Bearing device, motor, blower, and bearing device abnormality diagnosis method

Abstract

To provide a technology capable of improving accuracy of abnormality diagnosis in a bearing device.SOLUTION: A bearing device BE comprises: two rolling bearings 30 that rotatably support a rotational shaft; a bearing housing 40 that holds the two rolling bearings 30 from an outer peripheral surface side; and two strain gauges 100 each including a resistor 103 for detecting strain of the rolling bearing 30, and mounted on the bearing housing 40 so as to correspond to each of the two rolling bearings 30.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a bearing device and the like. [Background technology]

[0002] BACKGROUND ART Conventionally, there is known a technique for monitoring abnormalities in a bearing device that supports a rotating shaft with two bearings (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2007-108187 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the above-mentioned document, since the output of a single sensor is used, if the sensor output fluctuates due to, for example, changes in the rotational speed of the rotating shaft or the temperature of the bearing device, there is a possibility that the bearing device may be diagnosed as having an abnormality even though there is no abnormality. Furthermore, if the abnormality conditions are set too strict in order to prevent erroneous diagnosis of an abnormality, there is a possibility that an abnormality may not be diagnosed even though an abnormality actually exists. As a result, there is a possibility that the accuracy of the abnormality diagnosis may be reduced.

[0005] In view of the above-mentioned problems, an object of the present invention is to provide a technique that can improve the accuracy of abnormality diagnosis of a bearing device. [Means for solving the problem]

[0006] In one embodiment of the present disclosure, Two bearings that rotatably support the rotating shaft; a holding member that holds the two bearings from the outer circumferential surface side; attached to the holding member so as to correspond to each of the two bearings, Each of them detects the strain of the bearing. Ru2 one Sensorand, a diagnostic unit that diagnoses abnormalities in the bearing by comparing the outputs of the two sensors; Equipped with A bearing arrangement is provided. In another embodiment of the present disclosure, Two bearings that rotatably support the rotating shaft; a holding member that holds the two bearings from the outer circumferential surface side; two sensors attached to the holding member so as to correspond to the two bearings, respectively, and each detecting a strain of the bearing; the bearing includes an outer ring, an inner ring arranged coaxially with the outer ring on the inner peripheral side of the outer ring, and a plurality of rolling elements arranged between the outer ring and the inner ring, and is preloaded to form a predetermined contact angle; the holding member has a thick portion having a relatively large thickness and a thin portion having a relatively small thickness, the sensor is disposed in the thin portion, the thick-wall portion is arranged so as to contact the outer ring at least in a region between an intersection of a straight line indicating the contact angle and the outer peripheral surface of the outer ring to a preload-side end face, which is an end face of the outer ring closer to the intersection point. A bearing arrangement is provided.

[0007] In addition, the present disclosure Furthermore In other embodiments, the bearing device described above; a stator core fixed to the radially outer side of the holding member, A motor is provided. In still another embodiment of the present disclosure, a bearing device including two bearings that rotatably support a rotating shaft, a holding member that holds the two bearings from their outer circumferential surfaces, and two sensors that are attached to the holding member so as to correspond to the two bearings, respectively, and that detect strain in the bearings; a stator core fixed to the radially outer side of the holding member, A motor is provided.

[0008] In still another embodiment of the present disclosure, the motor described above; an impeller that is rotationally driven by the motor; A blower is provided.

[0009] In still another embodiment of the present disclosure, The rotating shaft is rotatably supported by two bearings, a holding member that holds the two bearings from the outer circumferential surface side, and two sensors that are attached to the holding member so as to correspond to the two bearings, and each sensor detects strain in the bearing. A method for diagnosing an abnormality in a bearing device, comprising: The information processing device Sensor and compares the outputs of the two sensors to diagnose an abnormality in the bearing. A method for diagnosing anomalies is provided. [Effects of the Invention]

[0010] According to the above-described embodiment, the accuracy of abnormality diagnosis of the bearing device can be improved. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a vertical cross-sectional view showing an example of an axial flow fan. [Figure 2] 1 is a vertical cross-sectional view showing an example of the structure of a motor (a first example of a bearing housing). [Figure 3] FIG. 2 is a perspective view showing a first example of a bearing housing. [Figure 4] FIG. 2 is a cross-sectional view showing an example of the structure of a bearing device. [Figure 5] FIG. 2 is a plan view showing an example of the structure of a motor. [Figure 6] FIG. 10 is a diagram showing an example of an output waveform of a strain gauge. [Figure 7] FIG. 10 is a diagram showing a comparative example of an output waveform of a strain gauge. [Figure 8] FIG. 10 is a vertical cross-sectional view showing another example of the motor structure (a second example of the bearing housing). [Figure 9] FIG. 10 is a vertical cross-sectional view showing still another example of the motor structure (third example of the bearing housing). [Figure 10] FIG. 10 is a vertical cross-sectional view showing another example of the structure of the bearing device (fourth example of the bearing housing). [Figure 11] FIG. 10 is a vertical cross-sectional view showing still another example of the structure of the bearing device (a fifth example of the bearing housing). [Figure 12] FIG. 10 is a vertical cross-sectional view showing still another example of the structure of the bearing device (sixth example of the bearing housing). [Figure 13] FIG. 10 is a vertical cross-sectional view showing still another example of the structure of the bearing device (seventh example of the bearing housing). [Figure 14] FIG. 10 is a vertical cross-sectional view showing still another example of the structure of the bearing device (eighth example of the bearing housing). [Figure 15] FIG. 10 is a vertical cross-sectional view showing still another example of the structure of the bearing device (ninth example of the bearing housing). [Figure 16] FIG. 20 is a perspective view showing a tenth example of a bearing housing. [Figure 17] FIG. 13 is a perspective view showing an eleventh example of a bearing housing. [Figure 18] FIG. 22 is a perspective view showing a twelfth example of a bearing housing. [Figure 19] FIG. 2 is a plan view showing an example of a strain gauge. [Figure 20] FIG. 1 is a cross-sectional view showing an example of a strain gauge. [Figure 21] FIG. 1 illustrates an example of the configuration of an abnormality monitoring system. [Figure 22] FIG. 2 illustrates an example of a hardware configuration of a monitoring device. [Figure 23] FIG. 10 is a diagram showing an example of time-series data of a strain gauge. [Figure 24] FIG. 1 is a diagram illustrating an example of a configuration of a machine learning system. [Figure 25] 1 is a flowchart illustrating an example of a machine learning procedure. [Figure 26] 10 is a flowchart illustrating an example of an abnormality diagnosis process. [Figure 27] FIG. 10 is a diagram illustrating an example of the contribution (variable importance) of each statistical feature amount by random forest (decision tree analysis). DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments will be described with reference to the drawings. In the drawings, the same or corresponding components are denoted by the same reference numerals, and redundant description may be omitted.

[0013] [Outline of the structure of an axial flow fan] The structure of an axial flow fan 1 will be outlined with reference to FIG.

[0014] FIG. 1 is a vertical cross-sectional view showing an example of an axial flow fan 1. As shown in FIG.

[0015] As shown in FIG. 1, an axial flow fan 1 (an example of a blower) includes an impeller 10, a motor EM, and a casing 70.

[0016] The impeller 10 (an example of an impeller) is driven to rotate by the motor EM to generate an air flow in a direction along the rotation axis m. Hereinafter, the direction along the rotation axis m may be referred to as the axial direction. The impeller 10 includes a rotor housing 11 and blades 12 provided on the outer peripheral surface of the rotor housing 11.

[0017] The motor EM rotates the impeller 10. The motor EM includes a rotor shaft 20, a bearing device BE, a stator 50, and a rotor 60.

[0018] The rotor shaft 20 (an example of a rotating shaft) has a rotation axis m and is fixed to approximately the center of the rotor housing 11 and the rotor 60 in the radial direction. The rotor shaft 20 is rotatably supported by two rolling bearings 30 arranged near both ends in the longitudinal direction (axial direction). This allows the rotor 60 to rotate around the rotor shaft 20, and the impeller 10 to rotate in accordance with the rotation of the rotor 60.

[0019] The bearing device BE includes a rolling bearing 30 , a bearing housing 40 , and a strain gauge 100 .

[0020] Two rolling bearings 30 (an example of a bearing) are provided, one at each end in the axial direction of the bearing housing 40. As described above, the two rolling bearings 30 support the rotor shaft 20 in a rotatable state.

[0021] The rolling bearing 30 includes an outer ring 31, an inner ring 32, and a plurality of rolling elements 33. The outer ring 31 is a cylindrical structure with its center axis coincident with the rotation axis m. The inner ring 32 is a cylindrical structure arranged coaxially with the outer ring 31 on the inner circumferential side of the outer ring 31. Each of the plurality of rolling elements 33 is arranged in a raceway OB (see FIG. 10, etc.) formed between the outer ring 31 and the inner ring 32. The rolling elements 33 are, for example, spherical or cylindrical objects. A lubricant such as grease is sealed within the raceway OB. Specifically, recesses 31x and 32x having arc-shaped cross sections are formed in the inner circumferential surface of the outer ring 31 and the outer circumferential surface of the inner ring 32 in the circumferential direction of the outer ring 31 and the inner ring 32, and the plurality of rolling elements 33 are guided in the circumferential direction by the recesses 31x and 32x (see FIG. 10, etc.). A cage 34 is disposed within the raceway OB, and the plurality of rolling elements 33 are held by the cage 34. Specifically, the cage 34 is an annular body coaxial with the rotation axis m, and has recesses 34x (see FIG. 4) on one side in the axial direction for accommodating the rolling elements 33. The raceway OB is isolated from the outside world by raceway seals 35, 36 that are provided so as to protrude from the inner peripheral surface of the outer ring 31 toward the inner ring 32 (see FIG. 10, etc.).

[0022] The rolling bearing 30 may be a so-called full spherical bearing that does not use a cage 34.

[0023] The rolling bearing 30 is fixed into the rolling bearing housing 40 by press fitting, bonding, or the like, and is held by the bearing housing 40 on its outer peripheral surface side. The bearing housing 40 presses the outer peripheral surface of the outer ring 31 over the entire circumference. The bearing housing 40 is made of a metal such as brass, for example.

[0024] The bearing housing 40 (an example of a holding member) holds two rolling bearings 30 at both ends in the axial direction.

[0025] The strain gauge 100 detects strain in the rolling bearing 30. Two strain gauges 100 are provided so as to correspond to the two rolling bearings 30, respectively.

[0026] The stator 50 is disposed on the outer circumferential side of the bearing housing 40 .

[0027] The stator 50 includes an insulator 51 , a stator core 52 , and a coil 53 .

[0028] The stator core 52 (an example of a stator core) is fixed to the outer periphery of the bearing housing 40 by press-fitting or the like, for example.

[0029] The rotor 60 includes a rotor yoke 61 that is integrally provided inside the rotor housing 11 , and a rotor magnet 62 that is attached inside the rotor yoke 61 .

[0030] In this example (FIG. 1), the rotor yoke 61 is provided integrally inside the rotor housing 11, but it may also be attached to the inside of the rotor housing 11. Furthermore, the rotor shaft 20 is attached to the rotor yoke 61 and fixed to the center of the rotor housing 11, but the rotor shaft 20 may also be fixed directly to the rotor housing 11.

[0031] The motor EM is a so-called outer rotor type motor, and when current is supplied to the coil 53 from a power source not shown, the impeller 10 rotates around the central axis of the rotor shaft 20, which is rotatably supported within the bearing housing 40, as the rotation axis m.

[0032] The casing 70 (an example of a housing) includes a casing outer frame 71 that covers the outer periphery of the impeller 10, a base hub 72 that fixes the bearing housing 40, and stator blades 73 that connect the casing outer frame 71 and the base hub 72.

