Device for detecting an anomaly when attaching a tool

Abstract

Device (1) for detecting an anomaly when attaching a tool to a tool magazine in a tool changer provided in a machining center (2), characterized in that it comprises: a data acquisition facility (30) for obtaining data relating to the machine processing center (2); a preprocessor (34) for generating data relating to at least chronological data of the vibration or sound produced in the tool changer when the tool is attached to the tool magazine, wherein the data relating to at least chronological data are generated on the basis of the data obtained by the data acquisition device (30); and a tool attachment anomaly detection device (36) for detecting an anomaly in the attachment of the tool to the tool magazine based on the data generated by the preprocessor (34).

Description

Background of the invention 1. Field of the invention:

[0001] The present invention relates to a device for detecting an abnormality in the attachment of a tool. 2. Description of the related art:

[0002] Machine tools equipped with an automatic tool changer, and in particular, typical machining centers, are widely used in the field of machine tools. Such a typical machining center includes a tool magazine that houses a plurality of tools, each of which is attached to a tool holder. The machining center selects a predetermined tool through an indexing action of the tool magazine to move the selected tool to a predetermined position, and releases the tool together with its tool holder from the tool magazine through a tool changing action by a tool changing arm or the like to bring the tool into proper engagement with a main spindle and use the tool in a machining operation (see, for example, International Publication No. WO 93 / 18884 A1, i.e.Japanese Patent Application Laid-Open No. JP H05 - 261 638 A (Patent Document 1)).

[0003] As examples of traditional technologies for detecting abnormalities in such a tool changing operation, Japanese Patent No. JP 4 501 918 B2 (Patent Document 2) and Japanese Patent Application Laid-Open No. JP H11 - 333 657 A (Patent Document 3), which are the patent documents described below, as well as Patent Document 1 described above, disclose technologies for detecting abnormalities associated with the position at which the tool is attached to the tool magazine, the tool changing operation, and the attachment state of the tool to the main spindle in machine tools including an automatic tool changer.

[0004] A tool magazine includes a plurality of mechanical hand grippers that grip the tool holder using elastic spring force. When a tool is to be attached to the tool magazine, the tool is positioned relative to the tool holder such that a keyway of the tool holder to which the tool is to be fitted corresponds to a positioning groove provided on the gripper. The tool holder is pushed into the gripper, and the tool holder is thereby gripped by the gripper.

[0005] At this point, in some cases, due to operator error or the like, the tool holder may not be correctly attached to the gripper. In such a case, the tool may be released from the tool magazine or collide with the main spindle as a result of the tool change operation, which may result in damage to a turning tool and the main body of the machine.

[0006] To solve such a problem, an abnormality in the attachment state of a tool relative to a main spindle can be detected according to the technology disclosed in Patent Document 2. However, a gripper provided in a tool magazine cannot detect the attachment state of the tool relative to its tool holder. Likewise, the technology disclosed in Patent Document 3 is only capable of detecting abnormalities associated with a tool change operation.

[0007] Likewise, according to the technology disclosed in Japanese Patent Application Laid-Open No. JP 2005-324262 A (Patent Document 4), it is possible to detect whether the posture of a tool attached to a tool magazine is correct or not. However, for example, the tool may appear to be attached to a tool magazine in the correct posture, while the gripper cannot actually fully grip the tool holder due to chips adhering to the attachment portion of the tool holder to the gripper. Accordingly, it is difficult for the technologies disclosed in the above-mentioned patent documents to detect such an abnormal state of tool attachment.

[0008] JP 2007 - 319 995 A describes a machine tool equipped with: at least one first air supply path arranged to penetrate the holding part toward the first abutment surface where the tool holder and the holding part abut each other when the tool holder is attached to the holding part; an air supply device that applies air at a prescribed air pressure and supplies the air through the first air supply path toward the first abutment surface; a measuring device; and an assessment device connected to the measuring device for assessing the foreign object bitten onto the first abutment surface based on measurement data acquired by the measuring device.

[0009] From DE 10 2017 003 165 A1, a machine learning device is known which learns the failure prediction of a main shaft of a machine tool or of a motor driving the main shaft, comprising a state observation unit which observes a state variable which includes data output by a motor controller controlling the motor and / or data output by a sensor detecting a state of the motor and / or data output by a measuring device measuring a state of the main shaft or the motor; a detection data acquisition unit which acquires the detection data for determining whether a failure has occurred on the main shaft or the motor or a degree of failure; and a learning unit which learns the failure prediction of the main shaft or the motor according to a data set generated based on a combination of the state variable and the detection data. Summary of the invention

[0010] In view of the circumstances described above, an object of the present invention is to provide an apparatus capable of accurately detecting an abnormality in the attachment of a tool to a tool magazine.

