Processing condition exploration device and processing condition exploration method

Abstract

Has: processing result collection part (12) that collects processing result information; evaluation value acquisition part (13) that calculates tentative evaluation value for implemented processing; convergence determination part (14) that estimates estimated convergence value in the case where tentative evaluation value does not converge; stop determination part (15) that determines whether to suspend processing before tentative evaluation value converges in the case where tentative evaluation value does not converge; evaluation decision part (16) that decides estimated convergence value as evaluation value in the case where processing is suspended, decides convergence value of tentative evaluation value as evaluation value after tentative evaluation value converges in the case where processing is not suspended; and exploration end determination part (113) that decides optimum processing condition in the case where exploration ends, generates processing condition that should be tried next until exploration ends is determined, repeats each processing by above-mentioned processing result collection part (12), evaluation value acquisition part (13), convergence determination part (14), stop determination part (15), evaluation decision part (16), exploration end determination part (113) until exploration ends is determined.

Description

Technical Field

[0001] This invention relates to a processing condition exploration apparatus and a processing condition exploration method for exploring processing conditions. Background Technology

[0002] Generally, machining centers used in industrial applications can set multiple control parameters. The machining result depends on the combination of the individual values ​​of these control parameters, i.e., the machining conditions. In other words, to obtain the desired machining result, appropriate machining conditions need to be set within the machining center.

[0003] However, there are multiple control parameters, and the values ​​of each parameter are either continuous or can be set through multiple levels. Therefore, if a person were to attempt to select the processing conditions that would actually enable the machine to perform the processing and achieve the desired result, it would require a tremendous amount of time. For example, in the case of a sheet metal laser processing machine, five key control parameters that significantly influence the processing result can be identified: laser output, cutting speed, beam diameter, focal position, and gas pressure. Each control parameter is selected from a range of values. For instance, if each of the five control parameters is selected from 10 different values, the total number of combinations becomes 10. 5 In this case, if it is set that it takes 5 minutes to test one processing condition, then in order to test 10... 5 The processing conditions require approximately 347 days.

[0004] Therefore, conventional techniques involve calculating evaluation values ​​corresponding to processing conditions based on processing results obtained by performing processing with a machine under several trial processing conditions generated from processing conditions with combinations of assumed control parameters, calculating the evaluation values ​​based on the calculated evaluation values ​​and the processing conditions corresponding to those evaluation values, predicting evaluation values ​​corresponding to processing conditions that were not trialed using Gaussian process regression, and exploring the optimal processing conditions from a large number of combinations of processing conditions based on the calculated and predicted evaluation values ​​(e.g., Patent Document 1). As a method for predicting evaluation values ​​corresponding to processing conditions that were not trialed using Gaussian process regression, for example, a method is described using a probabilistic model that assumes the evaluation values ​​for processing conditions are generated according to a probability variable with a specific distribution.

[0005] Patent Document 1: International Publication No. 2020 / 261572 Summary of the Invention

[0006] The processing results obtained when a processing machine performs processing under certain processing conditions sometimes fluctuate during the processing. For example, the processing speed obtained as a result may appear to be constant when observed over a long period of time, but it fluctuates when observed over a short period of time. If the processing result fluctuates, the evaluation value corresponding to that processing result also fluctuates.

[0007] In the exploration technology of optimal processing conditions, represented by the technology disclosed in Patent Document 1, for all the processing conditions tested, the processing machine is continuously processed for a certain period of time until the vibration change of the processing result stabilizes, and the evaluation value corresponding to the processing conditions is calculated.

[0008] Therefore, in the above-mentioned exploratory techniques, time is required to calculate the evaluation value corresponding to the trial processing conditions. As a result, there is a problem that time is required until the optimal processing conditions are found.

[0009] The present invention is proposed to solve the above-mentioned problems. Its purpose is to provide a processing condition exploration device and a processing condition exploration method that can shorten the time until the optimal processing conditions are found, compared with the prior art of performing processing under all trial processing conditions until the vibrational changes of the processing results stabilize.

[0010] The processing condition exploration apparatus of the present invention comprises: a processing condition calculation unit that generates processing conditions consisting of multiple control parameters that can be set on a processing machine; an actual processing command unit that causes the processing machine to perform processing according to the processing conditions generated by the processing condition calculation unit; a processing result collection unit that collects processing result information representing the processing performed by the processing machine caused by the actual processing command unit; an evaluation value acquisition unit that calculates a provisional evaluation value for the processed operation based on the processing result information collected by the processing result collection unit; and a convergence determination unit that, based on... The provisional evaluation value of the time series calculated by the evaluation value acquisition unit determines whether the provisional evaluation value has converged. If it is determined that the provisional evaluation value has not converged, the estimated convergence value, which is the convergence target of the provisional evaluation value, is estimated. The stop determination unit determines whether to suspend processing under the trial processing conditions before the provisional evaluation value converges if the convergence determination unit determines that the provisional evaluation value has not converged. The evaluation decision unit, if the stop determination unit determines that processing under the trial processing conditions should be suspended, causes the actual processing instruction unit to stop processing according to the processing machine. The processing conditions are as follows: processing is performed under certain conditions, and the estimated convergence value estimated by the convergence determination unit is determined as the evaluation value of the processing performed according to the processing conditions. If the stop determination unit determines that processing under the trial processing conditions should not be stopped, the convergence value of the provisional evaluation value is determined as the evaluation value after the convergence determination unit determines that the provisional evaluation value has converged; a prediction unit predicts the evaluation value corresponding to the untrialed processing conditions based on the evaluation value determined by the evaluation determination unit and the processing conditions corresponding to the evaluation value; and an exploration termination determination unit determines whether to terminate the processing conditions. The process involves determining whether to terminate the exploration. If the exploration ends, the optimal processing conditions are determined based on the evaluation value determined by the evaluation decision unit and the evaluation value predicted by the prediction unit. If the exploration does not end, the processing condition calculation unit generates the processing conditions to be tried next based on the predicted value predicted by the prediction unit. This process continues until the exploration ends, and the processing condition calculation unit, actual processing instruction unit, processing result collection unit, evaluation value acquisition unit, convergence decision unit, stop decision unit, evaluation decision unit, prediction unit, and exploration end decision unit processes are repeated.

[0011] The effects of the invention

[0012] According to the present invention, when exploring optimal processing conditions, compared with the prior art, which involves performing processing under all trial processing conditions until the vibrational changes of the processing results stabilize, the time until the optimal processing conditions can be explored can be shortened. Attached Figure Description

[0013] Figure 1This is a diagram illustrating a structural example of the processing condition exploration device involved in Embodiment 1.

[0014] Figure 2 This is a flowchart illustrating the operation of the processing condition exploration device involved in Embodiment 1.

[0015] Figure 3 This is a schematic diagram of an example of a method in Implementation 1 whereby the stop determination unit determines whether to stop processing under trial processing conditions by comparing the largest provisional evaluation value among the provisional evaluation values ​​within the interquartile range with a stop threshold.

[0016] Figure 4 This is a schematic diagram of an example of a method in Embodiment 1 whereby the stop determination unit determines whether to stop processing under trial processing conditions by comparing the provisional evaluation value contained in the range of the average value ± κσ of the provisional evaluation value with the stop threshold.

[0017] Figure 5 It is a graph that conceptually represents the relationship between the predicted value of the evaluation value in Implementation 1 and the index representing unreliability.

[0018] Figure 6A and Figure 6B This is a graph showing an example of a comparison between the time until the optimal processing conditions are discovered using existing optimal processing condition exploration techniques and the time until the optimal processing conditions are discovered using the processing condition exploration device involved in Embodiment 1.

[0019] Figure 7 This is a diagram illustrating an example of a method in Embodiment 1 where the stop determination unit sets a variable stop threshold based on the processing conditions completed during the trial and the evaluation value corresponding to those processing conditions.

[0020] Figure 8A and Figure 8B This is a diagram illustrating an example of the hardware structure of the processing condition exploration device involved in Embodiment 1. Detailed Implementation

[0021] Implementation Method 1

[0022] Figure 1 This is a diagram illustrating a structural example of the processing condition exploration device 1 involved in Embodiment 1.

[0023] The processing condition exploration device 1 according to Embodiment 1 is connected to the processing machine 2 and the display unit 3. The processing condition exploration device 1 explores the optimal processing conditions (hereinafter referred to as "optimal processing conditions") from a number of processing conditions that can be set on the processing machine 2. Optimal processing conditions are, for example, processing conditions that will yield processing results that meet the required specifications. Furthermore, the display unit 3 displays the processing conditions explored by the processing condition exploration device 1 according to requests from users such as processing operators. For example, the display unit 3 displays the processing conditions set on the processing machine 2 and the evaluation value of the processing performed by the processing machine 2 according to those processing conditions. Additionally, for example, the display unit 3 displays the predicted evaluation value of processing conditions that the processing machine 2 does not perform and the case where processing is envisioned to be performed by the processing machine 2 according to those processing conditions. Furthermore, for example, the exploration results, i.e., the optimal processing conditions, obtained through the exploration conducted by the processing condition exploration device 1 are displayed. Moreover, in... Figure 1 In this example, the display unit 3 is located outside the processing condition exploration device 1 and the processing machine 2, but this is just one example. The display unit 3 may also be located on the processing condition exploration device 1 or on the processing machine 2.

[0024] The processing machine 2 is an industrial apparatus that performs processing according to processing conditions. For example, the processing machine 2 removes unwanted parts to create a workpiece into a desired shape. Additionally, the processing machine 2 can also perform additional processing. Hereinafter, the workpiece will be referred to as the workpiece. The material of the workpiece is, for example, metal. However, this is merely an example, and the material of the workpiece is not limited to metal. The material of the workpiece can be, for example, ceramic, glass, or wood.

