State monitoring device and storage medium

Abstract

This state monitoring device comprises a first acquisition unit (201), a second acquisition unit (202), and an inference unit (203). The first acquisition unit is provided at a joint part of a turnable arm, and acquires a first physical quantity relating to a machine element provided at the joint part. The second acquisition unit is provided on a tip end-side, of the arm, different from a base end-side where the joint part is disposed, and acquires a second physical quantity relating to a motion state of the arm. The inference unit infers a state of the machine element on the basis of the first physical quantity acquired by the first acquisition unit and the second physical quantity acquired by the second acquisition unit.

Description

Condition monitoring device and storage medium

[0001] The present invention relates to a condition monitoring device and a storage medium. This application claims priority to Japanese Patent Application No. 2022-201123, filed on December 16, 2022, the contents of which are incorporated herein by reference.

[0002] Conventionally, reduction gears are mounted on the joints of the arms of construction machinery. As reduction gears deteriorate, they vibrate more than usual. Therefore, for example, an acceleration sensor is used to determine abnormalities (failure or deterioration) in the gears of reduction gears, etc. The vibration of gears is small both when deteriorated and under normal conditions. For this reason, it is necessary to focus on a target frequency (target frequency) that is suitable for estimating gear failure from among the various frequency vibrations detected by the acceleration sensor, and to determine abnormalities based on the magnitude of the vibration of the target frequency.

[0003] The target frequency is determined by, for example, the number of teeth and the rotation speed of the gear. The rotation speed of the gear can be obtained, for example, from the mother machine. As a related technique, a method is disclosed in which a vibration value at an arbitrary rotation speed is corrected to a vibration value at a reference rotation speed using an equation that approximates the relationship between the rotation speed of the rotating device in a normal state and the vibration value of the rotating device, and an abnormality is determined when this corrected value exceeds a threshold (see, for example, Patent Document 1).

[0004] Japanese Patent Publication No. 7-218333

[0005] However, the method of obtaining the target frequency from the mother machine requires wiring of a signal line from the mother machine to an external device, which is cumbersome. Therefore, it is not possible to obtain the target frequency with a simple configuration. For this reason, the conventional technology has a problem in that it is not possible to estimate a gear abnormality with a simple configuration.

[0006] An object of the present invention is to provide a condition monitoring device and a storage medium that can estimate the condition of a machine element (for example, the condition of a gear) with a simple configuration.

[0007] One aspect of the condition monitoring device of the present invention includes a first detection unit that detects a first physical quantity related to a mechanical element provided at a joint portion of a rotatable arm, a second detection unit that is provided at a tip end of the arm different from the base end where the joint portion is located and detects a second physical quantity related to the motion state of the arm, and an estimation unit that estimates the state of the mechanical element based on the first physical quantity detected by the first detection unit and the second physical quantity detected by the second detection unit.

[0008] According to the above configuration, the state of the machine element can be estimated with a simple configuration without acquiring information from the mother machine.

[0009] In the above configuration, the mechanical element may be a gear provided at the joint, the first physical quantity may be vibration of the gear, and the second physical quantity may be at least one of acceleration and angular velocity of the arm. The estimator may include a calculator that calculates a specific frequency of the vibration based on a detection result of the second physical quantity during a predetermined period in which the arm is moving and the number of teeth of the gear. Furthermore, the estimator may estimate the state of the gear using the calculated specific frequency and a detection result of the first physical quantity during the predetermined period.

[0010] According to the above configuration, the state of the gear (for example, an abnormality in the gear) can be estimated with a simple configuration.

[0011] In the aforementioned configuration, the estimation unit may extract data corresponding to the calculated specific frequency from time-series data of detection results of the first physical quantity in the predetermined period, and estimate the state of the gear based on the extracted data.

[0012] The above configuration eliminates the need for large-scale calculations such as FFT (Fast Fourier Transformation), reducing power consumption, and thus enabling the condition monitoring device to be applied to simple battery-powered devices (e.g., smart sensors).

[0013] In the aforementioned configuration, the estimation unit may perform filtering to extract data corresponding to the specific frequency from the time-series data of the detection result of the first physical quantity, and estimate the state of the gear based on a result of comparing the extracted data with a predetermined threshold.

[0014] According to the above configuration, the state of the gear, for example, whether or not there is an abnormality in the gear, can be estimated with a simpler configuration.

[0015] In the above configuration, the second physical quantity may be a triaxial acceleration of the arm. The estimator may include a determiner that determines whether the arm is in a constant velocity state, which indicates that the arm is moving at a constant velocity, based on a comparison result between a composite value of the triaxial acceleration and gravitational acceleration. The estimator may estimate the state of the mechanical element based on a detection result of the first detector at a time when the constant velocity state is determined.

[0016] According to the above configuration, it is possible to improve the accuracy of estimating the state of a machine element (for example, estimating an abnormality).

[0017] In the above configuration, the identification unit may identify the arm as being in the constant velocity state when the composite value of each axis within a predetermined period of time matches the acceleration of gravity and the value of each axis within the predetermined period of time is indefinite.

[0018] According to the above configuration, the constant velocity state of the arm can be easily identified.

[0019] In the above configuration, the specifying unit may specify a non-constant velocity state other than the constant velocity state based on the composite value of each axis and the gravitational acceleration. The estimating unit may not estimate the state of the mechanical element in the non-constant velocity state.

[0020] According to the above configuration, it is possible to improve the accuracy of estimating the state of a machine element (for example, estimating an abnormality).

[0021] In the above configuration, when the composite value of each axis does not match the gravitational acceleration, the identification unit may identify the non-constant velocity state as an acceleration / deceleration state indicating that the arm is operating at an accelerated or decelerated speed.

[0022] According to the above configuration, the acceleration / deceleration state of the arm can be easily identified.

[0023] In the above configuration, the mechanical element may be a gear provided at the joint, the first physical quantity may be vibration of the gear, and the second physical quantity may include a rotation speed of the arm and a triaxial acceleration of the arm. The estimator may include: an identifier that identifies whether the arm is in a constant velocity state, indicating that the arm is moving at a constant velocity, based on a comparison result between a composite value of each axis of the triaxial acceleration and gravitational acceleration; and a calculator that calculates a specific frequency of the vibration based on information including at least the rotation speed of the arm and the number of teeth of the gear at the time when the constant velocity state is identified. The estimator may extract data corresponding to the calculated specific frequency from time-series data of the detection result of the first physical quantity, and estimate the state of the gear based on the extracted data.

[0024] With the above configuration, the constant velocity state of the arm can be easily identified, large-scale calculations such as FFT are not required, and power consumption can be reduced. Therefore, the condition monitoring device can be applied to simple battery-powered devices (e.g., smart sensors).