[0011] According to the present invention, vibration or sound applied to the tool changer (specifically, a tool magazine) when an operator attaches a tool (specifically, a tool holder to which the tool is attached) to a tool magazine is used as state data. Machine learning is performed to learn, for example, either the vibration or sound when the tool is correctly attached and the vibration or sound when the tool is incorrectly attached, and the normality / abnormality of the tool attachment is estimated using the learning result to solve the above-described problem.At the time of attaching a tool to the tool magazine, when a mechanical gripper grips the tool holder by an elastic spring force or the like, its predetermined portions engage with each other, resulting in the generation of a specific vibration or sound. However, if chips adhere to the attachment portion of the tool holder and interfere with the gripper gripping the tool holder or the operator pushing the tool, and the operation is thus incomplete, then no such vibration or sound is generated, but a different vibration or sound is generated. As a result, if a learning model for learning the vibrations or sounds at the time of normal and abnormal operations is created through machine learning, it is possible to estimate the difference between them and detect normality / abnormality in the tool attachment.

[0012] Furthermore, an embodiment of the present invention provides an apparatus for detecting an abnormality in the attachment of a tool to a machine tool in a tool changer provided in a machining center, the apparatus comprising: data acquiring means for acquiring data regarding the machining center; a preprocessor for preparing data regarding at least chronological data of the vibration or sound generated in the tool changer when the tool is attached to the tool magazine, the data regarding the at least chronological data being prepared based on the data acquired by the data acquiring means; and tool attachment abnormality detecting means for detecting an abnormality in the attachment of the tool to the tool magazine based on the data prepared by the preprocessor.

[0013] Another embodiment of the present invention provides an apparatus for detecting an abnormality in the attachment of a tool to a tool magazine in a tool changer provided in a machining center, the apparatus comprising: a data acquisition device for acquiring data regarding the machining center; a preprocessor for creating state data including at least tool attachment vibration data regarding chronological data of a vibration or sound generated in the tool changer when the tool is attached to the tool magazine, the state data being created as learning data based on the data acquired by the data acquisition device;and a tool attachment anomaly detection device configured to detect an anomaly in the attachment of the tool to the tool magazine based on data created by the preprocessor. The tool attachment anomaly detection device includes a learning device for performing machine learning using learning data created by the preprocessor and further for creating a learning model for detecting an anomaly in the attachment of the tool to the tool magazine.

[0014] Another aspect of the present invention provides an apparatus for detecting an abnormality in the attachment of a tool to a tool magazine in a tool changer provided in a machining center, comprising data acquiring means for acquiring data regarding the machining center; a preprocessor for preparing state data including at least tool attachment vibration data regarding chronological data of vibration or sound generated in the tool changer when the tool is attached to the tool magazine, the state data being prepared based on the data acquired by the data acquiring means; and tool attachment abnormality detecting means for detecting an abnormality in the attachment of the tool to the tool magazine based on the data prepared by the preprocessor.The tool attachment anomaly detection device includes a learning model memory configured to store a learning model for learning the attachment state of the tool attached to the tool magazine with respect to chronological data of a vibration or sound generated in the tool changer when the tool is attached to the tool magazine; and an estimator configured to estimate the attachment state of the tool attached to the tool magazine using the learning model stored in the learning model memory, wherein the attachment state is estimated based on the state data created by the preprocessor.

[0015] According to the present invention, it is possible to estimate the attachment state of a tool with high accuracy, which in turn makes it possible to reduce the probability of erroneous operation at the time of tool attachment by an operator. Likewise, according to the present invention, errors in tool attachment can be detected at reduced cost because it is possible to estimate the attachment state of the tool using a single sensor or a small number of sensors. Short description of the drawings

[0016] The above-described and other objects and features will become apparent upon studying the following description of embodiments with reference to the accompanying drawings. Fig. 1 is a hardware configuration diagram schematically showing a tool attachment abnormality detecting apparatus according to a first embodiment of the present invention; Fig. 2 is a functional block diagram schematically showing the device according to the first embodiment; Fig. 3 is a hardware configuration diagram schematically showing a tool attachment abnormality detecting apparatus according to a second or third embodiment of the present invention; Fig. 4 is a functional block diagram schematically showing the device according to the second embodiment; Fig. 5 is a diagram showing an example of a learning model created by unsupervised learning in the device according to the second embodiment; Fig. 6 is a diagram showing an example of a learning model created by supervised learning in the apparatus according to the second embodiment; Fig. 7 is a functional block diagram schematically showing the apparatus according to the third embodiment of the present invention; Fig. 8 is a diagram showing an example of estimation of normality / anomaly in tool attachment using the learning model constructed by the unsupervised learning in the apparatus according to the third embodiment; Fig. 9 is a diagram showing an example of estimation of normality / abnormality in tool attachment using the learning model created by supervised learning in the apparatus according to the third embodiment; and Fig. 10 is a diagram showing an embodiment of the tool attachment abnormality detecting apparatus for estimating normality / abnormality in the attachment of a tool of a plurality of machining centers. Detailed description of the preferred embodiments

[0017] Some embodiments of the present invention will be described below with reference to the drawings.

