Wear Amount Prediction Device, Wear Amount Prediction Method, Control Program, and Recording Medium

Abstract

This wear amount predicting device comprises a predicting unit which uses a learning model to predict an amount of wear of a cutting tool corresponding to a machining time, and a display unit for displaying information based on the predicted result, wherein the learning model is generated by performing machine learning using, as teacher data, a dataset including: information relating to a cutting tool, conditions relating to cutting, and information relating to a workpiece material; and a machining time during which the cutting tool was used, an amount of wear of the cutting tool accompanying the machining, an initial wear time from when cutting by the cutting tool started until initial wear was complete, and an initial wear amount indicating an amount of wear of the cutting tool at a point in time at which the initial wear time had elapsed.

Description

Technical Field

[0001] The present disclosure relates to a wear amount prediction device for predicting the wear amount due to cutting of a cutting tool, a wear amount prediction method, a control program used in the wear amount prediction device, and a recording medium on which this control program is recorded.

Background Art

[0002] Cutting tools wear out with use. In the past, the degree of wear during use has often been predicted based on the operator's rules of thumb. Further, Patent Document 1 and Patent Document 2 describe configurations for predicting the wear amount using prediction formulas. Furthermore, Patent Document 3 describes a configuration for predicting wear by inputting an image of the cutting edge, machining conditions, and workpiece specifications into a learned model.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

[0004] A wear amount prediction device according to one aspect of the present disclosure includes a prediction unit that predicts the wear amount of a cutting tool according to the machining time using a learning model, and a display unit that displays information based on the result predicted by the prediction unit. The learning model uses information about the cutting tool, conditions related to cutting, and information related to the workpiece as input data, and a data set with the machining time using the cutting tool, the wear amount of the cutting tool associated with the machining, the initial wear time from the start of cutting by the cutting tool until the initial wear is completed, and the initial wear amount indicating the wear amount of the cutting tool at the time when the initial wear time has elapsed as output data is used as teacher data to perform machine learning to generate the model. The prediction unit inputs data including information about the cutting tool, conditions related to cutting, and information related to the workpiece into the learning model to predict the wear amount.

[0005] A wear amount prediction method according to one aspect of the present disclosure includes a prediction step of predicting the wear amount of a cutting tool according to the machining time using a learning model, and a display step of displaying information based on the result predicted in the prediction step. The learning model uses information about the cutting tool, conditions related to cutting, and information related to the workpiece as input data, and a data set with the machining time using the cutting tool, the wear amount of the cutting tool associated with the machining, the initial wear time from the start of cutting by the cutting tool until the initial wear is completed, and the initial wear amount indicating the wear amount of the cutting tool at the time when the initial wear time has elapsed as output data is used as teacher data to perform machine learning to generate the model. In the prediction step, data including information about the cutting tool, conditions related to cutting, and information related to the workpiece is input into the learning model to predict the wear amount.

[0006] The wear amount prediction device according to each aspect of the present disclosure may be realized by a computer. In this case, a control program for a wear amount prediction device that realizes the wear amount prediction device by operating the computer as each part (software element) included in the wear amount prediction device, and a computer-readable recording medium on which it is recorded also fall within the scope of the present disclosure.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Mode for Carrying Out the Invention

[0008] 〔Embodiment 1〕 Hereinafter, an embodiment of the present disclosure will be described in detail. The wear amount prediction device 1 according to this embodiment predicts the wear amount associated with the use of a cutting tool using a learning model 40 learned by machine learning.

[0009] First, with reference to FIG. 1, the main configuration of the wear amount prediction device 1 will be described. FIG. 1 is a functional block diagram showing the main configuration of the wear amount prediction device 1. As shown in FIG. 1, the wear amount prediction device 1 includes an input reception unit 10, a prediction unit 20, a display unit 30, and a learning model 40.

[0010] The input reception unit 10 receives the input to the wear amount prediction device 1. The input data in this embodiment includes data related to the cutting tool, data related to the machining conditions, and data related to the workpiece material.

[0011] The prediction unit 20 predicts the wear amount of the cutting tool based on the input data received by the input reception unit 10, and includes a wear amount prediction unit 21 and a graph generation unit 22.

[0012] The wear amount prediction unit 21 predicts the initial wear time, the initial wear amount, the final machining time, the final wear amount, and / or the defect probability from the input data received by the input reception unit 10 using the learning model 40. Here, the initial wear time is a time with a relatively large wear amount from the start of cutting until the cutting tool adapts to the workpiece material, that is, the time required for so-called initial wear (severe wear). In contrast to the relatively large wear amount of the initial wear, the wear that progresses relatively gently and stably is generally called steady wear (mild wear).

