441 and the update completion timing (sampling timing) of the measurement value of the angle sensor 443 are set to be synchronized. In other words, the update period (sampling period) of the acceleration sensor 441 is equal to the update period (sampling period) of the angle sensor 443. Then, the measurement value of the acceleration sensor 441 and the measurement value of the angle sensor 443 are stored in the memory unit 230 in a synchronized form (sharing the same time base or sharing at least one time base). In other words, the memory unit 230 associates and stores the instantaneous value of acceleration and the instantaneous value of angle measured simultaneously by the acceleration sensor 441 and the angle sensor 443. In this embodiment, the 25 readers of the angle sensor 443 simultaneously measure the angular position of the axis 24a for a predetermined measurement time. Therefore, the memory unit 230 of the measurement device 100 and the memory unit 13 of the gear failure determination device 10 store the instantaneous acceleration and angle values measured simultaneously by the acceleration sensor 441 and the 25 readers in association with each other. In this manner, in this embodiment, angle and acceleration measurements are performed in synchronization with time. The sampling frequency (ODR) of the acceleration sensor 441 and the angle sensor 443 is selected (set) in advance based on, for example, the gear meshing frequency.

11 is a flowchart for explaining the operation of the measuring apparatus 100. This operation starts when the measuring apparatus 100 is first powered on (first power on in FIG. 9(A)).
In S1, the CPU 455 judges whether or not the pairing time (for example, 5 seconds) has elapsed. If the judgement result is Yes, the operation proceeds to S2. If the judgement result is No, S1 is repeated.
In S2, the CPU 455 determines whether or not a sensor preparation period (for example, 50 ms) has elapsed. If the determination result is Yes, the operation proceeds to S3. If the determination result is No, S2 is repeated.

In S3, the CPU 455 judges whether or not a processing waiting time (for example, 30 seconds) has elapsed. If the judgment result is Yes, the operation proceeds to S4. If the judgment result is No, S3 is repeated.
In S4, the CPU 455 determines whether or not the Z-phase detector 444 has detected the Z-phase magnet 162. If the determination result is Yes, the operation proceeds to S5. If the determination result is No, S4 is repeated.
In S5, the CPU 455 judges whether or not it is the time when the measurement start signal is supplied. If the judgement result is Yes, the operation proceeds to S6. If the judgement result is No, S5 is repeated. Since the initial value of the time when the measurement start signal is supplied is zero, in the case of the first measurement, the judgement result of S5 becomes Yes immediately after S4, and the operation proceeds to S6.

In S6, the CPU 455 supplies a measurement start signal to the acceleration sensor 441 and the angle sensor 443. When the measurement start signal is supplied, as shown in FIG. 9, the acceleration sensor 441 and the angle sensor 443 are initialized (initialization time is, for example, 100 ms), and after a measurement start delay (about 480 ms), the acceleration sensor 441 and the angle sensor 443 perform measurement for a predetermined measurement time (for example, 25 ms), and the DMA controller 456 stores the instantaneous value of acceleration measured by the acceleration sensor 441 and the instantaneous value of angle measured by the angle sensor 443 in a synchronized form in the memory unit 230. As described above, during the measurement start delay, the CPU 455 performs the three processes (1) to (3) and goes into a sleep state.

Next, in S7, the CPU 455 judges whether it is time to transmit the measurement value. If the judgment result is Yes, the operation proceeds to S8. If the judgment result is No, S7 is repeated. The judgment result of S7 being Yes means that the CPU 455 will return from the sleep state to the operating state due to the expiration of the wake-up timer.
In S8, the CPU 455 transmits (transfers) the instantaneous values measured by the acceleration sensor 441 and the instantaneous values measured by the angle sensor 443 stored in the memory unit 230 from the measuring device 100 to the gear failure determination device 10 via BLE wireless communication.
In this manner, one measurement by the measuring device 100 is performed through steps S2 to S8.

Next, in S9, the CPU 455 instructs the power management IC 263 to stop power supply. As a result, the power management IC 263 stops power supply from the storage battery 150 to the acceleration sensor 441, the angle sensor 443, etc., and charges the storage battery 150 with electricity generated by the power generation unit 149.
As described above, when charging of the storage battery 150 is completed, the power management IC 263 sends a charge completion signal to the FET 264 to notify the FET 264 of the completion of charging, and resumes power supply from the storage battery 150 to the acceleration sensor 441, the angle sensor 443, etc. When charging of the storage battery 150 is completed, the power management IC 263 also sends a charge completion signal to the CPU 455.
In S10, the CPU 455 determines whether or not a charging completion signal has been received from the power management IC 263. If the determination result is Yes, the operation proceeds to S11. If the determination result is No, S8 is repeated.
In S11, the CPU 455 determines whether or not to execute the next measurement. In other words, it determines whether or not the number of measurements executed has reached a predetermined number. If the determination result is Yes, the process proceeds to S12. If the determination result is No, the operation ends.
In S12, the CPU 455 determines the time when the next measurement start signal is to be supplied (measurement start command time) for the Z-phase detection pulse based on the time when the measurement start signal was supplied (measurement start command time) immediately before the power supply was stopped. The next measurement start command time can be determined, for example, by adding or subtracting a predetermined time from the immediately previous measurement start command time. In this way, the CPU 455 changes the measurement start time every time it measures the acceleration and angle for a predetermined time.
After this, the operation returns to S1.

Next, the principle of gear failure determination in this embodiment will be described.
12 shows the change over time in acceleration and angle measured by the measurement device 100. Specifically, the upper graph shows the change in acceleration about one axis output by the three-axis acceleration sensor 441. The change in acceleration about the other two axes is not shown. The center graph shows the change in angle (electrical angle) output by one reader of the angle sensor 443, and the lower graph shows the change in angle output by another reader of the angle sensor 443 (the reader next to the reader related to the center graph). The change in angle output by the other 23 readers is not shown.
As described above, the acceleration sensor 441 repeats measurements at sampling intervals during a predetermined measurement time, so that the measurement results are a plurality of discrete instantaneous values. Also, each reader of the angle sensor 443 repeats measurements at sampling intervals during a predetermined measurement time, so that the measurement results are a plurality of discrete instantaneous values.
As described above, the measurement start time changes each time the acceleration and angle are measured for a predetermined time (a time equivalent to less than one rotation of the shaft 24a), so there is a measurement start time difference between the first and second measurements with respect to the Z-phase detection time. The same applies to the subsequent measurements.
As noted above, the 25 readers are equiangularly spaced about the axis of shaft 24a so that the readers output similar readings 14.4° apart from the readings of adjacent readers.

