Anomaly detection system and anomaly detection method

Abstract

We provide an anomaly detection system that can accurately detect abnormalities in speed reducers. [Solution] An abnormality detection system according to one embodiment of the present disclosure is an abnormality detection system for detecting an abnormality in a speed reducer, comprising: a first vibration sensor, a second vibration sensor, and a third vibration sensor attached to the surface of the housing of the speed reducer for detecting vibrations on the surface of the housing of the speed reducer; and an abnormality determination unit that determines whether or not there is an abnormality in the speed reducer based on the vibrations detected by each of the first vibration sensor, the second vibration sensor, and the third vibration sensor.

Description

Technical Field

[0001] The present technology relates to an abnormality detection system and an abnormality detection method.

Background Art

[0002] Extruders are widely used as industrial machines for manufacturing plastic products. For example, as one type of extruder, there is a twin-screw extruder that kneads plastic raw materials using a twin screw (see Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In an extruder, which is one type of industrial machine, a speed reducer that transmits the power generated by a motor to each of the twin screws is used. The speed reducer converts the power generated by the motor into a predetermined torque and transmits it to the twin screws. The speed reducer incorporates various components such as shafts, bearings, and gears. Since the speed reducer incorporates various components in this way, if a defect (such as the occurrence of scratches) occurs in a part of each component, there is a risk that this defect in the component will lead to a failure of the entire speed reducer. Therefore, there is a need for an abnormality detection system that can accurately detect abnormalities in the speed reducer (components).

[0005] The present disclosure has been made in view of such circumstances, and an object thereof is to provide an abnormality detection system and the like that can accurately detect abnormalities in a speed reducer.

Means for Solving the Problems

[0006] An abnormality detection system according to one embodiment of the present disclosure is an abnormality detection system for detecting an abnormality in a speed reducer, comprising: a first vibration sensor, a second vibration sensor, and a third vibration sensor attached to the housing surface of the speed reducer for detecting vibrations on the housing surface of the speed reducer; and an abnormality determination unit that determines whether or not there is an abnormality in the speed reducer based on the vibrations detected by each of the first vibration sensor, the second vibration sensor, and the third vibration sensor. [Effects of the Invention]

[0007] In an abnormality detection system according to one embodiment of the present disclosure, it is possible to accurately detect abnormalities in the speed reducer. [Brief explanation of the drawing]

[0008] [Figure 1] This is a block diagram showing an example configuration of a molding machine system (anomaly detection system) according to Embodiment 1. [Figure 2] This is a schematic diagram showing an example of the configuration of a molding machine according to Embodiment 1. [Figure 3] This is a block diagram showing an example configuration of a data acquisition device according to Embodiment 1. [Figure 4] This is a block diagram showing an example configuration of the abnormality detection device according to Embodiment 1. [Figure 5] This is a functional block diagram of the processing unit of the abnormality detection device involved in the abnormality detection process. [Figure 6] This is a perspective view of the gearbox illustrating the mounting position of the vibration sensor. [Figure 7] This is a perspective view of the gearbox illustrating the mounting position of the vibration sensor. [Figure 8] This flowchart shows an example of an anomaly detection process. [Figure 9] This graph shows the vibration waveforms detected by the first vibration sensor, the second vibration sensor, and the third vibration sensor. [Figure 10] This graph shows an example of statistical processing performed using the vibration acceleration detected by the first vibration sensor, the second vibration sensor, and the third vibration sensor. [Figure 11] A graph showing an example of frequency analysis of vibrations detected by a vibration sensor. [Figure 12] A graph for explaining abnormality determination. [Figure 13] A graph for explaining an example of the power spectrum obtained by frequency analysis. [Figure 14] An explanatory diagram showing an example of an abnormal frequency DB. [Figure 15] A conceptual diagram showing an example of the record layout of a collection data DB. [Figure 16] A conceptual diagram showing a life prediction model. [Figure 17] A flowchart showing the learning processing procedure of a life prediction model. [Figure 18A] A graph showing the learning result of a life prediction model. [Figure 18B] A graph showing the inference result using a life prediction model. [Figure 19] A flowchart showing the processing procedure of life prediction processing. [Figure 20] A flowchart showing the processing procedure of display processing. [Figure 21] A schematic diagram showing an example of an overall display screen displayed on a portal site. [Figure 22] A schematic diagram showing an example of a remaining life estimation result display screen displayed on a portal site. [Figure 23] A perspective view of a speed reducer for explaining the mounting position of a vibration sensor according to Embodiment 2. [Figure 24] A perspective view of a speed reducer for explaining the mounting position of a vibration sensor according to Embodiment 2. [Figure 25] An explanatory diagram for explaining the details of a mounting seat. [Figure 26] An explanatory diagram for explaining the correlation of the temporal change of vibration data related to each vibration sensor when vibration sensors are attached to all mounting seats. [Figure 27] A schematic perspective view of a speed reducer according to Embodiment 3. [Figure 28]This is a schematic perspective view of the gearbox according to Embodiment 3. [Modes for carrying out the invention]

[0009] An anomaly detection system and anomaly detection method according to embodiments of this disclosure will be described below with reference to the drawings. This disclosure is not limited to these examples, but is intended to include all modifications within the meaning and scope of the claims, as indicated by the claims. Furthermore, at least some of the embodiments described below may be combined in any way.

[0010] (Embodiment 1) Figure 1 is a block diagram showing an example configuration of a molding machine system (anomaly detection system) according to Embodiment 1. The molding machine system comprises a molding machine 1, a plurality of detectors 2, a data acquisition device 3, a router 4, an anomaly determination device 5, and terminal devices 6a and 6b. The molding machine 1 includes an injection molding machine and an extruder. Hereinafter, the molding machine 1 will be described as a twin-screw extruder, for example. In this embodiment, the molding machine 1 performs injection molding or extrusion molding of plastic, but it may also perform injection molding or extrusion molding of, for example, a magnesium alloy.

[0011] Figure 1 shows one molding machine 1 and a data acquisition device 3, but the anomaly detection device 5 is connected to multiple data acquisition devices 3 (not shown) via a network. One or more molding machines 1 are connected to the data acquisition device 3. The anomaly detection device 5 can collect information from each of the multiple molding machines 1 and predict the remaining lifespan of one or more components constituting each molding machine 1. The multiple molding machines 1 and data acquisition devices 3 are assumed to be installed in the factories of multiple users who own the molding machines 1. The users are not operators, but rather organizations such as corporations that own the molding machines 1.

[0012] Terminal devices 6a and 6b are communication terminals having a display unit, such as a computer, tablet, or smartphone. Terminal device 6a is a terminal used by the user. Terminal device 6b is a terminal used by service providers such as sales representatives and maintenance personnel related to the user's molding machine 1.

[0013] <Forming machine 1> Figure 2 is a schematic diagram showing an example of the configuration of a molding machine 1 according to Embodiment 1. The molding machine 1 comprises a cylinder 10 having a hopper 10a into which resin raw material is fed, two screws 11, and a die 12 (see Figure 1) provided at the outlet of the cylinder 10. The two screws 11 are arranged substantially parallel to each other in a meshed state and are rotatably inserted into the hole of the cylinder 10, transporting the resin raw material fed into the hopper 10a in the extrusion direction (to the right in Figures 1 and 2), where it is melted and kneaded. The molten resin raw material is discharged from the die 12 which has a through hole. Screw 11 is constructed by combining and integrating multiple types of screw pieces. For example, screw 11 is constructed by combining forward flight pieces in the shape of a flight screw for transporting resin raw materials in the forward direction, reverse flight pieces for transporting resin raw materials in the reverse direction, and kneading pieces for mixing resin raw materials, etc., in an order and position according to the characteristics of the resin raw materials.

[0014] The molding machine 1 also includes a motor 13 that outputs a driving force to rotate the screw 11, a reduction gear 14 that reduces and transmits the driving force of the motor 13, and a control device 15. The motor 13 is connected to the input shaft 81 of the reduction gear 14. The two screws 11 are connected to the output shafts 85 and 86 of the reduction gear 14, respectively. The screw 11 rotates due to the driving force of the motor 13, which is reduced and transmitted by the reduction gear 14.

