Machining condition adjustment device, machining system, machining condition adjustment method, machining method, and machining condition adjustment program

Abstract

The machining condition adjustment device (10A) includes a test item inference device (30) that derives machining test items from required specifications using a learned machining test item model for inferring machining test items from required specifications desired by an operator for electric discharge machining, a machining result model update device (20) that acquires machining results corresponding to the machining test, infers machining condition parameters from the machining results, and infers the machining results from the machining condition parameters, and generates an updated machining result model by updating the reference machining result model based on the acquired machining result data, a parameter adjustment unit (41) that calculates machining condition parameters for obtaining a machining result that satisfies the required specifications from the required specifications as calculated machining condition parameters using the updated machining result model, and a machining result estimation unit (42) that estimates an estimated machining result corresponding to the calculated machining condition parameters based on the updated machining result model and the calculated machining condition parameters.

Description

[Technical field]

[0001] The present disclosure relates to a machining condition adjustment device, a machining system, a machining condition adjustment method, a machining method, and a machining condition adjustment program that adjust machining condition parameters of an electric discharge machine. [Background technology]

[0002] Conventionally, the machining conditions applied to electric discharge machines such as wire electric discharge machines were composed of multiple machining condition parameters, and because there were a wide variety of combinations of the machining condition parameters, the know-how of an expert was required to adjust the machining condition parameters to improve machining accuracy.

[0003] The machining condition adjustment device described in Patent Document 1 performs predetermined test actual machining to update the calculation model used in the simulation processing of the wire EDM machine to suit the on-site machine, and adjusts the machining conditions using the updated calculation model. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2022 / 210472 Summary of the Invention [Problem to be solved by the invention]

[0005] However, the technology of Patent Document 1 does not take into account the machining results such as machining dimensions and surface roughness in actual machining for testing, so that even if the machining condition parameters are adjusted, it is difficult to correct the machining results with high precision, and there is a problem that adjustments to the machining condition parameters cannot be made to achieve high-precision machining.

[0006] The present disclosure has been made in consideration of the above, and has an object to provide a machining condition adjustment device that can adjust machining condition parameters to achieve high-precision machining. [Means for solving the problem]

[0007] In order to solve the above-mentioned problems and achieve the object, the machining condition adjustment device disclosed in the present disclosure acquires required specifications desired by an operator for electric discharge machining, and infers machining test items, which are items of a machining test, from the required specifications by using a trained machining test item model for inferring the machining test items, which are items of a machining test, from the acquired required specifications. The minimum number The machining condition adjustment device of the present disclosure includes a test item inference device that derives machining test items. The machining condition adjustment device of the present disclosure also includes a machining result model update device that acquires a machining result corresponding to a machining test as model update data, infers machining condition parameters from the machining result, and updates the reference machining result model based on the acquired model update data to generate an updated machining result model that is the updated reference machining result model. The machining condition adjustment device of the present disclosure also includes a parameter adjustment unit that uses the updated machining result model to calculate machining condition parameters for obtaining a machining result that satisfies the required specifications from the required specifications as calculated machining condition parameters, and a machining result estimation unit that estimates an estimated machining result corresponding to the calculated machining condition parameters based on the updated machining result model and the calculated machining condition parameters. Effect of the Invention

[0008] The machining condition adjustment device according to the present disclosure has an effect of enabling adjustment of machining condition parameters to realize highly accurate machining. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram showing a configuration of an electric discharge machining system having a machining condition adjustment device according to a first embodiment. [Diagram 2] FIG. 1 is a diagram showing a configuration of a processing condition adjustment device according to a first embodiment; [Diagram 3] 1 is a flowchart showing a processing procedure of a process executed by a machining condition adjustment device according to a first embodiment; [Figure 4]11 is a flowchart showing a processing procedure of a processing executed by a machining condition adjustment device according to a second embodiment. [Diagram 5] FIG. 13 is a diagram showing a configuration of a processing condition adjustment device according to a third embodiment. [Figure 6] 11 is a flowchart showing a processing procedure for a processing condition adjustment device according to a third embodiment to derive a processing test item. [Figure 7] 11 is a flowchart showing a processing procedure of a process in which a machining condition adjustment device according to a third embodiment updates a machining result model. [Figure 8] FIG. 13 is a diagram showing a configuration of a processing condition adjustment device according to a fourth embodiment. [Figure 9] FIG. 13 is a diagram showing a configuration of a processing condition adjustment device according to a fifth embodiment. [Figure 10] FIG. 13 is a diagram showing a configuration of a processing condition adjustment device according to a sixth embodiment. [Figure 11] FIG. 23 is a diagram for explaining a process in which the machining condition adjustment device according to the seventh embodiment updates a machining result model. [Figure 12] FIG. 13 is a diagram showing a configuration of a machining result model learning device according to an eighth embodiment. [Figure 13] FIG. 23 is a diagram for explaining a neural network used by the machining result model learning device according to the eighth embodiment. [Figure 14] 13 is a flowchart showing a procedure of processing executed by a processing result model learning device according to an eighth embodiment. [Figure 15] FIG. 23 is a diagram for explaining a plurality of steps from which a reference processing result model is derived by the processing result model learning device according to the eighth embodiment. [Figure 16] FIG. 13 is a diagram showing a configuration of a test item model learning device according to a ninth embodiment. [Figure 17] 13 is a flowchart showing a procedure of a process executed by a test item model learning device according to a ninth embodiment. [Figure 18] FIG. 1 is a diagram showing a configuration example of a processing circuit provided in a machining condition adjustment device according to a first embodiment when the processing circuit is realized by a processor and a memory. [Figure 19]FIG. 1 is a diagram showing an example of a processing circuit included in a machining condition adjustment device according to a first embodiment when the processing circuit is configured with dedicated hardware; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A machining condition adjustment device, a machining system, a machining condition adjustment method, a machining method, and a machining condition adjustment program according to embodiments of the present disclosure will be described in detail below with reference to the drawings.

[0011] Embodiment 1 FIG. 1 is a diagram showing a configuration of an electric discharge machining system having a machining condition adjustment device according to a first embodiment. In FIG. 1, a wire electric discharge machine 1 is described as an example of an electric discharge machine applied to the machining condition adjustment device 10A. The electric discharge machine applied to the machining condition adjustment device 10A is not limited to the wire electric discharge machine 1, and may be an electric discharge machine such as a small hole electric discharge machine or a die sinker electric discharge machine. In FIG. 1, two axes in a plane parallel to the upper surface of a base T and perpendicular to each other are defined as an X-axis and a Y-axis. An axis perpendicular to the X-axis and the Y-axis is defined as a Z-axis.

