Information update device and information update method for diagnosing processing abnormalities

Abstract

To provide an information updating device for diagnosing the processing abnormality of a machine tool capable of realizing the efficiency of diagnosis possibility judgement on unlearned data.SOLUTION: An information updating device 2 includes: information holding means 9 for holding a diagnosis model capable of updating through learning and processing condition information that can be supported by the diagnosis model; information inputting means 8 capable of inputting processing condition of processing of carrying out diagnosis; data acquisition means 3 for acquiring processing data; label providing means 4 for providing the processing data with a label; calculation means 5 for calculating the degree of processing abnormality; possibility judgment means 6 for judging, by using the abnormality degree and the label, the possibility of diagnosis support of processing abnormality diagnosis in the processing of carrying out the diagnosis; and information updating means 7 for updating the processing condition information held by the information holding means 9 based on results of the possibility judgment of the diagnosis support judged by the possibility judgment means 6 and the processing condition inputted by the information inputting means 8.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This disclosure relates to an information update device and information update method for diagnosing machining abnormalities in machine tools such as machining centers. [Background technology]

[0002] In machining using machine tools, damage such as breakage or chipping of the tool used can damage the workpiece being machined. Damaged workpieces become defective due to reduced precision and surface quality, leading to decreased productivity. Furthermore, if the material cost of the workpiece is high, this can result in significant cost losses. In this context, a technology has been disclosed that diagnoses machining abnormalities by measuring signals indicating the machining state and classifying whether the machining state is normal or abnormal. For example, Patent Document 1 discloses a technique that takes machine operation information as input, compares the features output by a diagnostic model created using machine learning with a threshold value, and diagnoses whether the tool's condition is normal or abnormal. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2019-67137 [Overview of the project] [Problems that the invention aims to solve]

[0004] The technology disclosed in Patent Document 1 allows for diagnosis of untrained data, such as when processing conditions change, and enables the separation of normal and abnormal conditions, thereby expanding the range of diagnosable processing conditions. However, while this technology does expand the range of diagnosable processing conditions, it does not involve machine learning, and the expanded range of diagnosable processing conditions is not retained. In other words, the diagnostic model is unable to learn the untrained data, even if it is clear from the work history that processing anomaly diagnosis is possible. Therefore, for untrained data, a decision on whether or not to perform a diagnosis, including machine learning and actual processing, is required before a diagnosis can be made. Consequently, a lot of man-hours are required to apply processing anomaly diagnosis, and there is a need to improve the efficiency of the decision-making process for whether or not to perform processing anomaly diagnosis on untrained data.

[0005] Therefore, the purpose of this disclosure is to provide an information update device and information update method for diagnosing machining abnormalities in machine tools that can improve the efficiency of determining whether or not a diagnosis is possible for unlearned data. [Means for solving the problem]