[0025] The processing machine 2 may be, for example, a laser processing machine, an electrical discharge machining machine, a cutting machine, a grinding machine, an electrolytic machining machine, an ultrasonic processing machine, an electron beam processing machine, or an auxiliary processing machine. In Embodiment 1 below, as an example, the processing machine 2 is set as a laser processing machine. However, this is only an example, and in Embodiment 1, the processing machine 2 may also be a processing machine other than a laser processing machine.

[0026] The processing machine 2 is capable of performing routine processing to shape the workpiece into a desired shape, and is also capable of performing experimental processing on the workpiece.

[0027] In the experimental processing, the processing condition exploration device 1 according to Embodiment 1 generates trial processing conditions, and the processing machine 2 performs experimental processing according to these processing conditions. The processing machine 2 performs the pre-set experimental processing on the workpiece according to the above processing conditions.

[0028] Here, the processing conditions are constituted by a combination of multiple control parameters used in the control of the processing machine 2. These control parameters include, for example, laser output, cutting speed, beam diameter, focal position, and gas pressure. Each control parameter included in the processing conditions can be adjusted. For example, in laser processing, there are five adjustable control parameters, and with the ability to select the value of each control parameter in 10 increments, the processing conditions constituted by the combination of these control parameters can be 10. 5 = 100,000 kinds.

[0029] The processing condition exploration device 1 generates trial processing conditions from among the numerous combinations of processing conditions described above, and causes the processing machine 2 to perform experimental processing. If the processing machine 2 performs experimental processing according to the processing conditions, the processing condition exploration device 1 collects information indicating the processing result (hereinafter referred to as "processing result information") from the processing machine 2. The processing result information includes, for example, information indicating the state of the processing machine 2 during processing, information indicating the state of the workpiece during processing, or information indicating the state of the workpiece after processing. The processing result information also includes information about the processing conditions accompanying the processing by the processing machine 2.

[0030] For example, the machining machine 2 has sensors that detect sound, light, or machining speed emitted during machining, and the machining condition exploration device 1 collects machining result information from these sensors. For example, the sensor could be an imaging device that captures an image of the machined workpiece, or a measuring device that measures the unevenness of the workpiece's cross-section. Furthermore, the sensor can be located outside the machining machine 2. The machining condition exploration device 1 only needs to be able to collect machining result information.

[0031] The processing condition exploration device 1 determines the evaluation value of the processing performed under those processing conditions based on the processing result information collected from the processing performed under those conditions. Furthermore, based on the combination of processing conditions and evaluation values, the processing condition exploration device 1 predicts the evaluation value corresponding to processing conditions that have not been tried, while simultaneously exploring the optimal processing conditions. Details regarding the method by which the processing condition exploration device 1 explores the optimal processing conditions will be described later.

[0032] Here, as described above, the processing result obtained when the processing machine 2 performs processing under certain processing conditions sometimes changes oscillatoryly during the processing. If the processing result changes oscillatoryly, the evaluation value calculated based on that processing result also changes oscillatoryly. Assuming that the processing condition exploration device 1 performs processing with the processing machine 2 under all tested processing conditions for a certain period of time until the oscillatory changes in the processing results under each processing condition stabilize, if the oscillatory changes in the processing results are allowed to stabilize, then calculating the evaluation value corresponding to each processing condition requires time.

[0033] Therefore, the evaluation value calculated by the processing condition exploration device 1 according to Embodiment 1 during the process until the change in vibration of the processing result stabilizes, even if it is an evaluation value before the change in vibration stabilizes, is assumed to have no effect when exploring the optimal processing conditions. Therefore, this evaluation value is also used when exploring the optimal processing conditions, and the processing conditions used for exploration are switched according to the processing stoppage in the trial processing conditions experiment. Thus, the processing condition exploration device 1 according to Embodiment 1 can shorten the time until the optimal processing conditions can be explored.

[0034] A detailed structural example of the processing condition exploration device 1 involved in Embodiment 1 will be described.

[0035] The processing condition exploration device 1 includes a processing condition generation unit 11, a processing result collection unit 12, an evaluation value acquisition unit 13, a convergence determination unit 14, a stop determination unit 15, an evaluation decision unit 16, and a machine learning unit 17. Additionally, the processing condition exploration device 1 includes a processing result storage unit 18A, an evaluation value storage unit 18B, a convergence result storage unit 18C, a stop determination storage unit 18D, an exploration result storage unit 18E, a prediction result storage unit 18F, and an unreliability storage unit 18G. Furthermore, all or part of the storage units 18A to 18G may be provided in an external device separately from the processing condition exploration device 1.

[0036] The exploration processing condition generation unit 11 generates processing conditions used in actual processing for experiments, enabling the processing machine 2 to perform processing according to the generated processing conditions. That is, the exploration processing condition generation unit 11 generates processing conditions for exploration through actual processing in a multi-dimensional space where the control parameters constituting the processing conditions are set as dimensions. For example... Figure 1 As shown, the exploration processing condition generation unit 11 includes a processing condition calculation unit 111, an actual processing instruction unit 112, and an exploration end determination unit 113.

[0037] The processing condition calculation unit 111 of the processing condition generation unit 11 generates processing conditions consisting of multiple control parameters that can be set on the processing machine 2. Specifically, the processing condition calculation unit 111 generates processing conditions used for experimental processing. For example, the processing condition calculation unit 111 selects a combination corresponding to the processing content from multiple control parameters of the processing machine 2 and combinations of the ranges of values ​​that these control parameters can obtain, and generates processing conditions based on the selected combination. Control parameters include, for example, laser output, cutting speed, beam diameter, focal point position, and gas pressure.

[0038] The processing condition calculation unit 111 outputs the generated processing conditions to the actual processing instruction unit 112.

[0039] The actual machining instruction unit 112 instructs the machining machine 2 to perform machining according to the machining conditions generated by the machining condition calculation unit 111. Furthermore, the actual machining instruction unit 112 continuously instructs the machining machine 2 to perform machining according to the machining conditions generated by the machining condition calculation unit 111. Specifically, the actual machining instruction unit 112 generates instructions to operate the machining machine 2 based on the machining conditions output from the machining condition calculation unit 111, and outputs the generated instructions to the machining machine 2. The machining machine 2 performs machining according to the machining conditions based on the instructions output from the actual machining instruction unit 112.

[0040] Furthermore, when the actual machining instruction unit 112 outputs an instruction from the evaluation decision unit 16 to end the machining under the trial machining conditions (hereinafter referred to as the "machining end instruction"), the experimental machining currently being performed on the machining machine 2 is ended. Details regarding the evaluation decision unit 16 will be described later.

[0041] The exploration end determination unit 113 determines whether to end the exploration of processing conditions based on the information stored in the prediction result storage unit 18F or the unreliability storage unit 18G.

[0042] If the exploration end determination unit 113 determines that no further processing conditions are required, it determines the optimal processing conditions based on the evaluation value determined by the evaluation determination unit 16. Specifically, the exploration end determination unit 113 sets the processing condition corresponding to the highest evaluation value among the evaluation values ​​stored in the exploration result storage unit 18E as the optimal processing condition. Details about the evaluation determination unit 16 will be described later.

[0043] In addition, if the exploration end determination unit 113 determines that it is necessary to conduct further exploration of processing conditions, the processing condition calculation unit 111 generates the processing conditions that should be tested next for exploration.

[0044] The processing result collection unit 12 collects processing result information from the processing machine 2, which represents the processing result performed according to the processing conditions.

[0045] The processing result collection unit 12 collects processing results each time processing is performed by the actual processing instruction unit 112. As described above, the actual processing instruction unit 112 continuously performs processing according to the processing conditions. During the processing performed by the processing machine 2, multiple processing steps are performed. Therefore, when the processing machine 2 performs experimental processing according to certain processing conditions, multiple processing result information is collected.

[0046] The processing result collection unit 12 stores the collected processing result information in the processing result storage unit 18A. The processing result collection unit 12 stores the processing result information in the processing result storage unit 18A in association with, for example, the time when the processing result information was obtained.

[0047] The processing result storage unit 18A stores the processing result information according to the time sequence.

[0048] The evaluation value acquisition unit 13 calculates an evaluation value for the processing performed by the processing machine 2 based on the processing result information collected by the processing result collection unit 12. In Embodiment 1, the evaluation value calculated by the evaluation value acquisition unit 13 based on the processing result information is also called a "provisional evaluation value". The evaluation value acquisition unit 13 calculates the provisional evaluation value for each processing result information. That is, the evaluation value acquisition unit 13 calculates the provisional evaluation value for each step of the processing. In addition, the evaluation value acquisition unit 13 obtains the processing result information collected by the processing result collection unit 12 from the processing result storage unit 18A.

[0049] In Implementation 1, the evaluation value is a value indicating whether the processing is satisfactory or not, defined as a value where a larger value indicates a better processing. The evaluation value is represented, for example, by values ​​from 0 to 1. In this case, the evaluation value is 1 for the best processing and 0 for the worst processing.

[0050] The evaluation value acquisition unit 13 stores information (hereinafter referred to as "provisional evaluation value information") that associates the acquisition time of the processing result information, the processing conditions, and the calculated provisional evaluation value in the evaluation value storage unit 18B. Furthermore, here, the provisional evaluation value information is assumed to associate the acquisition time of the processing result information with the processing conditions and the provisional evaluation value, but this is merely an example. For instance, the provisional evaluation value information could also associate the calculation time of the provisional evaluation value with the processing conditions and the provisional evaluation value itself.

[0051] The evaluation value storage unit 18B stores provisional evaluation value information according to time sequence.

[0052] The convergence determination unit 14 determines whether the provisional evaluation values ​​have converged based on the provisional evaluation values ​​of the time series calculated by the evaluation value acquisition unit 13. In Embodiment 1, "convergence" means the disappearance of the oscillating changes in the values. The convergence determination unit 14 determines whether the provisional evaluation values ​​have converged for each processing condition. Furthermore, the convergence determination unit 14 acquires the provisional evaluation values ​​of the time series calculated by the evaluation value acquisition unit 13 based on the provisional evaluation value information stored in the evaluation value storage unit 18B.