Figure 13 shows elements of the gear failure detection system 30 that are particularly related to gear failure judgment. As described above, in the measurement device 100, the DMA controller 456 stores in the memory unit 230 the instantaneous value of acceleration measured by the acceleration sensor 441 and the instantaneous value of angle measured by the angle sensor 443 in a synchronized form. The instantaneous value measured by the acceleration sensor 441 and the instantaneous value measured by the angle sensor 443 stored in the memory unit 230 are wirelessly transmitted by BLE from the wireless module 453 and received by the communication unit 15 of the gear failure judgment device 10. Since the BLE wireless transmission is repeated periodically, the communication unit 15 of the gear failure judgment device 10 periodically receives the instantaneous value of acceleration and the instantaneous value of angle.

The control unit 12 of the gear failure determination device 10 accumulates in the memory unit 13 the instantaneous values of acceleration measured by the acceleration sensor 441 and the instantaneous values of the angle measured by the angle sensor 443. In one accumulation, the instantaneous values of acceleration and angle for a predetermined time (a time equivalent to less than one rotation of the shaft 24a) are stored in the memory unit 13, and accumulation is performed periodically while the shaft 24a rotates multiple times. Therefore, the memory unit 13 stores the instantaneous values of acceleration measured by the acceleration sensor 441 and the instantaneous values of the angle measured by the angle sensor 443 while the shaft 24a rotates multiple times.
The control unit (acceleration profile generating unit) 12 reads out multiple instantaneous values of acceleration stored in the memory unit 13, rearranges the multiple instantaneous values of acceleration measured by the acceleration sensor 441 while the axis 24a rotates multiple times, and connects together the rearranged instantaneous values of acceleration to generate an acceleration profile that represents the change in acceleration per rotation of the axis 24a.
Furthermore, the control unit (angle profile generating unit) 12 reads out a plurality of instantaneous values of the angle stored in the memory unit 13, rearranges the plurality of instantaneous values of the angle measured by the angle sensor 443 while the shaft 24a rotates a plurality of times, and connects the rearranged instantaneous values of the angle to generate an angle profile that represents the change in the measured angle value per one rotation of the shaft 24a. As described above, the acceleration sensor 441 and the angle sensor 443 have the same sampling period, and the memory unit 230 associates and stores the instantaneous values of the acceleration and the instantaneous values of the angle measured simultaneously by the acceleration sensor 441 and the angle sensor 443, so that the generated angle profile is synchronized with the acceleration profile.

14 shows a graph of an acceleration profile and a graph of an angle profile generated by the control unit 12 of the gear fault determination device 10 which receives a plurality of discrete instantaneous values of acceleration and a plurality of discrete instantaneous values of angle from the measurement device 100. The gear 22 to be measured has 22 teeth, and the tooth numbers are depicted in FIG. 14 for reference.
The horizontal axis of the acceleration profile graph and the angle profile graph is time, which can also be thought of as mechanical angle (the actual angle around axis 24a).
The acceleration profile graph shows a change in acceleration about one axis measured by the triaxial acceleration sensor 441. The control unit 12 generates an acceleration profile about the three axes of the triaxial acceleration sensor 441, but graphs of the acceleration profiles about the other two axes are not shown.
12, the measurement results of the acceleration sensor 441 are a plurality of discrete instantaneous values. The measurement results of the acceleration sensor 441 are periodically received by the gear failure determination device 10 via BLE wireless transmission. Therefore, the control unit 12 rearranges the measurement results of the acceleration sensor 441 while the shaft 24a rotates multiple times, based on the time point at which the measurement results are received, and connects the rearranged instantaneous values of acceleration together, thereby generating an acceleration profile per one rotation (360°) of the shaft 24a.

The angle profile graph of FIG. 14 shows the change in angle (electrical angle) measured by the 25 readers of the angle sensor 443.
12, the measurement results of the angle sensor 443 are a plurality of discrete instantaneous values. The measurement results of the angle sensor 443 are periodically received by the gear failure determination device 10 via BLE wireless transmission. Therefore, the control unit 12 rearranges the measurement results of the angle sensor 443 while the shaft 24a rotates multiple times, based on the time point at which the measurement results are received, and connects the rearranged instantaneous values of the angles to generate an angle profile per one rotation (360°) of the shaft 24a.
Since the angle sensor 443 has 25 readers arranged at equal angular intervals around the axis of the shaft 24a, the angle profile generated from the measurement results of the 25 readers has 25 approximately linear diagonal segments during one rotation of the shaft 24a (angle displacement of 360° mechanical angle), with each segment having a displacement of an electrical angle from 0° to 360° composed of instantaneous values of the electrical angle (absolute angle) measured by the 25 readers.

15 shows graphs of an acceleration profile and an angle profile obtained by continuous measurement over one rotation (360°) of the shaft 24a using an expensive acceleration sensor and angle sensor with high resolution and high sampling frequency, which are not an embodiment of the present invention. The gear 22 to be measured has 22 teeth, and the tooth numbers are depicted in FIG. 15 for reference. The angle sensor has one reader.
As is clear from a comparison between Figures 14 and 15, the acceleration profile and angle profile obtained by repeated measurements using the low-performance acceleration sensor and angle sensor with low resolution and low sampling frequency in this embodiment have accuracy that is comparable to that of the acceleration profile and angle profile obtained by measurement using an expensive acceleration sensor and angle sensor with high resolution and high sampling frequency.

The control unit (fault determination unit) 12 of the gear fault determination device 10 determines a fault in the gear based on the acceleration profile and the angle profile.
Next, the fault determination process executed by the control unit 12 of the gear fault determination device 10 will be described with reference to Fig. 16. In this embodiment, a method called the Per-tooth method (Per-tooth evaluation method) is used to detect a fault for each tooth of a gear.
First, the control unit 12 converts the measured angle (electrical angle) of the angle profile ( FIG. 14 ) into a mechanical angle φ(t) by, but not limited to, dividing the accumulated value of the electrical angle differences by 360 and finding the remainder (process 490).
The control unit 12 associates the measured acceleration value a(t) of the acceleration profile, the mechanical angle φ(t) obtained in process 490, and the time t of the acceleration profile and the angle profile, and stores them in the storage unit 13 (process 500). In the figure, only the acceleration value a(t) for one axis is shown, and the acceleration values for the other two axes are omitted.

The user of the gear failure determination device 10 inputs the number of gear teeth using the input unit 11 (process 510). Although not required, a tooth angle offset value may also be input. The tooth angle offset value is a calibration value for indicating (understanding) at which absolute angle position a specific known tooth is located. The angle offset value will be described later with reference to Figures 17 and 18.

The control unit (map generating unit) 12 creates a map of angles and teeth (a mapping function of angles and teeth) (process 520).
Figures 17 and 18 are diagrams showing examples of angle vs. tooth maps. The vertical axis of the map in Figure 17 shows 22 teeth, and the horizontal axis shows shaft angle position (relative value). Figure 17 shows a map created without using an angle offset value. If an angle offset value is not used, the tooth at shaft angle = 0 degrees when the map is created would be tooth number 1.
The vertical axis of the map in Fig. 18 indicates the 22 teeth, and the horizontal axis indicates the shaft angle position (absolute value). That is, Fig. 18 is a map created using angle offset values. When creating the map in Fig. 18, the gear teeth are numbered (or marked) in advance (tooth numbers 1 to 22 in the example in Fig. 18). The tooth at the position where shaft angle = 0 degrees when the map is created is tooth number 20 in the example in Fig. 18. That is, tooth number 1 starts meshing at 44.6 degrees (absolute angle), so when using the map in Fig. 18, the user will input 44.6 degrees as the angle offset value into input unit 11 in process 510.