[0006] To achieve the above objective, the first configuration of this disclosure is an information update device for diagnosing machining anomalies in machining using a machine tool that processes a workpiece using a cutting tool, comprising: an information holding means that holds a learning-updatable diagnostic model and machining condition information that the diagnostic model can handle; an information input means that can input the machining conditions of the machining to be diagnosed; a data acquisition means that acquires machining data; a labeling means that assigns a normal or abnormal label to the machining data; a calculation means that calculates the degree of abnormality of the machining to be diagnosed based on the machining data; a feasibility determination means that determines whether or not a diagnostic response for machining anomaly diagnosis in the machining to be diagnosed is possible using the comparison result between the maximum value of the rate of change of the degree of abnormality in the machining to be diagnosed and a preset rate of change threshold, and the label; and an information update means that updates the machining condition information held by the information holding means based on the feasibility determination result of the diagnostic response determined by the feasibility determination means and the machining conditions input by the information input means. Another aspect of the first configuration of this disclosure is characterized in that, in the above configuration, the feasibility determination means, in determining whether or not to respond to a processing abnormality diagnosis, uses the result of comparing the maximum value of the rate of change of abnormality in the processing to be diagnosed with a preset rate of change threshold, and the label, and if the label is abnormal, the diagnosis is deemed possible if the maximum value of the rate of change of abnormality exceeds the rate of change threshold, and not possible if it does not exceed the rate of change threshold; if the label is normal, the diagnosis is deemed possible if the maximum value of the rate of change of abnormality does not exceed the rate of change threshold, and not possible if it exceeds the rate of change threshold. Another aspect of the first configuration of this disclosure is characterized in that, in the above configuration, if the result of the judgment on whether or not a diagnostic response is possible for the processing to be diagnosed is no, the information update means updates the diagnostic model by performing additional learning using processing data and labels, and then adds the processing conditions entered by the information input means to the processing condition information held by the information holding means; if the result of the judgment on whether or not a diagnostic response is possible, it compares the processing conditions entered by the information input means with the processing condition information held by the information holding means, and only if the processing condition information held by the information holding means does not include the processing conditions entered by the information input means, it adds the processing conditions entered by the information input means to the processing condition information held by the information holding means. To achieve the above objective, the second configuration of this disclosure is an information update device for diagnosing machining anomalies in machining using a machine tool that processes a workpiece using a cutting tool, comprising: an information holding means for holding a learning-updatable diagnostic model and machining condition information that the diagnostic model can handle; an information input means for inputting machining conditions for machining to be performed; a data acquisition means for acquiring machining data; a labeling means for assigning normal or abnormal labels to the machining data; a calculation means for calculating the degree of abnormality of the machining to be performed based on the machining data; a feasibility determination means for determining whether or not a diagnostic response to the machining anomaly diagnosis in the machining to be performed is possible using the degree of abnormality and the labels; and a determination of whether or not a diagnostic response is possible as determined by the feasibility determination means. The system includes an information update means that updates the processing condition information held by the information holding means based on the diagnosis result and the processing conditions entered by the information input means. The information update means updates the diagnostic model by performing additional learning using processing data and labels if the result of the diagnosis to determine whether a diagnostic response is possible for the processing to be diagnosed is negative, and then adds the processing conditions entered by the information input means to the processing condition information held by the information holding means. If the result of the diagnosis is positive, it compares the processing conditions entered by the information input means with the processing condition information held by the information holding means, and only adds the processing conditions entered by the information input means to the processing condition information held by the information holding means if the processing condition information held by the information holding means does not include the processing conditions entered by the information input means. To achieve the above objective, the third configuration of this disclosure is an information update method for diagnosing machining anomalies in machining using a machine tool that processes a workpiece using a cutting tool, characterized in that it reads a learning-updatable diagnostic model and machining condition information that the diagnostic model can handle, inputs the machining conditions for the machining to be diagnosed, acquires machining data, assigns a normal or abnormal label to the machining data, calculates the degree of abnormality of the machining to be diagnosed based on the machining data, uses the result of comparing the maximum value of the rate of change of the degree of abnormality in the machining to be diagnosed with a preset rate of change threshold and the label to determine whether or not a diagnostic response is possible for the machining anomaly diagnosis in the machining to be diagnosed, and updates the diagnostic model and / or machining condition information based on the determined result of whether or not a diagnostic response is possible and the input machining conditions. To achieve the above objective, the fourth configuration of this disclosure is an information update method for diagnosing machining anomalies in machining using a machine tool that processes a workpiece using a cutting tool, characterized in that it reads a learning-updatable diagnostic model and machining condition information that the diagnostic model can handle, inputs the machining conditions for the machining to be diagnosed, acquires machining data, assigns a normal or abnormal label to the machining data, calculates the degree of abnormality of the machining to be diagnosed based on the machining data, uses the degree of abnormality and the label to determine whether or not a diagnostic response for machining anomaly diagnosis in the machining to be diagnosed is possible, if the result of the judgment on whether or not a diagnostic response is possible for the machining to be diagnosed is negative, performs additional learning using the machining data and the label to update the diagnostic model, then adds the input machining conditions to the machining condition information, if the result of the judgment on whether or not a diagnostic response is possible is positive, compares the input machining conditions with the machining condition information, adds the input machining conditions to the machining condition information only if the input machining conditions are not included in the machining condition information, and updates the diagnostic model and / or machining condition information based on the determined result of the judgment on whether or not a diagnostic response is possible and the input machining conditions. [Effects of the Invention]