[0012] The electric discharge machining system includes a wire electric discharge machine 1 and a machining condition adjustment device 10A. The wire electric discharge machine 1 includes a control device 2 and a power supply device 3. The control device 2 is a numerical control device. The power supply device 3 may include, for example, a detection unit (not shown) that detects the state of the wire electric discharge machine 1.

[0013] Moreover, the wire electric discharge machine 1 includes a wire electrode 6, a wire electrode bobbin 81, a wire electrode recovery box 82, a wire electrode transport roller 83, a lower roller 84, and a wire traveling speed control motor 85. Moreover, the wire electric discharge machine 1 includes an upper guide 86, a lower guide 87, a machining section 88, an X-axis motor 89X, and a Y-axis motor 89Y.

[0014] The wire electrode 6 is supplied from a wire electrode bobbin 81, sent to the processing section 88 via a wire electrode transport roller 83, an upper power supply 8a, and an upper guide 86, and then collected in a wire electrode collection box 82 via a lower power supply 8b, a lower guide 87, a lower roller 84, and a wire traveling speed control motor 85.

[0015] Upper power supply 8a and lower power supply 8b are disposed between wire electrode transport roller 83 and lower roller 84. Upper guide 86 and lower guide 87 are also disposed between wire electrode transport roller 83 and lower roller 84.

[0016] The machining unit 88 is disposed between the upper power feeder 8a and the lower power feeder 8b, and between the upper guide 86 and the lower guide 87. The wire electrode 6 is held by the upper guide 86 and the lower guide 87. When power is supplied from the power supply device 3 to the upper power feeder 8a and the lower power feeder 8b, a machining current is supplied to the wire electrode 6 between the upper power feeder 8a and the lower power feeder 8b. As a result, a machining current is supplied to the wire electrode 6 in the machining unit 88.

[0017] After machining a workpiece (object to be machined) 4 placed on a platen T in machining section 88, wire electrode 6 is collected in wire electrode collection box 82. The transport speed of wire electrode 6 is controlled by wire traveling speed control motor 85.

[0018] The control device 2 inputs position commands to the X-axis motor 89X and the Y-axis motor 89Y, which change the position of the surface plate T on the X-axis and Y-axis. In this way, the control device 2 controls the relative position of the wire electrode 6 with respect to the workpiece 4 in the XY plane.

[0019] The wire electric discharge machine 1 may control the relative position of the wire electrode 6 in the XY plane with respect to the workpiece 4 by moving the wire electrode 6 in a plane parallel to the XY plane.

[0020] 2 is a diagram showing a configuration of a machining condition adjustment device according to the first embodiment. The machining condition adjustment device 10A is a computer that adjusts machining conditions (electric discharge machining conditions) by adjusting machining condition parameters. The machining condition adjustment device 10A is connected to an electric discharge machine such as a wire electric discharge machine 1.

[0021] The machining condition adjustment device 10A acquires various information used for adjusting the machining conditions from the wire electric discharge machine 1 and the operator 5. The machining condition adjustment device 10A includes an input unit 7, a machining result model update device 20, a test item inference device 30, a reference machining result model storage unit 45, a test item model storage unit 43, a parameter adjustment unit 41, a machining result estimation unit 42, and a display unit 44. The display unit 44 may be disposed outside the machining condition adjustment device 10A.

[0022] The input unit 7 receives the processing result input by the worker 5 and inputs it to the processing result model update device 20. The input unit 7 also receives the requirement specifications input by the worker 5 and inputs it to the test item inference device 30 and the parameter adjustment unit 41.

[0023] The machining results are the results obtained by machining using the wire electric discharge machine 1. The machining results include the dimensional error from the desired value of the workpiece 4 obtained by the machining test, surface roughness, machining speed, time required for machining (machining time), power consumption required for machining, machining specifications, etc. The machining specifications are information that affects the machining results. The machining specifications include, for example, information on the type of machining such as punch machining or die machining, the material of the workpiece 4, the material of the wire electrode 6, the diameter of the wire electrode 6, the model of the wire electric discharge machine 1, etc. There may be one type of machining result or multiple types.

[0024] The required specifications are the specifications desired by the operator 5 for the electric discharge machining. The type of information included in the required specifications is the same as the type of information included in the machining result. The required specifications include the range of machining parameters and the range of machining results (hereinafter referred to as the required machining result) that the operator 5 requires for the range of machining parameters.

[0025] The required machining results include the range of dimensional errors of the workpiece 4 that is desired to be achieved by machining with the wire electric discharge machine 1, the range of surface roughness, the range of machining speeds, the range of time required for machining, the range of power consumption required for machining, etc. The type of the required machining results may be one or more. The required machining results are not limited to information indicating a specific range and may be a specific value.

[0026] The required specifications are the specifications desired by the operator 5, and the required machining result is the result of the final machining desired by the operator 5. The machining result input to the input unit 7 is the result of actual machining performed by the wire electric discharge machine 1 through a machining test. Moreover, the machining result displayed on the display unit 44 is the machining result estimated by the machining condition adjustment device 10A or the like (hereinafter referred to as the estimated machining result). Hereinafter, the result of machining performed by the wire electric discharge machine 1 through a machining test will be referred to as the machining result, and is distinguished from the required specifications including the required machining result, and the estimated machining result.

[0027] The reference processing result model storage unit 45 stores a reference processing result model derived in advance. The reference processing result model is a trained model that represents the relationship between the processing result and the processing condition parameters. In other words, the reference processing result model is a trained model for inferring the processing condition parameters from the processing result and for inferring the learning result from the processing condition parameters.

[0028] The reference machining result model is generated in advance by a learning device or the like and stored in the wire electric discharge machine 1. The machining condition adjustment device 10A reads out the reference machining result model stored in the wire electric discharge machine 1 and stores it in the reference machining result model storage unit 45. The machining condition parameters include the tension of the wire electrode 6, the feed speed of the wire electrode 6, the machining current value supplied to the wire electrode 6, and the like.

[0029] The reference machining result model stored in the reference machining result model storage unit 45 may be sent to the reference machining result model storage unit 45 from a device other than the wire electric discharge machine 1. The reference machining result model may be stored in an external device other than the machining condition adjustment device 10A. In this case, the external device that stores the reference machining result model and the machining condition adjustment device 10A are connected by a network or the like.