[0025] One aspect of the storage medium of the present invention is a computer-readable storage medium storing a program that causes a computer to function as a condition monitoring device, the program storing the computer causing the computer to function as: a first acquisition unit that is provided at a joint portion of a rotatable arm and acquires a first physical quantity related to a mechanical element provided at the joint portion from a first detection unit that detects the first physical quantity, the first physical quantity being related to the mechanical element; a second acquisition unit that is provided at a tip end of the arm different from the base end where the joint portion is located and acquires a second physical quantity related to the motion state of the arm from a second detection unit that detects the second physical quantity, the second physical quantity being related to the motion state of the arm; and an estimation unit that estimates the state of the mechanical element based on the first physical quantity acquired by the first acquisition unit and the second physical quantity acquired by the second acquisition unit.

[0026] According to the above configuration, the state of the gears (for example, gear abnormalities) can be estimated with a simple configuration without acquiring information from the mother machine.

[0027] According to the present invention, the condition of a gear (for example, an abnormality in a gear) can be estimated with a simple configuration.

[0028] An explanatory diagram of a construction machine 1 according to an embodiment. A block diagram showing an example of the functional configuration of a state monitoring device 100. A schematic diagram showing an overview of the state monitoring device 100. An explanatory diagram of the hardware configuration of the state monitoring device 100. A flowchart showing an example of processing performed by an estimation control unit 200. A sequence diagram related to abnormality estimation performed by the estimation control unit 200 and a change amount sensor 24.

[0029] An embodiment of the present invention will be described with reference to the drawings.

[0030] (Construction Machine 1) Fig. 1 is an explanatory diagram of a construction machine 1 according to an embodiment. As shown in Fig. 1, the construction machine 1 is, for example, a type of excavator, a so-called power shovel. The construction machine 1 includes a main body 2, an arm unit 4 (an example of an arm), and a rotating shaft unit 20 (an example of a joint). The main body 2 has a swivel unit 3a and a traveling unit 3b.

[0031] (Arm section 4) The arm section 4 includes a boom 5, a shovel arm 7, and a bucket 9. The boom 5 is connected to the main body 2 and is rotatable relative to the main body 2. The shovel arm 7 is connected to the boom 5 and is rotatable relative to the boom 5. The bucket 9 is connected to the shovel arm 7 and is rotatable relative to the shovel arm 7.

[0032] (Rotating shaft portion 20) The rotating shaft portion 20 includes a first rotating shaft portion 20a, a second rotating shaft portion 20b, a third rotating shaft portion 20c, and a fourth rotating shaft portion 20d. The first rotating shaft portion 20a rotatably connects the traveling portion 3b and the swivel portion 3a. The second rotating shaft portion 20b rotatably connects the swivel portion 3a (support portion 3c provided on the swivel portion 3a) and the boom 5. The third rotating shaft portion 20c rotatably connects the boom 5 and the shovel arm 7. The fourth rotating shaft portion 20d rotatably connects the shovel arm 7 and the bucket 9.

[0033] The second rotating shaft 20b is provided on a portion of the support 3c that connects to the boom 5 (or a portion of the boom 5 that connects to the support 3c). The third rotating shaft 20c is provided on a portion of the boom 5 that connects to the shovel arm 7 (or a portion of the shovel arm 7 that connects to the boom 5). The fourth rotating shaft 20d is provided on a portion of the shovel arm 7 that connects to the bucket 9 (or a portion of the bucket 9 that connects to the shovel arm 7).

[0034] The rotation axis of the first rotating shaft 20a is perpendicular to the surface (e.g., the ground) that the running part 3b contacts, while the rotation axes of the second rotating shaft 20b, the third rotating shaft 20c, and the fourth rotating shaft 20d are parallel to the surface that the running part 3b contacts.

[0035] Each rotating shaft portion 20 is provided with a rotating electric machine (not shown) and a reducer 22. Specifically, the first rotating shaft portion 20a is provided with a first rotating electric machine (not shown) and a first reducer 22a. The second rotating shaft portion 20b is provided with a second rotating electric machine (not shown) and a second reducer 22b. The third rotating shaft portion 20c is provided with a third rotating electric machine (not shown) and a third reducer 22c. The fourth rotating shaft portion 20d is provided with a fourth rotating electric machine (not shown) and a fourth reducer 22d. The rotating electric machines are rotated by electricity (electric power).

[0036] (Reduction Gear 22) The reduction gear 22 is a gear mechanism having a plurality of gears that mesh with each other. The first reduction gear 22a rotates and drives the swivel unit 3a using power generated by the first rotating electric machine. Specifically, the input section of the first reduction gear 22a is connected to the output section of the first rotating electric machine. The output section of the first reduction gear 22a is connected to the swivel unit 3a. The first reduction gear 22a reduces the rotational speed of the power input to the input section from the first rotating electric machine, and outputs the reduced power from the output section.

[0037] The second reducer 22b rotates the boom 5 using power generated by the second rotating electric machine. Specifically, the input section of the second reducer 22b is connected to the output section of the second rotating electric machine. The output section of the second reducer 22b is connected to the boom 5. The second reducer 22b reduces the rotational speed of the power input to the input section from the second rotating electric machine, and outputs the reduced power from the output section.

[0038] The third reducer 22c uses power generated by the third rotating electric machine to rotationally drive the shovel arm 7. Specifically, an input portion of the third reducer 22c is connected to an output portion of the third rotating electric machine. An output portion of the third reducer 22c is connected to the shovel arm 7. The third reducer 22c reduces the rotational speed of the power input to the input portion from the third rotating electric machine, and outputs the reduced power from the output portion.

[0039] The fourth reducer 22d rotates and drives the bucket 9 using power generated by the fourth rotating electric machine. Specifically, an input portion of the fourth reducer 22d is connected to an output portion of the fourth rotating electric machine. An output portion of the fourth reducer 22d is connected to the bucket 9. The fourth reducer 22d reduces the rotational speed of the power input to the input portion from the fourth rotating electric machine, and outputs the reduced power from the output portion.

[0040] (Condition monitoring device 100) The construction machine 1 is provided with a condition monitoring device 100 (100b to 100d) that estimates an abnormality in the reducer 22. The condition monitoring device 100 (100b to 100d) estimates the state of the mechanical elements that make up the reducer 22. Specifically, the condition monitoring device 100 (100b to 100d) estimates the state (abnormal state) of the gears that make up the reducer 22. The abnormalities estimated by the condition monitoring device 100 include various targets such as failure, deterioration, risk of failure, degree of deterioration, and remaining usable life. Note that by associating a threshold value (described below) with each of the various targets, it is possible to distinguish between the various targets.