[0015] <Detector 2> Detector 2 detects physical quantities related to the state of the components constituting the molding machine 1 and outputs the detected physical quantity data directly or indirectly to the data acquisition device 3. The physical quantity data is time-series data of detected values ​​indicating the detected physical quantities. Detector 2 includes those installed in the molding machine 1 for operational control and those installed to estimate the lifespan of the components. Some of the multiple detectors 2 are connected to the data acquisition device 3, and the data acquisition device 3 acquires physical quantity data from these detectors 2. Some of the multiple detectors 2 are connected to the control device 15, and the data acquisition device 3 acquires physical quantity data from these detectors 2 via the control device 15.

[0016] Physical quantities include temperature, position, velocity, acceleration, current, voltage, pressure, time, image data, torque, force, strain, power consumption, and weight. These physical quantities can be measured using thermometers, position sensors, velocity sensors, vibration (acceleration) sensors, ammeters, voltmeters, pressure gauges, timers, cameras, torque sensors, power meters, weighing scales, etc. A vibration sensor may be composed of a velocity sensor. In this case, acceleration is calculated by differentiating the time-series data of velocity detected by the vibration sensor.

[0017] The multiple detectors 2 include, for example, a first detector 21 that detects a physical quantity related to the gearbox 14, a second detector 22 that detects a physical quantity related to the screw 11, a third detector 23 that detects a physical quantity related to the motor 13, and a fourth detector 24 that detects a physical quantity related to the die 12. The first detector 21 is, for example, a sensor (vibration sensor) that detects vibration (acceleration) of the reduction gear 14. Hereafter, the first detector 21 will also be referred to as the vibration sensor 21. The vibration sensor 21 includes a first vibration sensor 211, a second vibration sensor 212, and a third vibration sensor 213. Details of the vibration sensor 21 (first vibration sensor 211, second vibration sensor 212, and third vibration sensor 213) will be described later. The second detector 22 is a torque detector that detects the shaft torque of the screw 11, a tachometer that detects the rotational speed of the screw 11, a pressure gauge that detects the screw tip pressure, a thermometer that detects the temperature of the screw 11, a displacement sensor that detects the displacement of the rotation center of the screw 11, etc. The third detector 23 is an ammeter that detects the motor current, a tachometer that detects the motor rotational speed, etc. The fourth detector 24 is a pressure gauge that detects the die head pressure acting on the die 12. The values ​​detected by the detector 2 may be transmitted directly to the data acquisition device 3, or they may be transmitted to the data acquisition device 3 via the control device 15.

[0018] <Control device 15> The control device 15 is a computer that controls the operation of the molding machine 1 and includes a transmitting / receiving unit (not shown) and a display unit that send and receive information with the data acquisition device 3. Specifically, the control device 15 transmits operating data indicating the operating status of the molding machine 1 to the data acquisition device 3. The operating data includes, for example, motor current, screw 11 rotation speed, screw 11 tip pressure, die head pressure, feeder supply amount (amount of resin raw material supplied), extrusion amount, cylinder temperature, resin pressure, etc. The control device 15 receives information from the data acquisition device 3 indicating whether or not there is an abnormality in the reduction gear 14, various graph data, and remaining life data indicating the remaining life of the components constituting the molding machine 1. The control device 15 displays the contents of the received graph data and remaining life data. The control device 15 also outputs a warning according to the remaining life indicated by the received remaining life data.

[0019] <Data acquisition device 3> Figure 3 is a block diagram showing an example configuration of the data acquisition device 3 according to Embodiment 1. The data acquisition device 3 is a computer and comprises a control unit 31, a storage unit 32, a communication unit 33, and a data input unit 34, with the storage unit 32, the communication unit 33, and the data input unit 34 being connected to the control unit 31. The data acquisition device 3 is, for example, a PLC (Programmable Logic Controller).

[0020] The control unit 31 includes arithmetic processing circuits such as a CPU (Central Processing Unit), multi-core CPU, ASIC (Application Specific Integrated Circuit), and FPGA (Field-Programmable Gate Array), internal storage devices such as ROM (Read Only Memory) and RAM (Random Access Memory), and I / O terminals. The control unit 31 executes a control program stored in the storage unit 32 (described later) to collect physical quantity data and transmit it to the abnormality determination device 5. Note that each functional unit of the data acquisition device 3 may be implemented in software, or some or all of them may be implemented in hardware.

[0021] The storage unit 32 is a hard disk, EEPROM (Electrically Erasable Programmable Memory). This is a non-volatile memory such as ROM or flash memory. The storage unit 32 stores a control program that causes the computer to perform the physical quantity data collection process.

[0022] The communication unit 33 is a communication circuit that sends and receives information according to a predetermined communication protocol such as Ethernet (registered trademark). The communication unit 33 is connected to the control device 15 via a first communication network such as a LAN, and the control unit 31 can send and receive various types of information to and from the control device 15 via the communication unit 33. The control unit 31 acquires physical quantity data via the communication unit 33. A router 4 is connected to the first network, and the communication unit 33 is connected to the anomaly detection device 5 on the cloud, which is the second communication network, via the router 4. The control unit 31 can send and receive various information to and from the anomaly detection device 5 via the communication unit 33 and the router 4.

[0023] The data input unit 34 is an input interface to which the signal output from the detector 2 is input. The detector 2 is connected to the data input unit 34, and the control unit 31 acquires physical quantity data via the data input unit 34.

[0024] <Abnormality determination device 5> Figure 4 is a block diagram showing an example configuration of the abnormality detection device 5 according to Embodiment 1. The abnormality detection device 5 is a computer and comprises a processing unit 51, a storage unit 52, and a communication unit 53. The storage unit 52 and the communication unit 53 are connected to the processing unit 51.

[0025] The processing unit 51 is a processor and includes arithmetic processing circuits such as a CPU, multi-core CPU, GPU (Graphics Processing Unit), GPGPU (General-purpose computing on graphics processing units), TPU (Tensor Processing Unit), ASIC, FPGA, NPU (Neural Processing Unit), internal storage devices such as ROM and RAM, and I / O terminals. The processing unit 51 functions as an abnormality detection device 5 according to this embodiment by executing a computer program P stored in the storage unit 52, which will be described later. Note that each functional part of the abnormality detection device 5 may be implemented in software, or some or all of them may be implemented in hardware.

[0026] The communication unit 53 is a communication circuit that sends and receives information according to a predetermined communication protocol such as Ethernet (registered trademark). The communication unit 53 is connected to the data acquisition device 3 and terminal devices 6a and 6b via a second communication network, and the control unit 31 can send and receive various types of information to and from the data acquisition device 3 and terminal devices 6a and 6b via the communication unit 53.

[0027] The storage unit 52 is a non-volatile memory such as a hard disk, EEPROM, or flash memory. The storage unit 52 stores a computer program P for causing a computer to perform a process to estimate the lifespan of the components constituting the molding machine 1, an abnormal frequency DB 52a, collected data DB 52b, and a lifespan prediction model M.

[0028] The computer program P, etc., may be recorded on the recording medium 50 in a manner that is computer-readable. The storage unit 52 stores the computer program P, etc., read from the recording medium 50 by a reading device (not shown). The recording medium 50 is a semiconductor memory such as flash memory. The recording medium 50 may also be an optical disc such as a CD (Compact Disc)-ROM, DVD (Digital Versatile Disc)-ROM, or BD (Blu-ray® Disc). Furthermore, the recording medium 50 may be a magnetic disc such as a flexible disk or hard disk, or a magneto-optical disc. In addition, the computer program P, etc., may be downloaded from an external server (not shown) connected to a communication network (not shown) and stored in the storage unit 52.

[0029] <Anomaly detection process> Figure 5 is a functional block diagram of the processing unit 51 of the abnormality determination device 5 involved in abnormality determination processing. As shown in Figure 5, the processing unit 51 of the abnormality determination device 5 includes a statistical analysis unit 511, a preprocessing unit 512, a frequency analysis unit 513, a peak processing unit 514, and an abnormality determination unit 515. The processing unit 51 acquires the vibration detection values ​​(vibration data) detected by the first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213, and determines an abnormality in the reduction gear 14 based on the acquired detection values.