[0030] The test item model storage unit 43 stores a machining test item model derived in advance. The machining test item model is a learned model that represents the relationship between the required specifications and the machining test items. In other words, the machining test item model is a learned model for inferring the machining test items, which are the items of the machining test, from the required specifications. The machining test item model is generated in advance by a learning device or the like. The machining test item model may be stored in an external device other than the machining condition adjustment device 10A. In this case, the external device that stores the machining test item model and the machining condition adjustment device 10A are connected by a network or the like.

[0031] The test item inference device 30 derives the processing test items from the requirement specifications by using the processing test item model. That is, the test item inference device 30 infers the processing test items corresponding to the requirement specifications by applying the requirement specifications to the processing test item model. The test item inference device 30 sends the processing test items to the processing result model update device 20 and the display unit 44. As a result, processing corresponding to the processing test items is executed, and when the processing result of this processing is input to the input unit 7 by the operator 5, this processing result is sent to the processing result model update device 20. The processing test items are also used when the processing result model update device 20 updates the reference processing result model.

[0032] The machining test items include, for example, a combination of machining condition parameters, a nozzle separation amount (the distance between the workpiece 4 and the machining fluid nozzle), a placement position of the workpiece 4 on the surface plate T, and the like.

[0033] Based on the machining result and the machining test items, the machining result model update device 20 updates the reference machining result model stored in the reference machining result model storage unit 45. The machining result model update device 20 may update the reference machining result model stored in the reference machining result model storage unit 45 based on the machining result without using the machining test items.

[0034] In this way, the machining results corresponding to the machining test items are used to update the reference machining result model. The machining result model update device 20 sends the latest updated updated machining result model to the parameter adjustment unit 41, the machining result estimation unit 42, and the display unit 44.

[0035] The parameter adjustment unit 41 calculates machining condition parameters (calculated machining condition parameters) corresponding to the required specifications by applying the required specifications to the latest updated machining result model. That is, the parameter adjustment unit 41 calculates the machining condition parameters from the required specifications using the latest updated machining result model. In this way, the parameter adjustment unit 41 calculates the machining condition parameters for obtaining a machining result that satisfies the required specifications.

[0036] Furthermore, the parameter adjustment unit 41 calculates the adjustment amount of the machining condition parameter (hereinafter referred to as the parameter adjustment amount) based on the difference between the machining condition parameter that is set in advance as a reference and the calculated machining condition parameter. The parameter adjustment unit 41 may calculate the parameter adjustment amount based on the difference between the machining condition parameter calculated last time and the machining condition parameter calculated this time.

[0037] The parameter adjustment unit 41 calculates the machining condition parameters using an updated machining result model in which adjustments are made to the reference machining result model according to the machining result. Therefore, it can be said that the machining condition parameters calculated by the parameter adjustment unit 41 are machining condition parameters adjusted according to the machining result.

[0038] The parameter adjustment unit 41 sends the calculated machining condition parameters to the machining result estimation unit 42. In addition, the parameter adjustment unit 41 sends at least one of the parameter adjustment amount and the calculated machining condition parameters to the display unit 44. Below, a case where the parameter adjustment unit 41 sends both the parameter adjustment amount and the calculated machining condition parameters to the display unit 44 will be described.

[0039] The processing result estimating unit 42 estimates an estimated processing result when the processing condition parameters are adjusted (when the reference processing result model is updated) by applying the processing condition parameters calculated by the parameter adjusting unit 41 to the updated processing result model. That is, the processing result estimating unit 42 estimates an estimated processing result when processing is performed using the adjusted processing condition parameters.

[0040] The estimated machining result is an estimated value of the result obtained by machining using the wire electric discharge machine 1. This estimated machining result is the same type of information as the machining result and required specifications input by the operator 5. In other words, the estimated machining result is the dimensional error from the desired value of the workpiece 4 obtained by the machining test, surface roughness, machining speed, machining time, power consumption required for machining, etc. In addition, the estimated machining result may include the machining specifications described above. There may be one type of estimated machining result, or multiple types.

[0041] The required specifications that the parameter adjustment unit 41 inputs to the reference processing result model include the range of processing parameters and the required processing result, which is the range of the processing result. In this way, the required specifications are information indicating a range. The estimated processing result that the processing result estimation unit 42 outputs from the reference processing result model is information indicating a specific value of the processing result. The processing result estimation unit 42 sends the estimated estimated processing result to the display unit 44. Note that the processing condition adjustment device 10A does not have to be equipped with the processing result estimation unit 42.

[0042] The display unit 44 displays the machining test items, the adjusted machining condition parameters, and the estimated machining results. The display unit 44 may also display parameter adjustment amounts, a reference machining result model, an updated machining result model, and the like.

[0043] The machining result displayed by the display unit 44 corresponds to the updated machining result model and the parameter adjustment amount, and this parameter adjustment amount corresponds to the updated machining result model and the required specifications. The updated machining result model corresponds to the machining result when a machining test is performed using the machining test items, and the machining test items correspond to the required specifications. In other words, the machining result displayed by the display unit 44 corresponds to the required specifications input by the operator 5.

[0044] Note that the parameter adjustment unit 41 may transmit the calculated parameter adjustment amounts to the wire electric discharge machine 1. This enables the wire electric discharge machine 1 to execute machining using the parameter adjustment amounts.

[0045] 3 is a flowchart showing the procedure of the processing executed by the machining condition adjusting device according to the embodiment 1. Here, the procedure of the processing executed by the machining condition adjusting device 10A and the procedure of the processing executed by the operator 5 will be described.

[0046] When an instruction to start adjusting the machining conditions is input from the operator 5 to the machining condition adjustment device 10A, the display unit 44 of the machining condition adjustment device 10A displays the reference machining result model (step S10). The operator 5 checks the reference machining result model (step S20) and inputs the desired required specifications to the input unit 7 (step S30). The test item inference device 30 of the machining condition adjustment device 10A derives machining test items for updating the reference machining result model based on the test item model and the input required specifications (step S40). The display unit 44 displays the machining test items for updating the reference machining result model (step S50).

[0047] The operator 5 performs a machining test corresponding to the machining test items based on the displayed machining test items (step S60). That is, the operator 5 causes the wire electric discharge machine 1 to execute a machining test corresponding to the machining test items. The operator 5 measures the machining results of the machining by the wire electric discharge machine 1 (step S70) and inputs the machining results to the machining condition adjustment device 10A (step S80).