[0041] The condition monitoring device 100 (100b to 100d) is provided on the rotating shaft portion 20 (20b to 20d). As shown in FIGS. 1 and 2, the condition monitoring device 100 includes an acceleration sensor 23 (23b to 23d) and a change amount sensor 24 (24b to 24d). The condition monitoring device 100 is, for example, a smart sensor, and has the acceleration sensor 23 built in. The acceleration sensor 23 is an example of a first detection unit. The acceleration sensor 23 detects a first physical quantity related to wear of the gears that make up the reducer 22. The first physical quantity is, for example, triaxial acceleration. Note that the acceleration sensor 23 is a sensor that can detect triaxial angular velocity in addition to triaxial acceleration.

[0042] The change amount sensor 24 is an example of a second detection unit. The change amount sensors 24 (24b to 24d) are provided on the arm unit 4 (boom 5, shovel arm 7, bucket 9).

[0043] The change amount sensor 24 detects a second physical quantity related to the motion state of the arm unit 4. The second physical quantity is, for example, the angular velocity of the arm unit 4. In this embodiment, the change amount sensor 24 is a sensor capable of detecting triaxial acceleration and triaxial angular velocity. Note that the acceleration sensor 23 and the change amount sensor 24 may each be an integrated smart sensor.

[0044] (Second condition monitoring device 100b) The second rotating shaft portion 20b is provided with a second condition monitoring device 100b that estimates an abnormality in the second reducer 22b. The second condition monitoring device 100b includes a second acceleration sensor 23b. The second acceleration sensor 23b detects vibrations of a gear provided in the second reducer 22b. The second condition monitoring device 100b may be provided on a case of the second reducer 22b (a case that houses multiple gears of the second reducer 22b). The boom 5 is provided with a second change amount sensor 24b. The second change amount sensor 24b is provided on the tip side of the boom 5, which is different from the base end side where the second rotating shaft portion 20b is located. Specifically, the second change amount sensor 24b is provided on a portion of the boom 5 that is located closer to the base end than the third rotating shaft portion 20c, which is located at the tip. The second change amount sensor 24b detects the angular velocity of the boom 5.

[0045] (Third condition monitoring device 100c) The third rotating shaft portion 20c is provided with a third condition monitoring device 100c that determines a failure (deterioration state) of the third reducer 22c. The third condition monitoring device 100c includes a third acceleration sensor 23c. The third acceleration sensor 23c detects vibrations of a gear provided in the third reducer 22c. The third condition monitoring device 100c may be provided on a case of the third reducer 22c (a case that houses multiple gears of the third reducer 22c). The shovel arm 7 is provided with a third change amount sensor 24c. The third change amount sensor 24c is provided on the tip end side of the shovel arm 7, which is different from the base end side where the third rotating shaft portion 20c is located. Specifically, the third change amount sensor 24c is provided on a portion of the shovel arm 7 that is located closer to the base end than the fourth rotating shaft portion 20d, which is located at the tip end. The third change amount sensor 24 c detects the angular velocity of the shovel arm 7 .

[0046] (Fourth condition monitoring device 100d) The fourth rotating shaft portion 20d is provided with a fourth condition monitoring device 100d that determines a failure (deterioration state) of the fourth reducer 22d. The fourth condition monitoring device 100d includes a fourth acceleration sensor 23d. The fourth acceleration sensor 23d detects vibrations of a gear provided in the fourth reducer 22d. The fourth condition monitoring device 100d may be provided on the case of the fourth reducer 22d (a case that houses the multiple gears of the fourth reducer 22d). The bucket 9 is provided with a fourth change amount sensor 24d. The fourth change amount sensor 24d is provided on a portion of the bucket 9 that is located closer to the tip end side than the base end side where the fourth rotating shaft portion 20d is located. The fourth change amount sensor 24d detects the angular velocity of the bucket 9.

[0047] (Functional Configuration of State Monitoring Device 100) Fig. 2 is a block diagram showing an example of the functional configuration of the state monitoring device 100. Fig. 3 is a schematic diagram showing an overview of the state monitoring device 100. The estimation control unit 200 shown in Fig. 2 will be described below with reference to Fig. 3.

[0048] 2 , the state monitoring device 100 includes an estimation control unit 200, an acceleration sensor 23, and a change amount sensor 24. The estimation control unit 200 includes a first acquisition unit 201, a second acquisition unit 202, an estimation unit 203, a notification unit 204, and a storage unit 210.

[0049] The first acquisition unit 201 acquires vibration data (vibration values) of the gears provided in the reducer 22 from the acceleration sensor 23. Therefore, the acceleration sensor 23 and the first acquisition unit 201 cooperate with each other to acquire the vibration data of the gears. Therefore, the first acquisition unit 201 functions as a first detection unit together with the acceleration sensor 23. The vibration data is time-series data of vibrations detected by the acceleration sensor 23.

[0050] The second acquisition unit 202 acquires the number of rotations of the arm unit 4 calculated by the change amount sensor 24. The change amount sensor 24 and the second acquisition unit 202 cooperate with each other to acquire the number of rotations of the arm unit 4. Therefore, the second acquisition unit 202 functions as a second detection unit together with the change amount sensor 24. Here, the calculation of the number of rotations of the arm unit 4 will be described. The number of rotations N is expressed by the following equation 1 using the distance L between the acceleration sensor 23 and the change amount sensor 24 and the angular velocity V. The change amount sensor 24 stores the distance L in advance.

[0051] N=(V / 2πL)×60…(1)

[0052] When the change amount sensor 24 detects the angular velocity of the arm unit 4, it calculates the number of rotations of the arm unit 4 based on the detected angular velocity and Equation 1. Note that the calculation of the number of rotations of the arm unit 4 may be performed by, for example, the second acquisition unit 202.

[0053] The estimation unit 203 estimates the presence or absence of an abnormality related to wear of the gears in the reducer 22 based on the vibration data acquired by the first acquisition unit 201 and the rotation speed acquired by the second acquisition unit 202. Hereinafter, the estimation of the presence or absence of an abnormality related to wear of the gears in the reducer 22 may be referred to as "abnormality estimation."

[0054] The storage unit 210 stores various information. The information stored in the storage unit 210 includes, for example, the number of teeth of the reducer 22 (the number of teeth of multiple gears) and a threshold value used for comparison with the magnitude (amplitude) of vibration of the target frequency. These pieces of information are referenced in the anomaly estimation by the estimation unit 203. The storage unit 210 may also store the estimation result of the anomaly estimation together with the time of estimation.