[0007] According to this disclosure, the information update device and information update method for processing anomaly diagnosis determine whether diagnosis is possible using the rate of change in the degree of anomaly calculated when processing anomaly diagnosis is performed on unlearned data. If diagnosis is possible, the information on the diagnostic model's diagnosticable processing conditions is updated based on the processing conditions of the unlearned data. Therefore, the time required to determine whether processing anomaly diagnosis is possible for unlearned data can be shortened, and efficiency can be improved. [Brief explanation of the drawing]

[0008] [Figure 1] This is an explanatory diagram showing the main parts of a machine tool according to an embodiment. [Figure 2] This graph shows an example of the machining torque obtained. [Figure 3] This graph shows an example of the calculated time-domain anomaly rate change. [Figure 4]This flowchart shows the procedure for updating information for diagnosing processing anomalies when the label of processing data is abnormal. [Figure 5] This flowchart shows the procedure for updating information for diagnosing processing anomalies when the processing data label is normal. [Modes for carrying out the invention]

[0009] The embodiments of this disclosure will be described below with reference to the drawings. Figure 1 is an explanatory diagram showing the main parts of a machine tool according to an embodiment. Note that the machine tool shown in Figure 1 omits the cover and other equipment; however, in reality, it is equipped with the cover and other equipment not shown. A machine tool 1, such as a machining center, which processes a workpiece using a cutting tool, is equipped with an information update device 2 for diagnosing processing abnormalities. The information update device 2 may be provided independently or may be incorporated into a control device (not shown) of the machine tool 1. The information update device 2 includes a data acquisition means 3, a labeling means 4, a calculation means 5, a pass / fail determination means 6, an information update means 7, an information input means 8, and an information holding means 9. Furthermore, the control device and the information update device 2, whether or not they are included within the control device, are composed of a CPU, a memory connected to the CPU, and a predetermined device capable of inputting and outputting predetermined information, in order to enable various processes.

[0010] The data acquisition means 3 acquires machining data from the control device, various sensors (not shown) such as torque sensors installed at predetermined locations on the machine tool 1, and so on. The acquired machining data can also be sent to the labeling means 4, the calculation means 5, and the information update means 7. In this disclosure, machining data refers to information obtained when machining a workpiece, and includes control information of the machine tool 1 acquired from the control device and measured values ​​acquired from various sensors. The information storage means 9 stores a diagnosis model and the processing condition information that the diagnosis model can handle, that is, the processing condition information that can be diagnosed by the diagnosis model. Further, the stored diagnosis model and the processing condition information can be transmitted to the calculation means 5 and the information update means 7. In the present disclosure, the processing condition information refers to all information related to the processing of the workpiece, such as the rotational speed of the main shaft, the feed rate, etc. The diagnosis model is a mathematical model generated using a machine learning technique such as a neural network for diagnosing processing abnormalities. The diagnosis model can be updated by undergoing machine learning of processing conditions and the like.

[0011] The information input means 8 is provided so that an operator can input the processing conditions for performing the processing abnormality diagnosis. Further, the input processing conditions can be transmitted to the information update means 7. The labeling means 4 assigns a normal or abnormal label to the processing data acquired by the data acquisition means 3. Although it is assumed that the operator assigns the label based on the processing result of the actual processing, it may be any information that leads to the success or failure of the processing, such as a success or failure signal determined based on the measurement result of the tool length by a tool length measuring device or the processing result of an image obtained by imaging with a camera. Further, the labeling means 4 can transmit the assigned label to the information update means 7 and the approval / disapproval determination means 6.

