Facility Monitoring System

Abstract

To provide a facility monitoring system capable of efficiently monitoring the presence or absence of abnormalities in processed workpieces.SOLUTION: A facility monitoring system 1 includes: processing facilities A1-A4 that manufacture a plurality of processed workpieces by sequentially executing processing processes; a server device 2 that configures a network capable of communicating with the processing facilities A1-A4; and a client terminal 3 that configures a network capable of communicating with the processing facilities A1-A4 and the server device 2. The server device 2 uses facility-related data at the time of executing the processing processes this time and a trained model and generates an abnormality estimation result representing the presence or absence of an abnormality before the processing processes next time start. The client terminal 3 includes a display unit 3a that can display the abnormality estimation result for the plurality of processed workpieces including the abnormality estimation result generated by the processing processes this time, before the next processing processes start.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a facility monitoring system. [Background technology]

[0002] In factory facilities where workpieces are manufactured, inspections using inspection devices can be performed to detect the presence or absence of abnormalities in the manufactured workpieces. However, inspecting all manufactured workpieces using inspection devices takes a significant amount of time.

[0003] In addition, various types of processing equipment in factory facilities are monitored for mechanical abnormalities by analyzing sensing data obtained from the processing equipment and by having workers at each processing equipment and factory equipment managers patrol the facilities.

[0004] In recent years, attempts have been made to apply machine learning to the analysis of sensing data obtained from processing equipment. For example, Patent Document 1 describes estimating the molding quality of injection-molded products using a trained model. Furthermore, Patent Document 2 describes determining long-term changes in injection molding machines by performing machine learning based on data obtained from multiple injection molding machines via a network. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2020-142460 [Patent Document 2] Japanese Patent Publication No. 2020-52821 Summary of the Invention [Problem to be solved by the invention]

[0006] However, some abnormalities that occur in processing equipment are difficult to detect even if monitoring is performed using the five human senses or by direct sensing of the processing equipment. For example, it is not easy to directly detect abnormalities such as gate wear or gate clogging in the mold of an injection molding machine using the injection molding machine's sensors. While it may be possible to detect such abnormalities by applying a very expensive mechanism, the high cost makes it difficult to introduce.

[0007] Therefore, up until now, it has been the case that an abnormality was finally discovered when an inspection was carried out using an inspection device in the inspection process after the manufacturing process. In such cases, defective workpieces may continue to be manufactured until the abnormality is discovered, and many defective workpieces may have to be discarded.

[0008] Furthermore, in the production of workpieces using processing equipment in factory facilities, normal workpieces are produced overwhelmingly more often than abnormal workpieces. Therefore, even if workers or managers visit factory facilities to monitor them, they often only confirm that the equipment is in a normal state. Furthermore, there is a limit to the scope of monitoring that workers and managers can perform, making monitoring inefficient.

[0009] Therefore, monitoring by workers or managers only allows them to grasp the state at that moment, and they are unable to grasp changes in the state of the processing equipment or the state of the workpiece from the past to the present.

[0010] The present invention has been made in view of the above-mentioned problems, and aims to provide a facility monitoring system that can efficiently monitor whether or not there is an abnormality in a machined workpiece. [Means for solving the problem]

[0011] One aspect of the present invention is a processing facility that produces a plurality of processed workpieces by sequentially performing processing processes; a server device that configures a network capable of communicating with the processing equipment; a client terminal that configures a network capable of communicating with the processing facility and the server device; Equipped with The server device an equipment-related data acquisition unit that acquires equipment-related data indicating an equipment status or a processing status of the processing equipment; a model storage unit for storing a trained model that represents a relationship between the equipment-related data and the presence or absence of an abnormality, the trained model being generated by machine learning using a training data set including the equipment-related data and data on the presence or absence of an abnormality in the machined workpiece; an abnormality estimation unit that generates an abnormality estimation result indicating the presence or absence of an abnormality by the start of the next processing process using the equipment-related data and the trained model when the current processing process is executed; an abnormality estimation result storage unit that stores the abnormality estimation results generated by the abnormality estimation unit for the plurality of machined workpieces; a transmitting unit that transmits the abnormality estimation results of the plurality of machined workpieces stored in the abnormality estimation result storage unit to the client terminal; Equipped with The client terminal is in an equipment monitoring system that has a display unit that can display the abnormality estimation results for multiple processed workpieces, including the abnormality estimation results generated by the current processing, before the start of the next processing. [Effects of the Invention]

[0012] According to the above-described equipment monitoring system, the server device applies machine learning to generate an abnormality estimation result indicating the presence or absence of an abnormality in the machined workpiece. The processing equipment for which the server device generates the abnormality estimation result forms a network capable of communicating with the server device. In other words, the server device can be any processing equipment for which the server device generates the abnormality estimation result as long as it forms a network with the server device.

[0013] The abnormality estimation result generated by the server device is then transmitted from the server device to the client terminal. The client terminal displays the transmitted abnormality estimation result. Because the client terminal forms a network capable of communicating with the server device, the user of the client terminal can be located anywhere as long as the location is capable of network connection with the server device. In other words, even if the user of the client terminal is not located at the installation location of the processing equipment, the user can understand the abnormality estimation result of the processed workpiece that has been processed by the processing equipment.

[0014] Furthermore, the server device generates the abnormality estimation result before the start of the next processing. Therefore, the user of the client terminal can know whether or not there is an abnormality in the processed workpiece due to the current processing before the start of the next processing. Therefore, even if the user of the client terminal is not located at the installation site of the processing equipment, the user can know whether or not there is an abnormality in the processed workpiece in real time.

[0015] Furthermore, the server device stores the abnormality estimation results of a plurality of machined workpieces, including the abnormality estimation result of the machined workpiece of the current processing.The server device then transmits the abnormality estimation results of the plurality of machined workpieces to the client terminal.The client terminal displays the abnormality estimation results of the plurality of machined workpieces, including the abnormality estimation result of the current machined workpiece, before the start of the next processing.Therefore, the user of the client terminal can grasp the changes in the abnormality estimation results up to the present.As a result, the user of the client terminal can grasp the current situation with high accuracy.

[0016] As described above, according to the above aspect, it is possible to provide a facility monitoring system that can efficiently monitor whether or not there is an abnormality in a machined workpiece. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is an overall configuration diagram of a facility monitoring system according to a first embodiment. [Figure 2]1 is a configuration diagram of factory equipment that constitutes the equipment monitoring system of the first embodiment. [Figure 3] FIG. 1 is a diagram showing a processing facility as an injection molding machine. [Figure 4] FIG. 1 is a diagram showing a processing facility as a machine tool. [Figure 5] FIG. 1 is a diagram showing a functional block configuration of an equipment monitoring system according to a first embodiment. [Figure 6] 4 is a timing chart of processing by each component of the equipment monitoring system. [Figure 7] FIG. 10 is a diagram illustrating the relationship between the anomaly reliability value and classification. [Figure 8] 10 is data showing an abnormality estimation result by an abnormality estimation unit of the server device. [Figure 9] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 10] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 11] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 12] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 13] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 14] FIG. 10 is a configuration diagram of factory equipment that constitutes an equipment monitoring system according to a second embodiment. [Figure 15] FIG. 10 is a diagram showing a functional block configuration of a facility monitoring system according to a second embodiment. [Figure 16] 10 shows estimated quality data generated by a quality estimation unit of the server device. [Figure 17] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 18] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 19] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 20] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. [Figure 21] FIG. 2 is a diagram showing one of the display contents of the display unit of the client terminal. DETAILED DESCRIPTION OF THE INVENTION

[0018] (Embodiment 1) 1. Configuration of Equipment Monitoring System 1 The configuration of the facility monitoring system 1 will be described with reference to Fig. 1. The facility monitoring system 1 includes a plurality of factory facilities A, B, and C, a server device 2, and a client terminal 3.

[0019] Each of the multiple factory facilities A, B, and C includes multiple processing facilities A1-A4, B1-B4, and C1-C4. Each of the processing facilities A1-A4, B1-B4, and C1-C4 manufactures multiple machined workpieces by sequentially performing processing. The multiple factory facilities A, B, and C form a network that allows communication between them. The multiple factory facilities A, B, and C may be installed nearby or far away. The multiple factory facilities A, B, and C may be installed in different countries, for example.