[0055] (Regarding Abnormality Estimation) A specific example of abnormality estimation will be described below. The estimation unit 203 includes a calculation unit 203a. The calculation unit 203a calculates a specific frequency (target frequency) used for abnormality estimation. Specifically, the calculation unit 203a calculates the target frequency based on the number of rotations of the arm unit 4 and the number of teeth of the gear acquired by the second acquisition unit 202. More specifically, the calculation unit 203a calculates the target frequency by multiplying the number of rotations N acquired by the second acquisition unit 202 by the number of teeth T stored in the storage unit 210. That is, the target frequency F is expressed by the following equation 2.

[0056] F = N × T... (2)

[0057] The estimation unit 203 then extracts data corresponding to the target frequency calculated by the calculation unit 203a from the time-series data of vibration acquired by the first acquisition unit 201. Specifically, the estimation unit 203 extracts vibration data of the target frequency from the time-series data of vibrations of various frequencies, for example, by performing filtering. In the filtering, for example, a band-pass filter or a low-pass filter is used. These filters may be realized by analog circuits or software. In the filtering, the estimation unit 203 changes the sampling speed and the moving average interval to extract the vibration data of the target frequency.

[0058] The estimation unit 203 then estimates an abnormality based on the extracted data. Specifically, the estimation unit 203 estimates an abnormality based on a comparison result between the extracted vibration data (vibration data at the target frequency) and a threshold value stored in the storage unit 210. More specifically, the estimation unit 203 compares the amplitude (e.g., the root mean square of the amplitude) of the extracted vibration data with a threshold value, and when the comparison result indicates that the vibration data is equal to or greater than the threshold value, estimates that an abnormality related to gear wear exists.

[0059] Note that the number of rotations of the arm unit 4 is not limited to being calculated by the change amount sensor 24, but may also be calculated by the calculation unit 203a. In this case, the second acquisition unit 202 may acquire the angular velocity of the arm unit 4 from the change amount sensor 24. The calculation unit 203a can also calculate the number of rotations of the arm unit 4 based on the angular velocity acquired by the second acquisition unit 202 and the above-mentioned formula 1.

[0060] (Regarding the notification unit 204) The estimation unit 203 outputs the estimation result to the notification unit 204. Note that the estimation unit 203 may be configured to output the estimation result to the notification unit 204 when it estimates that an abnormality exists, and not output the estimation result to the notification unit 204 when it estimates that no abnormality exists. The estimation unit 203 may also store the estimation result in the storage unit 210. Note that the estimation unit 203 may be configured to store the estimation result in the storage unit 210 when it estimates that an abnormality exists, and not store the estimation result in the storage unit 210 when it estimates that no abnormality exists.

[0061] The notification unit 204 notifies a predetermined output device of the estimation result output from the estimation unit 203. Specifically, when the estimation unit estimates that an abnormality exists, the notification unit 204 notifies the predetermined output device of that fact. The predetermined output device includes a light-emitting unit such as an LED, a speaker, or an external device. That is, the notification unit 204 notifies the abnormality by light emission from an LED or the like, sound from a speaker, or communication with an external device. Note that the notification unit 204 may perform the notification in at least one of the light emission, sound, and communication modes, and may perform the notification in all modes, for example.

[0062] (Regarding the timing for estimating an abnormality) Here, when the arm unit 4 is in a stopped state, no vibration is generated in the reducer 22. Therefore, even if an abnormality estimation is performed, no abnormality estimation result can be obtained. Furthermore, when the arm unit 4 is in an accelerating / decelerating state, the acceleration / deceleration component is added to the vibration detected by the acceleration sensor 23, and the target frequency becomes unstable. In other words, when the arm unit 4 is in a stopped state or an accelerating / decelerating state, an abnormality estimation cannot be performed accurately. Therefore, in this embodiment, an abnormality estimation is performed when the arm unit 4 is in a constant velocity state. Below, how to determine whether the arm unit 4 is in a constant velocity state will be described.

[0063] (Regarding Determining a Constant-Velocity State) The change amount sensor 24 detects triaxial acceleration as a second physical quantity related to the motion state of the arm unit 4. The second acquisition unit 202 acquires the triaxial acceleration from the change amount sensor 24. The estimation unit 203 includes an identification unit 203b. The identification unit 203b identifies whether the arm unit 4 is in a constant-velocity state based on a comparison result between a composite value of each axis of the triaxial acceleration acquired by the second acquisition unit 202 and the acceleration of gravity. Note that the movement of the arm unit 4 is slower (lower frequency) than the vibration of the reducer 22. For this reason, the identification unit 203b extracts the acceleration related to the movement of the arm unit 4 using, for example, a low-pass filter, and determines whether the arm unit 4 is in a constant-velocity state.

[0064] The determination unit 203b determines a constant velocity state when the composite value of each axis and the gravitational acceleration match within a predetermined period, and the value of each axis within the predetermined period is indefinite. The predetermined period is a period set in advance, for example, several seconds. Here, if the vibrations of the X-axis, Y-axis, and Z-axis of the acceleration acquired by the second acquisition unit 202 are defined as Ax, Ay, and Az, respectively, the acceleration due to the gravitational acceleration is defined as Gx, Gy, and Gz, and the acceleration due to the movement of the reducer 22 is defined as Rx, Ry, and Rz, then the vibration A can be expressed by the following equations 3 to 5.

[0065] Ax=Gx+Rx…(3) Ay=Gy+Ry…(4) Az=Gz+Rz…(5)

[0066] The composite value of Gx, Gy, and Gz is expressed by the following equation 6.

[0067] (Gx ２ +Gy ２ +Gz ２ ) 1 / 2 = 9.8 m / s ２ …(6)

[0068] When the vehicle is in a constant velocity state, there is no acceleration or deceleration, so Rx = Ry = Rz = 0, and therefore the above equations 3 to 5 can be expressed as follows: Ax = Gx Ay = Gy Az = Gz

[0069] Therefore, the composite value of Ax, Ay, and Az is 9.8 m / s ２ Furthermore, since the arm unit 4 is moving, the directions of the accelerations of the three axes detected by the change amount sensor 24 become indefinite. That is, the values ​​of Ax, Ay, and Az (=Gx, Gy, Gz) become indefinite. Therefore, the specifying unit 203b determines that the composite value of Ax, Ay, and Az is 9.8 m / s ２ " and "each value of Ax, Ay, Az = indefinite" are satisfied, the state can be identified as a constant velocity state. When the identifying unit 203b identifies the state as a constant velocity state, the estimating unit 203 estimates an abnormality based on the detection result of the acceleration sensor 23. In other words, the estimating unit 203 estimates the presence or absence of an abnormality related to wear of the reducer 22 based on the detection result of the acceleration sensor 23 and the detection result of the change amount sensor 24.