[0053] When the convergence determination unit 14 determines that the provisional evaluation value has converged, it stores information relating the acquisition time of the processing result information, the main idea of ​​the provisional evaluation value convergence, the processing conditions, the provisional evaluation value, and the convergence value of the provisional evaluation value as convergence determination information in the convergence result storage unit 18C. Alternatively, instead of the acquisition time of the processing result information, it may be associated with the calculation time of the provisional evaluation value. For example, the convergence determination unit 14 sets the latest provisional evaluation value as the convergence value of the provisional evaluation value. Furthermore, this is just one example; for instance, if information defining how the convergence value of the provisional evaluation value is calculated based on the time series provisional evaluation value is predetermined (hereinafter referred to as "convergence value calculation information"), the convergence determination unit 14 can calculate the convergence value of the provisional evaluation value according to the convergence value calculation information.

[0054] On the other hand, if the convergence determination unit 14 determines that the provisional evaluation value has not converged, it estimates the value that serves as the convergence target for the provisional evaluation value (hereinafter referred to as the "estimated convergence value"). Furthermore, the convergence determination unit 14 stores information relating the acquisition time of the processing result information, the main point that the provisional evaluation value has not converged, the processing conditions, the provisional evaluation value, and the estimated convergence value as convergence determination information in the convergence result storage unit 18C. Alternatively, the acquisition time of the processing result information can be replaced by the calculation time of the provisional evaluation value.

[0055] If the convergence determination unit 14 determines that the provisional evaluation value has not converged, the stop determination unit 15 determines whether to suspend processing under the trial processing conditions before the provisional evaluation value converges. The stop determination unit 15 determines whether to suspend processing under each processing condition. Furthermore, the stop determination unit 15 can determine whether the provisional evaluation value has not converged, as determined by the convergence determination unit 14, based on the convergence determination information stored in the convergence result storage unit 18C. The stop determination unit 15 can also directly obtain the information indicating that the provisional evaluation value has not converged from the convergence determination unit 14. Furthermore, in Figure 1 The arrow pointing from the convergence determination unit 14 to the stop determination unit 15 is omitted in the text.

[0056] The stop determination unit 15 stores information that associates the determination result of whether to stop processing under the processing conditions in the trial (hereinafter referred to as "stop determination result") with the latest convergence determination information output from the convergence determination unit 14 (hereinafter referred to as "stop determination information") in the stop determination storage unit 18D.

[0057] The stop determination storage unit 18D stores the information after the stop determination.

[0058] If the evaluation decision unit 16 determines that processing under trial processing conditions should be stopped by the stop decision unit 15, the actual processing instruction unit 112 terminates the processing according to the processing conditions for the processing machine 2, and determines the estimated convergence value estimated by the convergence decision unit 14 as the evaluation value of the processing performed according to the processing conditions. If the evaluation decision unit 16 determines that processing under trial processing conditions should not be stopped by the stop decision unit 15, after the convergence decision unit 14 determines that the provisional evaluation value has converged, it determines the convergence value of the provisional evaluation value as the evaluation value of the processing performed according to the processing conditions. Furthermore, the evaluation decision unit 16 determines the evaluation value for the processing performed according to the processing conditions for each processing condition.

[0059] The evaluation decision unit 16 only needs to determine whether the stop decision unit 15 has determined to stop processing under the processing conditions in the trial run, based on the stop decision storage unit 18D's post-stop decision information, and the convergence value of the estimated convergence value or provisional evaluation value estimated by the convergence decision unit 14. For example, the evaluation decision unit 16 can directly obtain the post-stop decision information from the stop decision unit 15. Furthermore, in Figure 1 The arrow pointing from the stop determination unit 15 to the evaluation decision unit 16 is omitted in the text.

[0060] The evaluation decision unit 16 stores the combination of processing conditions and evaluation values ​​as exploration results in the exploration result storage unit 18E.

[0061] The exploration results storage unit 18E stores the exploration results.

[0062] The machine learning unit 17 uses the exploration results stored in the exploration result storage unit 18E to predict the evaluation value of the processing corresponding to the processing conditions that were not tried (processing was not implemented). In addition, the machine learning unit 17 calculates the unreliability of the predicted value for the evaluation value, that is, the degree of deviation of the prediction.

[0063] The machine learning unit 17 includes a prediction unit 171 and an unreliability evaluation unit 172.

[0064] The prediction unit 171 predicts the evaluation value corresponding to the processing conditions that have not been tested, based on the evaluation value determined by the evaluation decision unit 16 and the processing conditions corresponding to the evaluation value. The prediction unit 171 can obtain the evaluation value determined by the evaluation decision unit 16 and the processing conditions corresponding to the evaluation value from the exploration results stored in the exploration result storage unit 18E.

[0065] The prediction unit 171 stores information (hereinafter referred to as "prediction result information") that associates the predicted value of the evaluation value obtained through prediction with the processing conditions in the prediction result storage unit 18F. The prediction result information is information that associates the processing conditions that have not been tried with the predicted values ​​of the corresponding evaluation values.

[0066] The prediction result storage unit 18F stores the prediction result information.

[0067] The unreliability evaluation unit 172 calculates an index representing the unreliability of the prediction of the evaluation value made by the prediction unit 171. The unreliability evaluation unit 172 uses the exploration results stored in the exploration result storage unit 18E to calculate the unreliability of the predicted value for the evaluation value, that is, an index representing the difficulty of deviation from the prediction. The unreliability evaluation unit 172 stores information (hereinafter referred to as "unreliability information") that associates the calculated index value with the processing conditions in the unreliability storage unit 18G. The unreliability information is information that associates untested processing conditions with the index value representing the unreliability of the predicted evaluation value corresponding to them.

[0068] The unreliability storage unit 18G stores unreliability information.

[0069] Next, the operation of the processing condition exploration device 1 involved in Embodiment 1 will be explained.

[0070] Figure 2 This is a flowchart illustrating the operation of the processing condition exploration device 1 involved in Embodiment 1.

[0071] If the processing condition exploration process begins, firstly, the processing condition calculation unit 111 of the processing condition generation unit 11 generates initial processing conditions (step ST1). The processing condition calculation unit 111 selects a predetermined number of processing conditions from all combinations that can be set as processing conditions to serve as initial processing conditions, thereby generating the initial processing conditions. Methods for selecting initial processing conditions implemented by the processing condition calculation unit 111 include, for example, experimental planning, optimal planning, or random sampling. Furthermore, if the user finds what they consider to be the optimal processing conditions based on past usage performance, the processing condition calculation unit 111 can use the processing conditions input from the user as initial processing conditions. Moreover, these methods are merely examples; the processing condition calculation unit 111 can generate initial processing conditions using any method.

[0072] For example, if there are 5 control parameters constituting the processing conditions, and if the value set on the processing machine 2 is selected from 10 levels for each control parameter, then the total number of combinations of processing conditions is 10. 5 = 100,000 types. The processing condition calculation unit 111 selects, for example, 10 processing conditions from this combination as initial processing conditions. Furthermore, the number of control parameters constituting the processing conditions, the number of settings that can be set for each control parameter, or the number of processing conditions selected as initial processing conditions are not limited to these. The number of settings that can be set can vary depending on the control parameters.

[0073] Next, the processing condition exploration device 1 selects one initial processing condition from the initial processing conditions generated by the processing condition calculation unit 111, and causes the machining machine 2 to perform processing based on the selected initial processing condition (step ST2). Specifically, the processing condition calculation unit 111 selects one initial processing condition from the initial processing conditions and outputs the selected initial processing condition to the actual processing command unit 112 of the exploration processing condition generation unit 11. The actual processing command unit 112 generates a command to operate the machining machine 2 based on the initial processing condition output from the processing condition calculation unit 111, and outputs the generated command to the machining machine 2. Thus, the machining machine 2 performs processing based on the initial processing conditions selected by the processing condition calculation unit 111. As described above, the processing condition exploration device 1 according to Embodiment 1 first causes the machining machine 2 to perform processing based on the initial processing conditions. Hereinafter, processing based on the initial processing conditions will also be referred to as "initial processing".

[0074] The processing result collection unit 12 collects processing result information from the processing machine 2, which represents the processing result of the initial processing performed according to the initial processing conditions (step ST3).

[0075] The processing result collection unit 12 stores the collected processing result information in the processing result storage unit 18A.

[0076] The evaluation value acquisition unit 13 calculates the provisional evaluation value for the processing performed by the processing machine 2 according to the initial processing conditions in step ST2 based on the processing result information collected by the processing result collection unit 12 (step ST4).

[0077] The evaluation value acquisition unit 13 stores the provisional evaluation value information associated with the acquisition time of the processing result information, the processing conditions (here, the initial processing conditions), and the calculated provisional evaluation value in the evaluation value storage unit 18B.

[0078] The convergence determination unit 14 determines whether the provisional evaluation value has converged based on the provisional evaluation value of the time series calculated by the evaluation value acquisition unit 13. If the convergence determination unit 14 determines that the provisional evaluation value has converged, it stores convergence determination information in the convergence result storage unit 18C, which associates the acquisition time of the processing result information, the information that the provisional evaluation value has converged, the processing conditions (here, the initial processing conditions), the provisional evaluation value, and the convergence value of the provisional evaluation value. On the other hand, if the convergence determination unit 14 determines that the provisional evaluation value has not converged, it estimates the estimated convergence value and stores convergence determination information in the convergence result storage unit 18C, which associates the acquisition time of the processing result information, the information that the provisional evaluation value has not converged, the processing conditions (here, the initial processing conditions), the provisional evaluation value, and the estimated convergence value (step ST5).

[0079] Here, specific examples will be given to explain the method for determining whether the provisional evaluation value based on the provisional evaluation value of the time series by the convergence determination unit 14 in step ST5 has converged, and the method for estimating the estimated convergence value when it is determined that the provisional evaluation value has not converged.

[0080] The convergence determination unit 14, for example, determines whether the provisional evaluation value has converged and determines the estimated convergence value based on the degree of fluctuation of the provisional evaluation value of the time series.