As shown in Fig. 17 and Fig. 18, one revolution of the gear is 360 degrees, so 22 line segments S are created by dividing 360 degrees by 22. Fig. 17 and Fig. 18 can be said to represent a mapping function, because when the shaft angle position is determined, one of the 22 teeth is determined. In other words, the mapping function is generated by process 520. The control unit 12 uses the map in Fig. 17 or the map in Fig. 18. When the map in Fig. 17 is used, it is not possible to know which tooth of the gear is damaged, but it is possible to know that the gear contains a damaged tooth. When the map in Fig. 18 is used, it is possible to know which tooth of the gear is damaged. In the following description, it is assumed that the map in Fig. 18 is used.

After generating the map (mapping function) of angles and teeth, the control unit 12 calculates the average acceleration signal strength for each tooth (process 530). The average acceleration signal strength is hereinafter referred to as acceleration signal strength. The numerical values and information used as inputs in calculating the acceleration signal strength are the mechanical angle φ(t) and the measured acceleration value a(t) stored in the storage unit 13 in process 500, and the map in FIG. 18. The acceleration signal strength of the kth tooth is calculated using equation (1) according to the Per-tooth method.

Here, P _(tooth=k) is the acceleration signal power of the k-th tooth, and s(φ(i)) is the measured acceleration value at a certain sample angular position φ(i). tooth(φ(i)) is the value of the mapping function at the sample angular position φ(i). If the sample angular position φ(i) does not correspond to the k-th tooth, tooth(φ(i)) is not equal to k, so s(φ(i)) is 0. If the sample angular position φ(i) corresponds to the k-th tooth, tooth(φ(i)) is equal to k, so s(φ(i)) is the measured acceleration value (a(i)). _{n samples,k} indicates the number of samples of the measured acceleration value that meets the condition (*) (the number of samples of the measured acceleration value corresponding to the k-th tooth).

According to equation (1), all the measured acceleration values (sample values) corresponding to one tooth are squared and summed, and then divided by the number of measured acceleration value samples corresponding to one tooth to obtain the acceleration signal strength per tooth. That is, the acceleration signal strength of the kth tooth can be obtained. This calculation is performed for each tooth of the gear, so that a list of acceleration signal strength values P of 22 teeth can be created. That is, the acceleration signal strengths of the 22 teeth can be obtained by process 530.
Reference number 540 in FIG. 16 indicates the values used in equation (1) for each tooth.

In this embodiment, as described above, the acceleration sensor 441 is configured to obtain measured acceleration values for each of the three axes (X/Y/Z axes). The measured acceleration values obtained for the X-axis, Y-axis, and Z-axis are defined as x(φ(i)), y(φ(i)), and z(φ(i)), respectively. When the kth tooth (tooth(φ(i))=k), s(φ(i)) in equation (1) is defined as the synthesis (vector synthesis) of these measured acceleration values. In this case, s(φ(i)) is expressed as in the following equation (2).
Each of x(φ(i)), y(φ(i)), and z(φ(i)) may be the measured acceleration value for each axis after filtering.

Therefore, using the measured acceleration values x(φ(i)), y(φ(i)), and z(φ(i)) obtained for each of the three axes, the acceleration signal strength of equation (1) taking into account equation (2) is expressed as the following equation (3).
The control unit (first calculation unit) 12 calculates the acceleration signal strength by using the formula (3).

After the calculation of equation (3) is completed, process 550 in FIG. 16 is performed. This process calculates relative signal power values. These signal power values are indices that indicate damage to the teeth. When teeth mate with each other, vibrations occur even if there is no damage to the teeth. In this embodiment, process 550 is performed so that vibrations generated from normal teeth and vibrations generated from abnormal teeth can be easily distinguished. Specifically, in process 550, first, the control unit (second calculation unit) 12 calculates equation (4). That is, the average value P of the acceleration signal strengths of all the teeth is calculated.
Here, Z represents the number of samples of the acceleration signal strength, i.e., the number of gear teeth.

Next, the control unit (third calculation unit) 12 calculates the relative signal intensity coefficient P _rel (k) for each tooth using equation (5).

The relative signal strength coefficient P _rel (k) is a coefficient that represents the following characteristics.
If the signal strength of tooth k matches the average acceleration signal strength, then P _rel (k)=0.
If the signal strength at tooth k is twice the average acceleration signal strength, then P _rel (k)=1.
If the signal strength at tooth k is three times the average acceleration signal strength, then P _rel (k)=2.
Similarly, P _rel (k) becomes 3, 4, . . .
If the signal strength of tooth k is lower than the average acceleration signal strength, then the value of P _rel (k) can be less than 0 (ie, P _rel (k)<0 instead of P _rel (k)=0).
Therefore, if the average vibration standard is normalized to zero, the value of P _rel (k) can be used as a numerical value to easily evaluate whether each tooth is normal or abnormal. The evaluation can be based on the judgment of an operator or can be automatic (for example, a threshold value can be set and if the value of P _rel (k) is equal to or greater than the threshold value, it can be evaluated as abnormal).
Thus, according to this embodiment, the relative signal strength coefficient is used as a damage index value indicating the condition of each individual tooth, making it possible to evaluate the acceleration signal strength for each tooth (per-tooth evaluation).

Next, the effects of the above-described embodiment will be described.
In this embodiment, a large number of instantaneous values of acceleration measured by the acceleration sensor 441 and a large number of instantaneous values of angles measured by the angle sensor 443 are accumulated while the shaft 24a rotates multiple times. A faulty part of the gear is detected as an acceleration abnormality at any angle position in an angle range of 360°. In multiple rotations of the shaft 24a, an acceleration abnormality corresponding to the faulty part of the gear and an angle corresponding to the faulty part of the gear are repeatedly measured. In this embodiment, even if the performance of the acceleration sensor 441 is low, an acceleration profile representing the change in acceleration per rotation of the shaft 24a can be generated by rearranging a large number of instantaneous values of acceleration accumulated in multiple rotations of the shaft 24a and connecting the rearranged instantaneous values of acceleration. In addition, even if the performance of the angle sensor 443 is low, an angle profile representing the change in angle per rotation of the shaft 24a can be generated by rearranging a large number of instantaneous values of angles accumulated in multiple rotations of the shaft 24a and connecting the rearranged instantaneous values of angles. Based on the thus obtained angle profile and acceleration profile, which are synchronized with each other, a gear failure can be accurately determined. That is, it is possible to properly grasp what kind of vibration occurs at what angle during the rotation of the shaft 24a, and to properly identify the location (angle) where the abnormality occurs. In addition, since the angle profile and acceleration profile are synchronized, it is possible to properly identify the location where the abnormality occurs even if the rotation speed of the shaft 24a fluctuates.