[0018] Fig. 1 is a hardware configuration diagram schematically showing the main part of a tool attachment abnormality detection device, including a machine learning device according to a first embodiment (which may also be referred to as a "tool attachment abnormality detection device" or simply as "the device 1"). The device 1 according to this embodiment can be implemented, for example, in a control device that controls a machining center 2. Likewise, the device 1 according to this embodiment can be implemented as a personal computer equipped with the control device that controls the machining center, or as a computer, such as an edge computer, a cell computer, a host computer, and a cloud server, connected to the control device via a wired or wireless network.In this embodiment, the device 1 is implemented as a personal computer equipped with the control device that controls the machining center 2.

[0019] With reference to Fig. 1, the tool attachment anomaly detection device 1 may include a central processing unit (CPU) 11, a read-only memory (ROM) 12, a random access memory (RAM) 13, a non-volatile memory 14, and a plurality of interfaces 16 to 19.

[0020] The central processing unit (CPU) 11 included in the device 1 according to this embodiment is a processor responsible for the overall control of the device 1. The CPU 11 reads system programs stored in the read-only memory (ROM) 12 via a bus 20 and controls the entire device 1 according to the system programs. The random access memory (RAM) 13 can temporarily store temporary calculation data and other various data input by an operator via an input 71.

[0021] The non-volatile memory 14 is configured as a memory backed up by a battery (not shown), a solid-state drive (SSD), or the like, and retains its state even when the device 1 is turned off. The non-volatile memory 14 includes a setting area in which setting information regarding the operation of the device 1 is provided, and stores data input from the input 71, various data acquired from the machining center 2 (an operating state regarding tool change, unit types of the machining center, etc.), chronological data of various physical quantities acquired from the sensor 3 (vibrations, sounds, etc.).generated in a tool changer provided in the machining center 2), and data read via a network and / or from an external memory (not shown). The programs and various data stored in the non-volatile memory 14 can also be stored in the RAM 13 when they are running or being used. Likewise, system programs including a known analysis program for analyzing various data are stored in advance in the ROM 12.

[0022] The machining center 2 is a machine tool equipped with an automatic tool changer. The machining center 2 is capable of performing various machining operations, such as milling, punching, and drilling, while interchanging various turning tools with each other through the automatic tool changer. A plurality of tools, each loaded into a tool holder, are attached to the tool magazine by the operator. A tool gripped at a predetermined position of the tool magazine and attached to it is attached to the main spindle of the machine tool to be used in a designated machining operation according to a tool change command from a machining program or the like.The control device of the machining center 2 outputs information regarding the state of the machining center (such as the unit types of the machining center, an operating state of the machine tool, an open / closed operating state of the cover, and an operating state of tool change) to the device 1 via the interface 16.

[0023] The sensor 3 is configured to acquire chronological data of a vibration or sound generated in the tool changer provided in the machining center 2 (specifically, the tool magazine). An acceleration sensor, an acoustic emission (AE) sensor, a noise collector, an optical interferometer, and the like can be used as the sensor 3, for example. At least one sensor 3 can sufficiently detect the vibration or sound generated in the tool changer. When a plurality of sensors 3 are mounted on or near the tool changer, it becomes possible to detect the vibration or sound generated in the tool changer in a more diverse manner, and the accuracy of the machine learning described below can be further improved.Alternatively, if the sensors 3 are mounted at appropriate positions taking into account the structure of the tool magazine, then it is also possible to sufficiently detect the vibration and sound generated in the tool changer by a single or a small number of sensors 3.

[0024] Data read into the memory, data obtained as a result of executing a program or the like, data output from the machine learning device 100, which will be described below, and any other relevant data can be output to a display 70 via the interface 17 and displayed on the display 70. Likewise, the input 71, which is configured with a keyboard, a pointing device, and the like, sends a command, data, etc. based on the operator's operation to the CPU 11 via the interface 18.

[0025] Fig. Fig. 2 is a functional block diagram schematically showing the tool attachment anomaly detection device 1 according to the first embodiment. The respective Fig. The functional blocks shown in Figure 2 are effected by the CPU 11 provided in the device 1, which in Fig. 1 shows how it runs system programs and controls the operations of the individual components in the device 1.

[0026] The device 1 according to this embodiment includes a data acquisition device 30, a preprocessor 34, and a tool attachment anomaly detection device 36. The tool attachment anomaly detection device 36 includes an estimation device 120. Also provided in the non-volatile memory 14 are an acquired data memory 50 in which data acquired by the data acquisition device 30 is stored, and a normal vibration data memory 140.

[0027] The data acquisition device 30 is a functional unit that acquires various data input from the machining center 2, the sensor 3, the input 71, and the like. The data acquisition device 30 acquires various data, such as the operating status regarding tool change, the unit types of the machining center 2, chronological data of the vibration or sound generated in the tool changer when the tool is attached to the tool magazine by the operator, and information regarding the tool attachment input by the operator, and stores these pieces of information in the acquired data storage 50. The data acquisition device 30 can also be configured to acquire data from an external storage (not shown) or from other devices via a wired or wireless network.