[0081] For example, the convergence determination unit 14 calculates the interquartile range of the provisional evaluation values ​​based on the provisional evaluation values ​​of the time series. Furthermore, the convergence determination unit 14 determines whether the provisional evaluation values ​​have converged based on the range of values ​​within which the interquartile range of the provisional evaluation values ​​falls. For example, it pre-determines the range of values ​​for which the provisional evaluation values ​​are considered convergent (hereinafter referred to as the "first convergence determination range"). If the interquartile range of the provisional evaluation values ​​falls within the first convergence determination range, the convergence determination unit 14 determines that the provisional evaluation values ​​have converged. If the interquartile range of the provisional evaluation values ​​does not fall within the first convergence determination range, the convergence determination unit 14 determines that the provisional evaluation values ​​have not converged.

[0082] If the convergence determination unit 14 determines that the provisional evaluation value has not converged, it then estimates the estimated convergence value based on the interquartile range of the provisional evaluation value obtained from the time series. For example, the convergence determination unit 14 estimates the median value of the interquartile range of the provisional evaluation value as the estimated convergence value.

[0083] For example, the convergence determination unit 14 can treat the provisional evaluation values ​​of the time series as a specific distribution and estimate that distribution. Based on the value of the interval between the mean ± κσ of the provisional evaluation values ​​within the distribution of the provisional evaluation values, it determines whether the provisional evaluation values ​​have converged. For instance, it can pre-determine the range of values ​​for which the provisional evaluation values ​​are considered convergent (hereinafter referred to as the "second convergence determination range"). If the value of the interval between the mean ± κσ of the provisional evaluation values ​​within the distribution of the provisional evaluation values ​​falls within the second convergence determination range, then the convergence determination unit 14 determines that the provisional evaluation values ​​have converged. If the value of the interval between the mean ± κσ of the provisional evaluation values ​​within the distribution of the provisional evaluation values ​​does not fall within the second convergence determination range, then the convergence determination unit 14 determines that the provisional evaluation values ​​have not converged.

[0084] If the convergence determination unit 14 determines that the provisional evaluation value has not converged, it then estimates the estimated convergence value based on the distribution estimated from the provisional evaluation values ​​in the time series. For example, the convergence determination unit 14 estimates the average value of the provisional evaluation values ​​as the estimated convergence value.

[0085] Alternatively, the convergence determination unit 14 can, for example, take the evaluation value of the time series as input and output the estimated convergence value, and estimate the estimated convergence value based on the trained model (hereinafter referred to as the "first machine learning model"). The convergence determination unit 14 inputs the provisional evaluation value of the time series into the first machine learning model to obtain the estimated convergence value.

[0086] Alternatively, for example, the first machine learning model may be a model that outputs information related to the degree of fluctuation of the provisional evaluation value in addition to the estimated convergence value. The convergence determination unit 14 may determine whether the provisional evaluation value has converged based on information related to the degree of fluctuation of the provisional evaluation value obtained by inputting the provisional evaluation value of the time series into the first machine learning model.

[0087] If the convergence determination unit 14 determines that the provisional evaluation value has not converged, the stop determination unit 15 determines whether to stop the processing under the initial processing conditions in the trial before the provisional evaluation value converges (step ST6).

[0088] Here, a specific example will be given to explain the method for determining whether to suspend processing under trial processing conditions before the provisional evaluation value converges, which is performed by the stop determination unit 15.

[0089] The stop determination unit 15 determines, for example, whether to stop processing under trial processing conditions before the provisional evaluation value converges by comparing the fluctuation degree of the provisional evaluation value of the time series stored in the evaluation value storage unit 18B, which is calculated by the evaluation value acquisition unit 13, with a threshold (hereinafter referred to as the "stop threshold").

[0090] The stop threshold is, for example, pre-specified by the user and stored in the stop determination unit 15. For example, the user pre-specifies the evaluation value (hereinafter referred to as the "benchmark evaluation value") that serves as the stop reference for processing under the processing conditions that prevent the processing trial from being stopped if the value is not exceeded as the stop threshold. The user sets the benchmark evaluation value, for example, in accordance with the required performance of the processing machine 2.

[0091] For example, if the convergence determination unit 14 calculates the interquartile range of provisional evaluation values ​​based on the provisional evaluation values ​​of the time series, the stop determination unit 15 determines whether to stop processing under the trial processing conditions by comparing the largest provisional evaluation value among the provisional evaluation values ​​within the interquartile range with a stop threshold. In this case, if the largest provisional evaluation value among the provisional evaluation values ​​within the interquartile range is less than the stop threshold, the stop determination unit 15 determines to stop processing under the trial processing conditions. On the other hand, if the largest provisional evaluation value among the provisional evaluation values ​​within the interquartile range is greater than or equal to the stop threshold, the stop determination unit 15 determines to continue processing under the trial processing conditions.

[0092] Figure 3 This is a schematic diagram of an example of a method in Embodiment 1 where the stop determination unit 15 determines whether to stop processing under trial processing conditions by comparing the largest provisional evaluation value among the provisional evaluation values ​​within the interquartile range with a stop threshold.

[0093] Figure 3 The horizontal axis indicates the time span during which processing was performed according to certain processing conditions. Figure 3 The vertical axis shows the evaluation value (provisional evaluation value). Figure 3 The black dots indicate provisional evaluation values ​​calculated based on the processing results performed according to the processing conditions. Furthermore, in Figure 3 To facilitate understanding, a diagram is shown illustrating the convergence of the provisional evaluation values. Figure 3 In the table, 201a, 201b and 201c show the interquartile range of the provisional evaluation values.

[0094] At time t1, the interquartile range of the provisional evaluation value is the range shown in 201a; at time t2, the interquartile range of the provisional evaluation value is the range shown in 201b. Regarding the interquartile ranges shown in 201a and 201b, the largest provisional evaluation value within each interquartile range is greater than or equal to the termination threshold. Therefore, in this case, the stop determination unit 15 determines that processing under the trial processing conditions should continue.

[0095] If time t3 has elapsed, the interquartile range of the provisional evaluation value becomes the range shown in 201c, and the largest provisional evaluation value within the interquartile range is less than the termination threshold. In this case, the termination determination unit 15 determines to terminate the processing under the trial processing conditions.

[0096] To give another specific example, if the convergence determination unit 14 estimates the distribution of provisional evaluation values ​​of the time series, the stop determination unit 15 can determine whether to stop processing under the trial processing conditions by comparing the provisional evaluation values ​​contained in the interval of the average value ± κσ of the provisional evaluation values ​​with the stop threshold. In this case, if all provisional evaluation values ​​contained in the interval of the average value ± κσ of the provisional evaluation values ​​are less than the stop threshold, the stop determination unit 15 determines to stop processing under the trial processing conditions. On the other hand, if all provisional evaluation values ​​contained in the interval of the average value ± κσ of the provisional evaluation values ​​are not less than the stop threshold, the stop determination unit 15 determines to continue processing under the trial processing conditions.

[0097] Figure 4 This is a schematic diagram of an example of a method in Embodiment 1 where the stop determination unit 15 determines whether to stop processing under trial processing conditions by comparing the provisional evaluation value contained in the range of the average value ± κσ of the provisional evaluation value with the stop threshold.

[0098] Figure 4 The horizontal axis indicates the time span during which processing was performed according to certain processing conditions. Figure 4 The vertical axis shows the evaluation value (provisional evaluation value). Figure 4 The black dots indicate provisional evaluation values ​​calculated based on the processing results performed according to the processing conditions. Furthermore, in Figure 4 To facilitate understanding, a diagram is shown illustrating the convergence of the provisional evaluation values. Figure 4 In the figures 301a, 301b and 301c, the largest provisional evaluation value is shown among the provisional evaluation values ​​contained in the interval of the mean ± κσ of the provisional evaluation values.

[0099] At time t4, the largest provisional evaluation value among the provisional evaluation values ​​contained in the interval of the provisional evaluation value's average ± κσ is the value shown in 301a. At time t5, the largest provisional evaluation value among the provisional evaluation values ​​contained in the interval of the provisional evaluation value's average ± κσ is the value shown in 301b. Both the values ​​shown in 301a and 301b are greater than or equal to the termination threshold. That is, all provisional evaluation values ​​contained in the interval of the provisional evaluation value's average ± κσ containing the value shown in 301a are not less than the termination threshold. In addition, all provisional evaluation values ​​contained in the interval of the provisional evaluation value's average ± κσ containing the value shown in 301b are not less than the termination threshold. Therefore, in this case, the stop determination unit 15 determines that processing under the trial processing conditions should continue.

[0100] If time t6 has elapsed, the largest provisional evaluation value among the provisional evaluation values ​​contained in the interval of the average value ± κσ of the provisional evaluation values ​​becomes the value shown in 301c. The value shown in 301c is less than the termination threshold. That is, all provisional evaluation values ​​within the interval of the average value ± κσ of the provisional evaluation values ​​containing the value shown in 301c are less than the termination threshold. In this case, the stop determination unit 15 determines to terminate the processing under the trial processing conditions.

[0101] Furthermore, for example, the stop determination unit 15 can also determine whether to stop processing under trial processing conditions before the provisional evaluation value converges, based on a trained model (hereinafter referred to as the "second machine learning model") that outputs information indicating whether processing should be stopped, using the evaluation value of the time series as input. The stop determination unit 15 inputs the provisional evaluation value of the time series calculated by the evaluation value acquisition unit 13 into the second machine learning model to obtain information indicating whether processing should be stopped. In addition, the stop determination unit 15 can, for example, obtain the provisional evaluation value of the time series calculated by the evaluation value acquisition unit 13 from the convergence determination information stored in the convergence result storage unit 18C.

[0102] Even if the provisional evaluation value has not converged, when observing the fluctuation of the provisional evaluation value in the time series, if the provisional evaluation value falls within a low range, then even if the processing machine 2 continues processing under this condition, a high provisional evaluation value cannot be obtained; in other words, a low evaluation value is expected. Therefore, the stop determination unit 15, for example, determines, by the method described above, that if the provisional evaluation value falls within a low range based on the fluctuation of the provisional evaluation value in the time series, processing under the trial processing conditions will be stopped even before the provisional evaluation value converges.