[0013] The initial wear amount is the wear amount of the cutting tool at the initial wear time. By measuring the changes in the machining time and wear amount under predetermined conditions, the initial wear time and initial wear amount under these conditions can be evaluated in advance. The final machining time is the limit time during which the workpiece can be machined until a defect occurs in the cutting tool after steady wear. The final wear amount is the wear amount at the final machining time. A defect is a chip that occurs in the cutting tool. Based on the data obtained from the above measurements, in addition to the initial wear time and initial wear amount, the final machining time, final wear amount, and the presence or absence of defects under predetermined conditions can be evaluated.

[0014] Examples of defects include abrasive chipping, chipping, welding chipping, mechanical chipping, thermal cracking chipping, and flaking. The defect probability is the probability that a defect occurs in the cutting tool. By predicting the defect probability, the user can be made aware of the probability that the cutting tool will have a defect.

[0015] The graph generation unit 22 generates a graph with the machining time on the horizontal axis and the wear amount on the vertical axis from the initial wear time, initial wear amount, final machining time, and final wear amount predicted by the wear amount prediction unit 21.

[0016] Here, with reference to FIG. 5, the method by which the graph generation unit 22 generates a graph will be described. FIG. 5 is a graph with the machining time (min) on the horizontal axis and the wear amount (mm) on the vertical axis. Assuming that the point indicating the initial wear time and initial wear amount predicted by the wear amount prediction unit 21 is point 502, and the point indicating the final machining time and final wear amount is point 501, the graph generation unit 22 generates a straight line 511 connecting point 502 and point 501.

[0017] The prediction unit 20 notifies the display unit 30 of the graph generated by the graph generation unit 22 as a prediction result and causes it to be displayed.

[0018] The display unit 30 displays the prediction result in the prediction unit 20 or information based on the prediction result. That is, the display unit 30 may display information based on the prediction result without displaying the prediction result itself. The display unit 30 may not be provided in the wear amount prediction device 1 but may be provided outside the wear amount prediction device 1 as an external device. The external device may be any device equipped with a display unit capable of displaying information, such as a personal computer, a tablet, or a smartphone. Also, the screen examples described later can be adapted to a personal computer with a large screen size or a smartphone with a small screen size. The size of the display area, the display position, etc. may be appropriately adjusted according to the screen size.

[0019] Next, with reference to FIGS. 6 and 7, an example of the prediction result of the wear amount prediction device 1 displayed on the display unit 30 will be described. FIGS. 6 and 7 are diagrams showing display examples of graphs as the prediction results of the wear amount prediction device 1.

[0020] The graph 601 in FIG. 6 shows the prediction result of the wear amount prediction device 1. As shown in the graph 601, since the wear amount prediction device 1 shows the prediction result in a graph, the user can recognize the wear amount at an arbitrary machining time. Also, the graphs 611 and 612 in FIG. 6 show the error range of the prediction result. The graph 611 shows the lower limit of the error, and the graph 612 shows the upper limit of the error. Therefore, a wear amount that falls at least between the graph 611 and the graph 612 is predicted.

[0021] FIG. 7 is an example of displaying a graph 621 showing the extraction result from the database on the graph shown in FIG. 6. Although the details of the extraction process from the database will be described later, the graph 621 shows the result of extracting past cutting data in the same data as the input data from the database. By showing the graph 621, the user can recognize the situation when cutting was performed under the same conditions in the past.

[0022] As described above, in the display unit 30, the results predicted by the prediction unit 20 are displayed in a graph that can list the processing time and the wear amount. As a result, a graph that can list the processing time and the wear amount is displayed, so that the user can easily and intuitively recognize the relationship between the processing time and the wear amount.

[0023] 〔Details of the learning model 40〕 The learning model 40 is a learning model generated by a machine learning algorithm using the teacher data 41. The machine learning algorithm can be any method, and for example, the following methods or combinations thereof can be used. · Gradient Boosting Decision Tree (GBDT) · Support Vector Machine (SVM) · Clustering · Inductive Logic Programming (ILP) · Genetic Programming (GP) · Baysian Network (BN) · Neural Network (NN) When using a neural network, a Convolutional Neural Network (CNN) including a convolutional process may be used. More specifically, as one or more layers included in the neural network, a convolutional layer that performs a convolutional operation may be provided, and a filter operation (sum of products operation) may be performed on the input data input to the layer. When performing the filter operation, processes such as padding may be used in combination, or an appropriately set stride width may be adopted.