[0048] The machining result model update device 20 of the machining condition adjustment device 10A derives an updated machining result model based on the reference machining result model and the input machining result (step S90). The machining result model update device 20 updates the reference machining result model to an updated machining result model (step S100).

[0049] The parameter adjustment unit 41 inputs the required specifications to the updated machining result model to calculate machining condition parameters (adjusted machining condition parameters) corresponding to the required specifications (step S110). The display unit 44 displays the parameter adjustment amount and the adjusted machining condition parameters (step S120).

[0050] The processing result estimating unit 42 estimates an estimated processing result when processing is performed with the adjusted processing condition parameters based on the updated processing result model (step S130). That is, the processing result estimating unit 42 estimates an estimated processing result corresponding to the adjusted processing condition parameters by inputting the adjusted processing condition parameters to the updated processing result model. The display unit 44 displays the estimated processing result estimated by the processing result estimating unit 42 (step S140). After the parameter adjustment amount and the adjusted processing condition parameters are calculated, the display unit 44 may display the parameter adjustment amount and the adjusted processing condition parameters at any timing. Note that the display unit 44 does not have to display at least one of the parameter adjustment amount and the estimated processing result.

[0051] The operator 5 performs the actual machining using the displayed adjusted machining condition parameters (step S150). That is, the operator 5 causes the wire electric discharge machine 1 to perform the actual machining using the adjusted machining condition parameters.

[0052] In this way, the machining condition adjustment device 10A updates the reference machining result model using machining results such as dimensional errors and surface roughness when a machining test is performed, so that an updated machining result model corresponding to a complex shape can be derived.

[0053] In addition, the processing condition adjustment device 10A can adjust the processing condition parameters based on an updated processing result model that is updated from a previously derived reference processing result model to suit the usage pattern (required specifications) of the worker 5, so that a processing result that satisfies the required specifications can be obtained.

[0054] Furthermore, the machining condition adjustment device 10A can estimate an estimated machining result with high accuracy before the actual machining is performed, and the worker 5 can know the estimated machining result with high accuracy before the actual machining is performed.

[0055] The test item inference device 30 and the test item model storage unit 43 may be connected, for example, via a network. At least one of the test item inference device 30 and the test item model storage unit 43 may be built into the wire electric discharge machine 1. At least one of the test item inference device 30 and the test item model storage unit 43 may exist on a cloud server.

[0056] The machining result model updating device 20 and the reference machining result model storage unit 45 may be connected, for example, via a network. The machining result model updating device 20 and the reference machining result model storage unit 45 may be built into the wire electric discharge machine 1. At least one of the machining result model updating device 20 and the reference machining result model storage unit 45 may exist on a cloud server.

[0057] In this way, the machining condition adjustment device 10A of the first embodiment infers machining test items corresponding to the required specifications of the electric discharge machining, and generates an updated machining result model by updating the reference machining result model based on the machining result corresponding to the machining test. Then, the machining condition adjustment device 10A adjusts the machining condition parameters based on the updated machining result model and the required specifications. As a result, the machining condition adjustment device 10A can adjust the machining condition parameters to realize high-precision machining. In addition, since the machining condition adjustment device 10A estimates the machining result when the machining condition parameters are adjusted, it can estimate the estimated machining result with high precision before machining is performed.

[0058] In addition, the machining condition adjustment device 10A displays the machining condition parameters and parameter adjustment amounts that have been adjusted to achieve high-precision machining, so that the worker 5 can know the machining condition parameters and parameter adjustment amounts that will allow for high-precision machining results before the actual machining is performed.

[0059] Embodiment 2 Next, the second embodiment will be described with reference to Fig. 4. In the first embodiment, the machining condition adjustment device 10A displays a reference machining result model, and then the operator 5 inputs the required specifications. In the second embodiment, the machining condition adjustment device 10A displays a reference machining result model after the operator 5 inputs the required specifications. The machining condition adjustment device 10A of the second embodiment has the same configuration as the machining condition adjustment device 10A of the first embodiment.

[0060] 4 is a flowchart showing the procedure of the processing executed by the machining condition adjustment device according to the second embodiment. Here, the procedure of the processing executed by the machining condition adjustment device 10A according to the second embodiment and the procedure of the processing executed by the operator 5 will be described. Note that, among the processing executed by the machining condition adjustment device 10A according to the second embodiment, the same processing as the processing executed by the machining condition adjustment device 10A according to the first embodiment is given the same step number, and the description thereof will be omitted.

[0061] Compared with the machining condition adjustment device 10A of the first embodiment, the machining condition adjustment device 10A of the second embodiment executes processes of steps S5, S10, and S20 instead of the processes of steps S10 to S30.

[0062] Specifically, the operator 5 inputs desired required specifications to the input unit 7 in order to start adjusting the machining conditions (step S5). When the machining condition adjustment device 10A receives the desired required specifications, the display unit 44 displays a reference machining result model (step S10). This allows the operator 5 to check the reference machining result model (step S20).

[0063] Further, the test item inference device 30 of the machining condition adjustment device 10A derives machining test items for updating the reference machining result model based on the test item model and the input required specifications (step S40). Thereafter, the operator 5 executes the processes of steps S60 to S80 described in the first embodiment, and the machining condition adjustment device 10A executes the processes of steps S90 to S140 described in the first embodiment. Further, the operator 5 executes the process of step S150 described in the first embodiment.

[0064] Thus, according to the second embodiment, the processing condition adjustment device 10A receives the required specifications from the operator 5 before the operator 5 confirms the reference processing result model. After that, the processing condition adjustment device 10A starts the update process of the reference processing result model, and the operator 5 confirms the reference processing result model. This allows the operator 5 to auxiliary execute confirmation of the reference processing result model.

[0065] Embodiment 3 Next, a third embodiment will be described with reference to Fig. 5 to Fig. 7. In the third embodiment, a machining condition adjustment device 10B, which will be described later, infers machining test items based on various information (such as machine characteristics and machining environment, which will be described later) acquired from the wire electric discharge machine 1, and updates a reference machining result model.

[0066] Fig. 5 is a diagram showing the configuration of a machining condition adjustment device according to embodiment 3. Among the components in Fig. 5, components that achieve the same functions as those in the machining condition adjustment device 10A according to embodiment 1 shown in Fig. 2 are given the same reference numerals, and duplicated explanations will be omitted.