[0028] When the data acquisition device 30 acquires the chronological data of the vibration or sound generated in the tool changer when the tool is attached to the tool magazine by the operator, the data acquisition device 30 identifies vibration or sound data indicating vibration or sound generated in the tool changer obtained from the sensor 3 based on the operating state of the machine tool acquired from the machining center 2, the open / closed operating state of the cover, the operating state of the tool changer, and the like. The data acquisition device 30 further identifies the vibration or sound data indicating the vibration or sound generated in the tool changer when the tool is attached to the tool magazine by the operator.Generally, when an operator performs various operations such as the task of attaching a tool to a tool magazine, he / she operates a control panel or the like, stops the machining by the machine tool, opens the cover, positions a specific gripper of the tool magazine at the tool attachment position to attach the tool thereto, closes the cover, and thus completes the tool change. The data acquisition unit 30 can detect such a typical tool attachment operation to identify the vibration or sound detected by the sensor 3 during such an operation as the vibration or sound data indicative of the vibration or sound generated in the tool changer when the tool is attached to the tool magazine by the operator.Likewise, if the pressure or the like associated with the tool magazine can be detected, the operator's pressing of the tool against the tool magazine (tool attachment) can be detected based on the state of the pressure. Thus, the pressure and / or the detected pressing of the tool by the operator can be used to identify the vibration or sound data indicative of the vibration or sound generated in the tool changer when the operator attaches the tool to the tool magazine. Note that examples of the case where the pressure or the like of the tool magazine can be detected include, among others, a case where a pressure sensor is independently installed and a case where the current of a motor for rotating the tool magazine can be detected.If any other means are available to detect the time at which the operator adds the tool to the tool magazine, then the identification can also rely on the detected time.

[0029] The preprocessor 34 creates data used in a tool attachment anomaly detection process by the attachment anomaly detector 36 based on the data acquired by the data acquirer 30 (and the data stored in the acquired data storage 50). The preprocessor 34 creates data acquired by subjecting the data acquired by the data acquirer 30 to conversion (quantification, normalization, sampling, etc.) into a unified format to be managed by the tool attachment anomaly detector 36.Preprocessor 34 can be configured to represent the data obtained by processing the chronological data of the vibration or sound generated in the tool changer at the time of tool attachment and detected by sensor 3 as associated data that has been resampled at a predetermined cycle. The preprocessor can also be configured to use data indicating characteristics of the chronological data (e.g., a known mel-frequency cepstrum coefficient).

[0030] The tool attachment anomaly detection device 36 is a functional unit for detecting an abnormality in the attachment of a tool based on the data prepared by the preprocessor 34.

[0031] The estimator 120 provided in the tool attachment anomaly detection device 36 is a functional unit that refers to normal vibration data stored in a normal vibration data memory 140 and estimates an abnormality in the attachment of a tool to the tool magazine by an operator based on the data created by the preprocessor 34. In the normal vibration data memory 140, chronological data of the vibration or sound detected by the sensor 3 when the operator correctly attached the tool to the tool magazine provided in the machining center 2 in advance is set as the normal vibration data.

[0032] The estimator 120 calculates a degree of correspondence between the chronological data of the vibration or sound generated in the tool changer, which are acquired by the data acquisition device 30, and the normal vibration data stored in the normal vibration data storage device 140, using a known waveform pattern matching model, such as dynamic programming (DP) matching and a hidden Markov model (HMM). If the calculated degree of correspondence exceeds a predetermined threshold, the estimator 120 estimates that the attachment of the tool to the tool magazine was performed correctly. Otherwise, the estimator 120 estimates that the attachment of the tool to the tool magazine is in an abnormal state.The estimator 120 may also be configured to calculate the degree of abnormality according to how far the degree of agreement between the chronological data of the vibration generated in the tool changer or the sound generated in the tool changer, which are acquired by the data acquisition device 30, and the normal vibration data stored in the normal vibration data storage 140 is from the predefined threshold.

[0033] Furthermore, if the estimation device 120 estimates that the attachment of the tool to the tool magazine is in an abnormal state, the tool attachment abnormality detection device 36 may display the estimation result by the estimation device 120 (normality / abnormality in the tool attachment, and if the occurrence of the abnormality is estimated, the degree of abnormality) on the display 70 and output the estimation result to send the result to a host computer, a cloud computer, or the like via a wired or wireless network (not shown). The device 1 may also be configured to change the display state of the display 70 according to the magnitude of the degree of abnormality.

[0034] The normal vibration data storage 140 may store a plurality of pieces of normal vibration data acquired from the same machining center 2. In this case, the estimator 120 may perform the estimation process to estimate the attachment state of the tool relative to the tool magazine using each of the plurality of pieces of normal vibration data. The estimator 120 may then estimate the attachment state of the tool relative to the tool magazine as normal if the attachment state is estimated to be normal based on any one of the plurality of pieces of normal vibration data, or if the attachment state is estimated to be normal with respect to a predetermined number of the pieces of normal vibration data, the predetermined number being defined in advance.

[0035] Likewise, normal vibration data acquired from the unit types of the machining center 2 may be stored in the normal vibration data storage 140 together and in association with the unit types of the machining center 2. In this case, the estimation device 120 is to perform the estimation process to estimate the attachment state of the tool relative to the tool magazine between the chronological data of the vibration generated in the tool changer or the sound generated in the tool changer, which are acquired by the data acquisition device 30, and the normal vibration data associated with the unit types of the machining center 2 from which the chronological data of the vibration or sound were acquired.