In the measurement device 100, the measurement start time is changed each time the acceleration sensor 441 measures the acceleration for a predetermined time. Therefore, the acceleration sensor 441 measures the instantaneous value of the acceleration at a different angular position on the axis 24a each time a measurement is performed for a predetermined time. Therefore, by rearranging a large number of instantaneous values of acceleration accumulated over multiple rotations of the axis 24a and connecting the rearranged instantaneous values of acceleration, it is possible to easily generate an acceleration profile that represents the change in acceleration per rotation of the axis 24a.
Furthermore, in the measurement device 100, the measurement start time point is changed each time the angle sensor 443 measures the angle for a predetermined time. Therefore, the angle sensor 443 measures the instantaneous value of the angle at a different angular position on the shaft 24a each time a measurement is performed for a predetermined time. Therefore, by rearranging a large number of instantaneous values of the angle accumulated over multiple rotations of the shaft 24a and connecting the rearranged instantaneous values of the angle, it is possible to easily generate an angle profile that represents the change in angle per rotation of the shaft 24a.

In the measurement device 100, the acceleration sensor 441 and the angle sensor 443 have the same sampling period, and the instantaneous acceleration and angle values measured simultaneously by the acceleration sensor 441 and the angle sensor 443 are associated and stored in the memory unit 230. The raw data of the measurement values stored in the memory unit 230 (multiple discrete instantaneous acceleration values and multiple discrete instantaneous angle values associated on the time axis) is then transferred to the gear failure judgment device 10 and accumulated in the memory unit 13 of the gear failure judgment device 10. Therefore, in the gear failure judgment device 10, the control unit 12 can easily generate synchronized acceleration and angle profiles.

In this embodiment, the location of a fault in the gear can be determined with high accuracy using a three-axis acceleration sensor 441 that measures the acceleration of each of the X-axis, Y-axis, and Z-axis.

The angle sensor 443 has multiple readers arranged at equal angular intervals around the axis of the shaft 24a, so even if the resolution of each reader is low, the gear fault determination device 10 can generate a highly accurate angle profile.

The measuring device 100 of this embodiment (acceleration sensor 441, angle sensor 443, circuit board group 140, etc.) is built into the first bearing 25. Moreover, electricity obtained by power generating unit 149, which generates electricity by utilizing the relative rotation between outer ring 250A and inner ring 250B of bearing portion 250, becomes the energy source that supplies power to acceleration sensor 441, angle sensor 443, Z-phase detector 444, etc., so there is no need to provide a large energy source at or near bearing portion 250. Therefore, it is possible to realize a compact measuring device 100 with low power consumption.
In this way, the data wireless transmission type measurement device 100 having a self-power generation function can use the minute power generated by the self-power generation function to measure acceleration and angle and wirelessly transmit the measured data to the failure determination device 10. The self-power generation function allows the acceleration sensor 441 to be used to measure acceleration for a long period of time (at least for multiple rotations of the shaft 24a).
Furthermore, each time the acceleration sensor 441 and the angle sensor 443 perform measurements for a specified period, the power supply from the storage battery 150 to the acceleration sensor 441, the angle sensor 443, the Z-phase detector 444, etc. is stopped, and the electricity generated by the power generation unit 149 is charged to the storage battery 150. Therefore, the amount of charge that can be stored in the storage battery 150 may be small.
Furthermore, since the CPU 455 goes into a sleep state during measurements by the acceleration sensor 441 and the angle sensor 443, the amount of charge that can be stored in the storage battery 150 may be small.

The measurement device 100 contributes to miniaturization of the entire configuration by connecting an acceleration sensor 441, a temperature sensor 442, and an angle sensor 443 to a microcomputer 451. Therefore, even if the cover (device housing) 110 is small, the measurement device 100 can be stored in the same housing.
Furthermore, the measured values obtained from each sensor are stored in the storage unit 230 as unconverted data. Therefore, when storing data, data processing such as conversion from hexadecimal notation to decimal notation is not required, and high-speed data storage is possible. In addition, the load due to data processing can be reduced, and power consumption can be reduced.

The measuring device 100 can acquire and store data by DMA transfer, thereby reducing the load on the CPU 455 .
In other words, the measuring device 100 in this embodiment can be configured using an inexpensive MEMS sensor and a low-power consumption microcomputer.
Since the measurement device 100 can install the acceleration sensor 441 very close to the vibration source (can be installed at a position physically close to the vibration excitation point), it can acquire a signal with a higher S/N ratio than when a vibration sensor (for example, a high-performance sensor that performs sensing in the direction of the action line where the gears mesh) is installed on the outer surface of the housing 21 of the gear transmission mechanism 20. Therefore, reliable gear diagnosis is possible even when an inexpensive MEMS acceleration sensor is used.

The gear diagnosis algorithm used in this embodiment is the Per-tooth method, which has the following advantages:
Abnormal teeth can be identified. Diagnosis can be made with low calculation cost. Abnormalities can be determined from the evaluation results alone, without the need to compare with vibration data of normal gears, as was previously done in gear diagnosis.
Being able to perform diagnosis at low computational cost means that diagnosis can be performed using arithmetic operations. In other words, high computational costs such as FFT (Fast Fourier Transform) are not required. In addition, the process of FIG. 16 requires only a small number of mathematical calculations, which enables calculations at low power.
Furthermore, according to this embodiment, gear type judgment is performed by the Per-tooth method using acceleration values measured for each of the three axes. By using data from three axes, even if data from a single axis cannot detect an abnormality in the peak-peak value, data from the other axes can compensate for it, thereby achieving high diagnostic sensitivity.

Furthermore, since the acceleration sensor 441 is not attached to the surface of the housing 21, the system 30 can be used regardless of the shape of the housing 21, and the shape of the housing 21 is irrelevant to the accuracy of fault diagnosis.
If an acceleration sensor is installed in the housing 21 of the gear transmission mechanism 20, the housing 21 exists between the meshing point of the gears and the acceleration sensor, so it is affected by the vibration path, and a high-end, high-performance sensor is required. In this embodiment, the acceleration sensor 441 is built into the first bearing 25, so the acceleration sensor 441 can measure vibrations in close proximity to the first gear 22. The acceleration sensor 441 does not need to be a high-end, high-performance sensor because it is not affected by the vibration transmission path. In other words, an inexpensive acceleration sensor will suffice. Even an inexpensive acceleration sensor can accurately detect vibrations, so fault determination can also be performed accurately.
In this embodiment, since a MEMS acceleration sensor is used as an inexpensive acceleration sensor, it is excellent not only in terms of sensor position but also in terms of cost.

The acceleration sensor 441 is also installed on the outer ring 250A of the first bearing 25. This location is usually one of the non-rotating mechanical elements closest to the gear meshing portion, and the vibration that can be obtained can be said to be gear meshing vibration. By measuring the acceleration close to the vibration source, it is not affected by the vibration transmission path such as the housing.