[0012] The calculation means 5 calculates the degree of abnormality of the processing for which the diagnosis is performed based on the processing data acquired from the data acquisition means 3 and the diagnosis model acquired from the information storage means 9. Further, the calculation result can be transmitted to the approval / disapproval determination means 6. In the present disclosure, the degree of abnormality refers to a numerical value indicating the degree of abnormality in the diagnosed processing. When calculating the degree of abnormality of the processing for which the diagnosis is performed based on the processing data and the diagnosis model, for example, when using a binary classification calculation, the probability of belonging to the abnormal class is used as the degree of abnormality. When using a multi-class classification calculation, the total value of the probabilities of belonging to classes other than normal is used as the degree of abnormality. When using a one-class classification calculation, the degree of deviation from normal is used as the degree of abnormality. The three methods described here are merely examples of calculation methods for determining the degree of anomaly; other methods may be used to calculate the degree of anomaly as long as they allow for the calculation of a numerical value indicating the degree of anomaly.

[0013] The feasibility determination means 6 determines, based on the degree of abnormality obtained from the calculation means 5 and the label obtained from the labeling means 4, whether the diagnostic model held by the information holding means 9 can handle the diagnosis for the processing to be diagnosed. The information update means 7 updates the diagnostic model and the diagnostic processing condition information held by the information holding means 9 based on at least one of the following: processing data, labels, diagnostic model, diagnostic processing condition information, processing conditions for performing the diagnosis, and the result of the determination of whether the diagnosis can be performed. In this disclosure, tool breakage and tool wear are assumed as examples of machining abnormalities to be diagnosed, but other examples may also be diagnosed as abnormalities.

[0014] The information update device 2, configured as described above, is for diagnosing machining anomalies in machining using a machine tool 1 that processes a workpiece using a cutting tool, and comprises: an information holding means 9 that holds a learning-updateable diagnostic model and machining condition information that the diagnostic model can handle; an information input means 8 that can input the machining conditions of the machining to be diagnosed; a data acquisition means 3 that acquires machining data; a labeling means 4 that assigns a normal or abnormal label to the machining data; a calculation means 5 that calculates the degree of abnormality of the machining to be diagnosed based on the machining data; a feasibility determination means 6 that uses the degree of abnormality and the label to determine whether or not a diagnostic response for machining anomaly diagnosis in the machining to be diagnosed is possible; and an information update means 7 that updates the machining condition information held by the information holding means 9 based on the feasibility determination result of the diagnostic response determined by the feasibility determination means 6 and the machining conditions input by the information input means 8. The information update device 2 determines whether a diagnosis is possible using the rate of change in the degree of abnormality calculated when a processing abnormality diagnosis is performed on unlearned data. If it determines that a diagnosis is possible, it updates the information on the diagnostic model's diagnosticable processing conditions based on the processing conditions of the unlearned data. Therefore, the time required to determine whether a processing abnormality diagnosis is possible for unlearned data is reduced, and efficiency is improved.

[0015] The following describes the procedure for updating information for diagnosing processing abnormalities in the information update device 2. Figure 4 is a flowchart showing the procedure for updating information for diagnosing processing anomalies when the label of the processing data is abnormal. Figure 5 is a flowchart showing the procedure for updating information for diagnosing processing anomalies when the label of the processing data is normal. First, the information holding means 9 transmits the diagnostic model it holds and the processing condition information that can be diagnosed by the diagnostic model to the calculation means 5 and the information updating means 7. The calculation means 5 and the information updating means 7 read the diagnostic model and the processing condition information that can be diagnosed transmitted from the information holding means 9 (S1). Next, the operator uses the information input means 8 to input the processing conditions for the diagnosis and transmits them to the information update means 7 (S2). The information input means 8 may be part of the functions of the information update device 2 that can input and transmit the processing conditions for the diagnosis, or it may be a physical device such as a keyboard used to transmit the processing conditions for the diagnosis to the information update means 7.