[0024] Also, as the neural network, a multi-layer or ultra-multi-layer neural network with dozens to thousands of layers may be used.

[0025] Next, referring to FIG. 2, the teacher data 41 used for the machine learning of the learning model 40 will be described. FIG. 2 is a diagram showing an example of the teacher data 41. The teacher data 41 is a data set including teacher input data 41A which is the input data of the teacher data, and teacher output data 41B which is the output data of the teacher data. As shown in FIG. 2, the teacher input data 41A may include environmental information, tool information, cutting conditions, and workpiece material information. Further, the teacher output data 41B may include an initial wear amount, an initial wear time, a final wear amount, a final machining time, and the presence or absence of defects. As will be described in detail later, the tool information is information regarding a cutting tool. The cutting conditions are conditions related to cutting. The workpiece material information is information related to the workpiece material.

[0026] Next, the details of the teacher data 41 will be described with reference to FIG. 3. FIG. 3 is a diagram showing the details of the teacher data 41.

[0027] As shown in FIG. 3, the environmental information of the teacher input data 41A may include information regarding "processing machine", "cutting oil", "coolant pressure", and "continuous, intermittent machining".

[0028] The tool information may include information regarding "holder, insert", "breaker", "corner R", "material type", "coating thickness", "hardness", "fracture toughness", and "flexural strength". The cutting conditions may include information regarding "cutting speed", "feed per tooth", and "depth of cut". The workpiece material information may include information regarding "workpiece material name" and "composition". The details are as follows. · "Processing machine" is information indicating whether the machine for processing is a "lathe" or a "milling machine". Since the processing form varies significantly between lathe machining (turning machining) and milling machine machining (milling machining), it is useful information as teacher data. · "Cutting oil" is information indicating whether to lubricate during processing (dry), or blow air, and if lubricating, whether the oil is oil-based or water-soluble. · The "coolant pressure" is information indicating the pressure of the coolant output during machining. Since the coolant pressure causes a change in the life of the cutting tool, it is useful information as teaching data. · The "continuous and intermittent machining" is information indicating whether the cutting edge at the rear of the cutting tool is always in contact with the workpiece or in intermittent contact during machining. · The "holder, insert" is information indicating the model number of the holder of the cutting tool and the model number of the insert. · The "breaker" is information indicating the structure of the face (rake face) along the cutting edge of the insert. Depending on the shape of the rake face, the chips can be controlled. That is, since the progress of wear by the chips changes, it is useful information as teaching data for evaluating the wear of the cutting tool. · The "corner radius" is information indicating the roundness of the tip of the insert. · The "material type" is information indicating the composition of the insert. · The "coating thickness" is information indicating the thickness of the coating covering the insert. · The "hardness" is information indicating the hardness of the coating. · The "fracture toughness" is information indicating the toughness of the base material of the insert. The toughness of the base material affects the presence or absence of chipping of the insert. · The "flexural strength" is information indicating the toughness of the base material of the insert. As described above, the toughness of the base material affects the presence or absence of chipping of the insert. · The "cutting speed" is information indicating the rotational speed. · The "feed per tooth" is information for indicating the speed at which the cutting tool advances per revolution. The "depth of cut" is information indicating the amount of thickness that the cutting tool removes per pass. For example, the cutting width refers to the width of the chip and corresponds to the width of the chip removed per pass. · The "workpiece name" is information indicating the name of the material to be cut by the cutting tool. · The "composition" is information indicating the composition of the workpiece and shows the contents such as carbon, silicon, and manganese.

[0029] In addition, the teacher output data 41B may include an "initial wear time", an "initial wear amount", a "final machining time", a "final wear amount", and a "presence or absence of defect". · The "initial wear time" is the time required for so-called initial wear from the start of cutting until the cutting tool conforms to the workpiece. · The "initial wear amount" is the wear amount of the cutting tool during the initial wear time. · The "final machining time" is the time required for machining the workpiece. In the case of a defect, it is the time until immediately before that. · The "final wear amount" is the wear amount during the final machining time. · The "presence or absence of defect" indicates whether there is a defect in the chip after machining.