[0067] The machining condition adjustment device 10B of the third embodiment is a computer that adjusts machining condition parameters, similar to the machining condition adjustment device 10A. The machining condition adjustment device 10B is connected to an electric discharge machine such as the wire electric discharge machine 1.

[0068] Compared to the machining condition adjustment device 10A, the machining condition adjustment device 10B includes a required specification input unit 11, a machining result input unit 12, a machining state acquisition unit 13, a machine characteristic acquisition unit 14, a machining path analysis unit 15, and a machining environment acquisition unit 16 instead of the input unit 7. Note that the machining condition adjustment device 10B does not necessarily have to include any of the machining state acquisition unit 13, the machine characteristic acquisition unit 14, the machining path analysis unit 15, and the machining environment acquisition unit 16.

[0069] The required specification input unit 11 receives the required specifications input by the operator 5 and inputs them to the test item inference device 30 and the parameter adjustment unit 41. The processing result input unit 12 receives the processing results input by the operator 5 and inputs them to the processing result model update device 20.

[0070] The machining state acquisition unit 13 acquires the machining state during the machining test from the wire electric discharge machine 1 and inputs it to the machining result model update device 20. The machining state is information such as the machining speed, voltage, and number of discharge pulses during the machining test by the wire electric discharge machine 1.

[0071] The machine characteristic acquisition unit 14 acquires the machine characteristics from the wire electric discharge machine 1 and inputs them to the test item inference device 30 and the machining result model update device 20. The machine characteristics are information on differences in machine configuration between the wire electric discharge machine 1 and other wire electric discharge machines, differences between machines, etc. The differences between machines are caused by assembly errors of members in the wire electric discharge machine 1, etc.

[0072] The machine configuration is, for example, the size of a table T and a column (not shown) equipped in the wire electric discharge machine 1. Depending on the user of the wire electric discharge machine 1, there may be cases where parts are added to or removed from the wire electric discharge machine 1 at a later date, so the machine characteristic acquisition unit 14 acquires information on differences in the machine configuration from other wire electric discharge machines.

[0073] The assembly error of the components is the mounting angle of the rollers of the wire transport system (wire electrode transport roller 83, lower roller 84, etc.). Since the angle of the roller changes over time, the mechanical property acquisition unit 14 acquires information on the assembly error of the components. The mechanical property acquisition unit 14 acquires pre-stored mechanical properties (initial values) and changed mechanical properties (latest values).

[0074] The machining trajectory analysis unit 15 acquires the machining trajectory and a position command program corresponding to the machining trajectory from the wire electric discharge machine 1. The machining trajectory is a machining trajectory input to the wire electric discharge machine 1 to obtain a desired workpiece 4 (finished product) in a machining test by the wire electric discharge machine 1. The machining trajectory analysis unit 15 analyzes the size and complexity of the machining trajectory based on the machining trajectory.

[0075] The complexity of the machining trajectory includes, for example, the angle of the acute angle machining bend contained in the machining trajectory, the number of acute angle machining bends, the number of consecutive arc shapes contained in the machining trajectory, and the number of locations where multiple machining trajectories are adjacent.

[0076] In addition, the machining trajectory analysis unit 15 extracts the machining offset amount and machining direction defined in the position command program by analyzing the position command program corresponding to the machining trajectory. The machining trajectory analysis unit 15 inputs the analysis result to the test item inference device 30 and the machining result model update device 20. Specifically, the machining trajectory analysis unit 15 inputs at least one of the size of the machining trajectory, the complexity of the machining trajectory, the machining offset amount, and the machining direction to the test item inference device 30 and the machining result model update device 20.

[0077] The machining environment acquisition unit (machining environment receiving unit) 16 receives the machining environment, which is the machining environment, from the wire electric discharge machine 1, and inputs it to the test item inference device 30 and the machining result model update device 20. The machining environment includes the ambient temperature of the wire electric discharge machine 1, the degree of deterioration of parts, the setup state, the nozzle separation amount, the material of the workpiece 4, the wire type of the wire electrode 6, etc., during a machining test using the wire electric discharge machine 1.

[0078] When the machining environment is the deterioration degree of a part, the deterioration degree of the part may be estimated based on, for example, the total machining time and the machining contents, or may be detected using a sensor that detects information corresponding to the deterioration degree. Also, the deterioration degree of the part may be calculated based on, for example, a current value and a voltage value.

[0079] The test item inference device 30 of the machining condition adjustment device 10B includes a derivation data acquisition unit 31, a derivation data evaluation unit 32, a test item derivation unit 33, and a test item storage unit 34. The derivation data acquisition unit 31 receives the required specifications from the required specifications input unit 11, and receives the machine characteristics from the machine characteristics acquisition unit 14. The derivation data acquisition unit 31 also receives the analysis result of the machining trajectory (hereinafter referred to as the trajectory analysis result) from the machining trajectory analysis unit 15, and receives the machining environment from the machining environment acquisition unit 16. The derivation data acquisition unit 31 sends the required specifications, the machine characteristics, the trajectory analysis result, and the machining environment to the derivation data evaluation unit 32 as derivation data of the machining test item (test item derivation data).

[0080] The derivation data evaluation unit 32 weights each of the required specifications, machine characteristics, trajectory analysis results, and machining environment, and sends them to the test item derivation unit 33. That is, the derivation data evaluation unit 32 sends data obtained by weighting each element included in the derivation data of the machining test items to the test item derivation unit 33 as inference data.

[0081] The test item derivation unit 33 derives machining test items for updating the machining result model based on the inference data sent from the derivation data evaluation unit 32 and the machining test item model stored in the test item model storage unit 43 so as to obtain the minimum necessary (minimum number) of machining test items. That is, the test item derivation unit 33 derives machining test items based on the inference data and the machining test item model so as to reduce the number of machining test items as much as possible. In this way, the machining test item model is a learned model that increases or decreases the machining test items according to the usage mode (required specifications) of the operator 5. This allows the machining condition adjustment device 10B to infer machining test items with high accuracy according to the usage mode of the operator 5.

[0082] The machining test items derived by the test item derivation unit 33 are machining test items in which at least one of the machining condition parameters, the machine characteristics, the machining trajectory, and the machining environment has been changed from the reference value. Note that the data for deriving the machining test items only needs to include at least one of the machine characteristics, the trajectory analysis result, and the machining environment, and the required specifications.