[0020] The server device 2 forms a network capable of communicating with multiple factory facilities A, B, and C. The server device 2 is configured with a processor, a storage device, an interface, etc. The server device 2 may be installed within the area of ​​any of the multiple factory facilities A, B, and C, or may be installed in a location completely different from the multiple factory facilities A, B, and C.

[0021] The client terminal 3 forms a network capable of communicating with multiple pieces of factory equipment A, B, and C and the server device 2. The client terminal 3 may be, for example, a mobile terminal such as a tablet or laptop, or a desktop terminal. The client terminal 3 can be used by pre-authorized users. The client terminal 3 has a display unit 3a that can display the status of workpieces processed by processing equipment A1-A4, B1-B4, and C1-C4 installed in multiple pieces of factory equipment A, B, and C. The user of the client terminal 3 can be located anywhere as long as a network can be formed with the server device 2.

[0022] 2. Configuration of Factory Equipment A The configuration of the factory facility A will be described with reference to Fig. 2. Note that although the configuration of the factory facility A will be described below, the same applies to the other factory facilities B and C.

[0023] As shown in Figure 2, factory equipment A includes a plurality of processing equipment A1, A2, A3, and A4, a plurality of edge computer terminals 11, 12, 13, and 14 that are provided one-to-one with each of the processing equipment A1, A2, A3, and A4, and an inspection device 15.

[0024] The processing equipment A1, A2 includes a processing equipment main body 21, a control device 22, an operation panel 23, and an observation device 24. The processing equipment main body 21 constitutes, for example, an injection molding machine. The processing equipment main body 21 sequentially performs injection molding to manufacture injection-molded products as multiple machined workpieces.

[0025] The control device 22 controls the drive devices that make up the processing equipment main body 21. The operation panel 23 functions as an input device as well as a display device. The observation device 24 acquires equipment-related data that indicates the equipment status or processing status of the processing equipment A1, A2. The equipment-related data acquired by the observation device 24 is time-series data. The observation device 24 is, for example, various sensors such as a temperature sensor, a pressure sensor, an acceleration sensor, and a displacement sensor. In other words, the observation device 24 acquires equipment-related data as various time-series data while injection molding is being performed.

[0026] The processing equipment A3 and A4 each include a processing equipment main body 31, a control device 32, an operation panel 33, and an observation device 34. The processing equipment main body 31 constitutes, for example, a machine tool. In FIG. 2, a grinding machine is illustrated as an example of the processing equipment main body 31. A lathe, a machining center, or the like can be used as the processing equipment main body 31. The processing equipment main body 21 produces a plurality of machined workpieces by sequentially performing cutting or grinding on unmachined workpieces.

[0027] The control device 32 controls the drive devices that make up the processing equipment main body 31. The operation panel 33 functions as an input device as well as a display device. The observation device 34 acquires equipment-related data that indicates the equipment status or processing status of the processing equipment A3, A4. The equipment-related data acquired by the observation device 34 is time-series data. The observation device 34 is, for example, various sensors such as a temperature sensor, a pressure sensor, an acceleration sensor, and a displacement sensor. In other words, the observation device 34 acquires equipment-related data as various time-series data when cutting or grinding is being performed.

[0028] The edge computer terminals 11 and 12 are provided in a one-to-one correspondence with the processing equipment A1 and A2. The edge computer terminals 11 and 12 form a network capable of communicating with the processing equipment A1 and A2 and the server device 2. Therefore, the edge computer terminals 11 and 12 acquire equipment-related data as time-series data acquired by the observation device 24 that constitutes the processing equipment A1 and A2 as injection molding machines.

[0029] Furthermore, the edge computer terminals 11 and 12 extract feature quantities from the acquired equipment-related data as time-series data. Specifically, the edge computer terminals 11 and 12 extract feature quantities representing abnormalities in injection molding from the equipment-related data as time-series data. In this way, the edge computer terminals 11 and 12 generate equipment-related data representing feature quantities. Feature quantities include maximum values, average values, first quartiles, third quartiles, variances, and standard deviations. Therefore, the equipment-related data representing feature quantities has a significantly smaller data size than the equipment-related data as time-series data.

[0030] The edge computer terminals 13, 14 are provided in a one-to-one correspondence with the processing equipment A3, A4. The edge computer terminals 13, 14 form a network capable of communicating with the processing equipment A3, A4 and the server device 2. Therefore, the edge computer terminals 13, 14 acquire equipment-related data as time-series data acquired by the observation device 34 that constitutes the processing equipment A3, A4 as machine tools.

[0031] Furthermore, the edge computer terminals 13 and 14 extract feature quantities from the acquired equipment-related data as time-series data. Specifically, the edge computer terminals 13 and 14 extract feature quantities representing abnormalities in the cutting or grinding process from the equipment-related data as time-series data. In this way, the edge computer terminals 13 and 14 generate equipment-related data representing feature quantities. Feature quantities include maximum values, average values, first quartiles, third quartiles, variance values, and standard deviations. Therefore, the equipment-related data representing feature quantities has a significantly smaller data size than the equipment-related data as time-series data. Furthermore, an abnormality means that the machined workpiece is defective.

[0032] In addition to the above, the processing equipment A1, A2, A3, and A4 can be applied with machines that perform various processing processes, such as forging machines, casting machines, electric discharge machines, press machines, etc. Furthermore, the processing equipment A1, A2, A3, and A4 can also be a transport device that transports processed workpieces.

[0033] The inspection device 15 can inspect machined workpieces manufactured by multiple processing equipment A1, A2, A3, and A4. The inspection device 15 forms a network capable of communicating with the edge computer terminals 11, 12, 13, and 14 and the server device 2. The inspection device 15 can, for example, inspect whether the machined workpiece is abnormal. When inspecting the appearance of the machined workpiece, the inspection device 15 serves as an appearance inspection device. When inspecting the dimensions, surface roughness, etc. of the machined workpiece, the inspection device 15 serves as a various measuring device. In other words, the inspection device 15 can inspect whether the machined workpiece is abnormal, i.e., whether the machined workpiece is a defective product.

[0034] 3. Configuration of processing equipment A1 as an injection molding machine The processing equipment A1 as an injection molding machine will be described with reference to Fig. 3. The processing equipment A1 is a device for molding, for example, a resin molded product by injection molding. The processing equipment A1 includes a processing equipment main body 21, a control device 22, an operation panel 23, and an observation device 24.

[0035] The processing equipment main body 21 includes a bed 41, an injection unit 42, a mold 43, and a mold clamping unit 44. The bed 41 is a member that is installed on an installation surface.

[0036] The injection device 42 is placed on the bed 41. The injection device 42 is a device that melts resin, which is a molding material, and applies pressure to the molten resin to supply it to a molded product cavity Ca of a mold 43. The injection device 42 includes a hopper 42a, a cylinder 42b, a screw 42c, a nozzle 42d, a heater 42e, and a drive device 42f.

[0037] The hopper 42a is an inlet for resin pellets (granular molding material), which are the raw material for the molding material. The cylinder 42b stores the molten resin produced by heating and melting the pellets fed into the hopper 42a. The cylinder 42b is provided so as to be movable in the axial direction of the cylinder 42b relative to the bed 41. The screw 42c is disposed inside the cylinder 42b and is provided so as to be rotatable and movable in the axial direction. The nozzle 42d is an outlet provided at the front end of the cylinder 42b, and discharges the molten resin inside the cylinder 42b as the screw 42c moves forward.

[0038] The heater 42e is provided, for example, on the outer circumferential surface of the cylinder 42b or inside the cylinder 42b. The heater 42e heats the resin inside the cylinder 42b. That is, the heater 42e melts the pellets and maintains the molten resin in a molten state. The driver 42f moves the cylinder 42b in the axial direction (forward and backward) and rotates and moves the screw 42c in the axial direction (forward and backward).

[0039] The mold 43 includes a first mold 43a, which is a fixed side, and a second mold 43b, which is a movable side. The mold 43 forms a molded product cavity Ca between the first mold 43a and the second mold 43b by clamping the first mold 43a and the second mold 43b together. The first mold 43a and the second mold 43b include a resin flow path P between the molded product cavity Ca and a portion that abuts against the nozzle 42d of the injection device 42. The resin flow path P is a flow path (spool, runner, gate) that guides the molten material supplied from the nozzle 42d of the injection device 42 to the molded product cavity Ca.