[0103] The stop determination unit 15 stores the stop determination post-stop information, which associates the stop determination result with the latest convergence determination post-information output from the convergence determination unit 14, in the stop determination storage unit 18D.

[0104] If the evaluation decision unit 16 determines that the processing under the initial processing conditions during the trial run will be stopped before the provisional evaluation value converges (in the case of "YES" in step ST6), the actual processing instruction unit 112 will end the processing according to the initial processing conditions for the processing machine 2. Specifically, the evaluation decision unit 16 outputs a processing end instruction to the actual processing instruction unit 112. If a processing end instruction is output from the evaluation decision unit 16, the actual processing instruction unit 112 will end the processing currently performed by the processing machine 2 according to the initial processing conditions generated in step ST1. In addition, the evaluation decision unit 16 determines the estimated convergence value estimated by the convergence decision unit 14 as the evaluation value of the processing performed according to the initial processing conditions. Moreover, the evaluation decision unit 16 stores the combination of processing conditions and evaluation value as an exploration result in the exploration result storage unit 18E (step ST8). In detail, the evaluation decision unit 16 stores the combination of initial processing conditions and evaluation value, which is the estimated convergence value in this case, as an exploration result in the exploration result storage unit 18E.

[0105] If the stop determination unit 15 determines that processing under the initial processing conditions during the trial should not be stopped (in the case of "NO" in step ST6), the convergence determination unit 14 determines whether the provisional evaluation value has converged (step ST7). If the convergence determination unit 14 determines that the provisional evaluation value has not converged (in the case of "NO" in step ST7), the operation of the processing condition exploration device 1 returns to the processing in step ST2. If the convergence determination unit 14 determines that the provisional evaluation value has converged (in the case of "YES" in step ST7), the evaluation determination unit 16 determines the convergence value of the provisional evaluation value as the evaluation value. Furthermore, the evaluation determination unit 16 stores the combination of processing conditions and evaluation value as the exploration result in the exploration result storage unit 18E (step ST8). In detail, the evaluation determination unit 16 stores the combination of the initial processing conditions and the convergence value of the evaluation value, which is here the provisional evaluation value, as the exploration result in the exploration result storage unit 18E.

[0106] The processing condition calculation unit 111 checks whether the initial processing has ended based on all the processing conditions selected as initial processing conditions (step ST9).

[0107] In the case of initial processing conditions where the initial processing has not ended (in the case of "NO" in step ST9), the processing from step ST1 to step ST8 is performed sequentially for the initial processing conditions where the initial processing has not ended. In the second and subsequent steps ST1, the processing condition calculation unit 111 selects initial processing conditions that have not been selected in the previous steps ST1. As a result, the exploration results that associate all initial processing conditions (e.g., 10 initial processing conditions) with the combination of evaluation values ​​are stored in the exploration result storage unit 18E.

[0108] If, for example, initial processing under 10 initial processing conditions has ended, the prediction unit 171 of the machine learning unit 17 uses the exploration results (processing conditions and their corresponding evaluation values) stored in the exploration result storage unit 18E. In other words, based on the evaluation value determined by the evaluation decision unit 16 and the processing condition corresponding to that evaluation value, it predicts the evaluation value corresponding to the untried processing conditions (step ST10). Regarding the processing conditions that have been processed and completed, the evaluation value is determined through the aforementioned step ST8. On the other hand, the processing conditions that have been processed are part of a combination of all processing conditions. For example, if there are 100,000 possible combinations of processing conditions, and 10 initial processing conditions have been generated, there are 99,990 untried processing conditions after the initial processing has ended. Therefore, in this case, the prediction unit 171 calculates the predicted values ​​for the 99,990 evaluation values. Furthermore, as described later, steps ST15 to ST22 also involve selecting processing conditions, implementing processing, collecting processing results, calculating a provisional evaluation value, predicting the convergence value of the provisional evaluation value, determining whether to stop processing before the provisional evaluation value converges, and deciding the evaluation value. Step ST10 is implemented after the processing in step ST22. When step ST10 is implemented via steps ST15 to ST22, processing conditions set in step ST15 that have never been tested are excluded.

[0109] As an example of the method by which the prediction unit 171 calculates the predicted value of the evaluation value corresponding to the untested processing conditions, that is, the method of predicting the evaluation value corresponding to the untested processing conditions, the method of using Gaussian process regression is given. When the prediction unit 171 uses Gaussian process regression to predict the evaluation value corresponding to the untested processing conditions, the calculation is performed as follows. The method of using Gaussian process regression is an example of a method that uses a probabilistic model of the processing conditions for which the evaluation value is generated according to a probability variable of a specific distribution. If the number of observations, that is, the number of processing conditions for which the evaluation value is calculated, is set to N, and the Gram matrix is ​​set to C... NThe values ​​of this control parameter under each processing condition stored in the exploration result storage unit 18E will be set to x1 to x2. N For processing conditions x that have not been tested N＋1 The predicted value m(x) of the evaluation value N＋1 K can be calculated using the following equation (1). K, as shown in the following equation (2), is the processing conditions x1, ..., x1 that have been explored. N Each and x N＋1 This is a vector formed by arranging the values ​​of the kernel function when set as the independent variable. Furthermore, the superscript T indicates transpose, and the superscript -1 indicates the inverse matrix.

[0110] m(x) N＋1 )=k T • (C) N －1 )・t・・・(1)

[0111]

[0112] Furthermore, while an example of using Gaussian process regression for prediction in prediction unit 171 has been described here, the method for predicting evaluation values ​​used by prediction unit 171 is not limited to this. For example, prediction unit 171 may use teacher-learning methods such as decision trees, linear regression, boosting methods, or neural networks to predict evaluation values.

[0113] If the prediction unit 171 predicts an evaluation value corresponding to the processing conditions that have not been tested, it stores the predicted value of the evaluation value (step ST11). In detail, the prediction unit 171 stores the prediction result information that associates the predicted value of the evaluation value predicted in step ST10 with the processing conditions in the prediction result storage unit 18F.

[0114] Furthermore, the unreliability evaluation unit 172 of the machine learning unit 17 uses the exploration results stored in the exploration result storage unit 18E to calculate an index representing the unreliability of the prediction of the evaluation value corresponding to the untried processing conditions (step ST12). As an example of an unreliability index, the standard deviation calculated using Gaussian process regression, which is an example of a probability model, is given. When the unreliability evaluation unit 172 outputs an unreliability index using Gaussian process regression, the calculation is performed as follows, for example. The number of observations, i.e., the number of processing conditions from which the evaluation value is calculated, is set to N, and the Gram matrix is ​​set to C. N The vector of the processing conditions stored in the exploration results storage unit 18E will be set as k, and the untried processing conditions x will be... N＋1The scalar value obtained by adding the kernel values ​​of each other to the accuracy parameters of the prediction model is denoted as c. Now, if one of the control parameters constituting the processing conditions is set as x... i (where i is a natural number), the value of this control parameter in each processing condition stored in the exploration result storage unit 18E will be set to x1 to x2. N This indicates that the processing condition x has not been tested. N＋1 The unreliability index corresponding to the predicted evaluation value is the standard deviation σ(x). N＋1 The variance σ can be calculated using the following equation (3). Furthermore, in equation (3), the variance σ is calculated. 2 (x) N＋1 However, by calculating the square root of the variance, the standard deviation σ(x) can be obtained. N＋1 ).

[0115] σ 2 (x) N＋1 ) = c - k T • (C) N －1 )・k・・・(3)

[0116] Furthermore, while an example of using Gaussian process regression to calculate an index representing the unreliability of a prediction has been described here, the method for calculating the unreliability index is not limited to this. For example, the unreliability evaluation unit 172 may also use methods such as density estimation or mixed density networks to calculate the aforementioned index.

[0117] Here, the predicted value of the evaluation value in Implementation 1 and the unreliability of the predicted value are explained.

[0118] Figure 5 It is a graph that conceptually represents the relationship between the predicted value of the evaluation value in Implementation 1 and the index representing unreliability.

[0119] exist Figure 5 The example shown illustrates the calculation of predicted values ​​and indices representing unreliability using Gaussian process regression. Figure 5 The horizontal axis represents the value x, which is a control parameter for machining conditions. Figure 5 The vertical axis represents the evaluation value. Figure 5The black dots represent the evaluation values ​​calculated based on actual processing using the initial processing conditions (hereinafter also referred to as the evaluation values ​​of the actual processing). In predictions using Gaussian process regression, the evaluation values ​​are predicted according to a Gaussian distribution. Therefore, if the predicted evaluation value is set as the mean m(x) of a Gaussian distribution, and the index representing the unreliability of the prediction is set as the standard deviation σ(x) of the Gaussian distribution, then the actual evaluation value is statistically shown to fall within the range greater than or equal to m(x) - 2σ(x) and less than or equal to m(x) + 2σ(x) with approximately 95% probability. Figure 5 In the diagram, the solid line represents the predicted value of the evaluation, m(x). Additionally, in... Figure 5 In the diagram, the dashed lines represent the curves for m(x) - 2σ(x) and m(x) + 2σ(x).

[0120] like Figure 5 As shown, the unreliability index decreases at locations close to the actual processing evaluation value, and increases at locations far from the actual processing evaluation value.

[0121] Return to Figure 2 The flowchart illustrates the operation of the processing condition exploration device 1.

[0122] The unreliability evaluation unit 172 stores the indicators representing the unreliability of the predicted values ​​(step ST13). In detail, the unreliability evaluation unit 172 stores unreliability information that associates the calculated values ​​of the indicators with the processing conditions in the unreliability storage unit 18G.