[0030] Figures 6 and 7 are perspective views of the gearbox 14 illustrating the mounting position of the vibration sensor 21. The vibration sensor 21 is mounted on the housing of the gearbox 14 and detects vibrations on the surface of the housing of the gearbox 14. In this embodiment, the first vibration sensor 211 detects vibrations in the H (Horizontal) direction. The second vibration sensor 212 detects vibrations in the A (Axis) direction. The third vibration sensor 213 detects vibrations in the V (Vertical) direction.

[0031] As shown in Figures 6 and 7, the housing (case) of the reduction gear 14 is a roughly rectangular (cuboidal) box shape and is composed of six faces f. However, the shape of the housing of the reduction gear 14 is not limited to a roughly rectangular (cuboidal) box shape; for example, it may be semi-circular, or a shape in which multiple rectangular (cuboidal) boxes are combined. As shown in Figure 6, the input side face (input surface) f1 of the reduction gear 14, that is, the input surface f1 on which power is input from the motor 13 (see Figure 1), has an input shaft 81 and a bearing (thrust bearing) 82. The reduction gear 14 also has a first side surface f2 connected to the left side of the input surface f1 when viewed from the outside, an upper surface f3 connected to the upper side of the input surface f1, a second side surface f4 connected to the right side of the input surface f1, an output surface f5 facing the input surface, and a lower surface f6 connected to the lower side of the input surface. As shown in Figure 7, the output side output surface f5 of the reduction gear 14 has output shafts 85 and 86. The reduction gear 14 converts the power transmitted from the motor 13 (see Figure 1) to the input shaft 81 into a predetermined torque, and transmits the power of the predetermined torque to the two screws 11 (see Figure 1) via the output shafts 85 and 86.

[0032] As shown in Figure 6, the input shaft 81 is positioned on the input surface f1 closer to the second side surface f4 than the bearing 82. The bearing 82 is positioned on the input surface f1 closer to the first side surface f2 than the input shaft 81. The bearing 82 supports the thrust shafts that transmit the power input to the input shaft 81 to the output shafts 85 and 86. The output shafts 85 and 86 are positioned on the output surface f5 closer to the first side surface than the center. Note that the arrangement of the input shaft 81, bearing 82, and output shafts 85 and 86 is not limited to that described above. For example, the input shaft 81, bearing 82, or output shafts 85 and 86 may be arranged in a reversed configuration in the H direction on the input surface f1 or output surface f5.

[0033] In this embodiment, direction A corresponds to the axial direction of the input shaft 81. Direction V corresponds to the vertical direction of the reducer 14. Direction H corresponds to the direction perpendicular to the axial direction of the input shaft 81 (direction A) and the vertical direction of the reducer 14 (direction V).

[0034] As shown in Figures 6 and 7, each vibration sensor 21 is mounted on one of the three surfaces constituting a corner of the housing of the reduction gear 14. The first vibration sensor 211 is mounted on the first side surface f2 near corner C1, which is formed by the input surface f1, the first side surface f2, and the top surface f3. The second vibration sensor 212 is mounted on the input surface f1 near corner C2, which is formed by the input surface f1, the top surface f3, and the second side surface f4. The third vibration sensor 213 is mounted near corner C3, which is formed by the first side surface f2, the top surface f3, and the output surface f5. If the output shafts 85 and 86 are located closer to the second side surface f4 than to the first side surface f2, the third vibration sensor 213 may be provided near corner C4, which is formed by the top surface f3, the second side surface f4, and the output surface f5. In other words, the third sensor is mounted near the corner C3 formed by the first side surface f2, the top surface f3, and the output surface f5, or near the corner C4 formed by the top surface f3, the second side surface f4, and the output surface f5, whichever is closer to the output shafts 85 and 86. Also, as shown in Figures 6 and 7, the first vibration sensor 211 is mounted near the bearing 82, the second vibration sensor 212 is mounted near the input shaft 81, and the third vibration sensor 213 is mounted near the output shafts 85 and 86.

[0035] The statistical analysis unit 511 shown in Figure 5 performs statistical analysis of vibrations detected by the vibration sensor 21. Specifically, the statistical analysis unit 511 performs statistical analysis using the vibration acceleration detected by the vibration sensor 21. Here, statistical analysis is a method of statistically analyzing vibration acceleration data, for example, a method of determining the frequency distribution of the magnitude of vibration acceleration detected by the vibration sensor 21. Details of the statistical analysis will be described later. In this embodiment, statistical analysis may also be performed using velocity or displacement in addition to vibration acceleration. The same applies to the frequency analysis described below.

[0036] The preprocessing unit 512 performs preprocessing on the vibration data detected by the vibration sensor 21. For example, as preprocessing, the preprocessing unit 512 removes noise contained in the vibration data detected by the vibration sensor 21. Noise removal can be performed using a filter such as a low-pass filter. Alternatively, as preprocessing, the preprocessing unit 512 may perform envelope processing on the vibration data detected by the vibration sensor 21.

[0037] The frequency analysis unit 513 performs frequency analysis of the vibrations detected by the vibration sensor 21. Specifically, the frequency analysis unit 513 performs frequency analysis on vibration data from which noise has been removed by the preprocessing unit 512. For example, the frequency analysis unit 513 generates a spectrum (Fourier spectrum or power spectrum) of the vibrations detected by the vibration sensor 21 that shows the magnitude of acceleration as a function of frequency. The Fast Fourier Transform (FFT) can be used for frequency analysis. Details of the frequency analysis will be described later. In the following description, the frequency analysis unit 513 is assumed to generate a power spectrum, but the frequency analysis unit 513 may also generate a Fourier spectrum with a linear scale on the vertical axis.

[0038] The peak processing unit 514 performs a process to clarify the power spectrum generated by the frequency analysis unit 513, that is, a process to emphasize the peaks. For example, the peak processing unit 514 can emphasize the peaks by adding the data (power spectrum) after frequency analysis a predetermined number of times and then dividing it by a predetermined value (averaging process).

[0039] The abnormality determination unit 515 determines abnormalities in the speed reducer 14 based on the analysis results of the statistical analysis unit 511 and the analysis results of the frequency analysis unit 513 (output of the peak processing unit 514). Specifically, the abnormality determination unit 515 determines the abnormality level of the speed reducer 14 based on the results of the statistical analysis in the statistical analysis unit 511. For example, the abnormality determination unit 515 can determine the abnormality level of the speed reducer 14 based on the frequency distribution of the magnitude of vibration acceleration detected by the vibration sensor 21.

[0040] Furthermore, the abnormality detection unit 515 can identify the component of the reducer 14 where an abnormality has occurred based on the power spectrum generated by the frequency analysis unit 513. The type of component identified by the abnormality detection unit 515 may include, for example, thrust bearings, gears, or various radial bearings. The abnormality detection unit 515 may also identify the type of abnormality, such as scratches, cracks, wear, or corrosion. In addition, the abnormality detected by the abnormality detection unit 515 may be an oil abnormality.

[0041] For example, the memory unit 52 stores abnormal frequency DB52a, which associates a component in the speed reducer 14 that has malfunctioned with a frequency at which the power value increases in the event of that malfunction. The abnormality determination unit 515 can identify the component in the speed reducer 14 that has malfunctioned by comparing the power spectrum generated by the frequency analysis unit 513 with the abnormal frequency DB52a previously stored in the memory unit 52. Details of the abnormal frequency DB52a will be described later.

[0042] Figure 8 is a flowchart illustrating an example of the abnormality detection process. The steps in the flowchart will be explained below, with reference to Figures 9-14 as appropriate. The processing unit 51 of the abnormality detection device 5 acquires vibration data from the vibration sensor 21 attached to the reduction gear 14, specifically from the surface of the reduction gear 14's housing (S1). At this time, the processing unit 51 acquires vibration data measured a predetermined number of times (e.g., 16 times) at a predetermined period (e.g., 520 μs).