[0040] The mold clamping device 44 is disposed on the bed 41 opposite the injection device 42. The mold clamping device 44 opens and closes the attached mold 43, and when the mold 43 is clamped, it prevents the mold 43 from opening due to the pressure of the molten material injected into the molded product cavity Ca.

[0041] The mold clamping device 44 includes a fixed platen 44a, a movable platen 44b, a diver 44c, and a drive device 44d. A first mold 43a is fixed to the fixed platen 44a. A second mold 43b is fixed to the movable platen 44b. The movable platen 44b can move toward and away from the fixed platen 44a. The diver 44c supports the movement of the movable platen 44b. The drive device 44d is configured by, for example, a cylinder device, and moves the movable platen 44b.

[0042] The control device 22 controls the drive device 42f of the injection device 42 and the drive device 44d of the mold clamping device 44. The operation panel 23 functions as an input device that is operated by an operator to input various programs and parameters. Furthermore, the operation panel 23 functions as a display device that displays various input information, the equipment status of the processing equipment A1, the processing status of the machined workpiece, etc.

[0043] The observation device 24 is a sensor that observes the processing equipment main body 21 for control by the control device 22. In addition to the control by the control device 22, the observation device 24 also acquires equipment-related data that indicates the equipment state or processing state of the processing equipment main body 21 in order to allow the server device 2 to estimate the presence or absence of an abnormality in the machined workpiece.

[0044] For example, the observation device 24 includes a screw pressure measuring device 51, a nozzle pressure measuring device 52, a flow path pressure measuring device 53, and a measuring device for a mold clamping device 54. The screw pressure measuring device 51 and the nozzle pressure measuring device 52 are provided in the injection device 42.

[0045] The screw pressure measuring device 51 is provided, for example, near the base end of the screw 42c and acquires pressure data that the screw 42c receives from the molten resin in the cylinder 42b. The nozzle pressure measuring device 52 is provided in the nozzle 42d and acquires pressure data that the nozzle 42d receives from the molten resin when the molten resin flows through the nozzle 42d. In addition to the above, the injection unit 42 also includes an observation device 24 that includes sensors that acquire the position of the cylinder 42b, the position of the screw 42c, the moving speed of the screw 42c, the temperature of the heater 42e, the state of the drive unit 42f, etc.

[0046] The flow path pressure measuring device 53 is provided in the mold 43 and acquires pressure data in the resin flow path P. The pressure data in the resin flow path P is pressure data that the inner wall surface of the resin flow path P receives from the molten resin flowing through the resin flow path P. The mold clamping unit measuring device 54 is provided in the mold clamping unit 44 and acquires the mold clamping force, mold temperature, the state of the drive unit 44d, etc.

[0047] Abnormalities in the processed workpiece molded by the processing equipment A1 as an injection molding machine include, for example, underfill, remaining burrs, surface burns, etc. These abnormalities can be found by inspection using, for example, a visual inspection device.

[0048] 4. Configuration of processing equipment A3 as a machine tool The processing equipment A3 as a machine tool will be described with reference to Fig. 4. The processing equipment A3 is a device for manufacturing a machined workpiece by cutting or grinding an unmachined workpiece. Fig. 4 shows a grinding machine as an example of the processing equipment A3. In addition to grinding machines, lathes, machining centers, etc. can also be used as the processing equipment A3.

[0049] 4, the processing equipment A3 can be a cylindrical grinding machine, a cam grinding machine, etc. Also, the processing equipment A3 can be a table traverse type grinding machine, a wheelhead traverse type grinding machine, etc. In this embodiment, the processing equipment A3 is exemplified by a table traverse type grinding machine.

[0050] The processing equipment A3 as a grinding machine includes a processing equipment main body 31, a control device 32 that controls the processing equipment main body 31, an operation panel 33 that constitutes an input device and a display device, and an observation device 34. The processing equipment main body 31 includes a bed 61, a workpiece table 62, a headstock 63, a tailstock 64, a grinding wheel head 65, and a rest device 66.

[0051] In the processing equipment main body 31, a workpiece table 62, a headstock 63, a tailstock 64, and a grinding wheel head 65 are arranged on a bed 61. A table guideway 61a extending in the Z-axis direction is provided on the bed 61. The workpiece table 62 is supported by the table guideway 61a so as to be movable in the Z-axis direction. The workpiece table 62 moves in the Z-axis direction by being driven by a Z-axis motor 61b provided on the bed 61.

[0052] A headstock 63 and a tailstock 64 are arranged on the workpiece table 62 so as to face each other in the Z-axis direction. The headstock 63 and the tailstock 64 rotatably support both ends of the workpiece W. A spindle motor 63a is provided on the headstock 63, and the workpiece W is rotated by driving the spindle motor 63a.

[0053] Additionally, a wheel head guide surface 61c extending in the X-axis direction is provided on the bed 61 at a position spaced apart from the table guide surface 61a in the X-axis direction. A wheel head 65 is supported by the wheel head guide surface 61c so as to be movable in the X-axis direction. The wheel head 65 is moved in the X-axis direction by the drive of an X-axis motor 61d provided on the bed 61.

[0054] The wheel head 65 supports the grinding wheel T so that it can rotate around an axis parallel to the Z axis. The grinding wheel T is driven to rotate by a wheel rotation motor 65a provided on the wheel head 65. The wheel head 65 moves in the X axis direction, causing the grinding wheel T to move closer to or away from the workpiece W.

[0055] The rest device 66 is disposed on the bed 61 so as to sandwich the workpiece W between it and the grinding wheel T. In other words, the rest device 66 is disposed on the opposite side of the grinding wheel head 65 in the X-axis direction with the workpiece table 62 as the reference. The rest device 66 supports, for example, the back side of the workpiece W relative to the position where it is ground by the grinding wheel T, and also supports the lower side in the direction of gravity. In other words, the rest device 66 has the function of suppressing deflection and deformation of the workpiece W during grinding. However, the rest device 66 may also support the back side and the lower surface at the axial center of the workpiece W.

[0056] The observation device 34 is a sensor that observes the processing equipment main body 31 for control by the control device 32. In addition to the control by the control device 32, the observation device 34 also acquires equipment-related data that indicates the equipment state or processing state of the processing equipment main body 31 in order to allow the server device 2 to estimate the presence or absence of an abnormality in the machined workpiece.

[0057] The observation device 34 includes, for example, a sensor 71 that detects vibrations of the headstock 63, a sensor 72 that detects vibrations of the tailstock 64, and a sensor 73 that detects vibrations of the wheel head 65. The observation device 34 may also include a sensor for measuring machining resistance.

[0058] Abnormalities in the machined workpieces manufactured by the processing equipment A3 as a machine tool include the occurrence of a processed layer on the surface, poor surface roughness, poor shape, etc. Some of these abnormalities can be discovered by inspection using a visual inspection device, while others can be discovered by inspection using a shape measurement device.

[0059] 5. Functional block configuration of equipment monitoring system 1 The functional block configuration of the facility monitoring system 1 will be described with reference to Figures 5 to 8. In particular, the functions of the server device 2 that constitutes the facility monitoring system 1 will be described in detail.

[0060] Observation devices 24, 34 in the processing equipment A1 to A4 (shown in FIG. 2) acquire equipment-related data as time-series data indicating the equipment status or processing status of the processing equipment A1 to A4. As shown in FIG. 6, the observation devices 24, 34 perform observations while processing is being performed by the processing equipment A1 to A4. Note that, since the timing of processing differs depending on the processing equipment A1 to A4, the observation devices 24, 34 perform observations in accordance with the timing of processing of the corresponding processing equipment A1 to A4.

[0061] 5, the edge computer terminals 11 to 14 acquire facility-related data as time-series data from the observation devices 24 and 34, and extract feature quantities from the acquired facility-related data. In this way, the edge computer terminals 11 to 14 generate facility-related data representing the feature quantities. The edge computer terminals 11 to 14 then transmit the generated facility-related data representing the feature quantities to the server device 2.

[0062] 6, the edge computer terminals 11-14 execute processing immediately after the current processing by the processing equipment A1-A4 is completed. In other words, the edge computer terminals 11-14 execute processing when the processing equipment A1-A4 has completed the current processing and is preparing for the next processing. The processing by the edge computer terminals 11-14 involves acquiring equipment-related data as time-series data, extracting features, and transmitting the data to the server device 2.