[0036] In the apparatus 1 having the above-described configuration, detection of an abnormality in tool attachment is performed using the data prepared by the preprocessor 34 based on the data acquired from the machining center 2 and the sensor 3. The apparatus 1 according to this embodiment estimates the attachment state of the tool attached by the operator not based on the external appearance, but based on the vibration or sound generated at the time of tool attachment. Therefore, even if a tool appears to be properly attached but is actually improperly attached due to a barely noticeable misalignment or the like, in the apparatus 1, it is possible to detect the abnormality of the tool with high accuracy.

[0037] Fig. 3 is a hardware configuration diagram schematically showing the main part of the tool attachment abnormality detection apparatus equipped with a machine learning device according to a second or third embodiment.

[0038] According to this embodiment, the CPU 11 provided in the tool attachment anomaly detection device 1 is a processor responsible for the overall control of the device 1. The CPU 11 reads system programs stored in the ROM 12 via the bus 20 to control the entire device 1 according to the system programs. The RAM 13 can temporarily store temporary calculation data, various data input by an operator via the input 71, etc.

[0039] The non-volatile memory 14 can be configured as a memory backed up by a battery (not shown), a solid-state drive (SSD), or the like, and retains its storage state even when the device 1 is turned off. The non-volatile memory 14 includes a setting area in which setting information regarding the operation of the device 1 is provided, and stores data input from the input 71, various data acquired from the machining center 2 (the operating status regarding tool change, unit types of the machining center, or the like), and chronological data of various physical quantities acquired from the sensor 3 (vibrations, sounds, etc. generated in the tool changer provided in the machining center 2).Furthermore, the non-volatile memory 14 can store data read from an external memory (not shown) and / or via a network. The programs and various data stored in the non-volatile memory 14 can be deployed to the RAM 13 when they are running or being used. Likewise, system programs, including a known analysis program for analyzing various data and a program for controlling communications with the machine learning device 100, which will be described below, can be pre-written in the ROM 12.

[0040] The device 1 according to this embodiment further includes an interface 21, which is an interface for interconnecting the tool attachment anomaly detection device 1 and the machine learning device 100 therein. The machine learning device 100 includes a processor 101 responsible for overall control of the machine learning device 100, the ROM 102 that stores system programs and the like, the RAM 103 for temporary storage of individual operations associated with machine learning, and the non-volatile memory 104 used to store a learning model and the like. The machine learning device 100 is capable of observing and monitoring various pieces of information that can be acquired by the device 1 via the interface 21.Here, the information that can be acquired by the device 1 may include, but is not limited to, the operating status of the tool change, the unit types of the machining center 2, and the chronological data of the vibration or sound generated in the tool changer, where the tool changer is provided in the machining center 2. Likewise, the device 1 can acquire the processing result output from the machine learning device 100 via the interface 21, store and display the acquired result, and transmit the result to another device via a network not shown.

[0041] Fig. Fig. 4 is a functional block diagram schematically showing the tool attachment anomaly detection device 1 and the machine learning device 100 according to the second embodiment. The device 1 according to this embodiment includes a structure required for the machine learning device 100 to perform learning (perform a "learning mode"). The individual functional blocks shown in Fig. 4 are effected by the CPU 11, which is in the Fig. 3, and the processor 101 in the machine learning device 100, which executes their respective system programs to control the operations of the individual components of the device 1 and the machine learning device 100.

[0042] The apparatus 1 according to this embodiment includes a data acquisition device 30, a preprocessor 34, and a tool attachment anomaly detector 36. The machine learning device 100, which constitutes the tool attachment anomaly detector 36, also includes a learning device 110. Similarly, an acquired data memory 50 is provided in the non-volatile memory 14. The acquired data memory 50 stores the data acquired by the data acquisition device 30. The non-volatile memory 104 in the machine learning device 100, which constitutes the tool attachment anomaly detector 36, includes a learning model memory 130. The learning model memory 130 stores the learning model constructed through machine learning by the learning device 110.

[0043] The data acquisition device 30 is a functional unit for acquiring various data input from the machining center 2, the sensor 3, the input 71, and the like. The data acquisition device 30 acquires various data including the operating status regarding tool change, the unit types of the machining center 2, the chronological data of the vibration or sound generated in the tool changer when the tool is attached to the tool magazine by the operator, and information regarding the attachment of the tool input by the operator, to store these pieces of information in the acquired data storage device 50. The data acquisition device 30 can also be configured to acquire the data from an external storage device (not shown) or from other devices via a wired or wireless network.