[0043] Figure 9 is a graph showing the vibration waveforms detected by the first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213. As shown in Figure 9, the vibration data from the vibration sensor 21 is waveform data where the horizontal axis represents time and the vertical axis represents acceleration. In other words, the vibration data shown in Figure 9 shows the change in the acceleration of vibration on the housing surface of the reduction gear 14 over time. In the vibration data shown in Figure 9, a larger acceleration indicates a larger amplitude of vibration on the housing surface. In the example shown in Figure 9, among the waveforms of the first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213, the acceleration of the vibration data from the third vibration sensor 213 is the largest, which indicates that the amplitude of vibration is large at the position where the third vibration sensor 213 is attached.

[0044] Next, the processing unit 51 of the abnormality detection device 5 performs statistical analysis using the vibration acceleration detected by the vibration sensor 21 via the statistical analysis unit 511 (S2).

[0045] Figure 10 is a graph showing an example of statistical processing using the vibration acceleration detected by the first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213. The graph in Figure 10 shows an example of determining the frequency distribution of the magnitude of the vibration acceleration detected by the first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213. The graph in Figure 10 is a histogram, with the horizontal axis representing the acceleration interval (class) and the vertical axis representing the frequency in each interval (class). In the graph in Figure 10, the width of each acceleration interval is set to 0.1. Since the acceleration of the vibration data from the first vibration sensor 211 and the second vibration sensor 212 shown in Figure 9 is in the range of ±1, the graphs for the first vibration sensor 211 and the second vibration sensor 212 shown in Figure 10 have a sharp shape in the frequency distribution. On the other hand, since the acceleration of the vibration data from the third vibration sensor 213 shown in Figure 9 is in the range of ±2, the graph for the third vibration sensor 213 shown in Figure 10 has a broad shape in the frequency distribution. The processing unit 51 calculates the variance value as a result of statistical analysis of the vibration data of each vibration sensor 21 from the results shown in Figure 10. In the example shown in Figure 10, the frequency distribution shapes are sharp in the graphs of the first vibration sensor 211 and the second vibration sensor 212, so the variance value is small. On the other hand, the frequency distribution shape in the graph of the third vibration sensor 213 is broad, so the variance value is high.

[0046] The processing unit 51 of the abnormality detection device 5 performs preprocessing on the vibration data detected by the vibration sensor 21 using the preprocessing unit 512 (S3). The processing unit 51 removes noise contained in the vibration data detected by the vibration sensor 21 using a low-pass filter. Next, the processing unit 51 performs frequency analysis on the noise-free vibration data using the frequency analysis unit 513 (S4). Specifically, the processing unit 51 generates a power spectrum showing the magnitude of acceleration with respect to the frequency of vibration detected by the vibration sensor 21 using the Fast Fourier Transform (FFT). At this time, the processing unit 51 generates the power spectrum by averaging the magnitude of acceleration related to each vibration data measured a predetermined number of times at a predetermined period.

[0047] Subsequently, the processing unit 51 of the abnormality detection device 5 performs peak processing by the peak processing unit 514 to clarify the power spectrum generated by the frequency analysis unit 513 (S5).

[0048] Figure 11 is a graph showing an example of frequency analysis of vibrations detected by the vibration sensor 21. The upper part of Figure 11 shows the power spectrum generated by the frequency analysis unit 513, and the lower part of Figure 11 shows the power spectrum after peak processing by the peak processing unit 514. In Figure 11, the power spectrum of the third vibration sensor 213 is shown as a representative example. The frequency analysis unit 513 generates a power spectrum showing the magnitude of acceleration against frequency, as shown in the upper part of Figure 11, by performing Fast Fourier Transform (FFT) processing on the vibration data after preprocessing (after noise reduction). The peak processing unit 514 then performs a process (averaging process) in which the data after frequency analysis (power spectrum) is added a predetermined number of times and then divided by a predetermined value. This makes it possible to emphasize the peaks, as shown in the lower part of Figure 11.

[0049] Next, the processing unit 51 of the abnormality detection device 5 performs an abnormality detection (identification of the abnormality level) of the speed reducer 14 using the abnormality detection unit 515, based on the results of the statistical analysis in S2 and the results of the frequency analysis in S3 to S5 (S6). Specifically, the abnormality detection unit 515 determines the abnormality level of the speed reducer 14 based on the frequency distribution of the vibration data from the vibration sensor 21 (see Figure 10).

[0050] Figure 12 is a graph illustrating the abnormality detection process. The horizontal axis of Figure 12 represents the degree of abnormality, and the vertical axis represents the monitoring (judgment) value. The monitoring (judgment) value corresponds to the variance value obtained from the frequency distribution shown in Figure 10. In other words, in Figure 12, a higher variance value indicates a higher degree of abnormality. Specifically, in Figure 10, the graphs for the first vibration sensor 211 and the second vibration sensor 212 have sharp frequency distribution shapes, resulting in small variance values. Therefore, in the graph shown in Figure 12, the variance values ​​for the first vibration sensor 211 and the second vibration sensor 212 are small, so the vibration data for the first vibration sensor 211 and the second vibration sensor 212 are judged to be normal. On the other hand, in Figure 10, the graph for the third vibration sensor 213 has a broad frequency distribution shape, resulting in a high variance value. Therefore, in the graph shown in Figure 12, the variance value for the third vibration sensor 213 is large, so the vibration data for the third vibration sensor 213 is judged to be abnormal (abnormality level 4).

[0051] Furthermore, in the graph shown in Figure 12, the variance of the frequency distribution is used as the monitoring (judgment) value, and abnormality levels 1 to 5 are set. In the graph shown in Figure 12, a higher abnormality level indicates a greater degree of abnormality. Note that abnormality level 1 indicates that the speed reducer 14 is functioning normally.

[0052] For example, by constructing a speed reducer 14 using pre-damaged parts and calculating the variance of the frequency distribution using the vibration data of this speed reducer 14, a graph like the one shown in Figure 12 can be created. Specifically, the speed reducer 14 is constructed using parts with a degree of "small," a degree of "medium," and a degree of "large." Then, vibration data is acquired for each of these constructed speed reducers 14, and the variance of the frequency distribution of the acceleration of each speed reducer 14 is calculated.

[0053] In other words, the degree of abnormality on the horizontal axis shown in Figure 12 corresponds to the degree of damage (small, medium, large) to the pre-damaged part, with a higher degree of abnormality corresponding to a larger degree of damage. Therefore, in the graph shown in Figure 12, the higher the degree of abnormality, the higher the variance value (monitoring (judgment) value) of the acceleration frequency distribution. The abnormality judgment unit 515 can accurately determine whether or not there is an abnormality in the reducer 14 by creating a graph (judgment criteria) in advance as shown in Figure 12.

[0054] The processing unit 51 identifies the component in the gearbox 14 where the malfunction occurred based on the power spectrum obtained from the frequency analysis (S3-S5) (S7).

[0055] Figure 13 is a graph illustrating an example of a power spectrum obtained by frequency analysis, showing the power spectrum in the normal case (upper part of Figure 13) and the power spectrum in the abnormal case (lower part of Figure 13). As shown in Figure 13, the power values ​​in the abnormal case (lower part of Figure 13) are higher than those in the normal case (upper part of Figure 13). In particular, the power values ​​are higher at specific frequencies in the abnormal case (lower part of Figure 13). In the example shown in the lower part of Figure 13, the power values ​​are especially high around 30 Hz. The abnormality detection unit 515 can identify the component in the reducer 14 where the abnormality occurred by using the specific frequency with high power values.

[0056] Figure 14 is an explanatory diagram showing an example of an abnormal frequency DB52a. The memory unit 52 of the abnormality detection device 5 stores an abnormal frequency DB52a that associates the component where the abnormality occurred with the frequency at which the power value becomes high in the case of that abnormality. The abnormal frequency DB52a has, for example, an "abnormal component" column and a "frequency" column. The "abnormal component" column stores the name of the component in the reducer 14 that is the target of abnormality detection (any of the thrust bearing, gear, or various radial bearings). The "frequency" column stores the frequency at which the power value becomes high in the power spectrum when an abnormality occurs in the component. The processing unit 51 of the abnormality detection device 5 can identify the component where the abnormality occurred by referring to the frequencies stored in the frequency table and matching it with a specific frequency at which the power value is high based on the power spectrum generated by the frequency analysis unit 513.