Next, a modified example of this embodiment will be described. Fig. 19 shows elements of the gear failure detection system 30 according to the modified example that are particularly related to gear failure judgment. In this modified example, a field programmable gate array (FPGA) 456A is provided instead of the DMA controller 456.
The acceleration sensor 441 sends the instantaneous value of acceleration to the FPGA 456A every time it measures the instantaneous value of acceleration, and when the FPGA (acceleration timestamp generating unit) 456A receives the instantaneous value of acceleration from the acceleration sensor 441, it generates an acceleration timestamp indicating the time when the acceleration sensor 441 measured the instantaneous value of acceleration. Then, the FPGA 456A associates the acceleration timestamp with the instantaneous value of acceleration and stores it in the storage unit 230. The angle sensor 443 sends the instantaneous value of angle to the FPGA 456A every time it measures the instantaneous value of angle, and when the FPGA (angle timestamp generating unit) 456A receives the instantaneous value of angle from the angle sensor 443, it generates an angle timestamp indicating the time when the angle sensor 443 measured the instantaneous value of angle. Then, the FPGA 456A associates the angle timestamp with the instantaneous value of angle and stores it in the storage unit 230.

The instantaneous values measured by the acceleration sensor 441 and the instantaneous values measured by the angle sensor 443 stored in the memory unit 230 are wirelessly transmitted via BLE from the wireless module 453 while being associated with the acceleration timestamp and the angle timestamp, and are received by the communication unit 15 of the gear failure determination device 10. Since the BLE wireless transmission is repeated periodically, the communication unit 15 of the gear failure determination device 10 periodically receives the instantaneous acceleration value and the instantaneous angle value.
The control unit 12 of the gear failure determination device 10 accumulates in the memory unit 13 the instantaneous value of acceleration measured by the acceleration sensor 441 and the instantaneous value of the angle measured by the angle sensor 443 while they are associated with the acceleration timestamp and the angle timestamp. In other words, the memory unit 13 associates the instantaneous value of acceleration measured by the acceleration sensor 441 with the acceleration timestamp, and stores the instantaneous value of the angle measured by the angle sensor 443 in association with the angle timestamp. The accumulation is performed periodically while the shaft 24a rotates multiple times.

In this modified example, even if the acceleration sensor 441 and the angle sensor 443 are not synchronized, synchronized acceleration profiles and angle profiles can be easily generated based on the time stamps related to acceleration and the time stamps related to angles.
In other words, as shown in FIG. 20, the sampling periods of the acceleration sensor 441 and the angle sensor 443 do not need to match.

As shown in FIG. 20, when the sampling period of the angle sensor 443 is longer than the sampling period of the acceleration sensor 441, the instantaneous value of the angle may not be measured by the angle sensor 443 at the moment when the instantaneous value of the acceleration is measured by the acceleration sensor 441. In other words, the number of instantaneous values of the angle measured is small compared to the number of instantaneous values of the acceleration. Therefore, as shown in FIG. 21, it is preferable that the control unit (linear interpolation calculation unit) 12 of the gear failure determination device 10 calculates the instantaneous value of the angle at the moment when the instantaneous value of the acceleration is measured by the acceleration sensor 441 and the instantaneous value of the angle is not measured by the angle sensor 443 by interpolating from the two instantaneous values of the angle measured by the angle sensor 443. Since the change in angle of the shaft 24a is approximately linear with respect to time, the interpolation may be linear. In FIG. 21, point A indicates the two instantaneous values of the angle measured by the angle sensor 443, and point B indicates the instantaneous value of the angle obtained by the interpolation.
The control unit 12 generates an angle timestamp corresponding to the instantaneous value of the angle obtained by linear interpolation from the angle timestamps corresponding to the two instantaneous values of the angle measured by the angle sensor 443, associates the angle timestamp with the instantaneous value of the angle obtained by linear interpolation, and stores it in the memory unit 13.

The control unit (acceleration profile generating unit) 12 reads out the multiple instantaneous values of acceleration and the acceleration time stamps stored in the memory unit 13, and uses the acceleration time stamps to rearrange the multiple instantaneous values of acceleration measured by the acceleration sensor 441 while the axis 24a rotates multiple times, and connects together the rearranged instantaneous values of acceleration to generate an acceleration profile that represents the change in acceleration per rotation of the axis 24a.
In addition, the control unit (angle profile generation unit) 12 reads out multiple instantaneous angle values and angle timestamps stored in the memory unit 13, and uses the angle timestamps to rearrange multiple instantaneous angle values measured by the angle sensor 443 while the shaft 24a rotates multiple times, and by connecting the rearranged instantaneous angle values together, generates an angle profile that represents the change in the measured angle value per rotation of the shaft 24a.
Even if the sampling period of the angle sensor 443 is longer than the sampling period of the acceleration sensor 441, the control unit 12 uses the instantaneous value of the angle obtained by linear interpolation and the angle timestamp in addition to the instantaneous value of the measured angle and the angle timestamp, so that the generated angle profile is synchronized with the acceleration profile.

Next, a preferable range of the performance of the acceleration sensor 441 will be described.
22 shows an example of an acceleration profile for the Y axis obtained in this embodiment. This acceleration profile was obtained under the conditions of a gear with 22 teeth, a shaft rotation speed of 3000 rpm, and a transmission torque of 1000 Nm. The gear had moderate damage.
In the acceleration profile of Figure 22, normal vibration (vibration without gear damage) is in the acceleration range of -2 G to +2 G. Abnormal vibration due to moderate gear damage is two times (-4 G to +4 G) to three times (-6 G to +6 G) the acceleration range of normal vibration.
The vibration acceleration of the gears is not significantly affected by the transmission torque, and can vary depending on the rotation speed of the shaft, the meshing state of the gears, and the type of gears. The acceleration profile in FIG. 22 corresponds to a condition where the vibration acceleration is low. Therefore, if the acceleration sensor 441 can measure acceleration in the range of -4G to +4G, which is at least twice the acceleration range of normal vibration, it can determine moderate damage to the gears. However, if the measurement value of the acceleration sensor 441 becomes saturated for a long period of time, the output of the acceleration sensor 441 becomes unstable. Therefore, it is preferable that the measurable acceleration range of the acceleration sensor 441 is at least three times the acceleration range of normal vibration, which is -6G to +6G.
Under conditions of high vibration acceleration, normal vibration is within the range of acceleration from -5.3G to +5.3G. Abnormal vibration due to moderate gear damage is twice (-10.6G to +10.6G) to three times (-16G to +16G) the acceleration range of normal vibration. Therefore, if the acceleration sensor 441 can measure acceleration from -10.6G to +10.6G, moderate damage to the gear can be determined. However, if the measurement value of the acceleration sensor 441 becomes saturated for a long period of time, the output of the acceleration sensor 441 becomes unstable. Therefore, it is preferable that the measurable acceleration range of the acceleration sensor 441 is -16G to +16G, which is three times the acceleration range of normal vibration.
Therefore, the range of acceleration that can be measured by the acceleration sensor 441 is preferably at least from −6G to +6G, and at most from −16G to +16G.
For a three-axis acceleration sensor, the range of acceleration that can be measured on all three axes is preferably at least -6 G to +6 G, and at most -16 G to +16 G. In the case of such a measurable acceleration range, depending on the meshing state of the gears and the type of gears, it may not be possible to identify a gear fault with only the acceleration measurement results on one axis, but a three-axis acceleration sensor makes it possible to accurately identify a gear fault.