[0044] The preprocessor 34 creates learning data for use in learning by the machine learning device 100 based on the data acquired by the data acquirer 30 (and the data stored in the acquired data storage 50). The preprocessor 34 creates state data obtained by converting the data acquired by the data acquirer 30 (e.g., quantification, normalization, and sampling) into a unified format to be managed by the machine learning device 100. For example, if the machine learning device 100 performs unsupervised learning, the preprocessor 34 creates, as the learning data, state data S having a predetermined format according to the unsupervised learning.If the machine learning device 100 performs supervised learning, the preprocessor 34 creates, as the learning data, a set of state data S and label data L having a predetermined format according to the supervised learning.

[0045] The state data S created by the preprocessor 34 includes at least tool attachment vibration data S1. The tool attachment vibration data S1 is chronological data of the vibration or sound generated in the tool changer and detected by the sensor 3 when the operator attaches the tool to the tool magazine. It is possible to use, as the tool attachment vibration data S1, data obtained by using the chronological data of the vibration or sound generated in the tool changer at the time of tool attachment, detected by the sensor 3, as associated data, which has been subjected to resampling at a predetermined cycle. It is also possible to use data indicating the characteristics of the chronological data (such as a known mel-frequency cepstrum coefficient).

[0046] If the label data L is included in the learning data created by the preprocessor 34, the label data L also includes at least tool attachment state data L1. The tool attachment state data L1 indicates the information regarding the normality / abnormality of the tool attachment at the time the operator attaches the tool. For example, an input value input from the input 71 indicating the result of manual confirmation of the tool attachment state by the operator after the operator attaches the tool can be used as the tool attachment state data L1.

[0047] The learning device 110 performs machine learning using the learning data created by the preprocessor 34. The learning device 110 performs machine learning using the data acquired from the machine processing center 2 according to a known machine learning scheme, such as unsupervised learning and supervised learning, to create a learning model, and stores the created learning model in the learning model storage 130. Examples of unsupervised learning schemes performed by the learning device 110 include the autoencoder and K-Means. Examples of supervised learning schemes include a multilayer perceptron, an iterative neural network, a long-term memory, and a convolutional neural network.

[0048] As an example of machine learning by the learning device 110, unsupervised learning can be performed based on the state data S created by the preprocessor 34 based on the data acquired when the tool was correctly attached to the tool changer provided in the machining center 2. In this way, the distribution (clusters) of the learning data acquired in a state where the attachment of the tool to the tool magazine was correctly performed can be generated as a learning model.

[0049] Fig. Fig. 5 is a diagram showing an example of a learning model created based on the learning data obtained by unsupervised learning in this embodiment, and obtained in a state where the attachment of the tool to the tool magazine was correctly performed. It should be noted that Fig. 5, for the sake of explanation and simplification, illustrates a learning model by way of example in which only parameters A, B, and C are involved as the state data S, but the actual state data S (for example, the state data S with values ​​representing the chronological data as its elements) is represented by a higher-order vector space. If the learning model created in this way is to be used, the estimator 120, which will be described below, estimates the normality / abnormality in the attachment of a tool according to whether or not the data acquired from the machining center 2 and the sensor 3 is included in the distribution of the learning data acquired in the state in which the attachment of the tool to the tool magazine was correctly performed (cluster 202 as the learning model) when the data is compared with the distribution.If it is determined that an anomaly exists, the degree of anomaly can be calculated as the result of the estimation according to how far the data is from the distribution of the training data.

[0050] Likewise, the learning device 110 may perform machine learning based on the data 204 obtained when the tool is correctly attached to the tool changer provided in the machining center 2 and the data 206 obtained when the tool is incorrectly attached to the tool changer provided in the machining center 2. For example, the learning device 110 performs supervised learning using the learning data (teaching data) created by the preprocessor 34, adding a label indicating normality to the data 204 and adding a label indicating abnormality to the data 206, and may generate, as the learning model, the discrimination boundary 208 between the normal data and the abnormal data.

[0051] Fig. Fig. 6 is a diagram showing an example of a learning model created based on the learning data obtained when the attachment of the tool to the tool magazine is performed according to the supervised learning in this embodiment. It should be noted that Fig. 6, for the sake of explanation and simplification, describes a learning model by means of an example in which only parameters A and B are involved. The current state data S (for example, the state data S with values ​​representing the chronological data as its elements) is represented by a high-order vector space. If the learning model created in this way is to be used, the estimator 120, which will be described below, estimates the normality / abnormality in the attachment of a tool according to which side the data 204, 206 newly acquired from the machining center 2 and the sensor 3 are plotted with respect to the discrimination boundary 208 as the learning model. If it is determined that there is an abnormality, the degree of abnormality can be calculated as the estimation result according to how far the data is from the discrimination boundary.

[0052] In the device 1 having the above-described configuration, the learning device 110 performs machine learning using the learning data created by the preprocessor 34 based on the data acquired from the machining center 2 and the sensor 3. The learning model 208 thus created can be used in estimation based on data regarding the vibration generated in the tool changer or the sound generated in the tool changer, which are obtained from the sensor 3 when the tool is attached to the tool magazine by the operator.