[0033] In this example (FIG. 1), the casing outer frame 71 and the base hub 72 are connected by the stator vanes 73, but the casing outer frame 71 and the base hub 72 may also be connected by a rod-shaped structure such as a connecting shaft. Also, the bearing housing 40 may be fixed so as to be integrated with the base hub 72 when the casing 70 is injection molded from resin, or may be fixed to the base hub 72 after the casing 70 has been molded first.

[0034] In this example (FIG. 1), the axial fan 1 creates an airflow from top to bottom, with the upper side in the figure being the air inlet side and the lower side being the air outlet side, and blows air. Therefore, in this example, the impeller 10 is provided on the air inlet side of the casing outer frame 71, and the base hub 72 is provided on the air outlet side. Hereinafter, the side of the rotor shaft 20 of the motor EM that is fixed to the impeller 10 (rotor 60) may be referred to as the "tip side," and the side of the motor EM that is fixed to the base hub 72 may be referred to as the "base side."

[0035] [Motor structure details] Next, the structure of the motor EM including the bearing device BE will be described in detail.

[0036] <First example of bearing device> FIG. 2 is a longitudinal sectional view showing an example of the structure of a motor EM. FIG. 3 is a perspective view showing a first example of a bearing housing 40. FIG. 4 is a transverse sectional view showing an example of the structure of a bearing device BE. FIG. 5 is a plan view showing an example of the structure of a motor EM. FIG. 6 is a diagram showing an example of the output waveform of a strain gauge 100. FIG. 7 is a diagram showing a comparative example of the output waveform of a strain gauge 100.

[0037] In FIG. 4, the rolling bearing 30 is illustrated in a simplified manner, and some of the components are not shown.

[0038] 2 and 3, the bearing housing 40 includes a relatively thick portion 41 and a relatively thin portion 42. In this example, the thick portion 41 has a relatively large radial thickness, while the thin portion 42 has a relatively small radial thickness.

[0039] The thin-walled portions 42 are provided at both axial ends of the bearing housing 40. Specifically, the thin-walled portions 42 are provided so as to include the axial positions where the rolling bearing 30 is held in the bearing housing 40. As shown in Fig. 3, the thin-walled portions 42 are recesses that are provided, for example, in the cylindrical outer peripheral surface 40a corresponding to the thick-walled portion 41, over a portion of the axial and radial ranges, and at positions away from the axial ends.

[0040] The two strain gauges 100 are provided in the two thin-walled portions 42 at both axial ends of the bearing housing 40. The strain gauges 100 are fixed to the thin-walled portions 42 of the bearing housing 40, for example, with an adhesive or the like. As a result, the two strain gauges 100 are fixed radially inward (toward the rotation axis m) of the inner diameter of the stator core 52.

[0041] The inner diameter of the stator core 52 refers to the portion of the stator core 52 that is in contact with the outer peripheral surface 40 a of the bearing housing 40 .

[0042] 3, the strain gauge 100 is a sensor having a resistor 103 that detects strain in the rolling bearing 30 (for example, strain in the outer ring 31). By disposing the strain gauge 100 in the thin-walled portion 42 of the bearing housing 40, the strain in the rolling bearing 30 is transmitted to the strain gauge 100 via the bearing housing 40 and can be detected by the strain gauge 100. Specifically, the strain gauge 100 can detect the strain in the rolling bearing 30 as a change in the resistance value of the resistor 103.

[0043] 3, the resistor 103 of the strain gauge 100 is arranged with its longitudinal direction (gauge length direction) facing the circumferential direction of the bearing housing 40. As a result, the bearing housing 40 expands and contracts more easily in the circumferential direction than in the axial direction, and the strain gauge 100 (resistor 103) can obtain a relatively large strain waveform.

[0044] 4, the gauge length L of the strain gauge 100 is preferably smaller than the distance between adjacent rolling elements 33 of the rolling bearing 30. Specifically, given the angle θ1 formed by two lines connecting the rotation axis m and the centers of two adjacent rolling elements 33, and the inner diameter R of the thin-walled portion 42 of the bearing housing 40, the gauge length L is preferably set to satisfy the condition of the following formula (1).

[0045] θ1 / 360×2π×R>L (1)

[0046] This allows the strain gauge 100 (resistor 103) to detect the strain of the outer ring 31 caused by a single rolling element 33.

[0047] As shown in FIG. 3, the strain gauge 100 has a pair of terminal portions 105 connected to both ends of the resistor 103 via wiring 104, and wiring 200 is electrically connected to each of the terminal portions 105 by solder or the like.

[0048] As shown in FIG. 2, the wiring 200 is drawn out from the axial lower side to the outside of the axial flow fan 1 (motor EM).

[0049] 2 and 5, the wiring 200 extending from the strain gauge 100 on the axial tip side passes through a notch 52x provided on the inner periphery of the stator core 52 and is drawn out to the axial base end side of the stator core 52. The notch 52x is, for example, an elongated recess provided approximately parallel to the rotation axis m.

[0050] In this example (FIG. 2), the wiring 200 electrically connected to the strain gauge 100 is drawn directly to the outside of the axial fan 1 (motor EM), but it may also be electrically connected to a circuit board arranged inside the axial fan 1 (motor EM). In this case, a separate wiring may be drawn from the circuit board to the outside of the axial fan 1 (motor EM).

[0051] The wiring 200 is preferably partially or entirely shielded because it is subject to interference from electromagnetic noise generated from the coil 53, etc. For example, it is preferable that at least the portion of the wiring 200 extending from the strain gauge 100 on the axial tip side that passes through the notch 52x is shielded. Also, for example, it is preferable that at least the portion of the wiring 200 extending from the strain gauge 100 on the axial base end side that passes through the inside of the axial fan 1 (motor EM) is shielded.

[0052] The waveform obtained from the output of the strain gauge 100 has a peak strain amount (output intensity) when the rolling element 33 passes directly below the resistor 103 of the strain gauge 100, and a bottom at the midpoint between adjacent rolling elements 33, resulting in a periodic waveform with repeated peaks and bottoms.

[0053] As shown in Fig. 7, when the wiring 200 is passed around the outer periphery of the stator core 52, electromagnetic noise is large and a correct distorted waveform cannot be obtained. In contrast, when the wiring 200 is passed through a notch 52x formed on the inner periphery and at least the portion of the wiring 200 that passes through the notch 52x is shielded as shown in Fig. 6, a correct distorted waveform can be obtained.

[0054] The wiring 200 may be, for example, a coaxial cable, or may have a structure in which a solid ground (GND) is formed on at least one side of a flexible substrate. From the viewpoint of suppressing the effects of electromagnetic noise, it is preferable that the bearing housing 40 to which the strain gauge 100 is fixed is a ground (GND) having the same potential as the shielding portion of the wiring 200. For example, by bonding the bearing housing 40 and the shielding portion of the wiring 200 with a conductive adhesive, both can be made to have the same ground (GND) potential. Examples of conductive adhesives include pastes in which particles of silver, nickel, gold, copper, carbon black, etc. are dispersed in the adhesive.

[0055] In this way, in this example, the strain gauges 100 are positioned radially inward of the inner diameter of the stator core 52, which is a position where strain can be easily detected, so even slight strain can be detected. In other words, if the outer ring 31 of the rolling bearing 30 is distorted even slightly, the bearing housing 40 will also be distorted, and this slight strain can be detected by the strain gauges 100 fixed to the recesses 40x provided in the outer peripheral surface 40a of the bearing housing 40. This makes it possible to improve the accuracy of abnormality diagnosis of the rolling bearing 30 based on the output of the strain gauges 100.

[0056] In this example, the strain gauge 100 is fixed to a recess (thin portion 42) provided in the outer peripheral surface 40a of the bearing housing 40, and the strain gauge 100 does not protrude beyond the outer peripheral surface 40a. Therefore, after the strain gauge 100 is fixed to the bearing housing 40, the stator core 52 can be press-fitted onto the outer periphery of the bearing housing 40. At this time, there is no interference between the inner peripheral surface of the stator core 52 and the strain gauge 100.

[0057] In this example, for example, the wiring 200 is passed through a notch 52x formed in the inner periphery of the stator core 52, and at least the portion of the wiring 200 that passes through the notch 52x is shielded. This makes it possible to significantly reduce the influence of electromagnetic noise compared to, for example, passing the wiring 200 outside the stator core 52. As a result, the accuracy of abnormality diagnosis of the rolling bearing 30 can be improved.

[0058] <Second example of bearing device structure> 8 is a vertical cross-sectional view showing another example of the structure of the motor EM (a second example of the bearing housing 40). The following description will focus on parts that are different from the first example described above, and descriptions of the same or corresponding content may be omitted.

[0059] 8, in this example, unlike the first example described above, the thin-walled portion 42 is a recess provided in a partial range including the axial end and a partial range in the radial direction on the cylindrical outer peripheral surface 40a corresponding to the thick-walled portion 41. In other words, the thin-walled portion 42 may be provided so as to reach the axial end.

[0060] <Third example of bearing device structure> 9 is a vertical cross-sectional view showing yet another example of the structure of the motor EM (a third example of the bearing housing 40). The following description will focus on the parts that are different from the first and second examples described above, and may omit a description of the same or corresponding content.

[0061] 9, unlike the first example and the like, the thin-walled portion 42 is a recess provided in a partial range including the axial end and in a range extending over the entire circumference in the radial direction on the cylindrical outer peripheral surface 40a corresponding to the thick-walled portion 41. In other words, the thin-walled portion 42 may be provided over the entire circumference in the circumferential direction.

[0062] <Fourth example of bearing device structure> 10 is a vertical cross-sectional view showing another example of the structure of the bearing device BE (a fourth example of the bearing housing 40). The following description will focus on the parts that are different from the first to third examples described above, and descriptions of the same or corresponding content may be omitted.

[0063] As shown in FIG. 10, in this example, the arrangement of the thick portion 41 is determined in relation to the preload angle θ2.

[0064] The preload angle θ2 is the contact angle of the rolling elements 33 with the outer ring 31 and the inner ring 32, determined by the preload applied to the outer ring 31 and the inner ring 32. Applying an appropriate preload to the outer ring 31 and the inner ring 32 can contribute to improving the runout accuracy of the rotating shaft m and reducing vibration and noise.

[0065] Specifically, the preload angle θ2 is the angle between a line A connecting the contact point between the outer ring 31 and the rolling element 33 and the contact point between the inner ring 32 and the rolling element 33, and a line B extending in the radial direction, when viewed in cross section.

[0066] The thick portion 41 of the bearing housing 40 is formed in at least the region D of the outer peripheral surface of the outer ring 31. O Area D is arranged so as to be in contact with O is the intersection point C in the cross section. O From the intersection point CO The area of ​​the outer peripheral surface of the outer ring 31 extends to the axial end face of the outer ring 31 closest to the intersection point C (hereinafter referred to as the "preload side end face"). O is the intersection of the extension of the straight line A and the outer peripheral surface of the outer ring 31 in a cross-sectional view. O is a region where the displacement during rotation of the rolling element 33 is relatively large. Hereinafter, the extension of the straight line A may be referred to as the "straight line indicating the contact angle."

[0067] In this way, the thick portion 41 of the bearing housing 40 is located in the region D where the displacement of the rolling elements 33 during rotation is relatively large. O This makes it possible to suppress a decrease in the rigidity of the shaft member (rotor shaft 20) inserted into the bearing device BE, which is caused by the structure of the portion where the strain gauge 100 is disposed, and ensures the rigidity of the shaft member.

[0068] In addition, the area D on the outer peripheral surface of the outer ring 31 O In the region excluding this, the displacement during rotation of the rolling element 33 is relatively small. This allows the thin-walled portion 42 of the bearing housing 40 to be disposed in this axial region. Therefore, the relatively thin thickness of the thin-walled portion 42 allows the strain gauge 100 to detect with high sensitivity the strain that occurs in the outer ring 31 when the rolling element 33 rotates.