The resolution of the acceleration sensor 441 is preferably 1/10 or less of the measurable acceleration range. In the example of Fig. 22, since the measurable acceleration range is from -6G to +6G, the resolution is preferably 1.2G or less. If the measurable acceleration range is from -16G to +16G, the resolution is preferably 3.2G or less.
Fig. 23 is a graph showing the relative signal strength coefficient P _rel for each of the three axes calculated by the Per-tooth method using a three-axis acceleration sensor whose resolution is one-fifth of the measurable acceleration range. As shown in Fig. 23, even though the resolution is one-fifth of the measurable acceleration range, it was determined that the 14th tooth was damaged in each of the three axes.
In order to further improve the accuracy of the measurement, it is preferable that the resolution of the acceleration sensor 441 is equal to or less than one tenth of the measurable range of acceleration. In this case, by using the Per-tooth method, it is possible to reliably detect a gear failure.
The higher the resolution of the acceleration sensor 441, the more accurately it is possible to detect a gear failure. However, a sensor with a high resolution is expensive. Therefore, it is preferable that the resolution of the acceleration sensor 441 is equal to or greater than 1/4096 (1/12 bits) of the measurable range of acceleration.

FIG. 24 is a graph showing the three-axis FFT results obtained by FFT (Fast Fourier Transform) from the three-axis acceleration profile. The Y-axis FFT results in the center of FIG. 24 were obtained from the acceleration profile in FIG. 22. These FFT results were obtained under the conditions of 22 gear teeth, 3000 rpm shaft rotation speed, and 1000 Nm transmission torque. There was moderate damage to the gears. Since the gear teeth were 22 and the shaft rotation speed was 3000 rpm, the meshing frequency was 1100 Hz.
In the three-axis FFT results in Fig. 24, the normal vibration components and the abnormal vibration components have local maximum values at integer multiples of the meshing frequency. The local maximum values are particularly high at 1 and 3 times the meshing frequency. However, depending on the shaft rotation speed and/or transmission torque, the local maximum value may be high at twice the meshing frequency.
It should be noted that for all three axes, the abnormal vibration components exceed the normal vibration components in the range from 1 to 2 times the meshing frequency. This characteristic remained the same even when the rotation speed and/or transmission torque of the axis was changed. Therefore, the measurable frequency band (detection band) of the acceleration sensor 441 is preferably 0.9 to 2.1 times the meshing frequency of the gears, and more preferably 1 to 2 times the meshing frequency of the gears.
As described above, the acceleration sensor 441 having low performance can be used in this embodiment.

Ideally, the sampling frequency of the acceleration sensor 441 that is actually set should be one time the meshing frequency. However, as long as it is within the above detection band, the sampling frequency of the acceleration sensor 441 is not important. This is because in this embodiment and the modified example described above with reference to Figures 19 to 21, the measured instantaneous values of acceleration and angle are synchronized, and the measurement start point is changed each time acceleration and angle are measured, making it possible to measure acceleration and angle for many angular positions of the axis.

Next, a preferred range of performance of the angle sensor 443 will be described.
The resolution of each reader of the angle sensor 443 is preferably equal to or less than one tenth of the angular pitch of the gears, as shown in Fig. 25. The repeatability of each reader of the angle sensor 443 is also preferably equal to or less than one tenth of the angular pitch of the gears. The inventors performed a simulation to investigate the upper limit of the resolution and the upper limit of the repeatability.
Fig. 26 is a graph showing an acceleration profile for the Y-axis, which was used as the basis for the simulation. This acceleration profile was obtained under the same conditions as the acceleration profile in Fig. 22, that is, the number of teeth of the gear was 22, the rotation speed of the shaft was 3000 rpm, and the transmission torque was 1000 Nm. The gear had moderate damage. While the acceleration profile in Fig. 22 was obtained in this embodiment, the acceleration profile in Fig. 26 was obtained from the results of continuous measurements during one rotation (360°) of the shaft 24a using expensive acceleration and angle sensors with high resolution and high sampling frequency.
In the simulation, a uniformly distributed error of 1.8° was applied to the mechanical angle of the high-precision acceleration profile actually measured as shown in Figure 26. Since the angular pitch of a gear with 22 teeth is 16.36°, 1.8° is 1/9.09 of the angular pitch. Figure 27 is a graph showing the acceleration profile obtained in this simulation. As is clear from a comparison with Figure 26, the simulation results are similar to the reference acceleration profile, and abnormal vibrations could be detected from these simulation results using the Per-tooth method.

In another simulation, a uniformly distributed error of 2° was added to the mechanical angle of the actually measured high-precision acceleration profile shown in Figure 26. Since the angular pitch of a gear with 22 teeth is 16.36°, 2° is 1/8.18 of the angular pitch. Figure 28 is a graph showing the acceleration profile obtained in this simulation. As is clear from a comparison with Figure 26, the simulation results were inferior to the reference acceleration profile, and abnormal vibrations could not be detected from this simulation result by the Per-tooth method.
Therefore, the resolution and repeatability of each reader of angle sensor 443 is preferably equal to or better than 1/9.09 of the angular pitch of the gears, and more preferably equal to or better than 1/10 to further improve the accuracy of the measurement.

The higher the resolution, the more accurately gear faults can be detected. However, high-resolution sensors are expensive. Therefore, it is preferable that the resolution of each reader of the angle sensor 443 is equal to or greater than 1/65536 (1/16 bits) of the gear's angular pitch.
Also, the higher the repeatability, the more accurately gear failures can be detected. However, sensors with high repeatability are expensive. Therefore, it is preferable that the repeatability of each reader of the angle sensor 443 is equal to or greater than 1/256 (1/8 bits) of the gear's angular pitch.
As described above, the angle sensor 443 with low performance can be used in this embodiment.