[0070] (Regarding Identification of Non-Constant Velocity State) Next, the identification of a non-constant velocity state will be described. The identification unit 203b identifies a non-constant velocity state other than a constant velocity state based on the combined value of the acceleration on each axis detected by the change amount sensor 24 and the acceleration of gravity. Specifically, when the combined value on each axis does not match the acceleration of gravity, the identification unit 203b identifies the non-constant velocity state as an acceleration / deceleration state indicating that the arm is operating while accelerating or decelerating.

[0071] (Regarding Identification of Acceleration / Deceleration State) In an acceleration / deceleration state, Rx, Ry, and Rz become indefinite. Furthermore, since the arm unit 4 moves, the direction of the acceleration of the three axes detected by the change amount sensor 24 becomes indefinite. That is, the values ​​of Ax, Ay, and Az (= Gx, Gy, Gz) become indefinite. Therefore, the identification unit 203b determines whether "the composite value of Ax, Ay, and Az is not equal to 9.8 m / s" or ... ２ If the condition "Acceleration / Deceleration state" is satisfied, it can be determined that the vehicle is in an acceleration / deceleration state. If the determination unit 203b determines that the vehicle is in an acceleration / deceleration state, the estimation unit 203 does not perform an abnormality estimation.

[0072] (Regarding identification of a stopped state) The identification unit 203b identifies a stopped state when the composite value of each axis within a predetermined period matches the gravitational acceleration and the value of each axis within the predetermined period is constant. More specifically, in a stopped state, there is no acceleration or deceleration, so Rx = Ry = Rz = 0. Therefore, the above formulas 3 to 5 can be expressed as follows: Ax = Gx Ay = Gy Az = Gz

[0073] Therefore, the composite value of Ax, Ay, and Az is 9.8 m / s ２ Furthermore, since the arm unit 4 does not move, the directions of the accelerations of the three axes detected by the change amount sensor 24 do not change. Therefore, the values ​​of Ax, Ay, and Az (=Gx, Gy, Gz) are constant. That is, the specifying unit 203b determines that the composite value of Ax, Ay, and Az is 9.8 m / s ２ " and "the values ​​of Ax, Ay, and Az are constant" are satisfied, the stop state can be identified. When the identification unit 203b identifies the stop state, the estimation unit 203 does not estimate an abnormality.

[0074] (Hardware Configuration of Condition Monitoring Device 100) Fig. 4 is an explanatory diagram of the hardware configuration of the condition monitoring device 100. As shown in Fig. 4, the condition monitoring device 100 includes a CPU (Central Processing Unit) 401, a memory 402, an I / F (Interface) 403, an acceleration sensor 23, and a change amount sensor 24. Each unit is connected by a bus 410.

[0075] The CPU 401 is responsible for overall control of the condition monitoring device 100. For example, the CPU 401 executes the processes of the first acquisition unit 201, the second acquisition unit 202, the estimation unit 203, and the notification unit 204 shown in Fig. 2. The CPU 401 may also control the operation of the acceleration sensor 23 and the operation of the change amount sensor 24.

[0076] The memory 402 is a general term for storage devices such as ROM, RAM, USB (Universal Serial Bus) flash memory, and SSD (Solid State Drive). For example, the storage unit 210 shown in FIG. 2 is configured with the memory 402. The I / F 403 is a general term for input I / F and output I / F (including communication I / F). For example, the second acquisition unit 202 acquires the angular velocity and acceleration of the arm unit 4 from the change amount sensor 24 via the I / F 403 (communication I / F).

[0077] Note that various programs including the abnormality estimation program according to this embodiment are stored in the memory 402. By executing the abnormality estimation program, the CPU 401 realizes the functions of the first acquisition unit 201, the second acquisition unit 202, the estimation unit 203, and the notification unit 204. In other words, the first acquisition unit 201, the second acquisition unit 202, the estimation unit 203, and the notification unit 204 are realized by the CPU 401.

[0078] (Processing Performed by Estimation Control Unit 200) Fig. 5 is a flowchart illustrating an example of processing performed by the estimation control unit 200. In Fig. 5, the estimation control unit 200 determines whether abnormality estimation has started (step S501). The timing to start abnormality estimation may be a predetermined timing (predetermined timing several times a day) or may be the timing when a start instruction is input from the operator. The estimation control unit 200 waits until abnormality estimation has started (step S501: NO). When abnormality estimation has started (step S501: YES), the estimation control unit 200 acquires the three-axis acceleration detected by the change amount sensor 24 (step S502).

[0079] Next, the estimation control unit 200 determines whether the arm unit 4 is in an accelerating / decelerating state (the composite value of Ax, Ay, and Az is not equal to 9.8 m / s ２If the arm unit 4 is in an accelerating / decelerating state (step S503: YES), the estimation control unit 200 returns to step S502. On the other hand, if the arm unit 4 is not in an accelerating / decelerating state (step S503: NO), the estimation control unit 200 determines whether the arm unit 4 is in a stopped state ("composite value of Ax, Ay, Az = 9.8 m / s"). ２ " and "the values ​​of Ax, Ay, and Az are constant" (step S504).

[0080] If the arm 4 is in a stopped state (step S504: YES), the estimation control unit 200 returns to step S502. On the other hand, if the arm 4 is not in a stopped state (step S504: NO), that is, if the arm 4 is in a constant velocity state (the composite value of Ax, Ay, and Az=9.8 m / s ２ " and "the values ​​of Ax, Ay, Az=indeterminate"), the abnormality estimation process (FIG. 6) is executed (step S505), and the series of processes is terminated.

[0081] (Processing Performed by Estimation Control Unit 200 and Change Amount Sensor 24) Fig. 6 is a sequence diagram relating to abnormality estimation performed by estimation control unit 200 and change amount sensor 24. The processing performed by estimation control unit 200 shown in Fig. 6 corresponds to the abnormality estimation processing (step S505) shown in Fig. 5.

[0082] 6, the estimation control unit 200 requests the change amount sensor 24 to transmit the number of rotations of the arm unit 4 (step S601). Upon receiving the transmission request from the estimation control unit 200, the change amount sensor 24 is activated (step S602) and detects the angular velocity of the arm unit 4 (step S603).

[0083] The change amount sensor 24 then calculates the number of rotations of the arm 4 based on the detected angular velocity and Equation 1 (step S604). The change amount sensor 24 then transmits the calculated number of rotations to the estimation control unit 200 (step S605), and ends the process. After making a request to transmit the number of rotations in step S601, the estimation control unit 200 acquires vibration data detected by the acceleration sensor 23 (step S606).