[0069] In this example, the two rolling bearings 30 are arranged with their preload-side end faces facing each other at a predetermined distance so that their rotation axes substantially coincide with the rotation axis m. Therefore, thin-walled portions 42 are arranged at both ends of the bearing housing 40 in the axial direction, and thick-walled portions 41 are arranged inside them. Therefore, as shown in Fig. 10, the regions in the bearing housing 40 between the rolling bearings 30 in the axial direction may be configured as thick-walled portions 41.

[0070] The combination of the two rolling bearings 30 is a so-called back-to-back combination (DB) in terms of the preload direction, and the intersection of the straight line A and the rotation axis m faces outward, resulting in high rigidity.

[0071] <Fifth example of bearing device structure> 11 is a vertical cross-sectional view showing yet another example of the structure of the bearing device BE (a fifth example of the bearing housing 40). The following description will focus on the parts that are different from the first to fourth examples described above, and descriptions of the same or corresponding content may be omitted.

[0072] As shown in FIG. 11, in this example, the arrangement of the thick portion 41 is determined in relation to the preload angle θ2, similar to the fourth example described above.

[0073] Specifically, similarly to the fourth example described above, the thick portion 41 of the bearing housing 40 is located in the region D where the displacement of the rolling element 33 during rotation is relatively large. O This provides the same effects and advantages as the fourth example described above.

[0074] In addition, the area D on the outer peripheral surface of the outer ring 31 O In the region other than this, the displacement of the rolling elements 33 during rotation is relatively small. This allows the thin-walled portion 42 of the bearing housing 40 to be disposed in this axial region, thereby achieving the same functions and effects as those of the fourth example described above.

[0075] On the other hand, in this example, unlike the above-described fourth example, the two rolling bearings 30 are arranged at a predetermined distance so that their rotation axes coincide with the rotation axis m, and their preload-side end faces face outward. Therefore, thick-walled portions 41 can be arranged at both ends of the bearing housing 40, and thin-walled portions 42 can be arranged inside them. Therefore, as shown in FIG. 11 , the areas in the bearing housing 40 between the rolling bearings 30 in the axial direction can be configured as thin-walled portions 42.

[0076] In the axial direction, the areas between the rolling bearings 30 in the bearing housing 40 may be configured as thick portions 41.

[0077] The combination of the two rolling bearings 30 is a so-called face-to-face combination (DF) in terms of the preload direction, and the intersection of the straight line A and the rotation axis m faces inward, so there is a large tolerance for mounting errors.

[0078] <Sixth example of bearing device structure> 12 is a vertical cross-sectional view showing yet another example of the structure of the bearing device BE (sixth example of the bearing housing 40). The following description will focus on the parts that are different from the first to fifth examples described above, and descriptions of the same or corresponding content may be omitted.

[0079] 12, unlike the first example and the like, the bearing housing 40 includes a cylindrical large-diameter portion 40A and a cylindrical small-diameter portion 40B. The small-diameter portion 40B is disposed axially over the entire range between the outer end faces of the two rolling bearings 30, and the large-diameter portion 40A is disposed on a portion of the outer periphery of the small-diameter portion 40B. The small-diameter portion 40B has an inner diameter approximately equal to the outer diameter of the outer ring 31 and an axial length approximately equal to the distance between the outer end faces of the two rolling bearings 30. The large-diameter portion 40A has an inner diameter approximately equal to the outer diameter of the small-diameter portion 40B and an axial length shorter than the axial length of the small-diameter portion 40B.

[0080] The large diameter portion 40A and the small diameter portion 40B are integrated together by, for example, press-fitting or bonding. In the bearing housing 40, the portion where the large diameter portion 40A is laminated on the outer peripheral surface of the small diameter portion 40B corresponds to the thick portion 41, and the portion consisting only of the small diameter portion 40B corresponds to the thin portion 42. A strain gauge 100 is disposed on the outer peripheral surface of the small diameter portion 40B as the thin portion 42, with an adhesive layer interposed between them.

[0081] In this way, the bearing housing 40 is not limited to being one piece, but may be formed by joining separate members together. For example, the bearing housing 40 may be formed by joining separate members with different diameters together.

[0082] <Seventh example of bearing device structure> 13 is a vertical cross-sectional view showing yet another example of the structure of the bearing device BE (seventh example of the bearing housing 40). The following description will focus on the parts that are different from the first to sixth examples described above, and descriptions of the same or corresponding content may be omitted.

[0083] 13 , unlike the first example and the like, the bearing housing 40 includes a cylindrical thick-walled portion 41 and a cylindrical thin-walled portion 42 that is thinner in the radial direction than the thick-walled portion 41. The thick-walled portion 41 has an inner diameter that is approximately equal to the outer diameter of the outer ring 31, and an axial length that is approximately equal to the distance between the outer end faces of the two rolling bearings 30 in the axial direction. In other words, the thick-walled portion 41 is arranged so as to contact the entire outer peripheral surface of the outer ring 31 of each of the two rolling bearings 30.

[0084] The thin portion 42 has an inner diameter substantially equal to the outer diameter of the outer ring 31, and protrudes in a direction substantially parallel to the rotation axis m (axial direction) from the radial side of the end face of the thick portion 41 closer to the rotation axis m. The thick portion 41 and the thin portion 42 are, for example, integrally molded. A strain gauge 100 is disposed on the outer peripheral surface of the thin portion 42 via an adhesive layer.

[0085] In this example, the combination of the two rolling bearings 30 is a so-called face-to-face combination (DF) in the preload direction, and both axial end faces of the thick-walled portion 41 are substantially aligned with the preload-side end faces of the rolling bearings 30 in the axial direction. This makes it easier for strain in the rolling bearings 30 to be transmitted to the thin-walled portion 42, and the strain gauge 100 can detect strain in the rolling bearings 30 with higher sensitivity.

[0086] The thin portion 42 may protrude from the end face opposite to the preload side end face. That is, the combination of the two rolling bearings 30 may be a so-called back-to-back (DB) combination in terms of the preload direction. The same applies to the eighth example described below.

[0087] In this way, the axial length of the thick portion 41 may be approximately equal to the distance between the outer end faces of the two rolling bearings 30, and the thin portion 42 may protrude radially from the end face of the thick portion 41 closer to the rotation axis m in a direction approximately parallel to the direction of the rotation axis m. As a result, as in the first example and the like described above, by arranging the strain gauge 100 in the thin portion 42, the strain gauge 100 can detect strain in the outer ring 31 via the thin portion 42, which is a strain transmission part. In addition, the thick portion 41 can be arranged so as to contact the entire outer peripheral surfaces of the outer rings 31 of the two rolling bearings 30. This makes it possible to sufficiently ensure the rigidity of the shaft member (rotor shaft 20) inserted into the bearing device BE.

[0088] <Eighth example of bearing device structure> 14 is a vertical cross-sectional view showing yet another example of the structure of the bearing device BE (an eighth example of the bearing housing 40). The following description will focus on the parts that are different from the above-described first to seventh examples, and descriptions of the same or corresponding content may be omitted.

[0089] 14 , in this example, the bearing housing 40, like the seventh example described above, includes a cylindrical thick-walled portion 41 and a cylindrical thin-walled portion 42 that is thinner in the radial direction than the thick-walled portion 41. Like the seventh example described above, the thick-walled portion 41 has an inner diameter that is approximately equal to the outer diameter of the outer ring 31, and an axial length that is approximately equal to the distance between the outer end faces of the two rolling bearings 30 in the axial direction. In other words, the thick-walled portion 41 is arranged so as to contact the entire outer peripheral surface of the outer ring 31 of each of the two rolling bearings 30.

[0090] On the other hand, unlike the seventh example described above, the thin portion 42 protrudes in a direction (axial direction) substantially parallel to the rotation axis m from the radially far side of the end face of the thick portion 41 from the rotation axis m. The thick portion 41 and the thin portion 42 are, for example, integrally molded. A strain gauge 100 is disposed on the inner peripheral surface of the thin portion 42 via an adhesive layer.

[0091] The strain gauge 100 may be disposed on the outer peripheral surface of the thin portion 42 .

[0092] In this way, the axial length of the thick portion 41 may be approximately equal to the distance between the outer end faces of the two rolling bearings 30, and the thin portion 42 may protrude in a direction approximately parallel to the direction of the rotation axis m from the radially farther side of the end face of the thick portion 41 from the rotation axis m. This provides the same functions and effects as the seventh example described above.

[0093] <Ninth example of bearing device structure> 15 is a vertical cross-sectional view showing yet another example of the structure of the bearing device BE (a ninth example of the bearing housing 40). The following description will focus on the parts that are different from the above-mentioned first to eighth examples, and descriptions of the same or corresponding content may be omitted.

[0094] 15 , the bearing housing 40, like the seventh and eighth examples described above, has a cylindrical thick-walled portion 41 and a cylindrical thin-walled portion 42 that is thinner than the thick-walled portion 41. The thick-walled portion 41 has an inner diameter that is approximately equal to the outer diameter of the outer ring 31, and an axial length that is approximately equal to the distance between the outer end faces of the two rolling bearings 30 in the axial direction. In other words, the thick-walled portion 41 is arranged so as to contact the entire outer peripheral surface of the outer ring 31 of each of the two rolling bearings 30.

[0095] On the other hand, unlike the seventh and eighth examples described above, the thin portion 42 extends from the end face of the thick portion 41 in a direction approximately perpendicular to the rotation axis m and is arranged in annular contact with a part of the end face of the outer ring 31. The axial thickness of the thin portion 42 is thinner than the radial thickness of the thick portion 41. The thick portion 41 and the thin portion 42 are, for example, integrally molded. A strain gauge 100 is arranged on the surface of the thin portion 42 opposite the side that contacts the end face of the outer ring 31, via an adhesive layer.

[0096] In this way, the thin-walled portion 42 of the bearing housing 40 may be disposed in contact with the end face of the outer ring 31. This provides the same functions and effects as the seventh and eighth examples described above.

[0097] <10th example of bearing device structure> 16 is a perspective view showing a tenth example of the bearing housing 40. The following description will focus on the parts that are different from the first to ninth examples described above, and descriptions of the same or corresponding content may be omitted.

[0098] As shown in FIG. 16, in this example, a fixing reinforcement portion 43 is provided on the base end side of both axial ends of the bearing housing 40, that is, on the end on the side fixed to the casing .

[0099] The fixing reinforcement portion 43 relatively increases the fixing strength between the bearing housing 40 and the casing 70. In this example, the fixing reinforcement portion 43 is a flange. The flange as the fixing reinforcement portion 43 can fix the bearing housing 40 to the casing 70, thereby improving the fixing strength.

[0100] The reinforcing fixing portion 43 is also provided with a wire passing portion 43x.

[0101] The wire passing portion 43x is a path portion for leading the wires 200 to the outside of the axial fan 1 (motor EM). In this example, the wire passing portion 43x is a through-hole that axially passes through the fixing reinforcement portion 43 (flange). This allows the wires 200 to be led from the inside to the outside of the axial fan 1 (motor EM) through the through-hole that serves as the wire passing portion 43x.

[0102] <11th example of bearing device structure> 17 is a perspective view showing an eleventh example of the bearing housing 40. The following description will focus on the differences from the first to tenth examples described above, and description of the same or corresponding content may be omitted.

[0103] As shown in Figure 17, in this example, a fixing reinforcement portion 43 is provided on the base end side of both axial ends of the bearing housing 40, i.e., the end side fixed to the casing 70, as in the above-mentioned tenth example.