[0063] The server device 2 acquires equipment-related data representing feature quantities from the edge computer terminals 11 to 14 and applies machine learning to estimate the presence or absence of an abnormality in the machined workpiece. Therefore, the server device 2 is configured to be able to execute the learning phase and estimation phase of machine learning.

[0064] As shown in FIG. 5, the server device 2 includes an equipment-related data acquisition unit 81, an abnormality information acquisition unit 82, a training dataset storage unit 83, a model generation unit 84, a model storage unit 85, an abnormality estimation unit 86, an abnormality estimation result storage unit 87, a transmission unit 88, a processing condition adjustment unit 89, and a post-processing unit 90.

[0065] The server device 2 uses a facility-related data acquisition unit 81, an abnormality information acquisition unit 82, a training data set storage unit 83, a model generation unit 84, and a model storage unit 85 in the learning phase.

[0066] The equipment-related data acquisition unit 81 acquires equipment-related data indicating the equipment status or processing status of the processing equipment A1 to A4 from the edge computer terminals 11 to 14 for the learning phase. In other words, the equipment-related data acquired by the equipment-related data acquisition unit 81 is not time-series data, but feature quantities extracted from the time-series data. Therefore, it is possible to prevent the network load from becoming high. The equipment-related data acquired by the equipment-related data acquisition unit 81 may include multiple types of data.

[0067] The anomaly information acquisition unit 82 acquires the results of inspection of the machined workpiece to be used in the learning phase by the inspection device 15. The inspection results include data on the presence or absence of anomalies in the machined workpiece. Therefore, the anomaly information acquisition unit 82 acquires data on the presence or absence of anomalies in the machined workpiece. The anomaly information acquisition unit 82 may be configured to acquire data on the presence or absence of anomalies for multiple types.

[0068] The training dataset storage unit 83 associates the equipment-related data acquired by the equipment-related data acquisition unit 81 with the data on the presence or absence of abnormalities in the machined workpiece acquired by the abnormality information acquisition unit 82 and stores them as a training dataset.

[0069] The model generation unit 84 generates a trained model by performing machine learning using a training data set. The model generation unit 84 performs machine learning using the equipment-related data as an explanatory variable and the data on the presence or absence of an abnormality in the machined workpiece as a target variable. Therefore, the trained model generated by the model generation unit 84 is a model that represents the relationship between the equipment-related data and the presence or absence of an abnormality.

[0070] A trained model is a model that can output the presence or absence of an abnormality when equipment-related data is input. Furthermore, when outputting the presence or absence of an abnormality, the trained model can also output the degree of abnormality (score). In other words, when equipment-related data is input, the trained model can output the degree of abnormality (score), in other words, a numerical value that represents the reliability of the abnormality. Note that various machine learning models, such as neural networks and support vector machines, can be applied to the model.

[0071] The model storage unit 85 stores the trained models generated by the model generation unit 84. The trained models stored in the model storage unit 85 may be models according to the types of processing equipment A1 to A4. For example, the trained model may be a common model for processing equipment A1 and A2 as injection molding machines, and a common model for processing equipment A3 and A4 as machine tools. Even if the processing equipment A1 to A4 are of the same type, it is possible to use trained models according to each of the processing equipment A1 to A4.

[0072] The learning phase is completed when the trained model is stored in the model storage unit 85. However, the trained model can also be updated while the estimation phase is being executed. In this case, the server device 2 executes the learning phase again.

[0073] Then, once the trained model has been stored in the model storage unit 85 of the server device 2, the server device 2 can execute the estimation phase. In the estimation phase, the server device 2 uses the equipment-related data acquisition unit 81, the model storage unit 85, the abnormality estimation unit 86, the abnormality estimation result storage unit 87, the transmission unit 88, the processing condition adjustment unit 89, and the post-processing unit 90.

[0074] For the estimation phase, the equipment-related data acquisition unit 81 acquires equipment-related data indicating the equipment status or processing status of the processing equipment A1 to A4 from the edge computer terminals 11 to 14. That is, the equipment-related data acquisition unit 81 acquires equipment-related data when the processing equipment A1 to A4 performed the current processing as the estimation target. Here, the equipment-related data acquired by the equipment-related data acquisition unit 81 for the estimation phase is the same type of data as the equipment-related data acquired for the learning phase described above.

[0075] The abnormality estimation unit 86 generates an abnormality estimation result indicating whether or not there is an abnormality in the machined workpiece that is the target of the current processing process, using the equipment-related data acquired by the equipment-related data acquisition unit 81 when the current processing process was performed and the trained model stored in the model storage unit 85. As shown in Fig. 6, the abnormality estimation unit 86 generates an abnormality estimation result when the target processing equipment A1 to A4 is preparing for the next processing process. In other words, the abnormality estimation unit 86 generates an abnormality estimation result before the start of the next processing process.

[0076] 7, the abnormality estimation unit 86 may generate a numerical value representing the reliability of the abnormality estimation result as the abnormality estimation result indicating the presence or absence of an abnormality. For example, the abnormality estimation unit 86 may express this as a numerical value on a 10-point scale, with a higher numerical value representing a higher reliability of the abnormality and a lower numerical value representing a lower reliability of the abnormality. In other words, a small numerical value representing the reliability of the abnormality means that the condition is normal, and a large numerical value representing the reliability of the abnormality means that the condition is abnormal.

[0077] 7, the abnormality estimation unit 86 classifies the result into one of three categories, for example, "normal," "abnormal," and "possibly abnormal," based on the numerical value representing the reliability of the abnormality. In this embodiment, when the numerical value representing the reliability of the abnormality is "1" to "5," it is classified as "normal," when it is "6" to "8," it is classified as "possibly abnormal," and when it is "9" or "10," it is classified as "abnormal." However, the number of classifications may be two, four, or more.

[0078] The abnormality estimation result storage unit 87 stores the abnormality estimation result generated by the abnormality estimation unit 86. In particular, the abnormality estimation result storage unit 87 stores the abnormality estimation result each time it is generated by the abnormality estimation unit 86. Therefore, the abnormality estimation result storage unit 87 stores the abnormality estimation result of the machined workpiece generated by the abnormality estimation unit 86 in the current processing, and also stores the abnormality estimation results of multiple machined workpieces generated by the abnormality estimation unit 86 in the past.

[0079] For example, when the abnormality estimation result is classified into one of the three categories described above, the abnormality estimation result storage unit 87 stores past abnormality estimation results including the current Ta, as shown in Fig. 8. In Fig. 8, black circles represent the abnormality estimation results of each past machined workpiece, and star marks represent the abnormality estimation result of the current machined workpiece.

[0080] 5, when a transmission request for displaying an abnormality is received from a client terminal 3, the transmission unit 88 transmits the abnormality estimation results of the multiple machined workpieces stored in the abnormality estimation result storage unit 87 to the target client terminal 3. The client terminal 3 displays the abnormality estimation results of the multiple machined workpieces transmitted from the transmission unit 88 on the display unit 3a.

[0081] 6, the transmission process by the transmission unit 88 to the client terminal 3 is performed before the start of the next processing. Therefore, the display unit 3a of the client terminal 3 will be in a state where the abnormality estimation results for multiple machined workpieces, including the current machined workpiece, are displayed before the start of the next processing.

[0082] The user of the client terminal 3 can grasp the abnormality estimation result of the currently machined workpiece before the start of the next processing. Furthermore, the user of the client terminal 3 can grasp the abnormality estimation results of past machined workpieces in addition to the currently machined workpiece, so that the user can grasp the changes up to the present.

[0083] Furthermore, the transmission unit 88 also transmits the multiple abnormality estimation results transmitted to the client terminal 3 to the control devices 22, 32 of the processing equipment A1 to A4, and causes the abnormality estimation results to be displayed on the display units of the operation panels 23, 33. When the transmission unit 88 transmits the abnormality estimation results to the control devices 22, 32, the abnormality estimation results are transmitted via the edge computer terminals 11 to 14. If an operator is located near the processing equipment A1 to A4, the operator can have the same understanding as the user of the client terminal 3.