[0016] Subsequently, the data acquisition means 3 acquires control information of the machine tool 1 obtained from the control device, and machining data including measured values ​​obtained from various sensors, such as machining torque as shown in Figure 2, and transmits the machining data to the labeling means 4, the calculation means 5, and the information update means 7 (S3). Figure 2 is a hypothetical graph for illustrating the embodiment. Then, the operator visually confirms the success or failure of the actual processing of the processing to be diagnosed, and then uses the labeling means 4 to assign (input) a label of normal or abnormal to the processing data. The labeling means 4 transmits the label to the pass / fail judgment means 6 and the information update means 7 (S4).

[0017] The following provides a detailed explanation of what happens when an abnormality label is assigned to processed data. In S4, an abnormality label is assigned to the processing data, and after the abnormality label is transmitted to the pass / fail determination means 6 and the information update means 7, the calculation means 5 calculates the abnormality degree E(t) of the processing to be diagnosed using the current diagnostic model obtained from the information holding means 9, based on the processing data obtained from the data acquisition means 3 (S5). The calculation means 5 transmits the calculation result of the abnormality degree E(t) to the pass / fail determination means 6. Next, the pass / fail determination means 6 calculates the rate of change dE(t) of the abnormality E(t) obtained from the calculation means 5 using equation 1. Then, the pass / fail determination means 6 determines the maximum value D of the rate of change dE(t) in a predetermined time domain. max Obtain (S6).

[0018]

number

[0019] The pass / fail determination means 6 determines the maximum value D of the rate of change in the time domain of the rate of change dE(t). max The system acquires the data and determines whether the label assigned to the processed data is normal or abnormal. If an abnormal label is assigned, the pass / fail determination means 6 determines the maximum value D of the change rate. max However, the predetermined threshold rate of change D S It is determined whether or not it exceeds (S7). When the label attached to the processed data is abnormal, the pass / fail determination means 6 determines the maximum value D of the rate of change. max However, the predetermined threshold rate of change D S If it exceeds this value, it can be diagnosed; the maximum value of the rate of change is D. max However, the predetermined threshold rate of change D SIf it does not exceed, it is determined that the diagnosis is impossible. In FIG. 3, the maximum value D of the change rate max is the change rate threshold D S When it exceeds, the waveform of the change rate dE(t) of the abnormality degree and the maximum value D of the change rate max is the change rate threshold D S When it does not exceed, an example of the waveform of the change rate dE(t) of the abnormality degree is shown as a graph. Note that FIG. 3 is a virtual graph for explaining the embodiment.

[0020] In S7, when the maximum value D of the change rate max is determined to exceed the change rate threshold D S That is, when the diagnosis determination means 6 determines that the processing data for which the diagnosis is to be performed is diagnosable, the information update means 7 compares the processing conditions for which the diagnosis is to be performed with the processing condition information read in S1, and determines whether the processing conditions for which the diagnosis is to be performed are included in the processing condition information read in S1 (S8). When the processing conditions for which the diagnosis is to be performed are included in the processing condition information read in S1, the procedure proceeds to S11, which will be described later. On the other hand, when the processing conditions for which the diagnosis is to be performed are not included in the processing condition information read in S1, the procedure proceeds to S10, which will be described later.

[0021] On the other hand, in S7, when the maximum value D of the change rate max is determined not to exceed the change rate threshold D S That is, when the diagnosis determination means 6 determines that the processing data for which the diagnosis is to be performed is not diagnosable, the information update means 7 performs additional learning of the diagnosis model read in S1 based on the processing data and the label, and updates the diagnosis model held by the information holding means 9 to an additional learned diagnosis model (S9). Thereafter, the information update means 7 compares the processing conditions for which the diagnosis is to be performed with the processing condition information read in S1, and determines whether the processing conditions for which the diagnosis is to be performed are included in the processing condition information read in S1 (S8). When the processing conditions for diagnosis are included in the processing condition information read in S1, the procedure proceeds to S11, which will be described later. On the other hand, when the processing conditions for which the diagnosis is to be performed are not included in the processing condition information read in S1, the procedure proceeds to S10.