[0104] In this example, the reinforced fixing portion 43 is a cylindrical press-fit portion with a knurled outer surface, whereby the reinforced fixing portion 43 as a press-fit portion can improve the fixation strength due to the press-fitting by the action of the knurling.

[0105] Furthermore, the reinforcing fixing portion 43 is provided with a wire passing portion 43x, similar to the tenth example described above.

[0106] The wire passing portion 43x is a cutout provided over the entire axial range and a part of the circumferential range of the cylindrical press-fit portion serving as the fixing reinforcement portion 43. This allows the wires 200 to be drawn from the inside to the outside of the axial fan 1 (motor EM) through the cutout of the press-fit portion serving as the wire passing portion 43x.

[0107] <12th example of bearing device structure> 18 is a perspective view showing a twelfth example of the bearing housing 40. The following description will focus on the differences from the first to eleventh examples described above, and description of the same or corresponding content may be omitted.

[0108] As shown in Figure 18, in this example, similar to the above-mentioned tenth and eleventh examples, a fixing reinforcement portion 43 is provided on the base end side of both axial end portions of the bearing housing 40, i.e., the end portion fixed to the casing 70, similar to the above-mentioned tenth example.

[0109] In this example, the fixing reinforcement portion 43 is a cylindrical press-fit portion, and its outer circumferential surface is knurled, similar to the eleventh example described above.

[0110] As in the eleventh example, the reinforced fixing portion 43 is provided with wire passing portions 43x as notches that are provided over the entire axial range and a portion of the circumferential range of the cylindrical press-fit portion serving as the reinforced fixing portion 43. Specifically, a plurality of (three) notches (wire passing portions 43x) are provided that are arranged at approximately equal intervals in the circumferential direction of the reinforced fixing portion 43. This allows the wires 200 to be drawn from the inside to the outside of the axial fan 1 (motor EM) through the plurality of notches in the press-fit portion serving as the wire passing portions 43x.

[0111] [Strain gauge details] Next, the strain gauge 100 will be described in detail with reference to FIGS.

[0112] Fig. 19 is a plan view showing an example of the strain gauge 100. Fig. 20 is a cross-sectional view showing an example of the strain gauge 100. Specifically, Fig. 20 shows a cross section taken along line AA in Fig. 19.

[0113] As shown in FIGS. 19 and 20 , the strain gauge 100 includes a substrate 101, a functional layer 102, a resistor 103, wiring 104, and a terminal portion 105. Hereinafter, for convenience, the side of the substrate 101 on which the resistor 103 is provided will be referred to as the upper side or one side, and the side on which the resistor 103 is not provided will be referred to as the lower side or the other side. Furthermore, the surface on which the resistor 103 is provided will be referred to as the one side or upper side, and the surface on which the resistor 103 is not provided will be referred to as the other side or lower side. Furthermore, a planar view refers to viewing the object from the normal direction of the upper surface 101 a of the substrate 101, and a planar shape refers to the shape of the object viewed from the normal direction of the upper surface 101 a of the substrate 101.

[0114] The strain gauge 100 can be used upside down and can be placed at any angle. The functional layer 102 may be provided as needed.

[0115] The substrate 101 is a flexible member that serves as a base layer for forming the resistor 103 and the like. The thickness of the substrate 101 is not particularly limited and can be appropriately selected depending on the purpose, but can be, for example, about 5 μm to 500 μm. In particular, a thickness of 5 μm to 200 μm is preferable in terms of the transferability of strain from the surface of the strain-generating element (e.g., the outer peripheral surface 40a of the bearing housing 40) joined to the lower surface of the substrate 101 via the adhesive layer 150 and dimensional stability against the environment, and a thickness of 10 μm or more is even more preferable in terms of insulation.

[0116] The substrate 101 can be formed from an insulating resin film such as PI (polyimide) resin, epoxy resin, PEEK (polyether ether ketone) resin, PEN (polyethylene naphthalate) resin, PET (polyethylene terephthalate) resin, PPS (polyphenylene sulfide) resin, or polyolefin resin.

[0117] The film refers to a flexible member having a thickness of about 500 μm or less.

[0118] Here, "formed from an insulating resin film" does not prevent the base material 101 from containing fillers, impurities, etc. in the insulating resin film. The base material 101 may be formed from an insulating resin film containing fillers such as silica or alumina, for example.

[0119] Examples of materials other than resin for the substrate 101 include crystalline materials such as SiO2, ZrO2 (including YSZ), Si, Si2N3, Al2O3 (including sapphire), ZnO, perovskite ceramics (CaTiO3, BaTiO3), and amorphous glass. Alternatively, the substrate 101 may be made of metal such as aluminum, aluminum alloy (duralumin), or titanium. In this case, an insulating film, for example, is formed on the metal substrate 101.

[0120] The functional layer 102 is formed on the upper surface 101a of the substrate 101 as a lower layer of the resistor 103. That is, the planar shape of the functional layer 102 is substantially the same as the planar shape of the resistor 103 shown in FIG.

[0121] In this application, the functional layer refers to a layer having a function of promoting the crystal growth of at least the upper layer, the resistor 103. The functional layer 102 preferably also has a function of preventing oxidation of the resistor 103 due to oxygen or moisture contained in the substrate 101, and a function of improving adhesion between the substrate 101 and the resistor 103. The functional layer 102 may also have other functions.

[0122] The insulating resin film that constitutes the substrate 101 contains oxygen and moisture, and since Cr forms a self-oxidized film, especially when the resistor 103 contains chromium (Cr), it is effective for the functional layer 102 to have the function of preventing oxidation of the resistor 103.

[0123] The material of the functional layer 102 is not particularly limited as long as it has the function of promoting the crystal growth of at least the upper layer, the resistor 103, and can be appropriately selected depending on the purpose. For example, Cr (chromium), Ti (titanium), V (vanadium), Nb (niobium), Ta (tantalum), Ni (nickel), Y (yttrium), Zr (zirconium), Hf (hafnium), Si (silicon), C (carbon), Zn (zinc), Cu (copper), Bi (bismuth), Examples of the metal include one or more metals selected from the group consisting of Fe (iron), Mo (molybdenum), W (tungsten), Ru (ruthenium), Rh (rhodium), Re (rhenium), Os (osmium), Ir (iridium), Pt (platinum), Pd (palladium), Ag (silver), Au (gold), Co (cobalt), Mn (manganese), and Al (aluminum), an alloy of any of the metals in this group, or a compound of any of the metals in this group.

[0124] Examples of the alloys include FeCr, TiAl, FeNi, NiCr, CrCu, etc. Examples of the compounds include TiN, TaN, Si3N4, TiO2, Ta2O5, SiO2, etc.

[0125] When the functional layer 102 is made of a conductive material such as a metal or alloy, the thickness of the functional layer 102 is preferably 1 / 20 or less of the thickness of the resistor. This range promotes the crystal growth of α-Cr and prevents a portion of the current flowing through the resistor from flowing through the functional layer 102, which would otherwise reduce the strain detection sensitivity.

[0126] When the functional layer 102 is made of a conductive material such as a metal or alloy, the thickness of the functional layer 102 is preferably 1 / 50 or less of the thickness of the resistor. This range promotes the crystal growth of α-Cr and further prevents a portion of the current flowing through the resistor from flowing through the functional layer 102, which would otherwise reduce the strain detection sensitivity.

[0127] When the functional layer 102 is made of a conductive material such as a metal or alloy, it is more preferable that the thickness of the functional layer 102 be 1 / 100 or less of the thickness of the resistor. In this range, it is possible to further prevent a portion of the current flowing through the resistor from flowing through the functional layer 102, thereby preventing a decrease in strain detection sensitivity.

[0128] When the functional layer 102 is made of an insulating material such as an oxide or nitride, the thickness of the functional layer 102 is preferably 1 nm to 1 μm, which not only promotes crystal growth of α-Cr but also allows the functional layer 102 to be easily formed without cracking.

[0129] When the functional layer 102 is made of an insulating material such as an oxide or nitride, it is more preferable that the thickness of the functional layer 102 be 1 nm to 0.8 μm. This range not only promotes crystal growth of α-Cr, but also makes it easier to form the functional layer 102 without cracking.

[0130] When the functional layer 102 is made of an insulating material such as an oxide or nitride, it is more preferable that the thickness of the functional layer 102 be 1 nm to 0.5 μm. This range promotes crystal growth of α-Cr and allows the functional layer 102 to be formed more easily without cracks.

[0131] The planar shape of the functional layer 102 is patterned to be substantially the same as the planar shape of the resistor shown in FIG. 10, for example. However, the planar shape of the functional layer 102 is not limited to being substantially the same as the planar shape of the resistor. If the functional layer 102 is formed from an insulating material, it does not need to be patterned to be the same as the planar shape of the resistor. In this case, the functional layer 102 may be formed in a solid state at least in the region where the resistor is formed. Alternatively, the functional layer 102 may be formed in a solid state over the entire upper surface of the substrate 101.

[0132] Furthermore, when the functional layer 102 is made of an insulating material, the functional layer 102 can be formed relatively thick, at a thickness of 50 nm to 1 μm, and formed in a solid state, thereby increasing the thickness and surface area of ​​the functional layer 102, allowing heat generated by the resistor to be dissipated to the substrate 101. As a result, the deterioration of measurement accuracy in the strain gauge 100 due to self-heating of the resistor can be suppressed.

[0133] The resistor 103 is a thin film formed in a predetermined pattern on the upper surface of the functional layer 102, and is a sensing part that generates a resistance change when subjected to strain.

[0134] The resistor 103 can be formed from, for example, a material containing Cr (chromium), a material containing Ni (nickel), or a material containing both Cr and Ni. That is, the resistor 103 can be formed from a material containing at least one of Cr and Ni. An example of a material containing Cr is a Cr mixed phase film. An example of a material containing Ni is Cu-Ni (copper-nickel). An example of a material containing both Cr and Ni is Ni-Cr (nickel-chromium).

[0135] Hereinafter, the resistor 103 will be described as an example in which it is a Cr mixed-phase film. Here, the Cr mixed-phase film is a film in which Cr, CrN, CrN, etc. are mixed. The Cr mixed-phase film may contain inevitable impurities such as chromium oxide. Furthermore, a portion of the material constituting the functional layer 102 may be diffused into the Cr mixed-phase film. In this case, the material constituting the functional layer 102 may form a compound with nitrogen. For example, if the functional layer 102 is made of Ti, the Cr mixed-phase film may contain Ti or TiN (titanium nitride).

[0136] The thickness of resistor 103 is not particularly limited and can be appropriately selected depending on the purpose, but can be, for example, about 0.05 μm to 2 μm. In particular, a thickness of resistor 103 of 0.1 μm or more is preferable in that the crystallinity of the crystals constituting resistor 103 (for example, the crystallinity of α-Cr) is improved, and a thickness of 1 μm or less is even more preferable in that cracks in the film constituting resistor 103 and warping from substrate 101 caused by internal stress in the film can be reduced.

[0137] By forming the resistor 103 on the functional layer 102, the resistor 103 can be formed using a stable crystalline phase, thereby improving the stability of the gauge characteristics (gauge factor, temperature coefficient of gauge factor TCS, and temperature coefficient of resistance TCR).

[0138] For example, when the resistor 103 is a Cr mixed phase film, the resistor 103 can be formed with α-Cr (alpha chromium) as the main component by providing the functional layer 102. Since α-Cr is a stable crystalline phase, the stability of the gauge characteristics can be improved.

[0139] Here, the term "main component" means that the target substance accounts for 50% by mass or more of all substances constituting the resistor. When the resistor 103 is a Cr mixed phase film, the resistor 103 preferably contains 80% by weight or more, and more preferably 90% by weight or more, of α-Cr in order to improve the gauge characteristics.

[0140] Incidentally, α-Cr is Cr with a bcc structure (body-centered cubic lattice structure).