[0057] The processing unit 51 of the abnormality detection device 5 transmits information indicating that an abnormality has occurred in the reduction gear 14 and the faulty part to the user's terminal device 6a and the service provider's terminal device 6b (S8), and then terminates the process. The abnormality detection process may also be performed by, for example, the control unit 31 of the data acquisition device 3.

[0058] <Lifetime prediction processing> Figure 15 is a conceptual diagram showing an example of the record layout of the collected data DB52b. The collected data DB52b includes a hard disk and a DBMS (Database Management System) and stores various physical quantity data collected from the molding machine 1. For example, the collected data DB52b has columns for "No." (record number), "Machine ID", "Operation Date and Time", "Operation Data", "First Vibration Data", "Second Vibration Data", "Third Vibration Data", and "Shaft Torque Data".

[0059] The "Equipment ID" column stores the equipment identifier of molding machine 1. The "Operating Date and Time" column stores information such as the year, month, and date and time when various data stored as records were obtained. The "Operating Data" column stores time-series physical quantities indicating the operating state of molding machine 1, such as motor current, screw 11 rotation speed, screw 11 tip pressure, die head pressure, feeder supply amount (amount of resin raw material supplied), extrusion amount, cylinder temperature, resin pressure, etc. The "First Vibration Data" column stores vibration data, which is time-series physical quantity data, detected by the first vibration sensor 211. The "Second Vibration Data" column stores vibration data, which is time-series physical quantity data, detected by the second vibration sensor 212. The "Third Vibration Data" column stores vibration data, which is time-series physical quantity data, detected by the third vibration sensor 213. The "Shaft Torque Data" column stores screw 11 torque data, which is time-series physical quantity data. In this embodiment, the values ​​stored in the "Operation Data" column, the "First Vibration Data" column, the "Second Vibration Data" column, the "Third Vibration Data" column, and the "Shaft Torque Data" column are the average values ​​of physical quantity data detected or measured a predetermined number of times (e.g., 16 times) at a predetermined period (e.g., 520 μs). A record in the collected data DB52b may be created each time physical quantity data is detected or measured.

[0060] Figure 16 is a conceptual diagram showing the life prediction model M. The life prediction model M comprises an RMS calculation unit 91 and a life estimation function 92. The RMS calculation unit 91 is an arithmetic processing unit that calculates the root mean square of physical quantity data. The life estimation function 92 is a function that estimates the remaining life of the components of the molding machine 1 based on the RMS of the physical quantity data. The life estimation function 92 is a regression equation that expresses the remaining life, which is the dependent variable, in terms of one or more independent variables.

[0061] Figure 17 is a flowchart showing the learning process for the lifetime prediction model M. The processing unit 51 calculates the RMS of a physical quantity based on the physical quantity data stored in the collected data DB 52b (S11), and labels the calculated RMS with the remaining lifetime of the component at the time the physical quantity data was measured (S12). Next, the processing unit 51 optimizes the coefficients of the lifetime estimation function 92 using the RMS with the remaining lifetime labeled as learning data (S13). The processing unit 51 can optimize the coefficients of the lifetime estimation function 92, for example, by Lasso regression. The storage unit 52 stores a pre-adjusted lifetime estimation function 92, and can generate the lifetime prediction model M by optimizing the coefficients of the lifetime estimation function 92.

[0062] Figure 18A is a graph showing the learning results of the life prediction model M, and Figure 18B is a graph showing the inference results using the life prediction model M. In Figures 18A and 18B, the horizontal axis represents the number of days elapsed, and the vertical axis represents the remaining life. The component to be predicted is, for example, the gear or bearing of the reduction gear 14, and the physical quantity is the vibration of the reduction gear 14. Machine learning of the life estimation function 92 was performed using the learning data obtained during the period shown in Figure 18A (elapsed days 1 to 16 months). The solid line shows the labeled remaining life. The dotted plot shows the remaining life predicted using the life estimation function 92.

[0063] In Figure 18B, the upper right dot plot shows the predicted remaining life based on the RMS of physical quantity data obtained from molding machine 1 equipped with a new component (long remaining life). In Figure 18B, the lower left dot plot shows the predicted remaining life based on the RMS of physical quantity data obtained from molding machine 1 equipped with a used component (short remaining life). As shown in Figures 18A and 18B, it can be seen that the remaining life of a component can be predicted using the trained life estimation function 92.

[0064] <Predicted remaining lifespan> Figure 19 is a flowchart showing the processing procedure for life prediction. The processing unit 51 of the abnormality detection device 5 acquires physical quantity data from the data acquisition device 3 (S21). Next, the processing unit 51 calculates the RMS of the physical quantity based on the acquired physical quantity data (S22). Then, the processing unit 51 predicts the remaining life of the component by substituting the calculated RMS of the physical quantity data into the life estimation function 92 (S23), and terminates the process. The processing unit 51 may also predict the remaining life of the component using a learning model that outputs the remaining life based on an image showing changes in physical quantity data, or a learning model that outputs the remaining life based on the relationship between physical quantity data and other state quantities other than physical quantities. Alternatively, the processing unit 51 may output the remaining life from multiple learning models and calculate the minimum, average, median, or weighted average of the remaining lifespans output by the multiple learning models. The life prediction process may also be performed by the control unit 31 of the data acquisition device 3, for example.

[0065] <Display Processing> Figure 20 is a flowchart showing the processing steps for the display process. The processing unit 51 of the abnormality detection device 5 calculates the characteristic frequency intensity representing the characteristics of each component constituting the molding machine 1 at each point in time in the time series, based on the acquired time-series physical quantity data (S31). Then, the processing unit 51 creates a time-series graph of the calculated characteristic frequency intensity (S32). The time-series graph has time on the horizontal axis and characteristic frequency intensity on the vertical axis, and represents the time-series characteristic frequency intensity calculated above.

[0066] Next, the processing unit 51 performs the process of displaying the remaining lifespan and time-series graph of each component constituting the molding machine 1 (S33). The processing unit 51 displays the remaining lifespan and time-series graph of each component on a portal site where, for example, the status of each molding machine 1 owned by the user can be viewed in a browser.

[0067] Furthermore, the processing unit 51 transmits to terminal devices 6a and 6b (S34) the data of the most recent three vibration events (1st vibration data, 2nd vibration data, and 3rd vibration data, hereinafter collectively referred to as vibration data), the calculated remaining lifespan of each component constituting the molding machine 1, the abnormality level of each component identified in S6 shown in Figure 8, and the time-series graph data. For example, the processing unit 51 transmits the remaining lifespan and time-series graph data to the terminal device 6a of the user or operator of the molding machine 1, and to the terminal device 6b of the service provider related to the user's molding machine 1. The processing unit 51 of the abnormality determination device 5 may also transmit data to terminal device 6a or 6b when it receives a request for various data from terminal device 6a or 6b.

[0068] The abnormality detection device 5 according to this embodiment provides a portal site where users can view time-series graphs of vibration data, remaining lifespan, abnormality level, and physical quantities related to the state of each component of the molding machine 1 for multiple molding machines 1 and components. The abnormality detection device 5 according to this embodiment manages the state of molding machines 1 owned by multiple users, and each user can view and confirm the state of their own molding machine 1 by accessing the portal site using a browser. In addition, service providers related to the molding machines 1 and components can also view and confirm the state of their assigned users or molding machines 1 by accessing the portal site using a browser.

[0069] Figure 21 is a schematic diagram showing an example of the overall display screen 101 displayed on the portal site. The overall display screen 101 displays a vibration data display unit 111, an operating speed display unit 112, an abnormal level display unit 113, a remaining life display unit 114, and screen transition buttons 115. Note that the display group consisting of the vibration data display unit 111, the operating speed display unit 112, the abnormal level display unit 113, the remaining life display unit 114, and the screen transition buttons 115 is displayed for each of the multiple gearboxes 14. The topmost display group will be explained below as an example.