[0030] In this embodiment, the teacher output data 41B does not include the case where the cutting tool is defective within a predetermined time from the start of cutting. As a situation where the cutting tool is defective at the timing of almost no initial wear from the start of cutting, for example, there are cases where incorrect cutting conditions are set and it cannot be said that cutting is appropriately performed, or cases where a sudden defect of the cutting tool occurs accidentally. Since these are all specific situations, they are not appropriate as teacher data for predicting the steady wear amount of the cutting tool by the prediction unit 20.

[0031] With reference to FIG. 4, the content of the teacher output data 41B will be described. FIG. 4 is a graph showing cutting data when actual cutting is performed, where the horizontal axis represents the machining time (min) and the vertical axis represents the wear amount (mm). Here, five cases from graph 401 to graph 405 are shown.

[0032] The time of point 411 on graph 401 is the initial wear time and the initial wear amount. Therefore, the machining time and wear amount indicated by point 411 become the teacher data of the initial wear time and the initial wear amount. Also, since it is defective at the time of point 413, the time of point 412 is the final machining time and the final wear amount. Therefore, the machining time and wear amount indicated by point 412 become the teacher data as the final machining time and the final wear amount.

[0033] Similar to graph 401, in graph 402, the time point of point 421 indicates the initial wear time and the initial wear amount. Therefore, the processing time and wear amount indicated by point 421 become the teacher data of the initial wear time and the initial wear amount. Also, since it is missing at the time point of point 423, the time point of point 422 will indicate the final processing time and the final wear amount. Therefore, the processing time and wear amount indicated by point 422 become the teacher data as the final processing time and the final wear amount.

[0034] Graph 403, graph 404, and graph 405 are missing before the processing time reaches 10 minutes (predetermined time), so these are not used for teacher data. In the generation device of the learning model 40, it may be determined whether to include it in the teacher data or not using a threshold value. That is, a configuration may be adopted in which data with a wear amount exceeding the threshold within a predetermined time is determined to be missing within the predetermined time and excluded from the teacher data.

[0035] In this way, by not using the cutting data when it is missing within a predetermined time as teacher data, it is possible to exclude data that is not desirable as data for predicting the wear amount accompanying the cutting of the cutting tool, and more accurate prediction can be performed.

[0036] As described above, the teacher data 41 does not include information corresponding to a cutting tool that has a defect within a time that is normally impossible after the start of processing. When a cutting tool has a defect within a time that is normally impossible, it is considered that, for example, incorrect cutting conditions are set. By using the teacher data 41 except for the cutting tools with initial defects, it is possible to exclude the influence of the initial defects and predict the wear amount. Therefore, the prediction accuracy can be improved. The time that is normally impossible is, for example, 10 minutes.

[0037] The teacher data 41 is an example and is not limited to those listed above.

[0038] 〔Screen example in the display unit 30〕 Next, referring to FIGS. 8 and 9, a screen example in the display unit 30 will be described. FIG. 8 is a diagram showing a screen example for inputting data. FIG. 9 is a diagram showing a screen example for showing prediction results.

[0039] As shown in FIG. 8, in the input screen, information regarding the cutting tool and the workpiece for which the wear amount is to be predicted can be input. In the screen example 801, in the input area 810, it is possible to input a machining machine, cutting oil material, coolant pressure, continuous / intermittent machining, holder model number, chip model number, and corner R, etc. The input to this input area 801 only needs to input information necessary as input data for the prediction unit 20 to predict the wear amount, and is not limited to that shown in the screen example 801.

[0040] As shown in FIG. 9, in the screen example 901 showing the results, the initial wear time (min) is displayed in the area 911, the initial wear amount (mm) is displayed in the area 912, the final machining time (min) is displayed in the area 913, the final wear amount (mm) is displayed in the area 914, and the defect probability is displayed in the area 915. Also, the area 916 accepts an input as to whether or not to perform error display. If there is an input such as a check in the area 916, an error range, that is, a graph showing the graphs 611 and 612 (see FIG. 6), is displayed in the graph display area 921.

[0041] In the area 917, the maximum value of the machining time (min) is displayed, and in the area 918, the maximum value of the wear amount (mm) is displayed. The maximum value of the machining time refers to the predicted final machining time. The maximum value of the wear amount refers to the wear amount at the predicted final machining time.