[0141] Furthermore, when the resistor 103 is a Cr mixed phase film, the Cr mixed phase film preferably contains 20 wt % or less of CrN and Cr2N, which can suppress a decrease in the gauge factor.

[0142] Furthermore, the ratio of Cr2N in CrN and Cr2N is preferably 80% by weight or more and less than 90% by weight, and more preferably 90% by weight or more and less than 95% by weight. When the ratio of Cr2N in CrN and Cr2N is 90% by weight or more and less than 95% by weight, the Cr2N has semiconducting properties, which leads to a more significant decrease in TCR (negative TCR). Furthermore, by reducing the amount of ceramic formation, brittle fracture is reduced.

[0143] On the other hand, if trace amounts of N2 or atomic N are mixed into or present in the film, they will escape to the outside of the film due to external conditions (such as high temperature environments), causing changes in film stress.By creating chemically stable CrN, the unstable N mentioned above will not be generated, and a stable strain gauge can be obtained.

[0144] Furthermore, the gauge characteristics can be improved by diffusing the metal (e.g., Ti) constituting the functional layer 102 into the Cr mixed phase film. Specifically, the gauge factor of the strain gauge 100 can be set to 10 or more, and the temperature coefficient of gauge factor TCS and the temperature coefficient of resistance TCR can be set to within the range of −1000 ppm / °C to +1000 ppm / °C.

[0145] The terminal portions 105 extend from both ends of the resistor 103 via the wiring 104, and are formed in a generally rectangular shape wider than the resistor 103 and the wiring 104 in a plan view. The terminal portions 105 are a pair of electrodes for outputting to the outside a change in the resistance value of the resistor 103 caused by strain. For example, the resistor 103 extends from one of the terminal portion 105 and the wiring 104 while folding back in a zigzag pattern, and is connected to the other wiring 104 and the terminal portion 105. The upper surface of the terminal portion 105 may be covered with a metal that has better solderability than the terminal portion 105.

[0146] Although the resistor 103, the wiring 104, and the terminal portion 105 are denoted by different reference numerals for convenience, they can be integrally formed from the same material in the same process.

[0147] A cover layer 106 (insulating resin layer) may be provided on the upper surface 101a of the substrate 101 so as to cover the resistor 103 and the wiring 104 and expose the terminal portion 105. By providing the cover layer 106, it is possible to prevent mechanical damage, etc. from occurring to the resistor 103 and the wiring 104. Furthermore, by providing the cover layer 106, it is possible to protect the resistor 103 and the wiring 104 from moisture, etc. Note that the cover layer 106 may be provided so as to cover the entire portion except for the terminal portion 105.

[0148] The cover layer 106 can be formed from an insulating resin such as PI resin, epoxy resin, PEEK resin, PEN resin, PET resin, PPS resin, or composite resin (e.g., silicone resin or polyolefin resin). The cover layer may contain a filler or pigment. There are no particular restrictions on the thickness of the cover layer, and it can be appropriately selected depending on the purpose, but it can be, for example, about 2 μm to 30 μm.

[0149] To manufacture the strain gauge 100, first, a substrate 101 is prepared, and a functional layer 102 is formed on an upper surface 101a of the substrate 101. The materials and thicknesses of the substrate 101 and the functional layer 102 are as described above. However, the functional layer 102 may be provided as needed.

[0150] The functional layer 102 can be formed in vacuum by conventional sputtering, for example, using a target made of a material capable of forming the functional layer 102 and introducing Ar (argon) gas into a chamber. By using conventional sputtering, the functional layer 102 is formed while etching the upper surface 101a of the substrate 101 with Ar, thereby minimizing the amount of the functional layer 102 formed and achieving an improvement in adhesion.

[0151] However, this is just one example of a method for forming the functional layer 102, and the functional layer 102 may be formed by other methods. For example, a method may be used in which the upper surface 101a of the substrate 101 is activated by plasma treatment using Ar or the like before forming the functional layer 102, thereby improving adhesion, and then the functional layer 102 is vacuum-formed by magnetron sputtering.

[0152] Next, a metal layer that will become the resistor 103, wiring 104, and terminal portion 105 is formed over the entire upper surface of the functional layer 102, and then the functional layer 102, resistor 103, wiring 104, and terminal portion 105 are patterned by photolithography into the planar shape shown in FIG. 10 . The materials and thicknesses of the resistor 103, wiring 104, and terminal portion 105 are as described above. The resistor 103, wiring 104, and terminal portion 105 can be integrally formed using the same material. The resistor 103, wiring 104, and terminal portion 105 can be formed by, for example, magnetron sputtering using a target material capable of forming the resistor 103, wiring 104, and terminal portion 105. The resistor 103, wiring 104, and terminal portion 105 may also be formed by reactive sputtering, vapor deposition, arc ion plating, pulsed laser deposition, or the like instead of magnetron sputtering.

[0153] There are no particular restrictions on the combination of the material of the functional layer 102 with the materials of the resistor 103, wiring 104, and terminal portion 105, and these can be selected appropriately depending on the purpose. For example, Ti can be used for the functional layer 102, and a Cr mixed phase film with α-Cr (alpha chromium) as the main component can be formed for the resistor 103, wiring 104, and terminal portion 105.

[0154] In this case, for example, the resistor 103, the wiring 104, and the terminal portion 105 can be formed by magnetron sputtering using a target made of a material capable of forming a Cr mixed phase film and introducing Ar gas into a chamber. Alternatively, the resistor 103, the wiring 104, and the terminal portion 105 can be formed by reactive sputtering using pure Cr as a target and introducing an appropriate amount of nitrogen gas together with Ar gas into a chamber. In this case, the ratios of CrN and CrN contained in the Cr mixed phase film and the ratio of CrN in CrN and CrN can be adjusted by changing the amount and pressure (nitrogen partial pressure) of the nitrogen gas introduced or by adjusting the heating temperature by providing a heating step.

[0155] In these methods, the Ti functional layer 102 defines the growth plane of the Cr mixed-phase film, allowing the formation of a Cr mixed-phase film primarily composed of α-Cr, which has a stable crystal structure. Furthermore, the Ti constituting the functional layer 102 diffuses into the Cr mixed-phase film, improving the gauge characteristics. For example, the gauge factor of the strain gauge 100 can be set to 10 or more, and the temperature coefficient of gauge factor (TCS) and temperature coefficient of resistance (TCR) can be set within the ranges of -1000 ppm / °C to +1000 ppm / °C.

[0156] When the resistor 103 is a Cr mixed phase film, the functional layer 102 made of Ti has all of the following functions: promoting crystal growth of the resistor 103, preventing oxidation of the resistor 103 due to oxygen and moisture contained in the substrate 101, and improving adhesion between the substrate 101 and the resistor 103. The same applies when Ta, Si, Al, or Fe is used as the functional layer 102 instead of Ti.

[0157] Thereafter, if necessary, a cover layer 106 that covers the resistor 103 and the wiring 104 and exposes the terminal portions 105 is provided on the upper surface 101a of the substrate 101, thereby completing the strain gauge 100. The cover layer 106 can be produced, for example, by laminating a semi-cured thermosetting insulating resin film on the upper surface 101a of the substrate 101 so as to cover the resistor 103 and the wiring 104 and expose the terminal portions 105, and then heating and curing the film. The cover layer 106 may also be produced by applying a liquid or paste-like thermosetting insulating resin to the upper surface 101a of the substrate 101 so as to cover the resistor 103 and the wiring 104 and expose the terminal portions 105, and then heating and curing the resin.

[0158] In this way, by providing the functional layer 102 below the resistor 103, it is possible to promote crystal growth of the resistor 103, and to produce a resistor 103 consisting of a stable crystalline phase. As a result, it is possible to improve the stability of the gauge characteristics of the strain gauge 100. Furthermore, by diffusing the material that constitutes the functional layer 102 into the resistor 103, it is possible to improve the gauge characteristics of the strain gauge 100.

[0159] The strain gauge 100, which uses a Cr mixed-phase film as the material for the resistor 103, achieves high sensitivity (500% or more compared to conventional strain gauges) and miniaturization (less than 1 / 10 the size of conventional strain gauges). For example, while the output of conventional strain gauges was approximately 0.04 mV / 2 V, the strain gauge 100 can obtain an output of 0.3 mV / 2 V or more. Furthermore, while the size of conventional strain gauges (gauge length x gauge width) was approximately 3 mm x 3 mm, the size of the strain gauge 100 (gauge length x gauge width) can be miniaturized to approximately 0.3 mm x 0.3 mm.

[0160] As described above, the strain gauge 100 using the Cr mixed-phase film as the material for the resistor 103 is small and can be easily attached to the recess 40x provided on the outer peripheral surface 40a of the bearing housing 40. Therefore, it is particularly suitable for use in a motor EM (bearing device BE) using a small rolling bearing 30 with a diameter (outer diameter of the outer ring 31) of 30 mm or less. Furthermore, the strain gauge 100 using the Cr mixed-phase film as the material for the resistor 103 is highly sensitive and can detect small displacements, making it possible to detect minute strains that were previously difficult to detect. In other words, the strain gauge 100 using the Cr mixed-phase film as the material for the resistor 103 can realize an axial fan 1 (motor EM) with the ability to accurately detect strain.

[0161] [Anomaly monitoring system] Next, the abnormality monitoring system 300 will be described with reference to FIGS.

[0162] Fig. 21 is a diagram showing an example of the configuration of the abnormality monitoring system 300. Fig. 22 is a diagram showing an example of the hardware configuration of the monitoring device 5. Fig. 23 is a diagram showing an example of time-series data of the strain gauge 100.

[0163] As shown in FIG. 21, the abnormality monitoring system 300 includes an axial fan 1, a driving device 2, an amplifier 3, a logger 4, and a monitoring device 5.

[0164] The abnormality monitoring system 300 monitors the bearing device BE for abnormalities based on the outputs of the two strain gauges 100 in the monitoring device 5.

[0165] The driving device 2 supplies a predetermined amount of electric power to the axial fan 1 (motor EM) to electrically drive the axial fan (motor EM).

[0166] The amplifier 3 amplifies the output of each of the two strain gauges 100 and outputs the amplified output to the logger 4 .

[0167] The logger 4 records the amplified outputs (hereinafter referred to as "amplified outputs") of the two strain gauges 100 in time series.

[0168] The monitoring device 5 (an example of an information processing device) monitors the presence or absence of an abnormality in the bearing device BE by diagnosing the abnormality in the bearing device BE based on the time-series data of the amplified outputs of the two strain gauges 100, which is recorded in the logger 4. The abnormality diagnosis of the bearing device BE by the monitoring device 5 may be performed in real time, for example, or may be performed after the fact when the axial fan 1 is stopped, based on the most recent operating data.

[0169] The functions of the logger 4 and the monitoring device 5 may be built into the axial fan 1 (motor EM). In this case, for example, a circuit board having the functions of the logger 4 and the monitoring device 5 may be included in the axial fan 1 (motor EM) as a product package set with the bearing device BE. The monitoring device 5 may also diagnose abnormalities in the bearing device BE for each of the multiple axial fans 1. In this case, for example, the monitoring device 5 may acquire time-series data of the amplified output of the two strain gauges 100 for each of the multiple axial fans 1 from multiple loggers 4 corresponding to the multiple axial fans 1. Alternatively, for example, one logger 4 may be installed for multiple axial fans 1, and the monitoring device 5 may acquire data of the amplified output of the two strain gauges 100 corresponding to the multiple axial fans 1 from that single logger 4.

[0170] The monitoring device 5 is, for example, a computer terminal. The computer terminal may be, for example, a stationary terminal device such as a desktop PC (Personal Computer), or a portable terminal device (mobile terminal) such as a laptop PC or a tablet terminal.