[0084] Then, upon acquiring the rotation speed of the arm unit 4 from the change amount sensor 24 (step S607), the estimation control unit 200 references the number of teeth stored in the storage unit 210 and calculates the target frequency by multiplying the number of teeth by the rotation speed acquired in step S607 (step S608). The estimation control unit 200 then extracts vibration data at the target frequency (step S609). Next, the estimation control unit 200 estimates an abnormality related to gear wear (step S610). The estimation control unit 200 then reports the estimation result (step S611) and ends the series of processes.

[0085] Effect of the embodiment As described above, the condition monitoring device 100 according to the present embodiment estimates an abnormality based on the detection result (vibration data of the reducer 22) of the acceleration sensor 23 provided on the rotating shaft portion 20 and the detection result (number of rotations of the arm portion 4) of the change amount sensor 24 provided on the arm portion 4. This makes it possible to estimate an abnormality in the reducer 22 with a simple configuration, without acquiring information from the mother machine.

[0086] Furthermore, the condition monitoring device 100 according to this embodiment estimates an abnormality based on the detection results of the acceleration sensor 23 provided on the rotating shaft 20 (vibration data of the reducer 22) and the detection results of the change amount sensor 24 provided on the arm 4 (acceleration on three axes of the arm 4). This makes it possible to estimate an abnormality based on the detection results of the acceleration sensor 23 at optimal timing obtained from the acceleration on three axes of the arm 4, without obtaining information from the mother machine. Therefore, an abnormality in the reducer 22 can be estimated with high accuracy with a simple configuration.

[0087] Furthermore, the condition monitoring device 100 according to this embodiment calculates a target frequency based on the rotation speed of the arm 4 and the number of teeth of the gear, extracts data corresponding to the target frequency from vibration data detected by the acceleration sensor 23, and performs abnormality estimation based on the extracted data. This eliminates the need for large-scale calculations such as FFT (Fast Fourier Transformation), thereby reducing power consumption. Therefore, the condition monitoring device 100 can be applied to simple battery-powered devices (e.g., smart sensors).

[0088] Furthermore, the condition monitoring device 100 according to this embodiment extracts data corresponding to a target frequency from vibration time-series data by filtering, and estimates the presence or absence of an abnormality based on the results of comparing the extracted data with a predetermined threshold value. This allows the presence or absence of an abnormality to be estimated with a simpler configuration.

[0089] Furthermore, the condition monitoring device 100 according to this embodiment determines whether the arm 4 is in a constant velocity state based on the comparison result between the combined value of each axis of the three-axis acceleration of the arm 4 and the acceleration of gravity. The condition monitoring device 100 then estimates the presence or absence of an abnormality based on the detection result of the acceleration sensor 23 at the time when the constant velocity state is determined. This improves the accuracy of abnormality estimation.

[0090] Furthermore, the condition monitoring device 100 according to this embodiment identifies a constant velocity state when the composite value of the three-axis acceleration of the arm 4 within a predetermined period of time matches the gravitational acceleration and the values ​​of each axis within the predetermined period of time are indefinite. This makes it possible to easily identify a constant velocity state.

[0091] Furthermore, the condition monitoring device 100 according to this embodiment identifies a non-uniform velocity state of the arm unit 4 based on the combined value of the three-axis acceleration of the arm unit 4 and the acceleration of gravity, and does not estimate the presence or absence of an abnormality in the non-uniform velocity state. Since an abnormality cannot be estimated accurately in a non-uniform velocity state, the accuracy of the abnormality estimation can be improved by not estimating an abnormality.

[0092] Furthermore, the state monitoring device 100 according to this embodiment identifies the acceleration / deceleration state of the arm unit 4 when the composite value of the three-axis acceleration of the arm unit 4 does not match the gravitational acceleration. This makes it possible to easily identify the acceleration / deceleration state.

[0093] Furthermore, the condition monitoring device 100 according to this embodiment determines whether the arm 4 is in a constant-velocity state based on a comparison between the composite value of the three-axis acceleration of the arm 4 and the gravitational acceleration. Furthermore, the condition monitoring device 100 calculates a target frequency based on the rotation speed of the arm 4 and the number of gear teeth when the arm 4 is in a constant-velocity state. The condition monitoring device 100 then extracts data corresponding to the target frequency from the time-series vibration data detected by the acceleration sensor 23 and estimates the presence or absence of an abnormality based on the extracted data. This allows for easy identification of a constant-velocity state and improved accuracy in abnormality estimation. Furthermore, the condition monitoring device 100 eliminates the need for large-scale calculations such as FFT, thereby reducing power consumption. Therefore, the condition monitoring device 100 can be applied to simple battery-powered devices (e.g., smart sensors). (Modifications of the Embodiments) Modifications of the embodiment are listed below. Note that in the following modifications, the content described in the above-described embodiment will be omitted as appropriate. Furthermore, it is also possible to combine the above-described embodiment with the configurations shown in each modification.

[0094] (Modification 1) In the above-described embodiment, an example has been described in which the condition monitoring device 100 calculates a target frequency and extracts data corresponding to the target frequency from time-series data of vibration detected by the acceleration sensor 23. The condition monitoring device 100 according to Modification 1 may extract data corresponding to the target frequency by performing FFT analysis on the time-series data of vibration detected by the acceleration sensor 23.

[0095] However, in the first modification, as in the above-described embodiment, the condition monitoring device 100 performs an abnormality estimation at optimal timing (for example, a constant velocity state) based on the detection results (the acceleration of the arm 4 on three axes) of the change amount sensor 24 provided on the arm 4. According to the first modification, an abnormality estimation based on FFT analysis can be performed at optimal timing obtained from the acceleration of the arm 4 on three axes, without acquiring information from the mother machine. Therefore, it is possible to improve the estimation accuracy when performing an FFT analysis to estimate an abnormality in the reducer 22.

[0096] (Variation 2) In the above-described embodiment, an example has been described in which the acceleration sensor 23 is used as the first detection unit. The condition monitoring device 100 according to Variation 2 may use an iron powder sensor as the first detection unit. The iron powder sensor is a sensor that detects the amount of iron powder (amount of wear) contained in the grease inside the reducer 22. Because iron powder inside the reducer 22 is likely to diffuse during, for example, acceleration and deceleration, it is expected that the accuracy of detecting iron powder during acceleration and deceleration will be improved.

[0097] Therefore, the condition monitoring device 100 identifies the acceleration / deceleration state based on the detection result of the change amount sensor 24 (the acceleration on three axes of the arm unit 4), and performs abnormality estimation based on the detection result of the iron powder sensor at the identified timing. According to the second modification, abnormality estimation can be performed based on the detection result of the iron powder sensor at the optimal timing obtained from the acceleration on three axes of the arm unit 4, without obtaining information from the mother machine. Therefore, the accuracy of abnormality estimation of the reducer 22 can be improved.