[0022] In S10, the information update means 7 adds the processing conditions to be diagnosed to the processing condition information read in S1, and the information holding means 9 updates the processing condition information it had been holding up to the processing condition information with the processing conditions to be diagnosed added. In S11, the information update means 7 determines whether to continue the diagnosis. If the diagnosis is to continue, it returns to S1; otherwise, it terminates.

[0023] If the label assigned to the processing data is normal, the pass / fail determination means 6 determines the maximum value D of the change rate in S7. max The rate of change threshold D S If the value does not exceed the limit, it is determined that the diagnosis is possible, and if it exceeds the limit, it is determined that the diagnosis is not possible. Even if the label attached to the processed data is normal, as shown in Figure 5, the diagnosis can be handled in the same way as when the label is abnormal, by having the diagnostic determination means 6 determine whether the diagnosis is possible, proceeding to step S8 if it is possible, and proceeding to step S9 if it is not possible.

[0024] The above is an explanation of the present disclosure based on illustrated examples, and the scope of the technology is not limited thereto. For example, the data acquisition means may perform predetermined processing on the acquired processed data, such as using a filter to smooth the acquired waveform as appropriate, if available in the calculation means and information update means. Furthermore, the data acquisition means, labeling means, calculation means, feasibility determination means, information update means, information input means, and information retention means may be provided as separate components connected to the information update device, or some may be provided as functions of the information update device, while others may be provided as separate components connected to the information update device, provided that each can realize the desired function. [Explanation of Symbols]

[0025] 1. Machine tool, 2. Information update device, 3. Data acquisition means, 4. Labeling means, 5. Calculation means, 6. Feasibility determination means, 7. Information update means, 8. Information input means, 9. Information retention means.

Claims

1. An information update device for diagnosing machining abnormalities in machining using a machine tool that processes a workpiece using a cutting tool, Information holding means for holding a diagnostic model that can be updated by learning, and processing condition information that the diagnostic model can handle, An information input means that allows input of processing conditions for the processing to be performed on the diagnostic, A data acquisition method for acquiring processed data, A labeling means for assigning a label of normal or abnormal to the aforementioned processing data, A calculation means for calculating the degree of abnormality of the processing to be diagnosed based on the aforementioned processing data, A means for determining whether or not a diagnostic response for the machining abnormality diagnosis in the machining process is possible, using the result of comparing the maximum value of the rate of change of the abnormality in the machining process in which the diagnosis is performed with a preset rate of change threshold, and the label, An information update device for diagnosing processing abnormalities, comprising: an information update means that updates the diagnostic model and / or processing condition information held by the information holding means based on the result of the feasibility determination of the diagnostic response determined by the feasibility determination means and the processing conditions input by the information input means.

2. An information update device for diagnosing machining abnormalities in machining using a machine tool that processes a workpiece using a cutting tool, Information holding means for holding a diagnostic model that can be updated by learning, and processing condition information that the diagnostic model can handle, An information input means that allows input of processing conditions for the processing to be performed on the diagnostic, A data acquisition method for acquiring processed data, A labeling means for assigning a label of normal or abnormal to the aforementioned processing data, A calculation means for calculating the degree of abnormality of the processing to be diagnosed based on the aforementioned processing data, A means for determining whether or not a diagnostic response to the processing abnormality diagnosis in the processing in which the diagnosis is performed is possible, using the aforementioned degree of abnormality and the aforementioned label. The system includes an information update means that updates the diagnostic model and / or processing condition information held by the information holding means based on the result of the diagnostic response determination made by the feasibility determination means and the processing conditions entered by the information input means, The information updating means updates the diagnostic model by performing additional learning using the processing data and the label if the result of the feasibility judgment for the processing to be diagnosed is negative, and then adds the processing conditions entered by the information input means to the processing condition information held by the information holding means. If the feasibility judgment result is positive, it compares the processing conditions entered by the information input means with the processing condition information held by the information holding means, and only if the processing condition information held by the information holding means does not include the processing conditions entered by the information input means, it adds the processing conditions entered by the information input means to the processing condition information held by the information holding means. This is an information updating device for diagnosing processing abnormalities.