[0083] For example, when the machining test item model is a linear model, the test item derivation unit 33 acquires two points separated by a specific distance as teacher data and derives machining test items by linear approximation. Note that when the two points obtained by the setting contents of the test items are close to each other, the test item derivation unit 33 derives machining test items by nonlinear approximation without linear approximation.

[0084] When performing nonlinear approximation, the test item derivation unit 33 derives machining test items and creates an experiment plan based on, for example, Bayesian optimization or constraint conditions. In this method, machining tests are performed in advance using various test patterns, and the test item derivation unit 33 derives as few machining test items as possible as understood from the test results of the machining tests as machining test items for updating the machining result model.

[0085] The test item storage unit 34 stores the machining test items derived by the test item derivation unit 33. The test item storage unit 34 sends the machining test items to the display unit 44 and the machining result model update device 20.

[0086] The machining result model updating device 20 of the machining condition adjusting device 10B has an update data acquiring section 21, an update data evaluating section 22, an updating section 23, and an update model storing section 24.

[0087] The update data acquisition unit 21 reads out machining test items from the test item storage unit 34. Moreover, after the machining test, the update data acquisition unit 21 receives the machining result from the machining result input unit 12 and receives the machining state from the machining state acquisition unit 13. Moreover, after the machining test, the update data acquisition unit 21 receives the machine characteristics from the machine characteristics acquisition unit 14, receives the trajectory analysis result from the machining trajectory analysis unit 15, and receives the machining environment from the machining environment acquisition unit 16.

[0088] The update data acquisition unit 21 sends the machining test items, machining results, machining states, machine characteristics, trajectory analysis results, and machining environment as model update data for the reference machining result model to the update data evaluation unit 22. Note that the model update data for the reference machining result model only needs to include at least one of the machining test items, machining states, machine characteristics, trajectory analysis results, and machining environment, and the machining results.

[0089] The update data evaluation unit 22 weights each of the machining test items, machining states, machine characteristics, trajectory analysis results, and machining environments. That is, the update data evaluation unit 22 weights each element included in the model update data. The update data evaluation unit 22 sends data including the weighted data and the machining results to the update unit 23 as inference data.

[0090] The update unit 23 updates the reference machining result model based on the inference data sent from the update data evaluation unit 22 and the reference machining result model stored in the reference machining result model storage unit 45. Specifically, the update unit 23 compares the inference data with the reference machining result model, and updates the reference machining result model to an updated machining result model so as to match the machining test items, machining results, machining state, machine characteristics, trajectory analysis results, and machining environment.

[0091] The updated model storage unit 24 stores the updated processing result model generated by the update unit 23. The updated model storage unit 24 sends the updated processing result model to the display unit 44, the parameter adjustment unit 41, and the processing result estimation unit 42. The updated processing result model updated by the processing result model update device 20 is displayed on the display unit 44 and presented to the worker 5.

[0092] The update unit 23 may further update the updated updated processing result model. In this case, the update unit 23 further updates the updated processing result model based on the inference data sent from the update data evaluation unit 22 and the updated processing result model stored in the update model storage unit 24. Then, when the updated processing result model is updated by the update unit 23, the update model storage unit 24 stores the updated reference processing result model as the latest updated processing result model. The update model storage unit 24 stores the latest updated processing result model every time the updated processing result model is updated by the update unit 23.

[0093] The operator 5 uses the wire electric discharge machine 1 to perform a machining test to update the reference machining result model in accordance with the machining test items displayed on the display unit 44, and inputs the machining results such as the dimensional error and surface roughness of the workpiece 4 obtained by the machining test into the machining result input unit 12.

[0094] The machining condition adjustment device 10B may have a storage unit having the functions of both the reference machining result model storage unit 45 and the updated model storage unit 24.

[0095] The parameter adjustment unit 41 of the machining condition adjustment device 10B calculates machining condition parameters corresponding to the required specifications by applying the required specifications to the updated machining result model. The parameter adjustment unit 41 also calculates parameter adjustment amounts based on the difference between the reference preset machining condition parameters and the calculated machining condition parameters.

[0096] The parameter adjustment unit 41 calculates the machining condition parameters using an updated machining result model in which adjustments are made to the reference machining result model according to the machining test items, machining results, machining states, machine characteristics, trajectory analysis results, and machining environment. Therefore, the machining condition parameters calculated by the parameter adjustment unit 41 in the third embodiment can be said to be adjusted machining condition parameters. That is, the parameter adjustment unit 41 calculates the machining condition parameters for obtaining machining results that satisfy the machining test items, machining results, machining states, machine characteristics, trajectory analysis results, machining environment, and required specifications. The machining result estimation unit 42 estimates the estimated machining result by the same process as in the first embodiment.

[0097] In addition, when the machining condition adjustment device 10B does not use the machining state, the machining condition adjustment device 10B may not have the machining state acquisition unit 13. In addition, when the machining condition adjustment device 10B does not use the machine characteristic, the machining condition adjustment device 10B may not have the machine characteristic acquisition unit 14. In addition, when the machining condition adjustment device 10B does not use the trajectory analysis result, the machining condition adjustment device 10B may not have the machining trajectory analysis unit 15. In addition, when the machining condition adjustment device 10B does not use the machining environment, the machining condition adjustment device 10B may not have the machining environment acquisition unit 16.

[0098] 6 is a flowchart showing a processing procedure of a process in which the machining condition adjustment device according to the third embodiment derives machining test items. Here, a process in which the machining condition adjustment device 10B derives machining test items using the test item inference device 30 will be described.

[0099] The derivation data acquisition unit 31 acquires the required specifications, machine characteristics, trajectory analysis results, and machining environment as derivation data for machining test items (step S210). The derivation data evaluation unit 32 evaluates the derivation data sent from the derivation data acquisition unit 31 (step S220). That is, the derivation data evaluation unit 32 weights the derivation data for the machining test items. The derivation data evaluation unit 32 sends the weighted data to the test item derivation unit 33 as inference data.

[0100] The test item derivation unit 33 inputs the inference data sent from the derivation data evaluation unit 32 into the processed test item model stored in the test item model storage unit 43 (step S230). The test item derivation unit 33 inputs the inference data into the processed test item model to derive processed test items (step S240).

[0101] The test item storage unit 34 stores the processing test items derived by the test item derivation unit 33 (step S250). The display unit 44 displays the processing test items stored in the test item storage unit 34 (step S260).