29 shows elements particularly related to gear fault determination in a gear fault detection system 30 according to yet another modified example of the embodiment. In this modified example, the angle sensor 443 is not provided, and the Z-phase detector 444 is used as a tachometer to generate a virtual angle profile based on the Z-phase detection pulse output by the Z-phase detector 444.
The acceleration sensor 441 sends the instantaneous value of acceleration to the FPGA 456A every time it measures the instantaneous value of acceleration, and when the FPGA (acceleration time stamp generating unit) 456A receives the instantaneous value of acceleration from the acceleration sensor 441, it generates an acceleration time stamp indicating the time when the acceleration sensor 441 measured the instantaneous value of acceleration. Then, the FPGA 456A associates the acceleration time stamp with the instantaneous value of acceleration and stores it in the storage unit 230. Every time the Z-phase detector 444 detects the Z-phase magnet 162 and generates a Z-phase detection pulse (every time the shaft 24a rotates once), it sends the Z-phase detection pulse to the FPGA 456A, and when the FPGA 456A receives the Z-phase detection pulse, it generates a Z-phase time stamp indicating the time when the Z-phase detector 444 generated the Z-phase detection pulse. Then, the FPGA 456A stores the Z-phase time stamp in the storage unit 230.

The instantaneous value measured by the acceleration sensor 441 associated with the acceleration timestamp stored in the memory unit 230 and the Z-phase timestamp are transmitted via BLE wireless from the wireless module 453 and received by the communication unit 15 of the gear failure determination device 10. Since the BLE wireless transmission is repeated periodically, the communication unit 15 of the gear failure determination device 10 periodically receives the instantaneous value of acceleration and the Z-phase timestamp.
The control unit 12 of the gear failure determination device 10 accumulates the instantaneous value measured by the acceleration sensor 441 associated with the acceleration time stamp and the Z-phase time stamp in the memory unit 13. That is, the memory unit 13 stores the instantaneous value of acceleration measured by the acceleration sensor 441 and the acceleration time stamp in association with each other, and further stores the Z-phase time stamp. The accumulation is performed periodically while the shaft 24a rotates multiple times.

In this modified example, even if there is no instantaneous value of the angle measured by the angle sensor, it is possible to generate a synchronized acceleration profile and angle profile based on the acceleration timestamp and the Z-phase timestamp.
The control unit (angle estimator) 12 of the gear fault determination device 10 estimates the instantaneous value of the angle of the shaft 24a from the time interval of the Z-phase time stamps (i.e., the interval of the Z-phase detection pulses). Specifically, as shown in FIG. 30, the control unit 12 estimates the moment identified by the Z-phase time stamp as an electrical angle of 360° at the end of a certain cycle of the mechanical angle, and an electrical angle of 0° at the beginning of the next cycle. Point A indicates the instantaneous value of the angle to which the Z-phase time stamp is thus applied. Furthermore, the control unit 12 calculates the instantaneous value of the angle at the moment when the acceleration sensor 441 measures the instantaneous value of the acceleration (the moment identified by the acceleration time stamp) by interpolating from the two instantaneous values of the angle at the moment identified by the Z-phase time stamp. When the shaft 24a rotates at a constant speed, the change in the angle of the shaft 24a is linear with respect to time, so the interpolation may be linear interpolation. In FIG. 30, point B indicates the instantaneous value of the angle obtained by the interpolation.
The control unit 12 generates angle timestamps corresponding to instantaneous values of all estimated angles from the Z-phase timestamps corresponding to the instantaneous value of the angle corresponding to point A. The angle timestamp corresponding to the instantaneous value of the angle corresponding to point A may be the Z-phase timestamp itself. The control unit 12 generates angle timestamps corresponding to the instantaneous value of the angle obtained by linear interpolation from the angle timestamps corresponding to two instantaneous values of the angle corresponding to point A, and stores the angle timestamps in the storage unit 13 in association with the instantaneous value of the angle.

The control unit (acceleration profile generating unit) 12 reads out the multiple instantaneous values of acceleration and the acceleration time stamps stored in the memory unit 13, and uses the acceleration time stamps to rearrange the multiple instantaneous values of acceleration measured by the acceleration sensor 441 while the axis 24a rotates multiple times, and connects together the rearranged instantaneous values of acceleration to generate an acceleration profile that represents the change in acceleration per rotation of the axis 24a.
In addition, the control unit (angle profile generation unit) 12 reads out multiple instantaneous angle values and angle timestamps stored in the memory unit 13, and uses the angle timestamps to rearrange multiple instantaneous angle values measured by the angle sensor 443 while the shaft 24a rotates multiple times, and by connecting the rearranged instantaneous angle values together, generates an angle profile that represents the change in the measured angle value per rotation of the shaft 24a.

When the shaft 24a rotates at a uniform speed, the generated angular profile is synchronous with the acceleration profile.
The preferred ranges of acceleration sensor performance described above also apply to this variant.

Although the present invention has been shown and described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that changes may be made in form and detail without departing from the spirit and scope of the invention as set forth in the appended claims. All such changes, modifications and alterations are intended to be embraced within the scope of the invention.

For example, the gear failure determination device 10 is a personal computer, but it may also be a tablet terminal, a smartphone, a wearable watch, etc.
In the above embodiment, the measurement device 100 transmits the measurement values to the gear fault judgment device 10 without processing them, but part of the processing performed by the gear fault judgment device 10 may be executed within the measurement device 100. In other words, the measurement device 100 may have part of the functions of the gear fault judgment device 10. For example, the processing shown in FIG. 16 may be performed by the microcomputer 451. The processing in FIG. 16 requires less calculation processing, so the power consumption is small. Therefore, if the power self-generated by the first bearing 25 is sufficiently large, the microcomputer 451 may perform the processing in FIG. 16. Such a configuration can be called on-bearing data processing. When such a configuration is adopted, the amount of data transmitted from the measurement device 100 to the gear fault judgment device 10 is significantly reduced. This is because thousands of angle data and acceleration data are condensed into one scalar value per tooth (for each axis) for each measurement.
The measurement time can be set appropriately depending on the capacity and power consumption of the storage unit 230 .

The temperature sensor 442 may not be provided.
In FIG. 6, a protection circuit may be provided between the smoothing circuit 262 and the power management IC 263. In FIG. 6, electromagnetic induction power generation is performed by the relative rotation of the magnet M and the coil C. Therefore, as the rotation speed of the coil C increases, the electromotive force also increases proportionally. If the device connected to the first axis 24a breaks down for some reason, the first axis 24a rotates at a rotation speed that exceeds the expected rotation speed, and the electromotive force of the magnet M and the coil C increases in proportion to the rotation, an electromotive force that exceeds the input limit of the power management IC 263 is input to the power management IC 263, and the measuring device 100 may be damaged. In order to avoid such damage, that is, to prevent the measuring device 100 from being damaged even if an electromotive force that exceeds the expected value is generated, a protection circuit may be provided between the smoothing circuit 262 and the power management IC 263. This protection circuit is an input protection circuit, and various types of protection circuits can be adopted. For example, a Zener diode is provided as a protection circuit between the smoothing circuit 262 and the power management IC 263. The Zener diode functions as a protection circuit that protects the measuring device 100 by converting an input voltage exceeding a predetermined value into heat and dissipating the heat.