[0123] The exploration end determination unit 113 of the exploration processing condition generation unit 11 uses the predicted values ​​of the evaluation values ​​of the processing conditions stored in the prediction result storage unit 18F and the unreliability index representing the predicted values ​​of the evaluation values ​​stored in the unreliability storage unit 18G to determine whether to end the exploration of the processing conditions (step ST14). For example, the exploration end determination unit 113 compares the value of the unreliability index representing the predicted values ​​of all processing conditions explored so far, stored in the unreliability storage unit 18G, with a threshold. If the value of the index is less than or equal to the threshold, it determines that the optimal processing conditions have been explored and ends the exploration of the processing conditions.

[0124] For example, the exploration completion determination unit 113 uses the processing condition x, the predicted value m(x) of the evaluation value for the processing condition x, and the index (standard deviation) σ(x) representing the unreliability of the prediction of the evaluation value. It can then determine that the larger the value of m(x) + κσ(x) becomes, the higher the value of exploring the processing condition. Furthermore, κ is a parameter determined before exploring the processing condition. The smaller the value of κ, the higher the predicted value of the evaluation value is selected; the larger the value of κ, the higher the possibility of selecting processing conditions that cause the prediction of the evaluation value to deviate significantly. The value of κ can be kept constant or changed midway.

[0125] If the exploration is determined to be a condition for ending processing (in the case of "YES" in step ST14), the exploration end determination unit 113 determines the processing condition associated with the highest evaluation value among all processing conditions stored in the exploration result storage unit 18E as the optimal processing condition. The exploration end determination unit 113, for example, extracts the optimal processing condition and outputs the extracted processing condition to the actual processing instruction unit 112. The actual processing instruction unit 112 outputs an instruction containing the processing condition output from the exploration end determination unit 113 to the machining machine 2, setting the processing condition to the machining machine 2. Thus, the actual processing instruction unit 112 causes the machining machine 2 to perform normal processing according to the set processing condition. Furthermore, this is merely one example; for instance, the exploration end determination unit 113 may also store the determined optimal processing condition in a storage unit not shown.

[0126] If it is determined that the exploration of processing conditions should not be terminated, in other words, if it is determined that the exploration of processing conditions needs to be further carried out (in the case of "NO" in step ST14), the exploration termination determination unit 113 instructs the processing condition calculation unit 111 to generate the processing conditions that should be tried next.

[0127] When instructed by the exploration end determination unit 113 to generate processing conditions to be tested next, the processing condition calculation unit 111 uses the predicted values ​​of the evaluation values ​​of the processing conditions stored in the prediction result storage unit 18F to generate the processing conditions to be tested next (step ST15). Specifically, the processing condition calculation unit 111 selects the processing conditions to be tested next, i.e., the new processing conditions, from all processing conditions. The processing conditions to be tested next generated by the processing condition calculation unit 111 are output to the actual processing instruction unit 112.

[0128] The actual machining instruction unit 112 outputs an instruction to the machining machine 2 containing the machining conditions to be tested next, generated by the machining condition calculation unit 111 in step ST15, and causes the machining machine 2 to perform machining under these machining conditions (step ST16). During machining performed by the machining machine 2, the machining result collection unit 12 collects machining result information (step ST17). The evaluation value acquisition unit 13 calculates a provisional evaluation value for the machining performed in step ST16 (step ST18). The convergence determination unit 14 determines whether the provisional evaluation value has converged and estimates the estimated convergence value based on the fluctuation of the provisional evaluation value in the time series (step ST19). The stop determination unit 15 determines whether to stop machining under the tested machining conditions (step ST20). If the stop determination unit 15 determines that processing is under the condition of suspending the trial processing, the evaluation decision unit 16 determines the estimated convergence value as the evaluation value. If the stop determination unit 15 determines that processing is under the condition of not suspending the trial processing, after the convergence determination unit 14 determines that the provisional evaluation value has converged (step ST21), the convergence value of the provisional evaluation value is determined as the evaluation value. Furthermore, the evaluation decision unit 16 stores the exploration results (step ST22). Next, it jumps to the processing of steps ST10 and ST12 and performs the aforementioned processing.

[0129] Display unit 3 displays information obtained during the above-described processing, the optimal processing conditions obtained as a result of the processing, etc. For example, display unit 3 displays the processing conditions obtained during the exploration of processing conditions by processing condition exploration device 1 and the evaluation value corresponding to those processing conditions. Additionally, display unit 3 displays the processing conditions and the predicted value of the evaluation value corresponding to those processing conditions. Furthermore, display unit 3 displays the optimal processing conditions of the exploration results. That is, display unit 3 displays at least one of the processing conditions read from exploration result storage unit 18E and the evaluation value corresponding to those processing conditions, the processing conditions read from prediction result storage unit 18F and the predicted value of the evaluation value corresponding to those processing conditions, or the optimal processing conditions of the exploration results output from processing condition calculation unit 111. Thus, by referring to the information displayed on display unit 3, the user can identify the exploration status and results of the processing conditions.

[0130] As described above, the processing condition exploration device 1 calculates a provisional evaluation value for the implemented processing based on processing result information collected by the processing machine 2 performing processing according to the generated processing conditions. Based on the calculated provisional evaluation value of the time series, the processing condition exploration device 1 determines whether the provisional evaluation value has converged. If it determines that the provisional evaluation value has not converged, it determines whether to suspend processing under the trial processing conditions before the provisional evaluation value converges. For example, the processing condition exploration device 1 compares the fluctuation of the provisional evaluation value of the time series (e.g., the interquartile range or distribution of the provisional evaluation value) with a termination threshold. If it determines that even if processing continues, a high evaluation value cannot be obtained—in other words, the obtained evaluation value is low—it determines to suspend processing under the trial processing conditions before the provisional evaluation value converges. In the case of a low evaluation value, this evaluation value is assumed to be one that has no impact on exploring the optimal processing conditions. If the processing condition exploration device 1 terminates processing under trial conditions before the provisional evaluation value converges, it sets the estimated convergence value as the evaluation value corresponding to the trial processing conditions. If the processing condition exploration device 1 predicts the evaluation value, it determines whether to end the exploration of processing conditions. If the exploration is terminated, the optimal processing conditions are determined based on the determined evaluation value. If the exploration continues, the next trial processing conditions are generated. The processing condition exploration device 1 repeats the above process until it is determined that the exploration of processing conditions should be terminated. Thus, the processing condition exploration device 1 determines the optimal processing conditions.

[0131] In existing techniques for exploring optimal processing conditions, for all tested processing conditions, the processing machine 2 is used for processing for a certain period of time until the vibrational changes of the processing result stabilize, and the evaluation value corresponding to the processing condition is calculated after waiting for the vibrational changes of the processing result to stabilize. Therefore, existing techniques for exploring optimal processing conditions have poor time efficiency until the optimal processing conditions can be explored.

[0132] In contrast, the processing condition exploration device 1 according to Embodiment 1, as described above, when calculating the evaluation value, if it is determined that a high evaluation value cannot be obtained even if processing continues, processing under the trial processing conditions is stopped before the evaluation value (provisional evaluation value) converges, and the estimated convergence value is set as the evaluation value corresponding to the trial processing conditions. Therefore, the processing condition exploration device 1 can omit the time from the time of suspension to the time until the processing result converges, specifically, the time required to explore the optimal processing conditions, for processing under a certain processing condition determined not to yield a high evaluation value. That is, the processing condition exploration device 1 can shorten the entire time required until the optimal processing conditions are explored using the aforementioned omitted time.

[0133] Figure 6A and Figure 6B This is a graph showing an example of the result obtained by comparing the time until the optimal processing conditions are discovered using existing optimal processing condition exploration techniques with the time until the optimal processing conditions are discovered using the processing condition exploration device 1 involved in Embodiment 1.

[0134] Figure 6A It is a graph representing the evaluation values ​​up to the point where the optimal processing conditions are discovered, using existing exploration techniques for optimal processing conditions. Figure 6B It is a graph representing the evaluation values ​​up to the point that the optimal processing conditions are discovered by the processing condition exploration device 1 involved in Embodiment 1.

[0135] exist Figure 6A and Figure 6B In the diagram, the black dots represent evaluation values ​​calculated based on the actual processing results performed until the processing results converge. Figure 6B In the diagram, the white circle indicates the estimated convergence value calculated based on the actual processing results that were stopped before the processing results converged.

[0136] also, Figure 6A and Figure 6B The results show that the optimal processing conditions were explored for the same processing machine 2, which will yield the same desired processing results.

[0137] Among the existing technologies for exploring optimal processing conditions, such as Figure 6A As shown, regardless of whether the evaluation value is good or bad, processing continues until the processing result, or in other words, the evaluation value, converges. Therefore, it takes time to explore the optimal processing conditions. Figure 6A In the example shown, it took 21 minutes to find the optimal processing conditions.

[0138] In contrast, in the processing condition exploration device 1 according to Embodiment 1, such as Figure 6B As shown, processing is stopped when the anticipated processing result, or in other words, the evaluation value, is low, thus allowing for the exploration of optimal processing conditions in a short time. Figure 6B In the example shown, the optimal processing conditions were discovered in 14 minutes. The time required until the optimal processing conditions were discovered in the processing condition exploration device 1 according to Embodiment 1 is compared with the time taken to... Figure 6A The existing technology for exploring optimal processing conditions reduces the time required until the optimal processing conditions are found by 7 minutes.

[0139] Furthermore, in Embodiment 1 described above, in the processing condition exploration device 1, the stop determination unit 15 uses a pre-specified benchmark evaluation value to determine whether to suspend processing under trial processing conditions before the provisional evaluation value converges. That is, the stop determination threshold is set to a fixed value. Moreover, the stop determination unit 15 determines whether to suspend processing under trial processing conditions before the provisional evaluation value converges by comparing the fluctuation of the provisional evaluation value in the time series with the stop determination threshold. However, this is only one example.