3. The information update device for diagnosing machining abnormalities according to claim 1, wherein the feasibility determination means, in determining whether or not to respond to the diagnosis of machining abnormality diagnosis, uses the result of comparing the maximum value of the rate of change of the degree of abnormality in the machining to be diagnosed with a preset rate of change threshold and the label, and if the label is abnormal, the diagnosis is deemed possible if the maximum value of the rate of change of the degree of abnormality exceeds the rate of change threshold, and impossible if it does not exceed the rate of change threshold; and if the label is normal, the diagnosis is deemed possible if the maximum value of the rate of change of the degree of abnormality does not exceed the rate of change threshold, and impossible if it exceeds the rate of change threshold.

4. The information update means, if the result of the judgment on whether or not a diagnostic response is possible for the processing to be diagnosed is negative, updates the diagnostic model by performing additional learning using the processing data and the label, and then adds the processing conditions entered by the information input means to the processing condition information held by the information holding means; if the result of the judgment on whether or not a diagnostic response is possible, compares the processing conditions entered by the information input means with the processing condition information held by the information holding means, and adds the processing conditions entered by the information input means to the processing condition information held by the information holding means only if the processing condition information held by the information holding means does not include the processing conditions entered by the information input means, as described in claim 1 or 3, for information update device for diagnosing processing abnormalities.

5. A method for updating information for diagnosing machining abnormalities in machining using a machine tool that processes a workpiece using a cutting tool, A diagnostic model that can be updated through learning, and processing condition information that the diagnostic model can handle are read, Enter the processing conditions for the process to be diagnosed. Obtain processing data, The aforementioned processing data is labeled as normal or abnormal. Based on the aforementioned processing data, the degree of abnormality of the processing to be diagnosed is calculated. Using the comparison result between the maximum value of the rate of change of the abnormality in the process in which the diagnosis is performed and a preset threshold for the rate of change, and the label, it is determined whether or not the process abnormality diagnosis in the process in which the diagnosis is performed can be performed. An information update method for diagnosing processing abnormalities, characterized by updating the diagnostic model and / or processing condition information based on the determined feasibility of the diagnostic response and the input processing conditions.

6. A method for updating information for diagnosing machining abnormalities in machining using a machine tool that processes a workpiece using a cutting tool, A diagnostic model that can be updated through learning, and processing condition information that the diagnostic model can handle are read, Enter the processing conditions for the process to be diagnosed. Obtain processing data, The aforementioned processing data is labeled as normal or abnormal. Based on the aforementioned processing data, the degree of abnormality of the processing to be diagnosed is calculated. Using the degree of abnormality and the label, it is determined whether or not the diagnostic response for the processing abnormality diagnosis in the processing in which the diagnosis is performed is appropriate. If the diagnostic assessment result for the processing to be diagnosed is negative, the diagnostic model is updated by performing additional learning using the processing data and the label, and then the input processing conditions are added to the processing condition information. If the assessment result is positive, the input processing conditions are compared with the processing condition information, and only if the input processing conditions are not included in the processing condition information are the input processing conditions added to the processing condition information. An information update method for diagnosing processing abnormalities, characterized by updating the diagnostic model and / or processing condition information based on the determined feasibility of the diagnostic response and the input processing conditions.

Images