[0084] The machining condition adjustment unit 89 adjusts the machining conditions for the next machining process for the processing equipment A1 to A4 based on the abnormality estimation results of the multiple machined workpieces stored in the abnormality estimation result storage unit 87. For example, when there is a high possibility of an abnormality, the machining condition adjustment unit 89 adjusts the machining conditions so that the next and subsequent machined workpieces will not become abnormal. The adjustment of the machining conditions can be performed automatically by setting an adjustment method in advance. When the machining condition adjustment unit 89 adjusts the machining conditions, it outputs the adjusted machining conditions to the control devices 22, 32 of the processing equipment A1 to A4, so that the control devices 22, 32 operate according to the adjusted machining conditions.

[0085] When the machining conditions are adjusted by the machining condition adjustment unit 89, the transmission unit 88 transmits the adjusted machining conditions to the client terminal 3 in addition to the abnormality estimation results for the multiple machined workpieces. In this case, the adjusted machining conditions are displayed on the display unit 3a of the client terminal 3. Furthermore, when the machining conditions are adjusted by the machining condition adjustment unit 89, the transmission unit 88 also transmits the adjusted machining conditions to the control devices 22, 32 of the target processing equipment A1 to A4, in the same way as the client terminal 3. Then, the adjusted machining conditions are displayed on the display units of the target operation panels 23, 33.

[0086] When the abnormality estimation unit 86 classifies the current abnormality estimation result of the machined workpiece as "abnormal," the post-processing unit 90 transmits an instruction to transport the target machined workpiece to a waste processing facility to the control devices 22, 32 of the target processing facilities A1 to A4. When the control devices 22, 32 receive the instruction to transport to the waste processing facility from the post-processing unit 90, they control the transport of the target machined workpiece to the waste processing facility. The waste processing facility is a storage case for disposal, a facility for crushing the machined workpiece that is an injection molded product, or the like.

[0087] Furthermore, when the abnormality estimation unit 86 classifies the abnormality estimation result of the current machined workpiece as "possibly abnormal," the post-processing unit 90 transmits an instruction to the control devices 22, 32 of the target processing equipment A1 to A4 to transport the target machined workpiece to the inspection device 15. When the control devices 22, 32 receive the instruction to transport to the inspection device 15 from the post-processing unit 90, they control the target machined workpiece to be transported to the inspection device 15.

[0088] The inspection device 15 starts inspecting the transported machined workpiece and outputs the inspection result. The inspection result by the inspection device 15 is either "normal" or "abnormal." If the inspection device 15 determines that the workpiece is "abnormal," it transports the workpiece to a waste disposal facility. On the other hand, if the inspection device 15 determines that the workpiece is "normal," it notifies the worker, manager, or client terminal 3.

[0089] 6. Example of display content on client terminal 3 Examples of display contents on the display unit 3a of the client terminal 3 will be described with reference to Figs. 8 to 13. Below, several types of display contents on the display unit 3a of the client terminal 3 are illustrated. The display contents on the display unit 3a of the client terminal 3 can be freely selected by the user of the client terminal 3.

[0090] The abnormality estimation result storage unit 87 stores the abnormality estimation results of a plurality of machined workpieces, including the current machined workpiece, as shown in Fig. 8. Therefore, the display unit 3a of the client terminal 3 clearly displays the current abnormality estimation result in large letters, and also displays a graph showing the progress of the abnormality estimation results of the machined workpieces, as shown in Fig. 9. This display content is displayed until the start of the next processing.

[0091] 1, a plurality of pieces of factory equipment A, B, and C form a network with the server device 2. Therefore, the abnormality estimation result storage unit 87 stores, for each piece of factory equipment A, B, and C, the abnormality estimation results of a plurality of processed workpieces manufactured by one or more pieces of processing equipment A1 to A4 installed in the factory equipment A, B, and C.

[0092] In this case, as shown in Fig. 10, the display unit 3a of the client terminal 3 can display the abnormality estimation results for multiple machined workpieces for each piece of factory equipment A, B, and C. For example, the display unit 3a of the client terminal 3 displays the number of pieces classified as "normal," "possibly abnormal," and "abnormal" for each piece of factory equipment A, B, and C, as well as the percentage of each. This display allows the user of the client terminal 3 to understand the status and performance of each piece of factory equipment A, B, and C. The user of the client terminal 3 can display this display content at any time.

[0093] Furthermore, the abnormality estimation result storage unit 87 stores the abnormality estimation results of the multiple machined workpieces for each of the processing equipment A1 to A4. In this case, as shown in Fig. 11, the display unit 3a of the client terminal 3 can display the abnormality estimation results of the multiple machined workpieces for each of the processing equipment A1 to A4. For example, the display unit 3a of the client terminal 3 displays the number of workpieces classified into "normal," "possibly abnormal," and "abnormal" for each of the processing equipment A1 to A4, as well as the percentage.

[0094] Such a display allows the user of the client terminal 3 to understand the status and performance of each of the processing equipment A1 to A4. The displayed content can be displayed by the user of the client terminal 3 at any timing.

[0095] The abnormality estimation result storage unit 87 can also store the abnormality estimation results of multiple machined workpieces for each material lot R1, R2. In this case, as shown in Fig. 12, the display unit 3a of the client terminal 3 can display the abnormality estimation results of multiple machined workpieces for each material lot R1, R2. For example, the display unit 3a of the client terminal 3 displays the number of workpieces classified as "normal," "possibly abnormal," or "abnormal" for each material lot R1, R2, as well as the percentage of those classified.

[0096] This display allows the user of the client terminal 3 to understand the status and results of each material lot R1, R2. For example, if the abnormality estimation result changes due to a change in material lot R1, R2, the user can understand that there may be a problem with the material lot R1, R2 in question. This display content can be displayed at any time by the user of the client terminal 3.

[0097] In addition, there is a case where the processing equipment A1 to A4 sequentially executes a series of different processing steps to manufacture one processed workpiece. In this case, the abnormality estimation result storage unit 87 can store the abnormality estimation result for each workpiece and for each process.

[0098] 13, the display unit 3a of the client terminal 3 can display the abnormality estimation results for the multiple machined workpieces W1, W2, and W3 for each workpiece and for each process. For example, the display unit 3a of the client terminal 3 displays one of the classified results of "normal," "possibly abnormal," or "abnormal" for each workpiece and for each process.

[0099] 7.Effects According to the equipment monitoring system 1 of this embodiment, the server device 2 applies machine learning to generate an abnormality estimation result indicating the presence or absence of an abnormality in the machined workpiece. The processing equipment A1 to A4, B1 to B4, and C1 to C4 for which the server device 2 generates an abnormality estimation result constitute a network capable of communicating with the server device 2. In other words, the server device 2 can generate an abnormality estimation result for all of the processing equipment A1 to A4, B1 to B4, and C1 to C4 for which the server device 2 generates an abnormality estimation result, as long as the processing equipment A1 to A4, B1 to B4, and C1 to C4 constitute a network with the server device 2.

[0100] The abnormality estimation result generated by the server device 2 is then transmitted from the server device 2 to the client terminal 3. The client terminal 3 displays the transmitted abnormality estimation result. Because the client terminal 3 forms a network capable of communicating with the server device 2, the user of the client terminal 3 can be located anywhere as long as the location is capable of network connection with the server device 2. In other words, even if the user of the client terminal 3 is not located at the installation location of the processing equipment A1 to A4, B1 to B4, C1 to C4, the user can understand the abnormality estimation result of the machined workpiece that has been processed by the processing equipment A1 to A4, B1 to B4, C1 to C4.

[0101] Furthermore, the server device 2 generates the abnormality estimation result by the time the next processing operation starts. Therefore, the user of the client terminal 3 can know whether or not there is an abnormality in the processed workpiece due to the current processing operation by the time the next processing operation starts. Therefore, even if the user of the client terminal 3 is not located at the installation location of the processing equipment A1 to A4, B1 to B4, C1 to C4, he or she can know in real time whether or not there is an abnormality in the processed workpiece.

[0102] Furthermore, the server device 2 stores the abnormality estimation results of a plurality of machined workpieces, including the abnormality estimation result of the machined workpiece of the current machining process. The server device 2 then transmits the abnormality estimation results of the plurality of machined workpieces to the client terminal. The client terminal 3 displays the abnormality estimation results of the plurality of machined workpieces, including the abnormality estimation result of the current machined workpiece, before the start of the next machining process. Therefore, the user of the client terminal 3 can grasp the changes in the abnormality estimation results up to the present. As a result, the user of the client terminal 3 can grasp the current situation with high accuracy.