[0140] For example, the stop determination unit 15 can also set a stop threshold based on the completed processing conditions and the corresponding evaluation value. The completed processing conditions and the corresponding evaluation value are stored as exploration results by the evaluation determination unit 16 in the exploration result storage unit 18E. The stop threshold set by the stop determination unit 15 based on the evaluation value of completion is also called a "variable stop threshold". Furthermore, in this case, if the stop determination unit 15 sets a variable stop threshold, it determines whether to stop processing under the processing conditions in the trial before the provisional evaluation value converges, for example, by comparing the estimated convergence value estimated by the convergence determination unit 14 with the variable stop threshold. The estimated convergence value estimated by the convergence determination unit 14 is the estimated convergence value in the latest convergence determination information stored in the convergence result storage unit 18C.

[0141] In detail, the stop determination unit 15 sets the variable stop threshold according to preset conditions (hereinafter referred to as "variable stop threshold setting conditions"), for example, based on the processing conditions completed in the trial and the evaluation value corresponding to the processing conditions.

[0142] In the variable abort threshold setting conditions, for example, the following conditions are set: <condition (1)>, <condition (2)>, or <condition (3)>.

[0143] <Condition (1)>

[0144] If the number of trials is less than X, the value used to prevent processing from being stopped will be set as the variable stop threshold. If the number of trials is greater than or equal to X, the Xth evaluation value among all evaluation values ​​corresponding to all processing conditions completed in the trial will be set as the variable stop threshold.

[0145] <Condition (2)>

[0146] The highest Y-value among all evaluation values ​​corresponding to all processing conditions completed in the trial is set as the variable termination threshold.

[0147] <Condition (3)>

[0148] The lowest value among the top Z% of all evaluation values ​​corresponding to all processing conditions completed in the trial is set as the variable termination threshold.

[0149] In addition, the values ​​of X, Y or Z in <condition (1)>, <condition (2)> or <condition (3)> can be set appropriately.

[0150] Additionally, in <condition (1)>, the "value for not stopping processing" is set to "0" for example. Furthermore, this is just an example; the "value for not stopping processing" can be set to a value that does not exceed the conceivable estimated convergence value.

[0151] Here, Figure 7 This is a diagram illustrating an example of a method in Embodiment 1 where the stop determination unit 15 sets a variable stop threshold based on the processing conditions completed during the trial and the evaluation value corresponding to those processing conditions.

[0152] Figure 7 This is a diagram illustrating an example of a variable stop threshold setting method in which the stop determination unit 15 sets a variable stop threshold according to the variable stop threshold setting condition of <condition (1)> based on the processing conditions completed during the trial and the evaluation value corresponding to those processing conditions. Figure 7 In the example, X in <condition (1)> is set to "5".

[0153] Figure 7 The horizontal axis indicates the number of trials for each processing condition. The number of trials represents the number of processing conditions that were successfully completed during the trials. Figure 7 The vertical axis shows the evaluation values ​​corresponding to each processing condition. Furthermore, when the processing conditions are under trial, Figure 7 The evaluation value on the vertical axis is the estimated convergence value. Figure 7 The dots in the black circles represent the evaluation values ​​or estimated convergence values ​​corresponding to each processing condition.

[0154] For ease of explanation, Figure 7 In this context, it's assumed that nine processing conditions have been tested, but for example, it's currently in the trial phase of the sixth processing condition. That is, in this case, in... Figure 7 In this context, the evaluation value corresponding to the 6th trial is the estimated convergence value.

[0155] In this case, according to Figure 7At the point where five trials have concluded, the fifth evaluation value among all evaluation values ​​corresponding to the processing conditions completed in the five trials is the evaluation value corresponding to the processing conditions of the third trial. Therefore, the stop determination unit 15 sets the evaluation value corresponding to the processing conditions of the third trial as a variable stop threshold. Furthermore, regarding the processing conditions in the trials, in other words, the estimated convergence value of the processing conditions in the sixth trial is less than the variable stop threshold, therefore the stop determination unit 15 determines to stop processing under the processing conditions in the trials.

[0156] Alternatively, for example, suppose we are currently in the trial phase of the 9th processing condition. That is, in this case, Figure 7 In this context, the evaluation value corresponding to the 9th trial is the estimated convergence value.

[0157] In this case, according to Figure 7 At the point where eight trials have concluded, the fifth evaluation value among all evaluation values ​​corresponding to the processing conditions completed in the eight trials is the evaluation value corresponding to the processing conditions of the fourth trial. Therefore, the stop determination unit 15 sets the evaluation value corresponding to the processing conditions of the fourth trial as a variable stop threshold. Furthermore, regarding the processing conditions in the trials, in other words, the estimated convergence value of the processing conditions in the ninth trial is less than the variable stop threshold, therefore the stop determination unit 15 determines to stop processing under the processing conditions in the trials.

[0158] As described above, the stop determination unit 15 can change the benchmark used to determine whether to stop processing under the processing conditions in the trial before the provisional evaluation value converges. In other words, the stop threshold is changed.

[0159] For example, if the termination threshold is too high, the processing condition exploration device 1 may stop processing midway through the process, waiting for the processing results to converge, potentially increasing the deviation of the predicted evaluation value. As a result, the processing condition exploration device 1 may fail to explore the optimal processing conditions. Conversely, if the termination threshold is too low, the processing condition exploration device 1 may need time to determine whether to terminate processing under the processing conditions corresponding to the low evaluation value before the provisional evaluation value converges, or it may wait until the provisional evaluation value converges without terminating processing. As a result, the processing condition exploration device 1 may need time until it can explore the optimal processing conditions.

[0160] In the processing condition exploration device 1, the stop determination unit 15 can change the stop threshold, thereby enabling the processing condition exploration device 1 to explore the optimal processing conditions and shorten the time until the optimal processing conditions are explored.

[0161] Furthermore, in this case, regarding the use Figure 2 The flowchart illustrates the operation of the processing condition exploration device 1. Between steps ST5 and ST6 and between steps ST19 and ST20, an additional step is added whereby the stop determination unit 15 sets a variable stop threshold.

[0162] The hardware structure for realizing the function of the processing condition exploration device 1 is as follows.

[0163] The processing condition exploration device 1's functions of the processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17 are implemented through a processing circuit. That is, the processing condition exploration device 1 has the functions of executing... Figure 2 The processing circuit from step ST1 to step ST22. The processing circuit can be dedicated hardware or a CPU (Central Processing Unit) that executes a program stored in memory.

[0164] Figure 8A This is a block diagram illustrating the hardware structure for realizing the functions of the processing condition exploration device 1. Furthermore, Figure 8B This is a block diagram representing the hardware structure of the software that performs the functions of the processing condition exploration device 1. Figure 8A and Figure 8B In this process, the input interface device 102 relays the processing result information output from the processing machine 2 to the processing condition exploration device 1, and relays the stored information output from each storage unit 18A to 18G to the processing condition exploration device 1. The output interface device 103 relays the information output from the processing condition exploration device 1 to the display unit 3, or the information output from the processing condition exploration device 1 to each storage unit 18A to 18G.

[0165] In the processing circuit Figure 8A In the case of the dedicated hardware processing circuit 101 shown, the processing circuit 101 may be, for example, a single circuit, a composite circuit, a programmable processor, a parallel programmable processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof. The functions of the exploration processing condition generation unit 11, the processing result collection unit 12, the evaluation value acquisition unit 13, the convergence determination unit 14, the stop determination unit 15, the evaluation decision unit 16, and the machine learning unit 17 in the processing condition exploration device 1 can be implemented by different processing circuits, or these functions can be combined and implemented by a single processing circuit.

[0166] In the processing circuit Figure 8B In the case of the processor 104 shown, the functions of the processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17 in the processing condition exploration device 1 are implemented by software, firmware, or a combination of software and firmware. Furthermore, the software or firmware is described as a program and stored in the memory 105.

[0167] The processor 104 reads and executes the program stored in the memory 105, thereby realizing the functions of the processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17 in the processing condition exploration device 1. For example, the processing condition exploration device 1 has a memory 105, which, when executed by the processor 104, is used for processing conditions generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17 in the processing condition exploration device 1. Figure 2 The program that ultimately executes the processes up to steps ST1 to ST22 in the flowchart shown is stored. These programs enable the computer to execute the processing sequence or method of the exploration processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17. The memory 105 may be a computer-readable storage medium storing programs that enable the computer to function as the exploration processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17.

[0168] The memory 105 includes, for example, non-volatile or volatile semiconductor memories such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), and EEPROM (Electrically-EPROM), as well as disks, floppy disks, optical disks, compact disks, mini disks, DVDs, etc.

[0169] The functions of the processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17 in the processing condition exploration device 1 are partly implemented by dedicated hardware and partly by software or firmware. For example, the processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, and evaluation decision unit 16 are implemented by the processing circuit 101 as dedicated hardware, and the machine learning unit 17 is implemented by the processor 104 reading and executing the program stored in the memory 105. As described above, the processing circuit can implement the above functions by hardware, software, firmware, or a combination thereof.

[0170] Furthermore, in Embodiment 1 described above, the processing condition exploration device 1 can be mounted on the processing machine 2, or it can be installed on a server connected to the processing machine 2 via a network. For example, some of the processing condition generation unit 11, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, and machine learning unit 17 can be mounted on the processing machine 2, while others can be installed on the server.