[0042] The screen example 901 is not limited to a configuration in which the maximum value of the processing time (min) is displayed in the area 917 and the maximum value of the wear amount (mm) is displayed in the area 918. The graph 601 that is linear may be extended so that a value larger than the maximum value of the processing time (min) is displayed in the area 917 and a value larger than the maximum value of the wear amount (mm) is displayed in the area 918. For example, when the probability of defect is output together with the final processing time and the final wear amount, when the probability of defect at the final processing time is lower than a predetermined threshold (for example, 3%, 5%, 10%), the graph 601 may be extended as described above.

[0043] Also, a graph redrawing button 919 is also displayed. When the graph redrawing button 919 is pressed, the graph displayed in the graph display area 921 is redrawn. This is used, for example, when a check is entered in the area 916 or when the check is removed, so that a graph including the error range is displayed or a graph not including the error range is displayed.

[0044] In the graph display area 921, the graph shown in FIG. 6 or FIG. 7 is displayed as a prediction result.

[0045] As described above, in the wear amount prediction device 1', the display unit 30 displays the result by the prediction unit 20 in a graph. At this time, as shown in FIG. 7, the extraction information by the database information extraction unit 50 described later may be superimposed and displayed on the graph of the result by the prediction unit 20. Thereby, it is possible to easily make the user determine whether the result predicted using the learning model is appropriate or inappropriate.

[0046] 〔Flow of processing〕 Next, with reference to FIG. 10, the flow of processing in the wear amount prediction device 1 will be described. FIG. 10 is a flowchart showing the flow of processing of the wear amount prediction device 1.

[0047] As shown in Fig. 10, first, in the wear amount prediction device 1, the input reception unit 10 receives the input of input data (S101). Next, the prediction unit 20 predicts the wear amount using the learning model 40 (S102, prediction step). Finally, the display unit 30 displays the prediction result of the prediction unit 20 (S103, display step).

[0048] As described above, the wear amount prediction device 1 according to the present embodiment includes a prediction unit 20 that predicts the wear amount of a cutting tool according to the machining time using the learning model 40, and a display unit 30 that displays information based on the result predicted by the prediction unit 20.

[0049] The learning model 40 is (1) at least any one of information on the cutting tool, conditions related to cutting, and information related to the workpiece, (2) the machining time using the cutting tool, the wear amount of the cutting tool accompanying the machining, the initial wear amount indicating the wear amount of the cutting tool at the time when the initial time has elapsed since the start of cutting by the cutting tool, and the initial time, and is generated by performing machine learning using a data set including the above as teacher data.

[0050] The prediction unit 20 inputs data including at least any one of information on the cutting tool, conditions related to cutting, and information related to the workpiece into the learning model 40 to predict the wear amount.

[0051] With the above configuration, the user can be made to recognize how much a cutting tool will wear when any cutting tool is used under any cutting conditions. Therefore, the user can recognize the wear amount of the cutting tool at an arbitrary machining time. Also, by changing the value of the input data, it is possible to recognize how the wear amount changes, so it is also possible to perform a simulation of the wear amount. As a result, it is possible to recognize how the wear amount will be if the cutting tool is used in a certain way, and it becomes easy to select a cutting tool suitable for the workpiece.

[0052] The wear amount predicted by the prediction unit 20 may be different depending on the type of cutting tool, even if the surface of the cutting tool for predicting the wear amount is different. For example, the surface for predicting the wear amount may be different between a tip formed of cemented carbide and a tip formed of CBN (Cubic Boron Nitride). As the surface for predicting the wear amount, a rake face of the tip, a flank face, a relief face, a corner R, etc. can be considered. Also, the wear amount such as crater wear on the rake face, boundary wear on the flank face, and boundary wear on the relief face may be predicted.

[0053] Depending on the type of cutting tool, the worn portion is different, but the wear amount of an appropriate surface can be predicted as the surface to be predicted.

[0054] Also, in the above-described embodiment, the configuration in which the result predicted by the prediction unit 20 is displayed by the display unit 30 has been described, but the result predicted by the prediction unit 20 may be provided to another external device. If an application capable of processing using the prediction result is installed in the external device, various processes using the prediction result can be performed in the external device.

[0055] [Embodiment 2] Another embodiment of the present disclosure will be described below. For convenience of explanation, members having the same functions as the members described in the above embodiment are given the same reference numerals, and the description thereof will not be repeated.

[0056] With reference to FIG. 11, the wear amount prediction device 1' according to the present embodiment will be described. FIG. 11 is a functional block diagram showing the main configuration of the wear amount prediction device 1'. As shown in FIG. 11, the wear amount prediction device 1' according to the present embodiment includes a database information extraction unit 50 and a DB (database) 60 in addition to the configuration of the wear amount prediction device 1 of the above-described Embodiment 1.