[0103] (Embodiment 2) Of the symbols used in the second embodiment, the same symbols as those used in the previous embodiments represent the same components as those in the previous embodiments, unless otherwise specified.

[0104] 1. Configuration of Equipment Monitoring System 101 The overall configuration of the facility monitoring system 101 is similar to the overall configuration of the facility monitoring system 1 of the first embodiment shown in Fig. 1. The facility monitoring system 101 will be described with reference to Fig. 14.

[0105] The equipment monitoring system 101 includes multiple factory equipment A, B, and C (shown in FIG. 1), a server device 102 (shown in FIG. 14), and a client terminal 3 (shown in FIG. 1). Compared to the factory equipment A of embodiment 1, the factory equipment A includes a first inspection device 111 and a second inspection device 112 instead of the inspection device 15. The first inspection device 111 and the second inspection device 112 can inspect machined workpieces manufactured by multiple processing equipment A1, A2, A3, and A4. The first inspection device 111 and the second inspection device 112 form a network capable of communicating with edge computer terminals 11, 12, 13, and 14 and the server device 2.

[0106] The first inspection device 111 can, for example, inspect whether a machined workpiece is abnormal. The first inspection device 111 is, for example, an appearance inspection device that inspects the appearance of the machined workpiece for abnormalities. In other words, the first inspection device 111 inspects whether a machined workpiece is abnormal, i.e., whether the machined workpiece is defective. However, the first inspection device 111 is not limited to an appearance inspection device, and various measuring devices can be applied. The second inspection device 112 is, for example, a measuring device that inspects the quality of the machined workpiece, such as its dimensions and surface roughness, which indicate the machining accuracy. In other words, the second inspection device 112 acquires quality data that indicates the machining accuracy of the machined workpiece when the machined workpiece is normal.

[0107] For example, qualities that represent the machining accuracy of machined workpieces molded by processing equipment A1, A2 as injection molding machines include inner and outer diameter dimensions, roundness, cylindricity, surface roughness, workpiece mass, etc. Qualities that represent the machining accuracy of machined workpieces manufactured by processing equipment A3, A4 as machine tools include dimensional accuracy, surface roughness, the presence or absence of chatter, the presence or absence of a machined layer on the surface, etc.

[0108] 2. Functional block configuration of equipment monitoring system 101 The functional block configuration of the equipment monitoring system 101 will be described with reference to Figures 15 and 16. As described above, the equipment monitoring system 101 includes a plurality of factory equipment A, B, and C (shown in Figure 1), a server device 102 (shown in Figure 14), and a client terminal 3 (shown in Figure 1).

[0109] In Fig. 15, the observation devices 24, 34 and the edge computer terminals 11-14 constituting the processing equipment A1-A4 are the same as those in the first embodiment shown in Fig. 5. The first inspection device 111 is substantially the same as the inspection device 15 in the embodiment shown in Fig. 5.

[0110] Furthermore, in the server device 102, the facility-related data acquisition unit 81, the anomaly information acquisition unit 82, the first training data set storage unit 83, the first model generation unit 84, the first model storage unit 85, the anomaly estimation unit 86, and the anomaly estimation result storage unit 87 are similar to the facility-related data acquisition unit 81, the anomaly information acquisition unit 82, the training data set storage unit 83, the model generation unit 84, the model storage unit 85, the anomaly estimation unit 86, and the anomaly estimation result storage unit 87 constituting the server device 2 in embodiment 1. Therefore, a description of the above configuration will be omitted.

[0111] Furthermore, the transmitting unit 88, the processing condition adjusting unit 89, and the post-processing unit 90 perform the same processes as in the first embodiment, and also perform the processes described below. The following describes the differences from the first embodiment.

[0112] The server device 102 further includes a quality data acquisition unit 121 , a second training dataset storage unit 122 , a second model generation unit 123 , a second model storage unit 124 , a quality estimation unit 125 , and an estimated quality data storage unit 126 .

[0113] The server device 102 is configured to be able to execute the learning phase and estimation phase of machine learning, similarly to embodiment 1. As described in embodiment 1, the server device 102 generates a trained model (first trained model) for outputting an anomaly estimation result in the first learning phase. Furthermore, the server device 102 generates a second trained model for estimating quality data in the second learning phase.

[0114] Furthermore, as described in embodiment 1, the server device 102 generates an anomaly estimation result using a trained model (first trained model) in the first estimation phase and performs processing according to the anomaly estimation result. Furthermore, as the second estimation phase, the server device 102 generates estimated quality data using a second trained model and performs processing according to the estimated quality data.

[0115] The following describes the processing related to the second learning phase and the second estimation phase. For the second learning phase, the server device 102 uses the facility-related data acquisition unit 81, the quality data acquisition unit 121, the second training dataset storage unit 122, the second model generation unit 123, and the second model storage unit 124.

[0116] For the second learning phase, the equipment-related data acquisition unit 81 acquires equipment-related data indicating the equipment status or processing status of the processing equipment A1 to A4 from the edge computer terminals 11 to 14. In other words, the equipment-related data acquired by the equipment-related data acquisition unit 81 is not time-series data but feature quantities extracted from the time-series data. The equipment-related data acquired by the equipment-related data acquisition unit 81 may include multiple types of data.

[0117] The quality data acquisition unit 121 acquires the results of inspection of the machined workpiece to be used in the second learning phase by the second inspection device 112. The inspection results by the second inspection device 112 include quality data that indicates the machining accuracy of the machined workpiece. Therefore, the quality data acquisition unit 121 acquires the quality data that indicates the machining accuracy of the machined workpiece. The quality data acquisition unit 121 may be configured to acquire quality data for multiple types.

[0118] The second training dataset storage unit 122 associates the equipment-related data acquired by the equipment-related data acquisition unit 81 with the quality data of the machined workpiece acquired by the quality data acquisition unit 121 and stores them as a second training dataset.

[0119] The second model generation unit 123 generates a second trained model by performing machine learning using the second training data set. The second model generation unit 123 performs machine learning using the equipment-related data as an explanatory variable and the quality data of the machined workpiece as a target variable. Therefore, the second trained model generated by the second model generation unit 123 is a model that represents the relationship between the equipment-related data and the quality data.

[0120] The second trained model is a model that can output quality data of the machined workpiece when equipment-related data is input. Furthermore, when outputting the quality data, the second trained model can also output a numerical value (score) that represents the reliability of the quality data. In other words, when equipment-related data is input, the second trained model can output a numerical value (score) that represents the reliability of the quality data. Note that various machine learning models, such as neural networks and support vector machines, can be applied to the model.

[0121] The second model storage unit 124 stores the second trained model generated by the second model generation unit 123. The second trained model stored in the second model storage unit 124 may be a model corresponding to the type of processing equipment A1 to A4. For example, the second trained model may be a common model for processing equipment A1 and A2 as injection molding machines, and a common model for processing equipment A3 and A4 as machine tools. Even if the processing equipment A1 to A4 are of the same type, the second trained model may be a model corresponding to each of the processing equipment A1 to A4.

[0122] The second learning phase is completed when the second trained model is stored in the second model storage unit 124. However, the second trained model can also be updated while the second estimation phase is being executed. In this case, the server device 102 executes the second learning phase again.

[0123] Then, after the second trained model is stored in the second model storage unit 124 of the server device 102, the server device 102 can execute the second estimation phase. For the second estimation phase, the server device 102 uses the equipment-related data acquisition unit 81, the second model storage unit 124, the quality estimation unit 125, the estimated quality data storage unit 126, the transmission unit 88, the processing condition adjustment unit 89, and the post-processing unit 90.

[0124] For the second estimation phase, the equipment-related data acquisition unit 81 acquires equipment-related data indicating the equipment status or processing status of the processing equipment A1 to A4 from the edge computer terminals 11 to 14. That is, the equipment-related data acquisition unit 81 acquires equipment-related data when the processing equipment A1 to A4 performed the current processing as the estimation target. Here, the equipment-related data acquired by the equipment-related data acquisition unit 81 for the second estimation phase is the same type of data as the equipment-related data acquired for the second learning phase described above.