[0070] The vibration data display unit 111 of the overall display screen 101 displays the numerical values ​​of the most recently detected first vibration data, second vibration data, and third vibration data, as well as horizontal bar graphs of the most recent three sets of first, second, and third vibration data. The vibration data display unit displays one horizontal bar graph for each set of vibration data, i.e., for each vibration sensor 21. In this embodiment, each vibration data displayed by bar graph and numerical value is data that shows the RMS value (double-mean-square) of acceleration (single: m / s / s) measured a predetermined number of times (e.g., 16 times) at a predetermined period (e.g., 520 μs). In the horizontal bar graph of vibration data, each of the most recent three sets of vibration data is displayed in a single graph for easy viewing. The most recent three sets of vibration data are values ​​detected when operating data including motor current, screw 11 rotation speed, screw 11 tip pressure, die head pressure, feeder supply amount (amount of resin raw material supplied), extrusion amount, cylinder temperature, or resin pressure are under the same conditions. The same conditions may also refer to indicating that the numerical values ​​for each item are within a predetermined range defined for each item. In the horizontal bar graph of vibration data, the three most recent vibration data points are displayed in a list format. For example, as shown in Figure 21, the most recent vibration data is displayed in color, the previous vibration data is displayed as a bar graph surrounded by a thick line, and the vibration data from two events prior is displayed as a bar graph surrounded by a thin line. In addition, the most recent vibration data for each vibration sensor 21 is displayed in a color corresponding to the abnormality level determined based on that vibration data. In the example shown in Figure 21, the abnormality level determined based on the first vibration data is 3 or less, so the first vibration data is displayed in a cool color. The abnormality level determined based on the second vibration data is 4, so the second vibration data is displayed in a light warm color. The abnormality level determined based on the third vibration data is 5, so the third vibration data is displayed in a dark warm color. The vibration data may also be displayed in order of increasing abnormality level, for example, in blue, yellow, or red.Furthermore, the color of the bar graph may change depending on whether the vibration data values ​​are increasing in the order of the two previous measurements, the previous measurement, and the most recent measurement, i.e., whether the vibration data values ​​are increasing over time or not. In addition, the color of the bar graph may change depending on whether the most recent vibration data value exceeds a predetermined value or whether the vibration data value is below a predetermined value.

[0071] The operating speed display section 112 on the overall display screen shows the most recently measured rotational speed (operating speed) of the screw 11.

[0072] The abnormal level display unit 113 displays the highest abnormal level among those determined based on the first vibration data, second vibration data, or third vibration data. In the top display group of Figure 21, the abnormal level determined based on the third vibration data is the highest, and the abnormal level is 5, so the abnormal level display unit displays Level 5. In addition, the abnormal level display unit also displays a gauge corresponding to the abnormal level. The gauge consists of multiple step gauges, and the abnormal level is represented by coloring a number of step gauges corresponding to the multiple abnormal levels. In the example shown in Figure 21, the step gauges are colored in descending order of abnormal level, starting with the smallest step gauge. The first, second, and third smallest step gauges are colored with cool colors, the fourth smallest step gauge is colored with a light warm color, and the fifth smallest step gauge is colored with a dark warm color. In this embodiment, the color assigned to each step gauge is the same regardless of the abnormal level, but the color of all step gauges may change according to the abnormal level. Furthermore, the background of the abnormal level display unit 113 and the display of the abnormal level are colored according to the abnormal level, making it easier to visually recognize the abnormal level. For example, when the abnormal level is 1 to 3, the background is white and the abnormal level is displayed in a cool color. When the abnormal level is 4, the background and the abnormal level are shown in a light warm color. When the abnormal level is 5, the background and the abnormal level are shown in a dark warm color. In other words, the gauge is colored the same as the color of the bar graph corresponding to the vibration data with the highest abnormal level. Note that the colors of each display are not limited to those described above. Also, in the above explanation, cool colors may correspond to blue, light warm colors to yellow, and dark warm colors to red.

[0073] The remaining life display unit 114 displays the predicted remaining life of the gearbox 14 based on the most recent physical quantity data. The background color of the remaining life display unit 114 changes according to the number of days of remaining life. For example, if the remaining life is 10 days or less, the background is shown in a dark warm color; if the remaining life is 11 days or more but 100 days or less, the background is shown in a light warm color; and if the remaining life is 101 days or more, or if the remaining life cannot be calculated, the background is shown in white. Note that the background color of the remaining life display unit 114 and the remaining life intervals at which the background color changes are not limited to these examples.

[0074] When the screen transition button 115 is pressed, the terminal device 6a or 6b displays the remaining lifespan estimation result display screen 102 (see Figure 22).

[0075] Figure 22 is a schematic diagram showing an example of the remaining lifespan estimation result display screen 102 displayed on the portal site. The remaining lifespan estimation result display screen 102 has an equipment name display unit 121 that displays the names of each of the multiple molding machines 1 owned or used by the user, and a component name display unit 122 that displays the names of the components that make up each molding machine 1. The remaining lifespan estimation result display screen 102 also has a component status display unit 123 that displays the status of each component, in particular the status based on the remaining lifespan. The component status display unit 123 displays the abnormality level of each component by displaying icons corresponding to each abnormality level. For example, if the abnormality level is 1 (normal), a "good" icon is displayed; if the abnormality level is 2, a "caution" icon is displayed; if the abnormality level is 3, a "check required" icon is displayed; if the abnormality level is 4, an "abnormal" icon is displayed; and if the abnormality level is 5, a "danger" icon is displayed. When an icon in the component status display unit 123 is operated, the processing unit 51 displays a time-series graph of physical quantities related to the status of the component. Furthermore, the remaining life estimate result display screen 102 has a remaining life display unit 124 that displays the remaining life of the abnormal component. The processing unit 51 displays the remaining life of the component calculated by the above processing.

[0076] Furthermore, the processing unit 51 of the abnormality detection device 5 may be configured to estimate the quality of the molded product produced by the molding machine 1. Abnormalities in the quality of the molded product may occur due to changes in the state of the components constituting the molding machine 1. For example, wear of the screw piece may cause the molded product to become unmelted. The molded product quality is assumed to be a continuous quantity indicating the degree of quality. When training the life prediction model M, if the molded product quality is trained instead of the remaining life, the molded product quality can be estimated in the same manner as in this embodiment.

[0077] (Embodiment 2) Figures 23 and 24 are perspective views of the gearbox 14 illustrating the mounting position of the vibration sensor 21 according to Embodiment 2. The gearbox 14 according to Embodiment 2 is provided with 12 mounting seats B for attaching the vibration sensor 21. Note that the number of mounting seats B provided on the gearbox 14 is not limited to 12.

[0078] Mounting seats B are provided on one of the faces of the reduction gear 14, near some of the corners formed by three of the faces of the reduction gear 14, or near the bearing 82. Mounting seats B also have boss holes drilled perpendicular to the surface on which they are provided (see Figure 25). The direction in which the boss holes of mounting seats B are drilled coincides with the direction of vibration detected by the vibration sensor 21 attached to the mounting seats B. Hereafter, multiple mounting seats B will be distinguished by the direction in which the boss holes are drilled (H, A, or V) and by their number.

[0079] Near the corner C1 formed by the input surface f1, first side surface f2, and top surface f3 of the reduction gear 14, a mounting seat BA1 is provided on the input surface f1, a mounting seat BH1 is provided on the first side surface f2, and a mounting seat BV1 is provided on the top surface f3. Near the corner C2 formed by the input surface f1, the upper surface f3, and the second side surface f4 of the reduction gear 14, a mounting seat BA2 is provided on the input surface f1, a mounting seat BV2 is provided on the upper surface f3, and a mounting seat BH2 is provided on the second side surface f4. Near the corner C3 formed by the first side surface f2, the top surface f3, and the output surface f5 of the reduction gear 14, a mounting seat BH3 is provided on the first side surface f2, a mounting seat BV3 is provided on the top surface f3, and a mounting seat BA3 is provided on the output surface f5. Near the corner C5 formed by the input surface f1, first side surface f2, and bottom surface f6 of the reduction gear 14, a mounting seat BA5 is attached on the input surface f1 and a mounting seat BH5 is attached on the first side surface f2. Furthermore, on the input surface f1 of the reduction gear 14, a mounting seat BA4 is provided near the lower part of the bearing 82.