[0057] In the DB 60, information indicating the wear amount associated with the use of the cutting tool is stored in association with various types of information.

[0058] Fig. 12 shows an example of the wear amount data stored in the DB60. The wear amount data is information in which information about the cutting tool, conditions related to cutting, information related to the workpiece material, machining time, wear amount, and presence or absence of defects are associated. As shown in Fig. 12, for example, in the DB60, information about the cutting tool, conditions related to cutting, information related to the workpiece material, machining time, wear amount, and presence or absence of defects are associated and stored. The information about the cutting tool includes a holder, a tip (holder model number, tip model number), a breaker, a corner R, a material type symbol, a film thickness, a hardness, a fracture toughness, and a flexural strength, etc. The conditions related to cutting include a machining machine, a cutting oil material, a coolant pressure, continuous, intermittent machining, a cutting speed, a feed per tooth, and a depth of cut, etc. The information related to the workpiece material includes a workpiece material name and a composition. Not all of these information are essential, and only arbitrary information may be associated.

[0059] The database information extraction unit 50 extracts wear amount data from the DB60 in which all or part of the input data received by the input reception unit 10 is the same as the information related to the workpiece material, etc., and notifies the display unit 30. The display unit 30 displays the extraction result by the database information extraction unit 50 together with the prediction result by the prediction unit 20.

[0060] 〔Example of a screen in the display unit 30〕 Next, with reference to Fig. 13, an example of a display screen of the extraction result by the database information extraction unit 50 will be described. Fig. 13 is a diagram showing an example of a display screen of the extraction result by the database information extraction unit 50.

[0061] In the screen example 1001 shown in Fig. 13, it is divided and displayed into an area 1011 and an area 1021. In the area 1011, "workpiece material", "material type symbol", and "tip model number" are displayed as selectable essential filter items, and "holder model number", "processing method", "machining machine", and "workpiece material classification" are displayed as selectable arbitrary filter items. Also, a search button 1012 and a clear button 1013 are displayed.

[0062] Search results are displayed in area 1021.

[0063] When the search button 1012 is pressed, wear amount data is extracted from the DB60 based on the items selected in the mandatory filter items and optional filter items, and the extraction result is displayed in the area 1021. Also, a graph showing the wear amount may be displayed, such as the graph 621 shown in FIG. 7 described above.

[0064] When the clear button 1022 is pressed, the selections in the mandatory filter items and optional filter items are canceled.

[0065] As described above, in the wear amount prediction device 1', from the DB60 that associates information on the cutting tool, conditions related to the cutting using this cutting tool, machining time using this cutting tool, and the wear amount of this cutting tool after machining in past use, it includes a database information extraction unit 50 that extracts information similar to the information on the cutting tool, conditions related to the cutting using this cutting tool, machining time using this cutting tool, and the wear amount of this cutting tool after machining. And the display unit 30 displays the information extracted by the database information extraction unit 50 in addition to the result predicted by the prediction unit 20.

[0066] Thereby, using the information extracted from the DB60, the user can recognize the actual state of the cutting tool in past use. And the user can use this result as a material for judging whether the result predicted using the learning model 40 is appropriate or inappropriate.

[0067] 〔Embodiment 3〕 Other embodiments of the present disclosure will be described below. For convenience of explanation, members having the same functions as the members described in the above embodiments are given the same reference numerals, and the description thereof will not be repeated.

[0068] Referring to FIG. 14, the wear amount prediction device 1' and the database update device 100 according to the present embodiment will be described. FIG. 14 is a functional block diagram showing the main configuration of the wear amount prediction device 1' and the database update device 100 according to the present embodiment. As shown in FIG. 14, in the present embodiment, in addition to the above-described wear amount prediction device 1', a database update device 100 is included.

[0069] The database update device 100 updates the DB60 and includes an input condition collection unit 110, a factor collection unit 120, an evaluation unit 130, and an update unit 140.

[0070] The input condition collection unit 110 collects the input data received by the input reception unit 10 and notifies the evaluation unit 130.

[0071] In the present embodiment, in addition to predicting the wear amount described above, the prediction unit 20 derives the presence or absence of defects, the final processing time, and the importance of factors that affected the prediction of the final wear amount. Then, the factor collection unit 120 collects from the prediction unit 20 the factors that affected the presence or absence of defects, the final processing time, and the final wear amount, and notifies the evaluation unit 130.