[0125] The quality estimation unit 125 generates estimated quality data, which is an estimate of quality data representing the machining accuracy of the machined workpiece that is the target of the current machining process, using the equipment-related data acquired by the equipment-related data acquisition unit 81 when the current machining process was performed and the second trained model stored in the second model storage unit 124. The quality estimation unit 125 generates the estimated quality data when the target processing equipment A1 to A4 is preparing for the next machining process. In other words, the quality estimation unit 125 generates the estimated quality data before the start of the next machining process.

[0126] The quality estimation unit 125 further predicts the transition of the estimated quality data of the next and subsequent machined workpieces based on the transition of the estimated quality data of the current and past multiple machined workpieces. The transition of the estimated quality data of the next and subsequent workpieces can be predicted by using an approximation formula such as a linear expression or a polynomial expression to calculate the rate of change of the past transition. Machine learning can also be applied to the prediction of the transition of the estimated quality data of the next and subsequent workpieces.

[0127] The estimated quality data storage unit 126 stores estimated quality data generated by the quality estimation unit 125. In particular, the estimated quality data storage unit 126 stores estimated quality data each time it is generated by the quality estimation unit 125. Therefore, the estimated quality data storage unit 126 stores the estimated quality data of the machined workpiece generated by the current processing process by the quality estimation unit 125, and also stores estimated quality data of multiple machined workpieces generated in the past by the quality estimation unit 125. Furthermore, the estimated quality data storage unit 126 stores the predicted transition of the estimated quality data of the machined workpieces from the next time onwards.

[0128] For example, the estimated quality data storage unit 126 stores past estimated quality data including the current Ta, as indicated by the star marks and black circles in FIG. 16. In FIG. 16, the black circles represent the estimated quality data of each past machined workpiece, and the star marks represent the estimated quality data of the current machined workpiece. Furthermore, the estimated quality data storage unit 126 stores the predicted trends in the estimated quality data of the next and subsequent machined workpieces as rewritable data, as indicated by the white circles in FIG. 16. In FIG. 16, the white circles represent the predicted trends in the estimated quality data of the next and subsequent machined workpieces. When a new predicted trend in the estimated quality data of the next and subsequent workpieces is obtained, it is rewritten.

[0129] 16, the quality data has a target value Tar for the quality data, an upper limit Th_max of the quality tolerance range (equivalent to the normal quality range), and a lower limit Th_min of the quality tolerance range. Here, the target value Tar means an ideal value in design, and may be set to the median value between the upper limit Th_max and the lower limit Th_min. The target value Tar may also be set to a value deviating from the median value, taking into account the quality tolerance range, machine characteristics, etc.

[0130] Furthermore, a threshold value Th1 is set for the quality data as a value greater than the target value Tar of the quality data and slightly smaller than the upper limit value Th_max. Also, a threshold value Th2 is set for the quality data as a value smaller than the target value Tar of the quality data and slightly greater than the lower limit value Th_min. The threshold values ​​Th1 and Th2 are used in the processing of the processing condition adjustment unit 89.

[0131] 15, when a transmission request for quality display is received from a client terminal 3, the transmission unit 88 transmits the estimated quality data of the multiple machined workpieces stored in the estimated quality data storage unit 126 to the target client terminal 3. At this time, the transmission unit 88 also transmits the predicted transition of the estimated quality data from the next time onwards. The client terminal 3 displays the estimated quality data of the multiple machined workpieces transmitted from the transmission unit 88 and the predicted transition of the estimated quality data on the display unit 3a.

[0132] Here, the transmission process by the transmission unit 88 to the client terminal 3 is performed before the start of the next processing. Therefore, the display unit 3a of the client terminal 3 will be in a state where estimated quality data of multiple machined workpieces including past, current, and subsequent machined workpieces is displayed before the start of the next processing.

[0133] The user of the client terminal 3 can grasp the estimated quality data of the currently machined workpiece before the start of the next processing. Furthermore, the user of the client terminal 3 can grasp the estimated quality data of past machined workpieces and the predicted transition of the estimated quality data for the next and subsequent periods in addition to the currently machined workpiece, so that the user can grasp the changes up to the present time and the predicted status for the next and subsequent periods.

[0134] Furthermore, the transmission unit 88 also transmits the multiple pieces of estimated quality data transmitted to the client terminal 3 to the control devices 22, 32 of the processing equipment A1-A4, and causes the estimated quality data to be displayed on the display units of the operation panels 23, 33. When the transmission unit 88 transmits the estimated quality data to the control devices 22, 32, the estimated quality data is transmitted via the edge computer terminals 11-14. If a worker is located near the processing equipment A1-A4, the worker can have the same understanding as the user of the client terminal 3.

[0135] The machining condition adjustment unit 89 (including the function of the second machining condition adjustment unit) adjusts the machining conditions for the next machining process for the processing equipment A1 to A4 based on the estimated quality data of the multiple machined workpieces stored in the estimated quality data storage unit 126. For example, when the difference between the estimated quality data of the current machined workpiece and the target value Tar within a predetermined quality tolerance range is equal to or greater than a predetermined value, the machining condition adjustment unit 89 adjusts the machining conditions so that the quality data of the next machined workpiece becomes the target value Tar.

[0136] 16, when the estimated quality data reaches a threshold value Th1 that is slightly smaller than the upper limit value Th_max of the acceptable quality range, the processing condition adjustment unit 89 adjusts the processing conditions so that the quality data becomes smaller. When the estimated quality data reaches a threshold value Th2 that is slightly larger than the lower limit value Th_min of the acceptable quality range, the processing condition adjustment unit 89 adjusts the processing conditions so that the quality data becomes larger.

[0137] The adjustment of the processing conditions can be performed automatically by setting the adjustment method in advance. When the processing condition adjustment unit 89 adjusts the processing conditions, the processing condition adjustment unit 89 outputs the adjusted processing conditions to the control devices 22 and 32 of the processing equipment A1 to A4, so that the control devices 22 and 32 operate according to the adjusted processing conditions.

[0138] When the machining conditions are adjusted by the machining condition adjustment unit 89, the transmission unit 88 transmits the adjusted machining conditions to the client terminal 3 in addition to the estimated quality data of the multiple machined workpieces. In this case, the display unit 3a of the client terminal 3 is in a state where the adjusted machining conditions are displayed. Furthermore, when the machining conditions are adjusted by the machining condition adjustment unit 89, the transmission unit 88 also transmits the adjusted machining conditions to the control devices 22, 32 of the target processing equipment A1 to A4, in the same way as the client terminal 3. Then, the adjusted machining conditions are displayed on the display units of the target operation panels 23, 33.

[0139] When the quality estimation unit 125 determines that the estimated quality data of the currently machined workpiece exceeds the quality tolerance range (Th_max to Th_min), the post-processing unit 90 transmits an instruction to the control devices 22, 32 of the target processing equipment A1 to A4 to transport the target machined workpiece to the second inspection device 112 or waste processing equipment. When the control devices 22, 32 receive the instruction to transport the target machined workpiece to the second inspection device 112 or waste processing equipment from the post-processing unit 90, they control the transport of the target machined workpiece to the second inspection device 112 or waste processing equipment. The determination of whether the target machined workpiece is to be transported to the second inspection device 112 or waste processing equipment is preset based on the size and type of estimated quality data. The waste processing equipment may be a storage case for disposal or equipment for crushing the machined workpiece, which is an injection-molded product.

[0140] When a machined workpiece is transported to the second inspection device 112, the second inspection device 112 starts inspecting the machined workpiece that has been transported and outputs the inspection results. If the quality data from the inspection by the second inspection device 112 is determined to be outside the quality tolerance range, the machined workpiece in question is transported to a waste disposal facility. On the other hand, if the quality data from the inspection by the second inspection device 112 is determined to be normal, the worker, manager, or client terminal 3 is notified of the normality as well as the quality data.

[0141] 3. Example of display content on client terminal 3 Examples of the display contents of the display unit 3a of the client terminal 3 will be described with reference to Figs. 17 to 21. As shown in Fig. 17, the display unit 3a of the client terminal 3 not only displays the current estimated quality data in a large size, but also can display a graph of the transition of the estimated quality data of the machined workpiece in the past, the current, and the next and subsequent times. This display content is displayed until the start of the next machining process.

[0142] As shown in Fig. 18, the display unit 3a of the client terminal 3 can display estimated quality data of multiple machined workpieces for each of the factory facilities A, B, and C. For example, the display unit 3a can display each of the quality types Qu1, Qu2, and Qu3. Also, as shown in Fig. 19, the display unit 3a of the client terminal 3 can display estimated quality data of multiple machined workpieces for each of the processing facilities A1 to A4.