[0053] Fig. Figure 7 is a functional block diagram schematically showing the tool attachment anomaly detection device 1 and the machine learning device 100 according to the third embodiment. The device 1 according to this embodiment includes a structure required for the machine learning device 100 to perform estimation (estimation mode). The individual functional blocks shown in Fig. 7 are executed by the CPU 11, which is in the Fig. 3, and causes the processor 101 in the machine learning device 100 to execute their respective system programs to control the operations of the individual components of the tool attachment anomaly detecting device 1 and the machine learning device 100.

[0054] The device 1 according to this embodiment includes the data acquisition device 30 and the preprocessor 34, as in the first embodiment. The machine learning device 100, which forms the tool attachment anomaly detection device 36, includes the estimation device 120. Likewise, the non-volatile memory 14 includes an acquired data storage 50. The acquired data storage 50 stores the learning data used in the state estimation by the machine learning device 100. The learning model storage 130 is provided on the non-volatile memory 104 of the machine learning device 100, which forms the tool attachment anomaly detection device 36. The learning model storage 130 stores the learning model constructed by machine learning by the learning device 110.

[0055] The data acquisition device 30 according to this embodiment is a functional unit for acquiring various data input from the machining center 2, the sensor 3, the input 71, and the like. The data acquisition device 30 acquires various data and stores the acquired data in the acquired data storage 50. Here, the various data that can be acquired by the data acquisition device 30 may include, but are not limited to, the operating status regarding tool change, the unit types of the machining center 2, the chronological data of the vibration or sound generated in the tool changer when the tool is attached to the tool magazine by the operator, and the information regarding the attachment of the tool input by the operator.The data acquisition device 30 may also be configured to acquire the data from an external storage (not shown) or from other devices via a wired or wireless network. The preprocessor 34 according to this embodiment creates state data S for use in estimation by the machine learning device 100 based on the data stored in the acquired data storage 50. The preprocessor 34 creates state data acquired by subjecting the acquired data to conversion (such as quantification, normalization, and sampling) into a unified format to be managed by the machine learning device 100. The state data S created by the preprocessor 34 includes at least tool attachment vibration data S1.The tool attachment vibration data S1 is chronological data of the vibration or sound generated in the tool changer, which was detected by the sensor 3 when the tool was attached to the tool magazine by the operator.

[0056] The estimator 120 performs an estimation of the state of the machining center using the learning model stored in the learning model memory 130 based on the state data S created by the preprocessor 34. In the estimator 120 of this embodiment, the state data S input from the preprocessor 34 is input into the learning model created by (its parameters determined by) the learning device 110 to estimate whether or not the tool has been correctly attached to the tool magazine.

[0057] Fig. Fig. 8 is a diagram showing an example of estimating the normality / anomaly of tool attachment using the learning model constructed by unsupervised learning in this embodiment. Note that Fig. 8, for the sake of explanation and simplification, describes a learning model by way of an example in which only parameters A, B, and C are involved as the state data S. The actual state data S (for example, the state data S with values ​​representing the chronological data as its elements) will be represented as a higher-order vector space. When estimation is performed using the learning model constructed by unsupervised learning, whether or not the tool attachment is normal or abnormal is estimated according to whether the state data S input as a target of estimation falls within the distribution of the learning data (the cluster 202) constructed as the learning model or lies outside the distribution. In the figures, the state data acquired when the tool attachment proceeds normally is indicated by reference numeral 210.In contrast, the condition data acquired when it is abnormal is indicated by reference numeral 212. If the estimator 120 acquires the condition data 212 and estimates that the tool attachment is abnormal, the estimator 120 may also calculate the degree of abnormality 214 based on the distance between the condition data and the cluster.

[0058] Fig. Fig. 9 is a diagram showing an example of estimating the normality / anomaly of tool attachment using the learning model created by supervised learning in this embodiment. It should be noted that Fig. 9, for the sake of explanation and simplification, describes a learning model using an example in which only parameters A and B are involved. The actual state data S (for example, the state data S with values ​​representing the chronological data as its elements) will be represented by a higher-order vector space. When the estimation is performed using the learning model created by supervised learning, whether the attachment of the tool is normal or abnormal or not is estimated according to which of the sides designated by the discrimination boundary 208 created as the learning model the state data S input as the target of the estimation lies. Similarly, if it has been estimated that an abnormality exists, the degree of anomaly 214 can be further calculated based on the distance between the state data and the discrimination boundary 208.

[0059] The result of the estimation by the estimator 120 (for example, normality / abnormality in the attachment of a tool, the degree of abnormality if its occurrence has been estimated) can be displayed and output on the display 70, and can be sent and output to a host computer, a cloud computer, or the like via a wired or wireless network (not shown) for use by the host computer, a cloud computer, or the like. The device 1 can also be configured to change the display state of the display 70 according to the magnitude of the degree of abnormality.

[0060] In the apparatus 1 having the above-described configuration, estimation of normality / abnormality in tool attachment is performed using the state data prepared by the preprocessor 34 based on the data acquired from the machining center 2 and the sensor 3. The apparatus 1 according to this embodiment performs estimation of the attachment state of the tool attached by the operator not based on the external appearance, but based on the vibration or sound generated at the time of tool attachment. This makes it possible to detect the abnormality of the tool with high accuracy even when a tool appears to be attached properly but is actually attached improperly due to a barely noticeable misalignment or the like.