[0098] (Modification 3) In the above-described embodiment, an example has been described in which the change amount sensor 24 is provided in a portion of the arm 4 that is located closer to the base end than the rotating shaft 20 that is located on the tip side. The change amount sensor 24 according to Modification 3 may be provided in the rotating shaft 20 that is located on the tip side of the arm 4.

[0099] More specifically, the third rotating shaft 20c is provided at one end (tip) of the boom 5 and is therefore included in the boom 5. For this reason, in Modification 3, the second change amount sensor 24b is provided on the third rotating shaft 20c included in the boom 5. Furthermore, the fourth rotating shaft 20d is provided at one end (tip) of the shovel arm 7 and is therefore included in the shovel arm 7. For this reason, in Modification 3, the third change amount sensor 24c is provided on the fourth rotating shaft 20d included in the shovel arm 7.

[0100] The second change amount sensor 24b and the third change amount sensor 24c are both sensors capable of detecting triaxial acceleration and triaxial angular velocity. As a result, the second change amount sensor 24b not only serves to detect the rotation speed of the boom 5 (the function of the second detection unit) but also serves to detect vibrations of the third reducer 22c (the gear at the joint between the boom 5 and the shovel arm 7) (the function of the first detection unit). Similarly, the third change amount sensor 24c not only serves to detect the rotation speed of the shovel arm 7 (the function of the second detection unit) but also serves to detect vibrations of the fourth reducer 22d (the gear at the joint between the shovel arm 7 and the bucket 9) (the function of the first detection unit). According to the third modification, the number of sensors serving as the first detection unit can be reduced, resulting in a simpler configuration.

[0101] (Modification 4) In the above-described embodiment, an example has been described in which the estimation result of an abnormality is reported as is. In Modification 4, an example will be described in which an abnormality is predicted based on performance data on which an abnormality estimation has been performed. Below, an example will be described in which an abnormality is predicted using a trained model.

[0102] In Variation 4, generation of a trained model will be described in which the rotation speed of the arm unit 4 and the vibration data of the reducer 22 are used as input samples, and the time until maintenance of the reducer 22 (usable period: 1 month remaining, 2 months remaining, ...) is used as an output sample. The trained model is generated by a learning device such as a personal computer. The learning device uses a pre-prepared data set to train parameters of a classification model such as a neural network. The data set uses actual data on various detection results and estimation results by the condition monitoring device 100. The classification model includes an input unit, a feature calculation unit, a classification unit, and an output unit. The input unit outputs the input rotation speed of the arm unit 4 and the vibration data of the reducer 22 to the feature calculation unit as vectors. The input unit constitutes the input layer of the neural network.

[0103] The feature calculation unit and classification unit are intermediate layers of the neural network. The output unit is the output layer of the neural network. The feature calculation unit converts the vector input from the input unit into a low-dimensional feature vector and outputs it to the classification unit. The classification unit converts the feature vector input from the feature calculation unit into a P-dimensional vector indicating the posterior probability of the usable period represented by the feature vector, where P is the number of usable periods to be estimated.

[0104] The learning device acquires a learning data set that associates the rotation speed of the arm unit 4, which is an input sample, with vibration data of the reducer 22. The rotation speed of the arm unit 4 and the vibration data of the reducer 22 are represented by a P-dimensional one-hot vector, where P is the number of usable periods in the data set.

[0105] The learning device uses the acquired learning dataset to train parameters of the classification model so that, when the rotation speed of the arm unit 4 and vibration data of the reducer 22 are input, a P-dimensional vector indicating the posterior probability of the usable period is output. Specifically, the learning device uses the acquired dataset to train parameters of the classification model. At this time, the learning device updates parameters of the feature calculation unit and the classification unit of the classification model.

[0106] More specifically, the learning device updates each parameter using the gradient descent method to minimize a loss function using the calculation results of the classification model. For example, the loss function represents the cross-entropy error between the output value of the classification model and the output samples of the dataset. The learning device terminates the learning process when the evaluation value of the loss function falls below a predetermined threshold or when the learning process has been repeated a predetermined number of times, and a trained model is generated. By inputting the rotation speed of the arm unit 4 and vibration data of the reducer 22 into this trained model, an estimated value can be calculated and a usable period with a high estimated value can be output.

[0107] The generated trained model may be stored in another device (e.g., a server) different from the condition monitoring device 100. In this case, when the condition monitoring device 100 detects the rotation speed of the arm section 4 and the vibration data of the reducer 22, it transmits the detection results to the other device that stores the trained model. The other device determines the usable period based on the detection results and outputs the determined usable period. The output usable period is displayed, for example, on a display provided in the construction machine 1 or on a personal computer located in a maintenance shop for the construction machine 1. The trained model may be stored in the condition monitoring device 100. In this case, the condition monitoring device 100 may determine the usable period based on the detection results and output the determined usable period.

[0108] According to the fourth modification, the usable period can be determined using the trained model based on the number of rotations of the arm portion 4 detected by the condition monitoring device 100 and the vibration data of the reducer 22. Furthermore, an abnormality in the reducer 22 can be predicted with a simple configuration without acquiring information from the mother machine. Note that, although the fourth modification has been described as an example in which the output sample represents the usable period, the present invention is not limited to this, and the output sample can also represent the replacement time (year and month, etc.).

[0109] (Other Examples) In the above-described embodiment, an example has been described in which the reducer 22 is a gear mechanism having multiple gears. However, the reducer 22 may have pins (internal pins) in addition to gears. For example, the reducer 22 may be configured such that multiple pin grooves are formed on the inner circumferential surface of the case to hold each of the multiple internal pins, and a gear mechanism (e.g., one or more oscillating gears) inside the case meshes with the multiple internal pins. The pin grooves or pins may be considered part of the gear mechanism. For example, multiple pin grooves may be referred to as gears, and each pin may be referred to as a tooth. For example, the number of teeth may include the number of pins. Note that if the number of teeth does not include the number of pins, the target frequency is obtained by multiplying the rotation speed of the arm unit 4 by the number of teeth of the gear and the number of pins.

[0110] In the above-described embodiment, an example has been described in which the condition monitoring device 100 is applied to a construction machine 1. However, the condition monitoring device 100 may also be applied to a robot (e.g., a six-axis robot). The robot is, for example, an industrial robot. In this case, the condition monitoring device 100 detects a first physical quantity related to wear of a gear provided at a joint portion of the robot arm and a second physical quantity related to the motion state of a gear provided at the tip side of the robot arm, and estimates the presence or absence of an abnormality related to gear wear based on these detection results.