10 Gear failure determination device 12 Control unit (acceleration profile generation unit, angle profile generation unit, failure determination unit, map generation unit, first calculation unit, second calculation unit, third calculation unit, angle estimation unit)
13 Memory unit 20 Gear transmission mechanism 24a First shaft 25 First bearing 30 Gear failure detection system 100 Measurement device 150 Storage battery (capacitor)
142 Angle sensor board 149 Power generation unit 250 Bearing unit 263 Power supply management IC (storage battery control unit, charging completion detection unit)
441 Acceleration sensor 443 Angle sensor 444 Z-phase detector (pulse generating unit)
455 CPU (measurement start time change unit)
456A FPGA (Acceleration time stamp generator, Angle time stamp generator)

Claims (15)

an acceleration sensor provided in a bearing that rotatably supports a shaft having a gear mounted thereon, the acceleration sensor measuring an instantaneous value of acceleration caused by vibration of the bearing for a predetermined period of time each time the shaft rotates, the period being shorter than the time required for one rotation of the shaft;
an acceleration profile generating unit that generates an acceleration profile representing a change in acceleration per one rotation of the axis by rearranging a plurality of instantaneous values of acceleration measured by the acceleration sensor while the axis rotates a plurality of times and connecting the rearranged instantaneous values of acceleration;
a fault determination unit that determines a fault in a gear based on the acceleration profile generated by the acceleration profile generation unit. A pulse generating unit generates one pulse for each rotation of the shaft,
After the pulse generating unit generates a pulse, the acceleration sensor measures an instantaneous value of acceleration for a predetermined time from a measurement start point to a time required for one rotation of the shaft,
2. The gear failure detection system according to claim 1, further comprising a measurement start time changing unit that changes the measurement start time each time the acceleration sensor measures the acceleration for a predetermined period of time. a power generating unit that generates power by utilizing relative rotation between an outer ring and an inner ring of the bearing unit;
a storage battery that stores the electric charge generated by the power generation unit and supplies the electric charge to the acceleration sensor;
a storage battery control unit that stops power supply from the storage battery to the acceleration sensor and charges the storage battery with electricity generated by the power generation unit each time the acceleration sensor finishes measuring the acceleration for a predetermined period of time;
a charging completion detection unit that detects the charging completion of the storage battery,
3. The gear failure detection system according to claim 2, wherein the acceleration sensor measures an instantaneous value of acceleration for a certain period of time from a measurement start point after the charge completion detection unit detects the charge completion of the storage battery and after the pulse generation unit generates a pulse. 4. The gear failure detection system according to claim 1, wherein the acceleration sensor is a three-axis acceleration sensor that measures acceleration along each of an X-axis, a Y-axis, and a Z-axis. an angle sensor provided in the bearing portion for measuring an instantaneous value of an angle of the shaft for a predetermined time shorter than a time required for one rotation of the shaft each time the shaft rotates;
an angle profile generating unit that generates an angle profile that represents a change in the angle measurement value per one rotation of the shaft and is synchronized with the acceleration profile by rearranging a plurality of instantaneous values of the angle measured by the angle sensor while the shaft rotates a plurality of times and connecting the rearranged instantaneous values of the angle,
4. The gear failure detection system according to claim 1, wherein the failure determination unit determines a gear failure based on the acceleration profile generated by the acceleration profile generation unit and the angle profile generated by the angle profile generation unit. The sampling period of the acceleration sensor and the sampling period of the angle sensor are the same,
6. The gear failure detection system according to claim 5, further comprising a storage unit that stores the instantaneous value of the acceleration and the instantaneous value of the angle measured simultaneously by the acceleration sensor and the angle sensor in association with each other. an acceleration time stamp generation unit that generates an acceleration time stamp indicating a time when the acceleration sensor measures the instantaneous value of the acceleration;
an angle timestamp generation unit that generates an angle timestamp indicating a time when the angle sensor measures an instantaneous value of the angle;
6. The gear failure detection system according to claim 5, further comprising a storage unit that stores the instantaneous value of the acceleration measured by the acceleration sensor in association with the acceleration time stamp, and stores the instantaneous value of the angle measured by the angle sensor in association with the angle time stamp. a sampling period of the angle sensor is longer than a sampling period of the acceleration sensor;
8. The gear fault detection system according to claim 7, further comprising a linear interpolation calculation unit that calculates an instantaneous value of an angle when an instantaneous value of acceleration is measured by the acceleration sensor and an instantaneous value of an angle is not measured by the angle sensor, by linear interpolation from a plurality of instantaneous values of the angle measured by the angle sensor. A pulse generating unit that generates one pulse for each rotation of the shaft;
an angle estimator that estimates an instantaneous value of an axis angle from the pulse interval;
an angle profile generating unit that generates an angle profile that represents a change in the angle measurement value per one rotation of the shaft and is synchronized with the acceleration profile by rearranging a plurality of instantaneous values of the angle estimated by the angle estimating unit and connecting the rearranged instantaneous values of the angle,
4. The gear failure detection system according to claim 1, wherein the failure determination unit determines a gear failure based on the acceleration profile generated by the acceleration profile generation unit and the angle profile generated by the angle profile generation unit. the acceleration sensor is a three-axis acceleration sensor that measures acceleration along each of an X-axis, a Y-axis, and a Z-axis;
The failure determination unit is
a map generator that generates a map that defines a relationship between the teeth of the gear and a rotation angle of the shaft;
a first calculation unit that calculates an acceleration signal intensity of each tooth using equation (1);
a second calculation unit that calculates an average value of the acceleration signal strengths of all teeth using equation (2);
A third calculation unit that calculates a damage index value for each tooth using formula (3);
Equipped with
The formula (1) is
where P _(tooth=k ) is the acceleration signal strength of the kth tooth, x(φ(i)), y(φ(i)), z(φ(i)) are the X-axis, Y-axis, and Z-axis acceleration values, respectively, when the sample angular position φ(i) corresponds to the kth tooth, and n _samples,k is the number of samples of the acceleration value corresponding to the kth tooth,
The formula (2) is
where P is the average value of the acceleration signal strength of all teeth, P(k) is the acceleration signal strength of the kth tooth, Z is the number of teeth on the gear,
The formula (3) is
Gear fault detection system according to any one of claims 5 to 9, characterized in that P _rel (k) is the damage index value of the kth tooth. 11. The gear failure detection system according to claim 1, wherein the acceleration sensor has a measurable acceleration range of at least -6G to +6G and at most -16G to +16G. 12. The gear fault detection system according to claim 11, wherein the resolution of the acceleration sensor is greater than or equal to 1/4096 and less than or equal to 1/10 of the measurable range of acceleration. 13. The gear failure detection system according to claim 1, wherein a measurable frequency band of the acceleration sensor is 0.9 to 2.1 times the meshing frequency of the gears. 9. A gear fault detection system according to claim 5, wherein the resolution of the angle sensor is greater than or equal to 1/65536 and less than or equal to 1/10 of the angular pitch of the gear. 15. A gear fault detection system according to any one of claims 5 to 8 and 14, characterized in that the repeatability of the angle sensor is greater than or equal to 1/256 and less than or equal to 1/10 of the angular pitch of the gear.