[0080] In this embodiment, the first vibration sensor 211 is mounted on mounting seat BH1, the second vibration sensor 212 is mounted on mounting seat BA2, and the third vibration sensor 213 is mounted on mounting seat BV3. While the number of vibration sensors 21 mounted on the reduction gear 14 is not limited to three, by limiting the number of vibration sensors 21 mounted on the reduction gear 14 to only the first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213, and not mounting vibration sensors 21 on other mounting seats B, it is possible to detect vibrations with sufficient accuracy while suppressing the cost of providing the vibration sensors 21. Furthermore, depending on the state or usage of the reduction gear 14, it is possible to change which of the multiple mounting seats B to which the vibration sensors 21 are mounted.

[0081] Figure 25 is an explanatory diagram illustrating the details of the mounting seat B. The mounting seat B is provided on one of the surfaces f of the reduction gear 14. The mounting seat B consists of a boss hole B1 and a smooth portion B2 provided around the boss hole B1 on the surface of surface f. The smooth portion B2 is formed, for example, by polishing surface f with sandpaper. A cylindrical boss hole B1, for example, with a depth of 9 mm, is formed approximately in the center of the smooth portion B2. A threaded portion B11 for fixing the vibration sensor 21 is formed from the surface of the boss hole B1 to a depth of, for example, 6 mm. Note that the depth of the boss hole B1 and the depth of the threaded portion B11 are not limited to those described above.

[0082] In this embodiment, the mounting seat B is composed of a boss hole B1 and a smooth portion B2, but is not limited to this. For example, if the mounting seat B is composed only of a smooth portion B2 and a vibration sensor 21 is attached to the mounting seat B, the boss hole B1 may be formed by drilling. It is also possible to form the mounting seat B by machining at a part or all of the surface of surface f. In other words, the vibration sensor 21 may be attached at any location on surface f.

[0083] Figure 26 is an explanatory diagram illustrating the correlation of the time-dependent changes in vibration data related to each vibration sensor 21 when a vibration sensor 21 is attached to all mounting seats B. In Figure 26, each vibration sensor 21 is distinguished by a code that distinguishes the mounting seat B to which the vibration sensor 21 is attached. For example, the vibration sensor 21 attached to mounting seat BA1 is vibration sensor A1. The table shown in Figure 26 shows the correlation coefficients of the time-dependent changes in vibration data (frequency intensity) detected by each vibration sensor 21 in matrix format. In Figure 26, cells with higher correlation coefficients are indicated by lighter colors.

[0084] As shown in Figure 26, when multiple vibration sensors 21 are divided into three groups: Group G1 including vibration sensors A3, H3, and V3; Group G2 including vibration sensors H2, A4, H5, A2, and A5; and Group G3 including vibration sensors A1, V1, H1, and V2, the correlation coefficient between vibration sensors 21 in the same group is higher than the correlation coefficient between vibration sensors 21 in different groups. Therefore, by selecting and using one vibration sensor 21 from each group, it is possible to reduce the number of vibration sensors 21 attached to the reducer 14 while maintaining the accuracy of abnormality detection of the reducer 14. Furthermore, by making the vibration detection directions of the three vibration sensors 21 attached to the reducer 14 all different, making all the surfaces f on which the vibration sensors 21 are attached different, or making all the corners on which the vibration sensors 21 are attached different, it is possible to obtain diverse vibration data while reducing the number of vibration sensors 21, and thus maintain the accuracy of abnormality detection of the reducer 14.

[0085] In this embodiment, by selecting vibration sensor V3 from group G1, vibration sensor A2 from group G2, and sensor H1 from group G3 and attaching them to the reduction gear 14, all of the above conditions are met, and the processing unit 51 of the abnormality detection device 5 can perform abnormality detection of the reduction gear 14 with high accuracy. Note that the combination of vibration sensors 21 attached to the reduction gear 14 is not limited to this. The combination of vibration sensors 21 attached to the reduction gear 14 may be, for example, vibration sensors A3, H2 and V1, or vibration sensors V3, H2 and A1, etc.

[0086] (Embodiment 3) The housing (case) of the speed reducer 14 may be in the shape of two rectangular (cuboidal) boxes combined (Enter key type). Below, the speed reducer 14 according to Embodiment 3 will be described with reference to the drawings.

[0087] Figures 27 and 28 are schematic perspective views of the gearbox 14 according to Embodiment 3. Compared to the gearbox 14 according to Embodiment 1, the gearbox 14 according to Embodiment 3 has a shape in which a portion including the edge extending in the V direction, including the corner C4, is cut out toward the center. That is, the gearbox 14 according to Embodiment 3 has a shape in which two substantially rectangular parallelepiped boxes with different lengths in the A direction are combined. In the example shown in Figure 27, when viewed from the direction directly facing the input surface f1, the rectangular parallelepiped box with the longer length in the A direction is on the left side (bearing 82 side), and the rectangular parallelepiped box with the shorter length in the A direction is on the right side (input shaft 81 side). Also, the upper surface f3 of the gearbox 14 has an L shape (enter key shape) when viewed from above.

[0088] The housing of the gearbox 14 according to Embodiment 3 includes a third side surface f7 and a second output surface f8. The third side surface f7 is arranged to be substantially parallel to the first side surface f2 and the second side surface f4, and is located closer to the first side surface f2 than to the second side surface f4. The second output surface f8 is arranged to be substantially parallel to the input surface f1 and the output surface f5, and is located closer to the input surface f1 than to the output surface f5. In Embodiment 3, the area of ​​the second side surface f4 is smaller than the area of ​​the first side surface f2, and the sum of the areas of the second side surface f4 and the third side surface f7 is substantially the same as the area of ​​the first side surface f2. In Embodiment 3, the area of ​​the output surface f5 is smaller than the area of ​​the input surface f1, and the sum of the areas of the output surface f5 and the second output surface f8 is substantially the same as the area of ​​the input surface f1.

[0089] In the examples shown in Figures 27 and 28, one edge of the output surface f5 extending in the V direction is connected to the first side surface f2. The other edge of the output surface f5 extending in the V direction is connected to one edge of the third side surface f7 extending in the V direction. The other edge of the third side surface f7 extending in the V direction is connected to one edge of the second side surface f4 extending in the V direction.

[0090] As shown in Figures 27 and 28, three vibration sensors 21 are mounted on the housing of the reduction gear 14 according to Embodiment 3. The first vibration sensor 211 is mounted on the first side surface f2 near corner C1, which is formed by the input surface f1, the first side surface f2, and the top surface f3. The second vibration sensor 212 is mounted on the input surface f1 near corner C2, which is formed by the input surface f1, the top surface f3, and the second side surface f4. The third vibration sensor 213 is mounted near corner C3, which is formed by the first side surface f2, the top surface f3, and the output surface f5. The vibration sensor 21 may also be mounted near corner C6, which is formed by the top surface f3, the third side surface f7, and the second output surface f8.

[0091] The positions in which the three vibration sensors 21 are mounted are not limited to those described above. The first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213 may be mounted on any of the input surface f1, the output surface f5, or the second output surface f8, any of the first side surface f2, the second side surface f4, or the third side surface f7, and the top surface f3, each on a different surface. Alternatively, the first vibration sensor 211, the second vibration sensor 212, and the third vibration sensor 213 may be mounted near different corners among the five corners formed by the top surface f3. In this embodiment, the second output surface f8 is located closer to the input surface f1 than the output surface f5, but may be located further away from the input surface f1 than the output surface f5.