[0171] The monitoring device 5 is, for example, a server. The server may be, for example, an on-premise server or a cloud server that is installed in a location different from where the axial fan 1 is installed, or an edge server that is installed in the same location as where the axial fan 1 is installed or in a location relatively close to that location.

[0172] The monitoring device 5 is communicatively connected via a predetermined communication line. The predetermined communication line includes, for example, a point-to-point communication line. The predetermined communication line also includes, for example, a local area network (LAN) in the location where the axial fan 1 is installed. The predetermined communication line also includes, for example, a wide area network (WAN). Wide area networks include, for example, mobile communication networks terminated at base stations, satellite communication networks using satellites, and the Internet. The predetermined communication line also includes, for example, a short-range communication line based on a predetermined communication standard such as Bluetooth (registered trademark) or WiFi.

[0173] The functions of the monitoring device 5 may be realized by any hardware or any combination of hardware and software. For example, as shown in Fig. 22, the monitoring device 5 includes, as a hardware configuration, an external interface 5A, an auxiliary storage device 5B, a memory device 5C, a CPU (Central Processing Unit) 5D, a high-speed arithmetic unit 5E, a communication interface 5F, an input device 5G, and a display device 5H, which are connected to each other via a bus.

[0174] The external interface 5A functions as an interface for reading data from the recording medium 5AA and writing data to the recording medium 5AA. Examples of the recording medium 5AA include a flexible disk, a CD (Compact Disc), a DVD (Digital Versatile Disc), a BD (Blu-ray (registered trademark) Disc), an SD memory card, and a USB (Universal Serial Bus) memory. This allows the monitoring device 5 to read various data used in processing from the recording medium 5AA through the external interface 5A, store the data in the auxiliary storage device 5B, and install programs that realize various functions.

[0175] The monitoring device 5 may acquire various data and programs from external devices through the communication interface 5F.

[0176] The auxiliary storage device 5B stores various installed programs, as well as files and data necessary for various processes. The auxiliary storage device 5B includes, for example, a hard disc drive (HDD), a solid state drive (SSD), a flash memory, etc.

[0177] When an instruction to start a program is received, the memory device 5C reads and stores the program from the auxiliary storage device 5B. The memory device 5C includes, for example, a dynamic random access memory (DRAM) or a static random access memory (SRAM).

[0178] The CPU 5D executes various programs loaded from the auxiliary storage device 5B to the memory device 5C, and realizes various functions related to the monitoring device 5 in accordance with the programs.

[0179] The high-speed arithmetic unit 5E performs arithmetic processing at a relatively high speed in conjunction with the CPU 5D. The high-speed arithmetic unit 5E includes, for example, a GPU (Graphics Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), etc.

[0180] The high-speed calculation device 5E may be omitted depending on the required calculation processing speed.

[0181] The communication interface 5F is used as an interface for connecting to an external device so that the monitoring device 5 can communicate with the external device, such as the logger 4, through the communication interface 5F. The communication interface 5F may have multiple types of communication interfaces depending on the communication method between the connected device and the like.

[0182] The input device 5G receives various inputs from a user. For example, the input device 5G includes an input device (remote control device) for an operator to perform remote control.

[0183] The input device 5G includes, for example, an operation input device that accepts mechanical operation input from a user. The operation input device includes, for example, a button, a toggle, a lever, etc. The operation input device also includes, for example, a keyboard, a mouse, a touch panel mounted on the display device 5H, a touch pad provided separately from the display device 5H, etc.

[0184] The input device 5G also includes, for example, a voice input device capable of receiving voice input from the user. The voice input device includes, for example, a microphone capable of collecting the user's voice.

[0185] The input device 5G also includes, for example, a gesture input device capable of receiving a gesture input from a user, and a camera capable of capturing an image of a user's gesture.

[0186] The input device 5G also includes, for example, a biometric input device capable of accepting biometric input from a user. The biometric input device includes, for example, a camera capable of acquiring image data containing information about a user's fingerprint or iris.

[0187] The display device 5H displays an information screen or an operation screen to the user. For example, the display device 5H includes the above-mentioned remote control display device. The display device 5H is, for example, a liquid crystal display or an organic EL (Electroluminescence) display.

[0188] 21, the monitoring device 5 includes, as functional units, a data acquisition unit 501, an abnormality determination unit 502, a storage unit 503, and an abnormality response unit 504. The functions of the data acquisition unit 501, the abnormality determination unit 502, the abnormality response unit 504, etc. are realized, for example, by loading a program installed in the auxiliary storage device 5B into the memory device 5C and executing the program on the CPU 5D. Furthermore, the function of the storage unit 503 is realized, for example, by a storage area defined in the auxiliary storage device 5B, etc.

[0189] The data acquiring unit 501 acquires data of the amplified outputs of the two strain gauges 100 at the same timing from the logger 4. For example, the data acquiring unit 501 acquires time series data of the amplified outputs of the two strain gauges 100 for the same period, i.e., for a predetermined time period (e.g., one minute) at the same time, from the logger 4 (see FIG. 23). The data acquiring unit 501 may acquire the data of the amplified outputs of the two strain gauges 100 transmitted from the logger 4 by periodically transmitting a request to the logger 4, or may acquire the data of the amplified outputs of the two strain gauges 100 transmitted automatically from the logger 4 at regular intervals.

[0190] The abnormality determination unit 502 (an example of a diagnosing unit) compares the amplified output data of the two strain gauges 100 over the same period acquired by the data acquisition unit 501 to determine (diagnose) an abnormality in the bearing device BE, specifically, an abnormality in the rolling bearing 30. This is because it is extremely rare for an abnormality to occur simultaneously in both rolling bearings 30, and it is possible to discover an abnormality or signs of an abnormality in one of the two rolling bearings 30 based on the difference in the strain states of the two rolling bearings 30 (outer rings 31). Specifically, the abnormality determination unit 502 diagnoses an abnormality in the rolling bearing 30 based on the ratio or difference between the statistical feature amounts of the amplified output data of the two strain gauges 100 over the same period.

[0191] Diagnosis of abnormalities in the rolling bearing 30 includes diagnosis of the presence or absence of abnormality and diagnosis of the degree (level) of abnormality. The following description will focus on the case where diagnosis of the presence or absence of abnormality is performed.

[0192] The abnormality determination unit 502 diagnoses (determines) whether both of the two rolling bearings 30 are normal or whether one of the two rolling bearings 30 is abnormal, based on, for example, a ratio or difference between statistical feature quantities of amplified output data from the two strain gauges 100 over the same period. In this case, the abnormality determination unit 502 diagnoses whether there is an abnormality in the bearing device BE by two-class classification. Alternatively, the abnormality determination unit 502 may diagnose (determine) whether there is an abnormality in the bearing device BE, whether both of the two rolling bearings 30 are normal, whether the rolling bearing 30 on the tip side is abnormal, or whether the rolling bearing 30 on the base side is abnormal, based on, for example, a ratio or difference between statistical feature quantities of similar data. In this case, the abnormality determination unit 502 diagnoses whether there is an abnormality in the bearing device BE by three-class classification.

[0193] The abnormality determination unit 502 diagnoses whether or not there is an abnormality in the bearing device BE using, for example, a determination model LM, which is a trained model that has been machine-learned using a predetermined algorithm. Specifically, the data acquisition unit 501 acquires time-series data of the amplified outputs of the two strain gauges 100 for the same period. The abnormality determination unit 502 then calculates at least one of the difference and ratio between the statistical feature quantities (hereinafter simply referred to as "statistical feature quantities") of both sets of time-series data. The abnormality determination unit 502 then inputs at least one of the difference and ratio between the statistical feature quantities into the determination model LM, thereby being able to diagnose whether or not there is an abnormality in the bearing device BE. Furthermore, by inputting both the difference and ratio between the statistical feature quantities of the time-series data into the determination model LM, the abnormality determination unit 502 can more accurately diagnose whether or not there is an abnormality in the bearing device BE.

[0194] The predetermined algorithm is, for example, a supervised learning algorithm such as a random forest, a support vector machine (SVM), deep learning, etc. Alternatively, the predetermined algorithm may be, for example, an unsupervised learning algorithm such as a local outlier factor (LOF) or an isolation forest.

[0195] The abnormality determination unit 502 may perform abnormality diagnosis using a method other than using the determination model LM as a trained model. For example, the abnormality determination unit 502 may perform abnormality diagnosis of the bearing device BE using data on the difference or ratio between statistical feature amounts of amplified output data for the same period of two strain gauges 100 or a predetermined threshold value related to statistical feature amounts of time-series data.

[0196] The statistical features used for abnormality diagnosis of the bearing device BE include, for example, the maximum value, minimum value, maximum-minimum difference (difference between the maximum and minimum values), kurtosis, skewness, mode, standard deviation, effective value, mean value, median value, 25th percentile value, 75th percentile value, etc. of the time series data of the amplified outputs of the two strain gauges 100 over the same period (predetermined time).

[0197] The abnormality determination unit 502 may diagnose the abnormality of the bearing device BE based on one statistical feature amount, or may diagnose the abnormality of the bearing device BE based on a plurality of statistical feature amounts.

[0198] A determination model LM is stored in the storage unit 503. As described above, the determination model LM may be installed in the storage unit 503 from a recording medium 5AA via the external interface 5A, or may be installed in the storage unit 503 from an external device such as the learning device 8 via the communication interface 5F. Furthermore, as will be described later, if the learning device 8 that generates the determination model LM and the monitoring device 5 are the same device, the determination model LM generated by the monitoring device 5 (learning device 8) is stored directly in the storage unit 503.

[0199] The abnormality response unit 504 responds when the abnormality determination unit 502 determines that there is an abnormality in the bearing device BE or that the degree of abnormality in the bearing device BE exceeds a predetermined standard, and performs abnormality processing.

[0200] For example, when the abnormality determination unit 502 determines that there is an abnormality in the bearing device BE or that the degree of abnormality in the bearing device BE exceeds a predetermined standard, the abnormality response unit 504 may restrict the operation of the axial fan 1 via the drive unit 2. Specifically, the abnormality response unit 504 may, for example, limit the rotation speed of the axial fan 1 to a predetermined upper limit or lower within a range that does not cause a rise in the temperature of the object to be cooled by the axial fan 1. This prevents the axial fan 1 from suddenly stopping in response to the occurrence of an abnormality in the bearing device BE or an increase in the degree of abnormality, and ensures sufficient time until the bearing device BE can be replaced.

[0201] Furthermore, for example, when the abnormality determination unit 502 determines that there is an abnormality in the bearing device BE or that the degree of abnormality in the bearing device BE exceeds a predetermined standard, the abnormality response unit 504 notifies the user to that effect via the display device 5H. Furthermore, when the abnormality response unit 504 determines that there is an abnormality in the bearing device BE or that the degree of abnormality in the bearing device BE exceeds a predetermined standard, the abnormality response unit 504 may send a signal notifying the user to that effect via the communication interface 5F to the user's terminal device. In this way, the abnormality response unit 504 can prompt the user to take action in response to the occurrence of an abnormality in the bearing device BE or the increase in the degree of abnormality (for example, replacing the bearing device BE).

[0202] [Machine Learning System] Next, the configuration of the machine learning system 400 will be described with reference to Fig. 24. Below, description of the same or corresponding configuration as that of the anomaly monitoring system 300 may be omitted.

[0203] FIG. 24 is a diagram illustrating an example of the configuration of a machine learning system 400.

[0204] The hardware configuration of the learning device 8 may be the same as, for example, the above-described monitoring device 5. The following description will be given on the assumption that the hardware configuration of the learning device 8 is the same as the above-described monitoring device 5.

[0205] As shown in FIG. 24, the machine learning system 400 includes an axial fan 1, an amplifier 3, a pulse generator 6, a logger 7, and a learning device 8.