[0061] While the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments and can be implemented in various modes with various modifications thereto as needed.

[0062] Although the second and third embodiments described above were described on the assumption that the tool attachment anomaly detecting device 1 and the machine learning device 100 are devices each having a central processing unit or processor different from each other, the machine learning device 100 may also be effected by the processor 11 provided in the tool attachment anomaly detecting device 1 and the system programs stored in the ROM 12.

[0063] While the second and third embodiments described above were also described assuming that the learning structure constitutes one embodiment and the estimation structure constitutes another embodiment different from the former, it is also possible to construct a tool attachment anomaly detection device 1 that combines both of these structures. In this case, the device 1 will be operated to update (learn) the learning model as needed while performing the estimation of the tool attachment state.

[0064] Furthermore, the tool attachment anomaly detection device 1 may be mounted on, for example, a host computer, a cloud server, and so on. As shown in Fig.10, if the tool attachment abnormality detection device 1 is connected via a wired or wireless network 5 to a plurality of machining centers 2 and the sensors 3, each of which is attached to a corresponding one of these machining centers 2, then the device 1 can be configured to collect data from the respective machining centers 2 and their sensors 3 and detect the normality / abnormality of the attachment state of the tool relative to the tool magazine in each of the machining centers 2.

Claims

[1] Device (1) for detecting an anomaly in the attachment of a tool to a tool magazine in a tool changer provided in a machining center (2), characterized by that it includes: a data acquisition device (30) for acquiring data relating to the machining center (2); a preprocessor (34) for creating data relating to at least chronological data of the vibration or sound generated in the tool changer when the tool is attached to the tool magazine, the data relating to the at least chronological data being created on the basis of the data acquired by the data acquisition device (30); and a tool attachment anomaly detection device (36) for detecting an anomaly in the attachment of the tool to the tool magazine based on the data prepared by the preprocessor (34). [2] Device (1) according to claim 1, characterized by , that the device is connected via a network to a plurality of machining centers (2), and based on data obtained from the plurality of machining centers (2), the abnormality in the attachment of the tool to the tool magazine is detected in the tool changer provided in a corresponding one of the plurality of machining centers (2). [3] Device (1) for detecting an anomaly in the attachment of a tool to a tool magazine in a tool changer provided in a machining center (2), characterized by that it includes: a data acquisition device (30) for acquiring data relating to the machining center (2); a preprocessor (34) for creating state data (S) comprising at least tool attachment vibration data regarding chronological data of a vibration or sound generated in the tool changer when the tool is attached to the tool magazine, the state data (S) being created as learning data based on the data acquired by the data acquisition device (30); and a tool attachment anomaly detection device (36) configured to detect an anomaly in the attachment of the tool to the tool magazine based on data created by the preprocessor (34), wherein the tool attachment anomaly detection device (36) comprises a learning device (110) for performing machine learning using learning data created by the preprocessor (34) and further for creating a learning model for detecting an anomaly in the attachment of the tool to the tool magazine. [4] Device (1) according to claim 3, characterized by that the learning model is generated by unsupervised learning. [5] Device (1) according to claim 3, characterized by that the learning model is generated through supervised learning. [6] Device (1) according to claim 3, characterized by , that the device (1) is connected via a network to a plurality of machining centers (2), and based on data obtained from the plurality of machining centers (2), a learning model is generated which is configured to detect the abnormality in the attachment of the tool to the tool magazine in the tool changer which is provided in a corresponding one of the plurality of machining centers (2). [7] Device (1) for detecting an anomaly in the attachment of a tool to a tool magazine in a tool changer provided in a machining center (2), characterized by that it includes: a data acquisition device (30) for acquiring data relating to the machining center (2); a preprocessor (34) for creating status data (S) comprising at least tool attachment vibration data regarding chronological data of a vibration or sound generated in the tool changer when the tool is attached to the tool magazine, the status data (S) being created based on the data acquired by the data acquisition device (30); and a tool attachment anomaly detection device (36) for detecting an anomaly in the attachment of the tool to the tool magazine based on the data created by the preprocessor (34), the tool attachment anomaly detection device (36) comprising: a learning model memory (130) configured to store a learning model for learning the attachment state of the tool attached to the tool magazine with respect to chronological data of a vibration or a sound generated in the tool changer when the tool is attached to the tool magazine; and an estimator (120) configured to estimate the attachment state of the tool attached to the tool magazine using the learning model stored in the learning model memory (130), wherein the attachment state is estimated based on the state data (S) created by the preprocessor (34). [8] Device (1) according to claim 7, characterized by that the learning model is generated by unsupervised learning. [9] Device (1) according to claim 7, characterized by that the learning model is generated through supervised learning. [10] Device (1) according to claim 7, characterized by , that the device (1) is connected via a network to a plurality of machining centers (2), and based on the data obtained from the plurality of machining centers (2), the abnormality in the attachment of the tool to the tool magazine of the tool changer provided in a corresponding one of the plurality of machining centers (2) is detected.

Images