[0072] Further, the display unit 30 may be configured to display the factors derived by the prediction unit 20. By displaying the factors, the user can recognize where the factors that affected the presence or absence of defects, the final processing time, and the final wear amount are.

[0073] Further, the display unit 30 may be configured to display a plurality of factors derived by the prediction unit 20 and the degree of influence exerted by each factor. Thereby, the user can recognize the degree of each of the factors that affected the presence or absence of defects, the final processing time, and the final wear amount.

[0074] Here, the factors will be described with reference to FIG. 15. FIG. 15 is a diagram showing the importance (degree) of the factors that affected the prediction of the presence or absence of defects, the final machining time, and the final wear amount. 1401 in FIG. 15 indicates the importance of the factors related to the presence or absence of defects, 1402 indicates the importance of the factors related to the final machining time, and 1403 indicates the importance of the factors related to the final wear amount.

[0075] As shown in 1401 of FIG. 15, the factors that affect the prediction of the presence or absence of defects, in descending order of importance, are hardness (HB), chip type number, ae (mm), workpiece material, Vc (m / min), ap (mm), breaker, fracture toughness (MPa·m 1 / 2 ), holder type number, fz (mm / t). Here, hardness (HB) is the hardness of the workpiece material. ae (mm) is the amount of width cut per pass. Vc (m / min) is the cutting speed. ap (mm) is the amount of thickness cut per pass. fz (mm / t) is the speed at which the cutting edge advances.

[0076] As shown in 1402 of FIG. 15, the factors that affect the prediction of the final machining time, in descending order of importance, are Vc (m / min), material type symbol, ap (mm), f (mm / rev), coefficient of linear expansion (K -1 ), fracture toughness (MPa·m 1 / 2 ), specific gravity (g / cm 3 ), film thickness (μm), holder type number, breaker. Here, f (mm / rev) is the speed at which the cutting tool advances per revolution and corresponds to the above-mentioned "feed per tooth" × "number of teeth". The coefficient of linear expansion (K -1 ) is the coefficient of expansion of the chip's base material. The specific gravity (g / cm 3 ) is the specific gravity of the chip's base material.

[0077] As shown in 1403 of FIG. 15, the factors that affect the prediction of the final wear amount, in descending order of importance, are Vc (m / min), ap (mm), ae (mm), breaker, chip type number, f (mm / rev), Cr, hardness (HB), holder type number, material type symbol. Here, Cr is information regarding the composition of the workpiece material and is the amount of chromium contained in the workpiece material.

[0078] The evaluation unit 130 evaluates whether the data to be stored in the DB 60 using the information on the input data acquired from the input condition collection unit 110 and the factors and their importance levels acquired from the factor collection unit 120. Then, it notifies the update unit 140 of the highly evaluated information.

[0079] For example, when "workpiece" is frequently input as the input data, the evaluation unit 130 determines that "workpiece" has a high evaluation and notifies the update unit 140. Also, for example, for factors with an importance level exceeding 0.5, it determines that they have a high evaluation and notifies the update unit 140.

[0080] The update unit 140 updates the DB 60 by storing the wear amount data including the information notified from the evaluation unit 130 in the DB 60. At this time, in addition to the DB 60, the learning model 40 may also be updated.

[0081] Also, the evaluation unit 130 may be executed by AI (Artificial Intelligence). In this case, the AI may operate in the database update device 100, or may operate in another device (for example, an edge computer or a cloud server, etc.).

[0082] Also, according to the above-described wear amount prediction devices 1 and 1', it can lead to efficient use of cutting tools. Thereby, it can contribute to the achievement of sustainable development goals (SDGs).

[0083] 〔Example of realization by software〕 The functions of the wear amount prediction device 1 (hereinafter referred to as "device") are programs for causing a computer to function as the device, and can be realized by programs for causing a computer to function as each control block of the device (especially each part included in the prediction unit 20).

[0084] In this case, as hardware for executing the above program, the above device includes a computer having at least one control device (for example, a processor) and at least one storage device (for example, a memory). By executing the above program with this control device and storage device, each function described in the above embodiments is realized.

[0085] The above program may be recorded on one or more computer-readable recording media, rather than temporarily. This recording medium may or may not be provided in the above device. In the latter case, the above program may be supplied to the above device via any wired or wireless transmission medium.