[0111] Although an embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs within the scope of the gist of the present invention.

[0112] For example, among the embodiments disclosed herein, those in which multiple functions are provided in a distributed manner may have some or all of the multiple functions integrated together, and conversely, those in which multiple functions are provided in a distributed manner may have some or all of the multiple functions integrated together. Regardless of whether the functions are integrated or distributed, it is sufficient that the configuration can achieve the object of the invention.

[0113] In the embodiment, a gear is used as an example of the mechanical element provided at the joint of the arm unit 4, but the present invention is not limited to a gear. For example, the mechanical element may be a bearing or a motor provided at the joint. Furthermore, in the embodiment, the vibration of a gear is used as an example of the first physical quantity related to the mechanical element, but the present invention is not limited to this. For example, the motor current of a motor (mechanical element) may be used as the first physical quantity. In this case, for example, the rotation speed of the arm unit 4 may be used as the second physical quantity related to the motion state of the arm unit 4. In this case, for example, the amplitude value of the frequency of the motor current determined from the rotation speed may be monitored to estimate a motor failure sign (the state of the mechanical element).

[0114] A program (e.g., a status monitoring program) for realizing the functions of the device (e.g., the status monitoring device 100) according to the above-described embodiment may be stored in a computer-readable storage medium, and the program stored in the storage medium may be read and executed by a computer system to perform processing. Note that the term "computer system" may also include hardware such as an operating system (OS) or peripheral devices. Also, the term "computer-readable storage medium" refers to a flexible disk, a magneto-optical disk, a read-only memory (ROM), a writable non-volatile memory such as a flash memory, a portable medium such as a DVD (Digital Versatile Disc), or a storage device such as a hard disk built into a computer system.

[0115] Furthermore, the term "computer-readable storage medium" also includes a medium that stores a program for a certain period of time, such as a volatile memory (e.g., DRAM (Dynamic Random Access Memory)) within an information processing device or a client computer system when the program is transmitted via a network such as the Internet or a communication line such as a telephone line. The program may also be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by transmission waves within the transmission medium. Here, the "transmission medium" that transmits the program refers to a medium that has the function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may also be a so-called differential file (differential program) that can realize the above-mentioned functions in combination with a program already stored in the computer system.

[0116] According to the present invention, the state of a machine element (for example, an abnormality in a gear) can be estimated with a simple configuration.

[0117] DESCRIPTION OF SYMBOLS 1...construction machine, 2...main body, 4...arm section, 5...boom, 7...shovel arm, 9...bucket, 20...rotating shaft section, 22...reduction gear, 23...acceleration sensor, 24...variation amount sensor, 100...condition monitoring device, 200...estimation control section, 201...first acquisition section, 202...second acquisition section, 203...estimation section, 203a...calculation section, 203b...identification section, 204...notification section

Claims

1. A first detection unit provided at a joint portion of a rotatable arm, for detecting a first physical quantity related to a mechanical element provided at the joint portion; A second detection unit provided at a tip side of the arm different from a base end side where the joint portion is disposed, for detecting a second physical quantity related to a motion state of the arm; An estimation unit for estimating a state of the mechanical element based on the first physical quantity detected by the first detection unit and the second physical quantity detected by the second detection unit; A condition monitoring device comprising the above.

2. The mechanical element is a gear provided at the joint portion; The first physical quantity is vibration of the gear; The second physical quantity is at least one physical quantity of acceleration and angular velocity of the arm; The estimation unit has a calculation unit for calculating a specific frequency of the vibration based on a detection result of the second physical quantity during a predetermined period in which the arm is moving and the number of teeth of the gear; Furthermore, the estimation unit estimates the state of the gear using the calculated specific frequency and the detection result of the first physical quantity during the predetermined period. The condition monitoring device according to Claim 1.

3. The estimation unit extracts data corresponding to the calculated specific frequency from time series data of the detection result of the first physical quantity during the predetermined period, and estimates the state of the gear based on the extracted data. The condition monitoring device according to Claim 2.

4. The estimation unit executes a filtering process for extracting data corresponding to the specific frequency from the time series data of the detection result of the first physical quantity, and estimates the state of the gear based on a comparison result between the extracted data and a predetermined threshold value. The condition monitoring device according to Claim 3.

5. The second physical quantity is a three-axis acceleration of the arm; The estimation unit has a specifying unit for specifying whether or not it is a constant speed state indicating that the arm is operating at a constant speed based on a comparison result between a combined value of each axis of the three-axis acceleration and gravitational acceleration; The estimation unit estimates the state of the mechanical element based on the detection result of the first detection unit at a time point when it is specified that it is the constant speed state. The condition monitoring device according to any one of Claims 1 to 4.

6. The specifying unit specifies that the arm is in the constant speed state when the combined value of each axis and gravitational acceleration match within a predetermined period and the values of each axis are indeterminate within the predetermined period. The condition monitoring device according to Claim 5. ​ ​ ​ ​ ​

7. The specific part identifies a non-constant speed state other than the constant speed state based on the combined value of each axis and the gravitational acceleration, and the estimation part does not estimate the state of the mechanical element in the non-constant speed state. The state monitoring device according to claim 5.

8. When the combined value of each axis does not match the gravitational acceleration, the specific part identifies an acceleration / deceleration state indicating that the arm is operating with acceleration or deceleration as the non-constant speed state. The state monitoring device according to claim 7.

9. The mechanical element is a gear provided at the joint part, the first physical quantity is the vibration of the gear, the second physical quantity includes the rotation speed of the arm and the three-axis acceleration of the arm, the estimation part includes a specific part that identifies whether it is a constant speed state indicating that the arm is operating at a constant speed based on a comparison result between the combined value of each axis of the three-axis acceleration and the gravitational acceleration, and a calculation part that calculates a specific frequency of the vibration based on information including at least the rotation speed of the arm and the number of teeth of the gear at the time when it is identified as the constant speed state, and the estimation part extracts data corresponding to the calculated specific frequency from the time-series data of the detection result of the first physical quantity, and estimates the state of the gear based on the extracted data. The state monitoring device according to claim 1.

10. A computer-readable storage medium storing a program that causes a computer to function as a state monitoring device, wherein the program causes the computer to function as a first acquisition part that acquires the first physical quantity from a first detection part that is provided at a joint part of a rotatable arm and detects the first physical quantity related to a mechanical element provided at the joint part, a second acquisition part that acquires the second physical quantity from a second detection part that is provided on a tip side different from a base end side where the joint part is arranged of the arm and detects the second physical quantity related to a motion state of the arm, and an estimation part that estimates the state of the mechanical element based on the first physical quantity acquired by the first acquisition part and the second physical quantity acquired by the second acquisition part.