[0143] As shown in Fig. 20, the display unit 3a of the client terminal 3 can display estimated quality data of multiple machined workpieces for each material lot R1, R2. As shown in Fig. 21, the display unit 3a of the client terminal 3 can display estimated quality data of multiple machined workpieces W1, W2, W3 for each workpiece and for each process.

[0144] Furthermore, the display unit 3a of the client terminal 3 can also display both the abnormality estimation result described in the first embodiment and the estimated quality data described in this embodiment.

[0145] 4.Effects According to the equipment monitoring system 101 of this embodiment, the server device 102 applies machine learning to generate estimated quality data for the machined workpieces. The estimated quality data generated by the server device 102 is transmitted from the server device 102 to the client terminal 3. The client terminal 3 displays the transmitted estimated quality data. Because the client terminal 3 forms a network capable of communicating with the server device 102, the user of the client terminal 3 can be located anywhere as long as the location is capable of network connection with the server device 2. In other words, even if the user of the client terminal 3 is not located at the installation locations of the processing equipment A1 to A4, B1 to B4, and C1 to C4, the user can grasp the estimated quality data of the machined workpieces that have been processed by the processing equipment A1 to A4, B1 to B4, and C1 to C4.

[0146] Furthermore, the server device 2 generates the estimated quality data before the start of the next processing. Therefore, the user of the client terminal 3 can grasp the status of the estimated quality data that indicates the processing accuracy of the processed workpiece by the current processing before the start of the next processing. Therefore, even if the user of the client terminal 3 is not located at the installation location of the processing equipment A1 to A4, B1 to B4, C1 to C4, he or she can grasp the estimated quality data of the processed workpiece in real time.

[0147] Furthermore, the server device 102 stores estimated quality data for a plurality of machined workpieces, including estimated quality data for the machined workpiece of the current machining process. Furthermore, the server device 102 also stores estimated quality data for the next and subsequent machining processes. The server device 102 then transmits the estimated quality data for the plurality of machined workpieces to the client terminal. The client terminal 3 displays the estimated quality data for the plurality of machined workpieces, including the estimated quality data for the current machined workpiece, before the start of the next machining process. Therefore, the user of the client terminal 3 can grasp the changes in the estimated quality data up to the present and the predicted trends in the estimated quality data for the next and subsequent machining processes. As a result, the user of the client terminal 3 can grasp the current situation with high accuracy. [Explanation of symbols]

[0148] 1. Facility monitoring system 2. Server device 3. Client terminal 3a Display of client terminal 81 Equipment-related data acquisition department 85 Model storage area 86 Anomaly estimation part 87 Abnormality estimation result storage section 88 Transmitter A1~A4, B1~B4, C1~C4 Processing equipment

Claims

1. a processing facility for manufacturing a plurality of processed workpieces by sequentially executing processing processes; a server device that configures a network capable of communicating with the processing equipment; a client terminal that configures a network capable of communicating with the processing facility and the server device; Equipped with The server device an equipment-related data acquisition unit that acquires equipment-related data indicating an equipment status or a processing status of the processing equipment; a model storage unit for storing a trained model that represents a relationship between the equipment-related data and the presence or absence of an abnormality, the trained model being generated by machine learning using a training data set including the equipment-related data and data on the presence or absence of an abnormality in the machined workpiece; an abnormality estimation unit that generates an abnormality estimation result indicating the presence or absence of an abnormality by the start of the next processing process using the equipment-related data and the trained model when the current processing process is executed; an abnormality estimation result storage unit that stores the abnormality estimation results generated by the abnormality estimation unit for the plurality of machined workpieces; a transmitting unit that transmits the abnormality estimation results of the plurality of machined workpieces stored in the abnormality estimation result storage unit to the client terminal; Equipped with The client terminal is an equipment monitoring system having a display unit capable of displaying the abnormality estimation results for a plurality of the machined workpieces, including the abnormality estimation results generated by the current processing, before the start of the next processing.

2. 2. The equipment monitoring system according to claim 1, wherein the server device further comprises a processing condition adjustment unit that adjusts processing conditions for a next processing operation for the processing equipment based on the abnormality estimation results of the plurality of machined workpieces stored in the abnormality estimation result storage unit.

3. the transmission unit of the server device transmits the machining conditions adjusted by the machining condition adjustment unit to the client terminal; The facility monitoring system according to claim 2 , wherein the display unit of the client terminal is further capable of displaying the machining conditions adjusted by the machining condition adjustment unit.

4. A plurality of the processing facilities are provided, the abnormality estimation result storage unit stores the abnormality estimation results of the plurality of machined workpieces for each of the processing facilities; 4. The equipment monitoring system according to claim 1, wherein the display unit of the client terminal is capable of displaying the abnormality estimation results for a plurality of the machined workpieces for each of the processing equipment.

5. The plurality of processing facilities are installed in a plurality of factory facilities, respectively; the abnormality estimation result storage unit stores, for each of the factory facilities, the abnormality estimation results of the plurality of machined workpieces manufactured by one or more of the processing facilities installed in the factory facilities; 5. The facility monitoring system according to claim 4, wherein the display unit of the client terminal is capable of displaying the abnormality estimation results for a plurality of the machined workpieces for each of the factory facilities.

6. the abnormality estimation result storage unit stores the abnormality estimation results of the plurality of machined workpieces for each material lot; 6. The equipment monitoring system according to claim 1, wherein the display unit of the client terminal is capable of displaying the abnormality estimation results for a plurality of the machined workpieces for each material lot.

7. a plurality of processing facilities for manufacturing one processed workpiece by sequentially performing different steps of the processing treatment; the abnormality estimation result storage unit stores the abnormality estimation result for each workpiece and for each process, 7. The equipment monitoring system according to claim 1, wherein the display unit of the client terminal is capable of displaying the abnormality estimation result for each work and for each process.

8. The abnormality estimation unit generating a numerical value representing the reliability of the abnormality estimation result as the abnormality estimation result representing the presence or absence of the abnormality; The equipment monitoring system according to any one of claims 1 to 7, wherein the equipment is classified into at least one of normal, abnormal, and possibly abnormal based on the numerical value representing the reliability.

9. The processing equipment comprises: If the abnormality estimation unit classifies the machined workpiece as abnormal, the machined workpiece is transported to a waste disposal facility; or 9. The equipment monitoring system according to claim 8, wherein when the abnormality estimation unit classifies the machined workpiece as possibly having the abnormality, the machined workpiece is transported to an inspection device.

10. The server device further a second model storage unit that stores a second learned model that represents a relationship between the equipment-related data and the quality data, the second learned model being generated by machine learning using a training data set that includes the equipment-related data and quality data that represents the machining accuracy of the machined workpiece; a quality estimation unit that generates estimated quality data that is an estimate of the quality data by the start of the next processing process, using the equipment-related data when the current processing process is executed and the second trained model; an estimated quality data storage unit that stores the estimated quality data generated by the quality estimation unit for a plurality of the machined workpieces; Equipped with the transmitting unit transmits the estimated quality data of the plurality of machined workpieces stored in the estimated quality data storage unit to the client terminal; An equipment monitoring system as described in any one of claims 1 to 9, wherein the display unit of the client terminal is capable of displaying the estimated quality data for multiple processed workpieces, including the estimated quality data generated by the current processing, before the start of the next processing.

11. moreover, A plurality of the processing facilities; a plurality of edge computer terminals provided in a one-to-one correspondence with each of the plurality of processing facilities; Equipped with the processing facility includes an observation device that detects the facility-related data as time-series data; the edge computer terminal forms a network capable of communicating with the processing facility and the server device, acquires the facility-related data from the observation device provided in the processing facility, and generates the facility-related data representing feature quantities; The facility monitoring system according to any one of claims 1 to 10, wherein the facility-related data acquisition unit of the server device acquires the facility-related data representing the feature amount from the edge computer terminal.

12. 12. The facility monitoring system according to claim 1, wherein the processing facility is an injection molding machine that manufactures the machined workpiece by injection molding.

13. 12. The facility monitoring system according to claim 1, wherein the processing facility is a machine tool that manufactures the machined workpiece by performing cutting or grinding on an unmachined workpiece.

Images