[0206] The machine learning system 400 acquires the amplified outputs of the two strain gauges 100 under various rotational speed and temperature conditions using an axial fan 1 and a pulse generator 6 placed in a thermostatic bath CTB. Then, the machine learning system 400 generates a data set for machine learning (hereinafter referred to as "learning data set") in a learning device 8, performs machine learning using the data set, and generates a determination model LM as a trained model.

[0207] The pulse generator 6 outputs a pulse signal that drives the axial flow fan 1 (motor EM).

[0208] Like the logger 4, the logger 7 records the amplified outputs of the two strain gauges 100 of the axial fan 1 in time series. The logger 7 accumulates the amplified outputs of the two strain gauges 100 under various conditions of rotation speed and temperature of the axial fan 1.

[0209] The learning device 8 generates a determination model LM as a trained model for diagnosing abnormalities in the axial flow fan 1.

[0210] The learning device 8 and the monitoring device 5 may be the same device.

[0211] The learning device 8 includes, as functional units, a data acquisition unit 801, a dataset generation unit 802, a model generation unit 803, and a storage unit 804. The functions of the data acquisition unit 801, the dataset generation unit 802, and the model generation unit 803 are realized, for example, by loading a program installed in an auxiliary storage device into a memory device and executing the program on a CPU. The function of the storage unit 804 is realized, for example, by a storage area defined in the auxiliary storage device.

[0212] The data acquisition unit 801 acquires time-series data of the amplified outputs of the two strain gauges 100 at the same predetermined time interval (same period) from the logger 4 (see FIG. 23).

[0213] The dataset generation unit 802 generates a training dataset to be used in machine learning to generate the determination model LM. The training dataset is a collection of data for machine learning. The machine learning data is data of the same dimension as the input of the determination model LM. When one statistical feature is used as the input to the determination model LM, the training data is also composed of one statistical feature of the same type, and when multiple statistical features are used as the input to the determination model LM, the training data is also composed of the same type and number of statistical features.

[0214] Specifically, the data set generation unit 802 calculates the difference or ratio between the above-mentioned statistical features based on the amplified outputs of the two strain gauges 100 at predetermined times during the same period acquired by the data acquisition unit 801, and generates a learning data set.

[0215] In the case of supervised learning, the learning dataset (i.e., the teacher dataset) further includes correct labels. For example, as described above, when the presence or absence of an abnormality in the bearing device BE is diagnosed by two-class classification, two types of correct labels (values) are prepared, and when the presence or absence of an abnormality in the bearing device BE is diagnosed by three-class classification, three types of correct labels (values) are prepared.

[0216] For example, in the logger 7, by classifying the storage locations (addresses) of time series data corresponding to each class of a two-class classification or a three-class classification, the dataset generation unit 802 can automatically assign correct label values ​​to the learning dataset.

[0217] The model generation unit 803 generates a determination model LM as a trained model by performing machine learning on the base learning model using the learning dataset generated by the dataset generation unit 802. The model generation unit 803 may also generate (update) the determination model LM by performing additional learning or re-learning on the determination model LM as a trained model.

[0218] The memory unit 804 stores (registers) the determination model LM generated by the model generation unit 803. When the determination model LM is updated, not only the updated determination model LM but also the old version of the determination model LM may remain in the memory unit 804. In this case, multiple versions of the determination model LM are accumulated in the memory unit 804. This allows multiple past versions of the determination model LM to be reused as starting points for re-learning or additional learning.

[0219] [Machine learning procedure] Next, with reference to FIG. 25, a machine learning procedure for generating the determination model LM will be described.

[0220] 25 is a flowchart showing an example of a machine learning procedure. Hereinafter, steps S102 to S112 represent the processing procedure performed by an operator using the machine learning system 400, and steps S114 to S118 represent the processing procedure performed by the learning device 8 in response to input from the operator.

[0221] In FIG. 25, the rolling bearing 30 on the tip side of the bearing device BE is denoted as "bearing (1)" and the rolling bearing 30 on the base end side is denoted as "bearing (2)".

[0222] In step S102, the worker prepares an axial flow fan 1 in which the rolling bearings 30 on both the base end side and the tip end side are normal.

[0223] Next, in step S104, the operator causes the logger 7 to acquire strain data (amplified output) of the two strain gauges 100 at predetermined time intervals while changing the temperature and rotation speed (rotational speed) of the axial fan 1 in step S102.

[0224] Next, in step S106, the worker prepares an axial flow fan 1 in which the rolling bearing 30 on the tip end side is abnormal and the rolling bearing 30 on the base end side is normal.

[0225] Next, in step S108, the operator changes the temperature and rotation speed (rotational speed) of the axial fan 1 in step S106, and causes the logger 7 to acquire strain data (amplified output) of the two strain gauges 100 at predetermined time intervals.

[0226] Next, in step S110, the worker prepares an axial flow fan 1 in which the rolling bearing 30 on the tip end side is normal and the rolling bearing 30 on the base end side is abnormal.

[0227] Next, in step S112, the operator causes the logger 7 to acquire strain data (amplified output) of the two strain gauges 100 at predetermined time intervals while changing the temperature and rotation speed (rotational speed) of the axial fan 1 in step S110.

[0228] The three groups of steps S102, S104, S106, S108, and S110, S112 may be performed in any order, or may be performed in parallel with one another.

[0229] Subsequently, in step S114, the data set generation unit 802 of the learning device 8 generates a learning data set based on the time-series data acquired in steps S104, S108, and S112.

[0230] Next, the model generation unit 803 of the learning device 8 performs machine learning on the base learning model using the learning data set generated in step S114, and generates a determination model LM as a learned model.

[0231] [Abnormality diagnosis processing] Next, with reference to FIG. 26, a process relating to abnormality diagnosis by the monitoring device 5 (hereinafter referred to as "abnormality diagnosis process") will be described.

[0232] FIG. 26 is a flowchart showing an example of an abnormality diagnosis process performed by the monitoring device 5.

[0233] This flowchart is repeatedly executed at predetermined time intervals, for example, while the axial fan 1 is operating and the abnormality processing in step S208, which will be described later, is not being executed.

[0234] As shown in FIG. 26, in step S202, the abnormality determination unit 502 of the monitoring device 5 calculates data on the difference or ratio between the statistical features based on the time series data of the amplified outputs of the two strain gauges 100 for the latest specified time period acquired by the data acquisition unit 501.

[0235] When the monitoring device 5 completes the process of step S202, the process proceeds to step S204.

[0236] In step S204, the abnormality determination unit 502 of the monitoring device 5 determines whether or not there is an abnormality in the bearing device BE based on the data of the differences and ratios between the statistical feature amounts calculated in step S202. Specifically, by inputting the data of the differences and ratios between the statistical feature amounts into the determination model LM, it is possible to obtain a determination result corresponding to the presence or absence of an abnormality in the bearing device BE.

[0237] When the processing of step S204 is completed, the monitoring device 5 proceeds to step S206.

[0238] In step S206, the abnormality determination unit 502 of the monitoring device 5 determines whether the determination result in step S204 is "abnormal." "Abnormal" means that there is an abnormality in one of the two rolling bearings 30 in a two-class classification, or that there is an abnormality in the rolling bearing 30 on the tip end side in a three-class classification, or that there is an abnormality in the rolling bearing 30 on the base end side. If the determination result is "abnormal," the abnormality determination unit 502 proceeds to step S208, and if the determination result is not "abnormal," the processing of this flowchart is terminated.

[0239] In step S208, the abnormality handling unit 504 of the monitoring device 5 starts abnormality processing.

[0240] When the monitoring device 5 completes the process of step S208, it ends this flow chart.

[0241] The abnormality processing may be cancelled, for example, based on an external human input. As a result, for example, if the bearing device BE is replaced after the abnormality processing associated with the occurrence of an abnormality in the bearing device BE, the abnormality processing is cancelled and the flowchart of Fig. 26 is executed again.

[0242] [Method for determining statistical features (types) to be used for abnormality diagnosis] Next, with reference to FIG. 27, an exemplary method for determining the type of statistical feature used for diagnosing an abnormality in the bearing device BE will be described.

[0243] 27 is a diagram showing an example of the contribution (variable importance) of each difference or ratio of statistical features obtained by random forest (decision tree analysis). The contribution (variable importance) is calculated in the range of 0 to 1 in random forest (decision tree analysis) and represents the contribution (importance) of each difference or ratio of statistical features as a factor in an abnormality of the bearing device BE.

[0244] 27 shows that the contributions to the abnormality of the bearing device BE are distributed among the differences or ratios of multiple, mutually different statistical feature quantities. Therefore, the monitoring device 5 can further improve the diagnostic accuracy by diagnosing the abnormality of the bearing device BE based on the differences or ratios of multiple, mutually different statistical feature quantities.

[0245] In this example (FIG. 27), seven statistical features are shown whose contribution to the abnormality of the bearing device BE exceeds the standard assumed to be significant. The statistical feature (1) with the greatest contribution is the ratio of the average values ​​of the amplified outputs of the two strain gauges 100 at the same timing over a given time period ("Bd mean"). The statistical feature (2) with the next greatest contribution is the ratio of the effective values ​​of the amplified outputs of the two strain gauges 100 at the same timing over a given time period ("Bd rms"). The statistical feature (3) with the next greatest contribution is the difference in standard deviations of the amplified outputs of the two strain gauges 100 at the same timing over a given time period ("Bs std"). The statistical feature (4) with the next greatest contribution is the difference in standard deviations of the amplified outputs of the two strain gauges 100 at the same timing over a given time period ("Bs kurt"). The statistical feature quantity (5) with the next highest contribution is the difference between the minimum values ​​of the amplified outputs of the two strain gauges 100 at the same timing over a given time period ("Bs min"). The statistical feature quantity (6) with the next highest contribution is the ratio of the standard deviations of the amplified outputs of the two strain gauges 100 at the same timing over a given time period ("Bd std"). The statistical feature quantity (7) with the next highest contribution is the difference between the 75th percentile values ​​of the amplified outputs of the two strain gauges 100 at the same timing over a given time period ("Bs 75%").

[0246] For example, the monitoring device 5 can further improve the accuracy of diagnosis by diagnosing abnormalities in the bearing device BE based on seven or more different types of statistical features including the above-mentioned statistical features (1) to (7).

[0247] [Effect] Next, the operation of the axial fan 1 (bearing device BE) and the abnormality monitoring system 300 (monitoring device 5) according to this embodiment will be described.

[0248] In this embodiment, the bearing device BE includes two rolling bearings 30, a bearing housing 40, and two strain gauges 100. Specifically, the two rolling bearings 30 rotatably support the rotor shaft 20. The bearing housing 40 holds the two rolling bearings 30 from their outer circumferential surfaces. The two strain gauges 100 each include a resistor 103 that detects strain in the rolling bearing 30, and are attached to the bearing housing 40 so as to correspond to each of the two rolling bearings 30.

[0249] For example, when using the output of one sensor to diagnose an abnormality in the bearing device BE (rolling bearing 30), if the sensor output fluctuates due to changes in the rotational speed of the rotating shaft or the temperature of the bearing device BE, the bearing device BE may be diagnosed as abnormal even though it is not. Furthermore, if the abnormality conditions are set too strict in order to prevent erroneous diagnosis of an abnormality, it may not be possible to diagnose an abnormality even though an abnormality actually exists. This may result in a decrease in the accuracy of the abnormality diagnosis.

[0250] In contrast to this, in this embodiment, it is possible to use the outputs of two strain gauges 100 corresponding to two rolling bearings 30 that are under the same conditions of rotational speed and temperature. Therefore, for example, by comparing the outputs of the two strain gauges 100, it is possible to diagnose with higher accuracy any abnormality that may have occurred in one of the two rolling bearings 30.