[0086] Also, part or all of the functions of the above control blocks can also be realized by a logic circuit. For example, an integrated circuit in which a logic circuit functioning as each of the above control blocks is formed is also included in the scope of the present disclosure. In addition to this, for example, it is also possible to realize the functions of the above control blocks by a quantum computer.

[0087] The invention according to the present disclosure has been described above based on the drawings and examples. However, the invention according to the present disclosure is not limited to the above-described embodiments. That is, the invention according to the present disclosure can be variously modified within the scope shown in the present disclosure, and embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the invention according to the present disclosure. Note that it is easy for those skilled in the art to make various deformations or corrections based on the present disclosure. Also, note that these deformations or corrections are included in the scope of the present disclosure.

Explanation of Signs

[0088] 1 Wear amount prediction device 10 Input reception unit 20 Prediction unit 21 Wear amount prediction unit 22 Graph generation unit 30 Display unit 40 Learning model 50 Database information extraction unit 60 DB 100 Database update device 110 Input condition collection unit 120 Cause collection unit 130 Evaluation unit 140 Update unit

Claims

1. A prediction unit that predicts the wear amount of a cutting tool according to the processing time using a learning model, A display unit that displays information based on the result predicted by the prediction unit, The learning model uses information related to the cutting tool, conditions related to cutting, and information related to the workpiece as input data, and uses the processing time using the cutting tool, the wear amount of the cutting tool accompanying the processing, the initial wear time from the start of cutting by the cutting tool until the initial wear is completed, and the initial wear amount indicating the wear amount of the cutting tool at the time when the initial wear time has elapsed as output data. It is generated by performing machine learning using a data set as teacher data, The prediction unit inputs data including information related to the cutting tool, conditions related to cutting, and information related to the workpiece into the learning model to predict the wear amount, a wear amount prediction device.

2. The data set includes information on factors contributing to the wear of the cutting tool, and the display unit displays the factors contributing to the wear of the cutting tool predicted by the prediction unit. The wear amount prediction device according to claim 1.

3. The display unit displays a plurality of factors contributing to the wear of the cutting tool predicted by the prediction unit and the degree to which each factor affects the wear. The wear amount prediction device according to claim 2.

4. The data set includes information on the presence or absence of defects in the cutting tool, and the display unit displays the probability of defects in the cutting tool predicted by the prediction unit. The wear amount prediction device according to any one of claims 1 to 3.

5. The data set does not include information corresponding to a cutting tool that has developed a defect within a time that is normally impossible after the start of processing. The wear amount prediction device according to any one of claims 1 to 3.

6. Depending on the type of the cutting tool, the surface of the cutting tool for predicting the wear amount is different. The wear amount prediction device according to any one of claims 1 to 3.

7. The display unit displays the result predicted by the prediction unit in a graph where the processing time and the wear amount can be listed. The wear amount prediction device according to any one of claims 1 to 3.

8. From a database associating information on a cutting tool, conditions related to cutting using the cutting tool, machining time using the cutting tool, and the amount of wear of the cutting tool after machining, extract information similar to the information on the cutting tool, conditions related to cutting using the cutting tool, machining time using the cutting tool, and the amount of wear of the cutting tool after machining, and include a database information extraction unit. The wear amount prediction device according to any one of claims 1 to 3, wherein the display unit displays the information extracted by the database information extraction unit in addition to the result predicted by the prediction unit.

9. The wear amount prediction device according to claim 8, wherein the display unit displays the result by the prediction unit and the extraction information by the database information extraction unit on the same graph.

10. A prediction step of predicting the wear amount of a cutting tool according to machining time using a learning model, And a display step of displaying information based on the result predicted in the prediction step. The learning model uses information on a cutting tool, conditions related to cutting, and information related to a workpiece as input data, and uses machining time using the cutting tool, the amount of wear of the cutting tool due to machining, the initial wear time from the start of cutting by the cutting tool until the initial wear is completed, and the initial wear amount indicating the amount of wear of the cutting tool at the time when the initial wear time has elapsed as output data. It is generated by performing machine learning using a data set as teacher data. In the prediction step, data including information on a cutting tool, conditions related to cutting, and information related to a workpiece is input to the learning model to predict the wear amount. A wear amount prediction method.

11. A control program for causing a computer to function as the wear amount prediction device according to claim 1, and a control program for causing a computer to function as the prediction unit.

12. A computer-readable recording medium recording the control program according to claim 11.

Images