[0102] 7 is a flowchart showing a processing procedure of a process in which the machining condition adjusting device according to the third embodiment updates a machining result model. Here, a process in which the machining condition adjusting device 10B updates a reference machining result model to an updated machining result model using the machining result model update device 20 will be described.

[0103] The update data acquisition unit 21 acquires machining test items, machining results, machining states, machine characteristics, trajectory analysis results, and machining environments as model update data for the reference machining result model (step S310).

[0104] The update data evaluation unit 22 evaluates the model update data sent from the update data acquisition unit 21 (step S320). That is, the update data evaluation unit 22 weights the model update data of the reference processing result model. The update data evaluation unit 22 sends the weighted data to the update unit 23 as inference data.

[0105] The update unit 23 compares the inference data sent from the update data evaluation unit 22 with the reference processing result model stored in the reference processing result model storage unit 45 (step S330).

[0106] The update unit 23 updates the reference machining result model to an updated machining result model so as to match the machining test items, machining results, machining states, machine characteristics, trajectory analysis results, and machining environment (step S340).

[0107] The updated model storage unit 24 stores the updated processing result model generated by the update unit 23 (step S350). The updated model storage unit 24 sends the updated processing result model to the display unit 44, the parameter adjustment unit 41, and the processing result estimation unit .

[0108] The parameter adjustment unit 41 calculates the machining condition parameters (adjusted machining condition parameters) based on the updated machining result model and the required specifications (step S360). The machining result estimation unit 42 estimates the estimated machining result based on the updated machining result model and the adjusted machining condition parameters (step S370). The display unit 44 displays the updated machining result model, the parameter adjustment amount, the adjusted machining condition parameters, and the estimated estimated machining result (step S380).

[0109] In addition, the display unit 44 may display the parameter adjustment amount and the adjusted machining condition parameter at any timing after the parameter adjustment amount and the adjusted machining condition parameter are calculated. In addition, the display unit 44 may display the updated machining result model at any timing after the updated machining result model is generated.

[0110] Incidentally, the machining conditions applied to the wire electric discharge machine 1 need to be prepared for each wire diameter and thickness of the workpiece 4. In addition, the machining conditions of the wire electric discharge machine 1 are made up of a plurality of machining condition parameters, and the combinations of the machining condition parameters are diverse.

[0111] Furthermore, the machining environment, such as the ambient temperature of the wire electric discharge machine 1, the degree of deterioration of parts, and the setup state, as well as the machining trajectory for obtaining a desired machined shape, vary widely depending on the type of business of the worker 5. For this reason, each wire electric discharge machine 1 has different mechanical characteristics, such as differences between models due to differences in mechanical configuration and the assembly accuracy of each component, making it difficult to adjust the machining condition parameters taking all usage patterns into consideration.

[0112] In the third embodiment, the machining condition adjustment device 10B adjusts the machining condition parameters based on the machine characteristics, machining trajectory, machining environment, etc., so that the machining condition parameters can be easily adjusted in consideration of all usage patterns.

[0113] Thus, according to the third embodiment, the machining condition adjustment device 10B derives machining test items based on the required specifications, machine characteristics, trajectory analysis results, and machining environment. Then, when a machining test corresponding to the machining test item is performed, the machining condition adjustment device 10B updates the reference machining result model to an updated machining result model based on the machining test item, machining result, machining state, machine characteristics, machining trajectory, and machining environment. In addition, the machining condition adjustment device 10B adjusts the machining condition parameters based on the updated updated machining result model. Therefore, the machining condition adjustment device 10B can estimate an estimated machining result that satisfies the required specifications based on the required specifications, machining test items, machining results, machining state, machine characteristics, machining trajectory, and machining environment.

[0114] Moreover, the machining condition adjustment device 10B can estimate the estimated machining result with high accuracy before the actual machining is performed, and the operator 5 can know the machining condition parameters that can obtain the highly accurate machining result before the actual machining is performed. Moreover, the machining condition adjustment device 10B can minimize the machining test items required for updating the reference machining result model, and can reduce the burden on the operator 5 for the machining test.

[0115] In addition, the machining condition adjustment device 10B updates the reference machining result model to the updated machining result model based on the machining offset amount. Therefore, the machining condition adjustment device 10B can calculate accurate machining condition parameters corresponding to the required specifications using the updated machining result model.

[0116] Embodiment 4 Next, a fourth embodiment will be described with reference to Fig. 8. In the fourth embodiment, a machining condition adjustment device 10C described later receives a machining environment input from an operator 5, infers machining test items, and updates a reference machining result model to an updated machining result model.

[0117] Fig. 8 is a diagram showing the configuration of a machining condition adjustment device according to embodiment 4. Among the components in Fig. 8, components that achieve the same functions as those in the machining condition adjustment device 10B according to embodiment 3 shown in Fig. 5 are given the same reference numerals, and duplicated explanations will be omitted.

[0118] The machining condition adjustment device 10C of the fourth embodiment is a computer that adjusts machining condition parameters, similar to the machining condition adjustment device 10B. The machining condition adjustment device 10C is connected to an electric discharge machine such as the wire electric discharge machine 1.

[0119] Compared to the machining condition adjustment device 10B, the machining condition adjustment device 10C includes a machining environment input unit 17 instead of the machining environment acquisition unit 16. The machining condition adjustment device 10C may not include any of the machining state acquisition unit 13, the machine characteristic acquisition unit 14, the machining trajectory analysis unit 15, and the machining environment input unit 17.

[0120] The machining environment input unit (machining environment receiving unit) 17 receives the machining environment input from the operator 5 and inputs it to the test item inference device 30 and the machining result model update device 20. In this way, the machining condition adjustment device 10B acquires the machining environment from the wire electric discharge machine 1, while the machining condition adjustment device 10C acquires the machining environment from the operator 5.

[0121] In addition, the process in which the machining condition adjustment device 10C derives machining test items and updates the machining result model is the same as the process in which the machining condition adjustment device 10B derives machining test items and updates the reference machining result model, so the description thereof will be omitted.

[0122] Thus, according to the fourth embodiment, the machining condition adjustment device 10C can derive machining test items and update the machining result model based on the machining environment input by the operator 5. Therefore, like the machining condition adjustment device 10B, the machining condition adjustment device 10C can estimate an estimated machining result that satisfies the required specifications based on the machining test items, machining results, machining state, machine characteristics, machining trajectory, and machining environment.