[0092] The embodiments disclosed herein should be considered in all respects as illustrative and not restrictive. The technical features described in each embodiment can be combined with each other, and the scope of the present invention is intended to include all modifications within the claims and scope equivalent to the claims. Furthermore, the independent and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. In addition, the claims use a multi-claim format in which claims refer to two or more other claims (multi-claim format), but are not limited to this. They may also be described using a multi-claim format in which at least one multi-claim refers to another multi-claim (multi-multi-claim format). [Explanation of Symbols]

[0093] 1: Molding machine 2: Detector 3: Data acquisition device 4: Router 5: Abnormality determination device 6a: Terminal device 6b: Terminal device 10: Cylinder 10a: Hoppers 11: Screw 12: Dice 13: Motor 14:Reducer 15: Control device 21: First detector (vibration sensor) 22: Second detector 23: Third detector 24: Detector No. 4 50: Recording media 51: Processing Unit 52: Storage section 53: Communications Department 81: Input axis 82: Bearing 85: Output shaft 86: Output shaft 101: Summary display screen 102:Remaining life estimation result display screen 211: First vibration sensor 212: Second vibration sensor 213: Third vibration sensor 511:Statistical Analysis Department 512: Pre-processing unit 513: Frequency Analysis Unit 514: Peak Processing Unit 515: Abnormality determination section DB52a: Abnormal frequency DB52b: Collected data M: Life prediction model P: Computer program B: Mounting base B1: Boss hole B2: Smooth part f :face f1: Input surface f2 :1st side f3:Top surface f4: 2nd side f5: Output surface f6 :Bottom surface

Claims

1. An anomaly detection system for detecting abnormalities in a gearbox, A first vibration sensor, a second vibration sensor, and a third vibration sensor are attached to the surface of the housing of the reduction gear and are used to detect vibrations on the surface of the housing of the reduction gear. An abnormality determination unit that determines whether or not there is an abnormality in the reduction gear based on the vibration detected by each of the first vibration sensor, the second vibration sensor, and the third vibration sensor. An anomaly detection system equipped with the following features.

2. The first vibration sensor detects acceleration in the axial direction of the input shaft of the reduction gear and in a direction perpendicular to the vertical direction of the reduction gear. The second vibration sensor detects the axial acceleration of the input shaft, The third vibration sensor detects the vertical acceleration of the reduction gear. The anomaly detection system according to claim 1.

3. The housing of the aforementioned reduction gear is roughly in the shape of a rectangular box. The aforementioned reduction gear is The input surface on the input axis side, A first side surface connected to either the left or right side of the input surface, The upper surface connected to the upper edge of the input surface, The second side facing the first side, The output surface on the output shaft side, which is opposite the input surface, The lower surface connected to the lower edge of the input surface and Equipped with, The first vibration sensor, the second vibration sensor, and the third vibration sensor are, It is mounted on either the input surface or the output surface, either the first side surface or the second side surface, and on each of the upper surfaces, Of the four corners formed by the aforementioned upper surface, each is attached near a different corner. Each device detects acceleration in a direction perpendicular to the surface to which it is attached. An anomaly detection system according to claim 1 or 2.

4. The housing of the aforementioned reduction gear is roughly in the shape of a rectangular box. The first vibration sensor, the second vibration sensor, and the third vibration sensor are mounted near any corner of the housing, on one of the three surfaces constituting the corner. The first vibration sensor, the second vibration sensor, and the third vibration sensor each detect acceleration in a direction perpendicular to the surface on which they are mounted. The first vibration sensor, the second vibration sensor, and the third vibration sensor are used to detect the acceleration in the direction perpendicular to the axial and vertical directions of the input shaft of the reduction gear, the axial direction of the input shaft, and the vertical direction of the reduction gear. An anomaly detection system according to claim 1 or 2.

5. The aforementioned reduction gear is The input surface on the input axis side, A first side surface connected to either the left or right side of the input surface, The upper surface connected to the upper edge of the input surface, The second side facing the first side, The output surface on the output shaft side, which is opposite the input surface, The lower surface connected to the lower edge of the input surface and Equipped with, The first vibration sensor is mounted near the corner formed by the input surface, the first side surface, and the top surface. The second vibration sensor is mounted near the corner formed by the input surface, the top surface, and the second side surface. The third vibration sensor is mounted near the corner formed by the first side surface, the top surface, and the output surface, or near the corner formed by the top surface, the second side surface, and the output surface that is closer to the output axis. The anomaly detection system according to claim 4.

6. The first vibration sensor is mounted on the first side surface, The second vibration sensor is mounted on the input surface, The third vibration sensor is mounted on the upper surface. The anomaly detection system according to claim 5.

7. The reduction gear is provided on the input surface and includes a bearing that supports a shaft that receives power input to the input shaft of the reduction gear, The bearing is located closer to the first vibration sensor than to the second vibration sensor. The anomaly detection system according to claim 6.

8. A boss hole is provided on the first side surface, which is drilled in a direction perpendicular to the axial direction and vertical direction of the input shaft of the reduction gear for mounting the first vibration sensor. The input surface is provided with a boss hole drilled in the axial direction of the input shaft for attaching the second vibration sensor. A boss hole is provided on the upper surface, which is drilled in the vertical direction of the reduction gear for mounting the third vibration sensor. The anomaly detection system according to claim 6.

9. The housing of the aforementioned reduction gear has a shape in which two roughly rectangular boxes are combined. The aforementioned reduction gear is The input surface on the input axis side, A first side surface connected to either the left or right side of the input surface, The upper surface connected to the upper edge of the input surface, A second side facing the first side and connected to the input surface, The output surface on the output shaft side, which is opposite the input surface and connected to the first side surface, The lower surface connected to the lower edge of the input surface and A third side facing the first side and connected to the output surface, A second output surface facing the input surface and connected to the third and second sides, Equipped with, The first vibration sensor, the second vibration sensor, and the third vibration sensor are, It is mounted on any of the input surface, the output surface, or the second output surface, any of the first side surface, the second side surface, or the third side surface, and each of the top surface, Of the five corners formed by the aforementioned upper surface, each is attached near a different corner. Each device detects acceleration in a direction perpendicular to the surface to which it is attached. An anomaly detection system according to claim 1 or 2.

10. The first vibration sensor is mounted near the corner formed by the input surface, the first side surface, and the top surface. The second vibration sensor is mounted near the corner formed by the input surface, the top surface, and the second side surface. The third vibration sensor is mounted near the corner formed by the first side surface, the top surface, and the output surface, or near the corner formed by the top surface, the second side surface, and the output surface that is closer to the output axis. The anomaly detection system according to claim 9.

11. The aforementioned reduction gear is The input surface on the input axis side, A first side surface connected to either the left or right side of the input surface, The upper surface connected to the upper edge of the input surface, The second side facing the first side, The output surface facing the input surface, The lower surface connected to the lower edge of the input surface, A bearing provided on the input surface supports a shaft that receives power input to the input shaft of the reduction gear, Equipped with, The vicinity of the corner formed by the input surface, the first side surface, and the top surface, respectively, The vicinity of the corner formed by the input surface, the upper surface, and the second side surface, respectively, The vicinity of the corner formed by the first side surface, the top surface, and the output surface, in each of the first side surface, the top surface, and the output surface, The vicinity of the corner formed by the input surface, the first side surface, and the lower surface on each of the input surface and the first side surface, Furthermore, the vicinity of the bearing on the input surface. Multiple mounting seats are provided for attaching the first vibration sensor, the second vibration sensor, or the third vibration sensor. The first vibration sensor, the second vibration sensor, and the third vibration sensor are each attached to any three of the aforementioned mounting seats. The anomaly detection system according to claim 4.

12. An anomaly detection method for detecting an abnormality in a gearbox, A step of detecting vibrations on the surface of the housing of the reduction gear using a first vibration sensor, a second vibration sensor, and a third vibration sensor attached to multiple locations on the surface of the housing of the reduction gear, A step of determining an abnormality in the gearbox using vibrations detected by the first vibration sensor, the second vibration sensor, and the third vibration sensor, respectively. Equipped with Anomaly detection method.

Images