[0251] In addition, in this embodiment, in the bearing device BE, one end (base end) of the axial ends of the bearing housing 40 that is fixed to the casing 70 may be provided with a fixing reinforcement portion 43 that increases the fixing strength with the casing 70.

[0252] As a result, by placing the strain gauge 100 on the base end side of the bearing housing 40, the fixing area between the bearing housing 40 and the casing 70 is limited, but the necessary fixing strength between the bearing housing 40 and the casing 70 can be ensured.

[0253] In this embodiment, the fixing reinforcement portion 43 may be a flange portion or a knurled press-fit portion.

[0254] As a result, the bearing device BE can ensure the necessary fixing strength between the bearing housing 40 and the casing 70, specifically.

[0255] In this embodiment, the monitoring device 5 also includes an abnormality determination unit 502. Specifically, the abnormality determination unit 502 may diagnose an abnormality in the rolling bearing 30 by comparing the outputs of the two strain gauges 100.

[0256] Specifically, in this embodiment, the abnormality determination unit 502 may diagnose an abnormality in the rolling bearing 30 based on the difference between the outputs of two strain gauges 100 during the same period.

[0257] More specifically, in this embodiment, the abnormality determination unit 502 may diagnose abnormalities in the rolling bearing 30 based on at least one of the difference and ratio of statistical features of time-series data of the outputs of two strain gauges 100 at a predetermined time during the same period.

[0258] This allows the monitoring device 5 to specifically diagnose abnormalities in the bearing device BE based on the outputs of the two strain gauges.

[0259] In addition, in this embodiment, the abnormality determination unit 502 may diagnose an abnormality in the rolling bearing 30 based on the difference or ratio between a plurality of mutually different statistical feature amounts.

[0260] Specifically, in this embodiment, the difference or ratio between different statistical features may include the ratio of the average values ​​at a predetermined time of the time series data of two strain gauges 100, the ratio of the effective values ​​at a predetermined time, the difference in standard deviations at a predetermined time, the difference in kurtosis at a predetermined time, the difference in minimum values ​​at a predetermined time, the ratio of standard deviations at a predetermined time, and the difference in 75th percentile values ​​at a predetermined time.

[0261] This allows the monitoring device 5 to further improve the accuracy of diagnosis regarding abnormalities in the bearing device BE.

[0262] In addition, in this embodiment, the abnormality determination unit 502 may diagnose abnormalities in the rolling bearing 30 using a trained model (determination model LM) that has been machine-learned using a predetermined algorithm.

[0263] Specifically, in this embodiment, the predetermined algorithm may be a random forest.

[0264] As a result, the monitoring device 5 can specifically diagnose abnormalities in the bearing device BE using the determination model LM as a learned model.

[0265] [Transformation / Change] Although the embodiments have been described in detail above, the present disclosure is not limited to such specific embodiments, and various modifications and variations are possible within the scope of the gist described in the claims.

[0266] For example, in the above-described embodiment, a motor EM that drives an axial fan 1 is exemplified, but this is not limited to this, and the present disclosure is broadly applicable to any motor EM used other than an axial fan 1.

[0267] Furthermore, for example, in the above-described embodiment, a bearing device BE that rotatably supports the shaft member (rotor shaft 20) of an axial fan 1 (motor EM) is exemplified, but this is not limited to this, and the present disclosure is also broadly applicable to bearing devices BE that rotatably support the shaft member of any object other than an axial fan 1 (motor EM). [Explanation of symbols]

[0268] 1 Axial flow fan (blower) 5. Monitoring equipment (information processing equipment) 8 Learning Device 10 Impeller 11 rotor housing 12 Feathers 20 Rotor shaft (rotating shaft) 30 Rolling bearings (bearings) 40 Bearing housing (retaining member) 41 Thick part 42 Thin section 43 Fixed reinforcement part 43x wiring passage section 50 Stator 51 Insulator 52 Stator core 52x notch 53 Coil 60 rotors 61 Rotor yoke 62 rotor magnet 70 Casing 71 Casing outer frame 72 Base hub 73 Stator blade 100 strain gauges 101 Base material 102 Functional Layer 103 Resistor 104 Wiring 105 Terminal section 106 Cover Layer 150 Adhesive layer 200 Wiring 300 Abnormality Monitoring System 400 Machine Learning Systems 501 Data Acquisition Department 502 Abnormality judgment unit (diagnosis unit) 503 Storage section 504 Abnormality Response Department 801 Data Acquisition Unit 802 Dataset Generation Unit 803 Model Generation Unit 804 Storage section BE bearing device EM motor LM decision model (pre-trained model)

Claims

1. two bearings that rotatably support the rotating shaft; a holding member that holds the two bearings from the outer circumferential surface side; two sensors attached to the holding member so as to correspond to the two bearings, respectively, and each detecting a strain of the bearing; a diagnosis unit that compares the outputs of the two sensors to diagnose an abnormality in the bearing. Bearing device.

2. Two bearings that rotatably support a rotating shaft; a holding member that holds the two bearings from the outer circumferential surface side; two sensors attached to the holding member so as to correspond to the two bearings, respectively, and each detecting a strain of the bearing; the bearing includes an outer ring, an inner ring arranged coaxially with the outer ring on the inner peripheral side of the outer ring, and a plurality of rolling elements arranged between the outer ring and the inner ring, and is preloaded to form a predetermined contact angle; the holding member has a thick portion having a relatively large thickness and a thin portion having a relatively small thickness, the sensor is disposed in the thin portion, the thick-wall portion is arranged so as to contact the outer ring at least in a region between an intersection of a straight line indicating the contact angle and the outer peripheral surface of the outer ring to a preload-side end face, which is an end face of the outer ring closer to the intersection point. Bearing device.

3. a diagnostic unit that diagnoses an abnormality in the bearing by comparing outputs of the two sensors; The bearing device according to claim 2 .

4. the diagnosing unit diagnoses an abnormality in the bearing based on a difference between outputs of the two sensors for the same period. The bearing device according to claim 1 or 3.

5. the diagnosing unit diagnoses an abnormality in the bearing based on at least one of a difference and a ratio of feature amounts of time-series data of outputs from the two sensors at a predetermined time during the same period. The bearing device according to claim 4.

6. the diagnosing unit diagnoses an abnormality in the bearing based on a difference or ratio between a plurality of the feature amounts that are different from one another. The bearing device according to claim 5 .

7. The difference or ratio of the different feature amounts includes a ratio of average values ​​at the predetermined time between the time series data of the two sensors, a ratio of effective values ​​at the predetermined time, a difference in standard deviations at the predetermined time, a difference in kurtosis at the predetermined time, a difference in minimum values ​​at the predetermined time, a ratio of standard deviations at the predetermined time, and a 75th percentile value at the predetermined time. The bearing device according to claim 6.

8. The diagnosing unit diagnoses an abnormality in the bearing using a trained model that has been machine-learned using a predetermined algorithm. A bearing device according to any one of claims 3 to 7.

9. The predetermined algorithm is a random forest. The bearing device according to claim 8.

10. One end of the holding member in the rotation axis direction that is fixed to the housing is provided with a reinforcing portion that enhances the fixing strength to the housing. A bearing device according to any one of claims 1 to 9.

11. The reinforced portion is a flange portion or a knurled press-fit portion. The bearing device according to claim 10.

12. a length of the thick-walled portion in the rotational axis direction is shorter than lengths of the outer ring and the inner ring in the rotational axis direction; The bearing device according to claim 2 .

13. a length of the thick-walled portion in the rotational axis direction is substantially equal to a length of the outer ring and the inner ring in the rotational axis direction, the thick-walled portion is disposed so as to contact the entire outer peripheral surface of the outer ring. The bearing device according to claim 2 .

14. the thin portion protrudes in a direction substantially parallel to the rotation axis direction from a side of an end surface of the thick portion that is closer to the rotation axis; The bearing assembly according to claim 13.

15. the thin portion protrudes in a direction substantially parallel to the rotation axis direction from a side of the end surface of the thick portion that is farther from the rotation axis; The bearing assembly according to claim 13.

16. the thin-walled portion extends from an end face of the thick-walled portion in a direction substantially perpendicular to the rotation axis and is disposed in contact with a part of the end face of the outer ring, a thickness of the thin portion in the direction of the rotation axis is smaller than a thickness of the thick portion in the radial direction; The bearing assembly according to claim 13.

17. The thick-walled portion and the thin-walled portion are cylindrical.

17. A bearing arrangement according to any one of claims 2 and 12 to 16.

18. The holding member is formed by joining separate members having different diameters together.

18. The bearing assembly of claim 17.

19. The thick portion is cylindrical, The thin portion is a recess provided in the thick portion.

19. A bearing arrangement according to any one of claims 2 and 12 to 18.

20. The sensor is a strain gauge including a resistor that detects strain in the bearing, The resistor is formed from a Cr mixed phase film.

20. A bearing arrangement according to any one of claims 1 to 19.

21. A bearing device according to any one of claims 1 to 20; a stator core fixed to the radially outer side of the holding member, Motor.

22. A bearing device having two bearings that rotatably support a rotating shaft, a holding member that holds the two bearings from the outer circumferential surface side, and two sensors attached to the holding member so as to correspond to each of the two bearings, each of which detects strain in the bearing; a stator core fixed to the radially outer side of the holding member, Motor.

23. The sensor is fixed radially inward of the inner diameter of the stator core.

23. A motor according to claim 21 or 22.

24. The sensor is fixed in a recess provided in the outer peripheral surface of the holding member.

24. The motor of claim 23.

25. Wiring extending from the sensor; At least a portion of the wiring is shielded.

25. A motor according to any one of claims 21 to 24.

26. The wiring passes through a recess provided on the inner periphery of the stator core and is drawn out to the outside of the motor, At least a portion of the wiring passing through a recess provided on the inner periphery of the stator core is shielded.

26. The motor of claim 25.

27. The holding member is a ground having the same potential as the shield portion of the wiring.

27. A motor according to claim 25 or 26.

28. A motor according to any one of claims 21 to 27; an impeller that is rotationally driven by the motor; Blower.

29. A method for diagnosing abnormalities in a bearing device having two bearings that rotatably support a rotating shaft, a holding member that holds the two bearings from the outer circumferential surface side, and two sensors attached to the holding member so as to correspond to each of the two bearings, each sensor detecting strain in the bearing, comprising: an information processing device that compares the outputs of the two sensors to diagnose an abnormality in the bearing; Abnormality diagnosis method.

30. the information processing device diagnoses an abnormality in the bearing based on a difference between outputs of the two sensors for the same period.

30. The abnormality diagnosis method according to claim 29.

31. the information processing device diagnoses an abnormality in the bearing based on at least one of a difference and a ratio of feature amounts of time-series data of outputs from the two sensors at a predetermined time during the same period; The abnormality diagnosis method according to claim 30.

32. the information processing device diagnoses an abnormality in the bearing based on a difference or ratio between a plurality of the feature amounts that are different from one another. The abnormality diagnosis method according to claim 31.

33. The difference or ratio of the different feature amounts includes a ratio of average values ​​at the predetermined time between the time series data of the two sensors, a ratio of effective values ​​at the predetermined time, a difference in standard deviations at the predetermined time, a difference in kurtosis at the predetermined time, a difference in minimum values ​​at the predetermined time, a ratio of standard deviations at the predetermined time, and a 75th percentile value at the predetermined time. The abnormality diagnosis method according to claim 32.

34. The information processing device diagnoses an abnormality in the bearing using a trained model that has been machine-learned using a predetermined algorithm.

34. The abnormality diagnosis method according to any one of claims 29 to 33.

35. The predetermined algorithm is a random forest. The abnormality diagnosis method according to claim 34.

Images