[0171] As described above, the processing condition exploration device 1 according to Embodiment 1 includes: a processing condition calculation unit 111 that generates processing conditions consisting of multiple control parameters that can be set on the processing machine 2; an actual processing instruction unit 112 that causes the processing machine 2 to perform processing according to the processing conditions generated by the processing condition calculation unit 111; a processing result collection unit 12 that collects processing result information representing the processing performed by the processing machine 2 caused by the actual processing instruction unit 112; and an evaluation value acquisition unit 13 that, based on the processing result information collected by the processing result collection unit 12, evaluates the processing condition of the processing machine 2. The following steps are performed: A provisional evaluation value is calculated for the implemented processing; a convergence determination unit 14 determines whether the provisional evaluation value has converged based on the provisional evaluation value of the time series calculated by the evaluation value acquisition unit 13; if the provisional evaluation value is determined not to have converged, a presumed convergence value that serves as the convergence target of the provisional evaluation value is estimated; a stop determination unit 15 determines whether to suspend processing under the trial processing conditions before the provisional evaluation value converges if the convergence determination unit 14 determines that the provisional evaluation value has not converged; and an evaluation decision unit 16 determines whether to suspend processing under the trial processing conditions if the stop determination unit 15 determines that the trial processing is to be suspended. In the case of processing under the processing conditions in progress, the actual processing instruction unit 112 terminates the processing according to the processing conditions for the processing machine 2, and determines the estimated convergence value estimated by the convergence determination unit 14 as the evaluation value of the processing performed according to the processing conditions. If the stop determination unit 15 determines that the processing under the processing conditions in progress should not be terminated, and after the convergence determination unit 14 determines that the provisional evaluation value has converged, the convergence value of the provisional evaluation value is determined as the evaluation value; and the exploration end determination unit 113 determines whether to end the exploration of the processing conditions. If the exploration ends, the basic The optimal processing conditions are determined by the evaluation value determined by the evaluation decision unit 16. Without ending the exploration, the processing condition calculation unit 111 generates processing conditions for the next trial based on the predicted value predicted by the prediction unit 171. This process is repeated until the exploration ends, and the processing condition calculation unit 111, actual processing instruction unit 112, processing result collection unit 12, evaluation value acquisition unit 13, convergence determination unit 14, stop determination unit 15, evaluation decision unit 16, prediction unit 171, and exploration end determination unit 113 processes are repeated. As a result, when the processing condition exploration device 1 explores the optimal processing conditions, compared with the prior art of performing processing under all processing conditions until the vibrational changes of the processing results stabilize, the time until the optimal processing conditions are explored is shortened.

[0172] Furthermore, it is possible to modify any structural element of the implementation method or omit any structural element of the implementation method.

[0173] Industrial applicability

[0174] The processing condition exploration device involved in this invention can be used, for example, to explore the processing conditions of a laser processing machine.

[0175] Explanation of the label

[0176] 1 Processing condition exploration device, 2 Processing machine, 3 Display unit, 11 Processing condition generation unit, 111 Processing condition calculation unit, 112 Actual processing instruction unit, 113 Exploration end determination unit, 12 Processing result collection unit, 13 Evaluation value acquisition unit, 14 Convergence determination unit, 15 Stop determination unit, 16 Evaluation decision unit, 17 Machine learning unit, 171 Prediction unit, 172 Unreliability evaluation unit, 18A Processing result storage unit, 18B Evaluation value storage unit, 18C Convergence result storage unit, 18D Stop determination storage unit, 18E Exploration result storage unit, 18F Prediction result storage unit, 18G Unreliability storage unit, 101 Processing circuit, 102 Input interface device, 103 Output interface device, 104 Processor, 105 Memory.

Claims

1. A processing condition exploration device, comprising: The machining condition calculation unit generates machining conditions consisting of multiple control parameters that can be set on the machining machine; The actual processing instruction unit causes the processing machine to perform processing according to the processing conditions generated by the processing condition calculation unit; The processing result collection unit collects processing result information, which represents the processing result of the processing performed by the processing machine caused by the actual processing instruction unit; The evaluation value acquisition unit calculates a provisional evaluation value for the processed product based on the processing result information collected by the processing result collection unit. The convergence determination unit determines whether the provisional evaluation value has converged based on the provisional evaluation value of the time series calculated by the evaluation value acquisition unit. If it is determined that the provisional evaluation value has not converged, it estimates the estimated convergence value that becomes the convergence target of the provisional evaluation value. The stop determination unit determines whether to suspend the processing under the processing conditions in the trial before the provisional evaluation value converges if the convergence determination unit determines that the provisional evaluation value has not converged. The evaluation decision unit, when the stop determination unit determines that the processing under the processing conditions in the trial should be stopped, causes the actual processing instruction unit to end the processing according to the processing conditions for the processing machine, and determines the estimated convergence value estimated by the convergence determination unit as the evaluation value of the processing performed according to the processing conditions. When the stop determination unit determines that the processing under the processing conditions in the trial should not be stopped, after the convergence determination unit determines that the provisional evaluation value has converged, the convergence value of the provisional evaluation value is determined as the evaluation value. The prediction unit predicts a value for the evaluation value corresponding to the processing conditions that have not been tried, based on the evaluation value determined by the evaluation decision unit and the processing conditions corresponding to the evaluation value. as well as The exploration end determination unit determines whether to end the exploration of the processing conditions. If the exploration ends, it determines the optimal processing conditions based on the evaluation value determined by the evaluation determination unit. If the exploration does not end, the processing condition calculation unit generates the processing conditions to be tried next based on the predicted value predicted by the prediction unit. Until the exploration is determined to be terminated by the exploration termination determination unit, the processes of the processing condition calculation unit, the actual processing instruction unit, the processing result collection unit, the evaluation value acquisition unit, the convergence determination unit, the stop determination unit, the evaluation decision unit, the prediction unit, and the exploration termination determination unit are repeated.

2. The processing condition exploration device according to claim 1, characterized in that, The convergence determination unit estimates the estimated convergence value based on the degree of fluctuation of the provisional evaluation value of the time series calculated by the evaluation value acquisition unit.

3. The processing condition exploration device according to claim 1, characterized in that, The convergence determination unit estimates the estimated convergence value based on the provisional evaluation value of the time series calculated by the evaluation value acquisition unit and a first machine learning model that takes the evaluation value of the time series as input and outputs the estimated convergence value.

4. The processing condition exploration device according to claim 1, characterized in that, The stop determination unit determines whether to stop the processing under the processing conditions in the trial before the provisional evaluation value converges by comparing the fluctuation degree of the provisional evaluation value of the time series calculated by the evaluation value acquisition unit with the stop threshold.

5. The processing condition exploration device according to claim 4, characterized in that, The stop determination unit sets a variable stop threshold based on the processing conditions completed in the trial and the evaluation value corresponding to the processing conditions. By comparing the estimated convergence value estimated by the convergence determination unit with the set variable stop threshold, it determines whether to stop the processing under the processing conditions in the trial before the provisional evaluation value converges.

6. The processing condition exploration device according to claim 1, characterized in that, The stop determination unit determines whether to stop the processing under the processing conditions in the trial before the provisional evaluation value is converged, based on the provisional evaluation value of the time series calculated by the evaluation value acquisition unit and a second machine learning model that takes the evaluation value of the time series as input and outputs information indicating whether to stop the processing.

7. The processing condition exploration device according to claim 1, characterized in that, It has an unreliability evaluation unit that calculates an index representing the unreliability of the predictions made by the prediction unit. The processing condition calculation unit generates the processing conditions that should be tried next based on the predicted value of the evaluation value predicted by the prediction unit and the index representing the unreliability of the prediction.

8. The processing condition exploration device according to claim 7, characterized in that, The exploration end determination unit uses the predicted value of the evaluation value and the index representing the unreliability of the evaluation value to determine whether to end the exploration under the processing condition. If it is determined that the exploration under the processing condition should end, the processing condition corresponding to the highest evaluation value among the evaluation values ​​determined by the evaluation determination unit is set as the optimal processing condition.

9. The processing condition exploration device according to claim 7, characterized in that, The prediction unit uses a probability model of the processing conditions for which the evaluation value is generated according to a specific distribution to predict the predicted value. The unreliability evaluation unit uses the probability model to calculate the index representing the unreliability of the prediction.

10. The processing condition exploration device according to claim 1, characterized in that, The device has a display unit that displays at least one of the processing conditions and the evaluation value corresponding to the processing conditions, the processing conditions and the predicted value of the evaluation value corresponding to the processing conditions, or the processing conditions of the exploration results.

11. A method for exploring processing conditions, comprising the following steps: The machining condition calculation unit generates machining conditions consisting of multiple control parameters that can be set on the machining machine; The actual machining instruction unit causes the machining machine to perform machining according to the machining conditions generated by the machining condition calculation unit; The processing result collection unit collects processing result information, which represents the processing result of the processing performed by the processing machine caused by the actual processing instruction unit; The evaluation value acquisition unit calculates a provisional evaluation value for the processed product based on the processing result information collected by the processing result collection unit. The convergence determination unit determines whether the provisional evaluation value has converged based on the provisional evaluation value of the time series calculated by the evaluation value acquisition unit. If it is determined that the provisional evaluation value has not converged, the unit estimates the estimated convergence value that is the convergence target of the provisional evaluation value. If the convergence determination unit determines that the provisional evaluation value has not converged, the stop determination unit determines whether to suspend the processing under the processing conditions in the trial before the provisional evaluation value converges. If the evaluation decision unit determines that the processing under the processing conditions in the trial run should be stopped, the actual processing instruction unit shall end the processing performed according to the processing conditions for the processing machine, and determine the estimated convergence value estimated by the convergence decision unit as the evaluation value of the processing performed according to the processing conditions. If the stop decision unit determines that the processing under the processing conditions in the trial run should not be stopped, after the convergence decision unit determines that the provisional evaluation value has converged, the convergence value of the provisional evaluation value shall be determined as the evaluation value. The prediction unit predicts a value for the evaluation value corresponding to the processing conditions that have not been tried, based on the evaluation value determined by the evaluation decision unit and the processing conditions corresponding to the evaluation value. as well as The exploration end determination unit determines whether to end the exploration of the processing conditions. If the exploration ends, it determines the optimal processing conditions based on the evaluation value determined by the evaluation decision unit and the evaluation value predicted by the prediction unit. If the exploration does not end, the processing condition calculation unit generates the processing conditions to be tried next based on the predicted value predicted by the prediction unit. The processes of the processing condition calculation unit, the actual processing instruction unit, the processing result collection unit, the evaluation value acquisition unit, the convergence determination unit, the stop determination unit, the evaluation decision unit, the prediction unit, and the exploration end determination unit are repeated until the exploration end determination unit determines that the exploration has ended.

Images