[0123] Embodiment 5. Next, a fifth embodiment will be described with reference to Fig. 9. In the fifth embodiment, a machining condition adjusting device 10D described later determines whether or not the reference machining result model needs to be updated based on the machining result when machining is performed under standard machining conditions derived based on the reference machining result model and the required specifications.

[0124] Fig. 9 is a diagram showing the configuration of a machining condition adjustment device according to embodiment 5. Among the components in Fig. 9, components that achieve the same functions as those in the machining condition adjustment device 10B according to embodiment 3 shown in Fig. 5 are given the same reference numerals, and duplicated explanations will be omitted.

[0125] The machining condition adjustment device 10D of the fifth embodiment is a computer that adjusts machining condition parameters, similar to the machining condition adjustment device 10B. The machining condition adjustment device 10D is connected to an electric discharge machine such as the wire electric discharge machine 1.

[0126] The machining condition adjustment device 10D includes an update necessity determination unit 46 in addition to the components included in the machining condition adjustment device 10B. In the machining condition adjustment device 10D, the required specification input unit 11 inputs the required specification input by the operator 5 to the update necessity determination unit 46. In addition, the machining result input unit 12 inputs the machining result input by the operator 5 to the update necessity determination unit 46.

[0127] The processing result input by the worker 5 to the processing result input unit 12 is a processing result obtained when processing is performed under standard processing conditions derived based on the reference processing result model.

[0128] The update necessity determination unit 46 determines whether or not the reference processed result model needs to be updated based on the required specifications and the processing result. The update necessity determination unit 46 determines that the reference processed result model does not need to be updated when the difference between the required specifications and the processing result is equal to or less than a specific value. On the other hand, the update necessity determination unit 46 determines that the reference processed result model needs to be updated when the difference between the required specifications and the processing result is greater than a specific value.

[0129] When the update necessity determination unit 46 determines that the reference processing result model needs to be updated, the operator 5 inputs an instruction to start updating the reference processing result model to the processing condition adjustment device 10D. As a result, the processing condition adjustment device 10D starts updating the reference processing result model and estimates the estimated processing result. The processing condition adjustment device 10D executes, for example, the process described in FIG. 7.

[0130] The update necessity determination unit 46 may determine whether or not further updating of the updated processing result model is necessary based on the required specifications and the processing result. The update necessity determination unit 46 determines that updating of the updated processing result model is unnecessary when the difference between the required specifications and the processing result is equal to or less than a specific value. On the other hand, the update necessity determination unit 46 determines that further updating of the updated processing result model is necessary when the difference between the required specifications and the processing result is greater than a specific value.

[0131] The update necessity determination unit 46 sends an update necessity signal indicating the determination result to the display unit 44. The display unit 44 displays whether or not the reference processing result model needs to be updated based on the update necessity signal. When it is displayed that the reference processing result model needs to be updated, the reference processing result model is updated. In this case, the update process of the reference processing result model described in any one of the first to fourth embodiments is executed. For example, the processing condition adjustment device 10D executes the process described in FIG. 6 and FIG. 7.

[0132] Thus, according to embodiment 5, the processing condition adjustment device 10D can determine whether or not the reference processing result model needs to be updated based on the required specifications and the processing result. Therefore, if the reference processing result model does not need to be updated, the worker 5 does not need to have the processing condition adjustment device 10D execute the adjustment process of the processing condition parameters.

[0133] Embodiment 6 Next, a sixth embodiment will be described with reference to Fig. 10. In the sixth embodiment, a machining condition adjusting device 10E, which will be described later, automatically measures the machining results obtained by the wire electric discharge machine 1.

[0134] Fig. 10 is a diagram showing the configuration of a machining condition adjustment device according to embodiment 6. Among the components in Fig. 10, components that achieve the same functions as those in the machining condition adjustment device 10B according to embodiment 3 shown in Fig. 5 are given the same reference numerals, and duplicated explanations will be omitted.

[0135] The machining condition adjustment device 10E of the sixth embodiment is a computer that adjusts machining condition parameters, similar to the machining condition adjustment device 10B. The machining condition adjustment device 10E is connected to an electric discharge machine such as the wire electric discharge machine 1.

[0136] Compared to the machining condition adjustment device 10B, the machining condition adjustment device 10E includes a machining result measuring device 18 instead of the machining result input unit 12. The machining result measuring device 18 is connected to the wire electric discharge machine 1 and automatically measures the machining results by the wire electric discharge machine 1. The machining result measuring device 18 measures, for example, the dimensional error from the desired value of the workpiece 4 obtained by a machining test, the surface roughness, the time required for machining, the power consumption required for machining, and the like. The machining result measuring device 18 sends the measured machining results to the machining result model update device 20.

[0137] In addition, the process of deriving the machining test items and updating the machining result model by the machining condition adjustment device 10E is the same as the process of deriving the machining test items and updating the reference machining result model by the machining condition adjustment device 10B, so the description thereof will be omitted. In addition, the third to sixth embodiments may be combined.

[0138] As described above, according to the sixth embodiment, the processing result measuring device 18 automatically measures the processing result without the operator 5 having to measure the processing result, so that the processing condition adjustment device 10E can reduce the burden on the operator 5.

[0139] Embodiment 7 Next, a seventh embodiment will be described with reference to Fig. 11. In the seventh embodiment, a machining condition adjusting device 10B updates a reference machining result model to an updated machining result model so as to obtain a machining result according to a desired required specification.

[0140] Fig. 11 is a diagram for explaining a process in which the machining condition adjustment device according to the seventh embodiment updates a machining result model. The horizontal axis of the graph shown in Fig. 11 represents the machining condition parameter, and the vertical axis represents the machining result. In the following, a case will be described in which the machining condition adjustment device according to the seventh embodiment is the machining condition adjustment device 10B, but the machining condition adjustment device according to the seventh embodiment may be any of the machining condition adjustment devices 10C to 10E.

[0141] In the seventh embodiment, a machining test is performed in advance before the machining condition parameters are adjusted. As a result, a reference machining result model Bm showing the relationship between the machining result and the machining condition parameters is derived. In this case, depending on the machine characteristics, machining trajectory, and machining environment, the machining result may not be obtained according to the reference machining result model Bm. Therefore, the machining condition adjustment device 10B updates the machining result model from the reference machining result model Bm to the updated machining result model Um. That is, the machining condition adjustment device 10B updates the relationship between the machining condition parameters and the machining result. As a result, the machining condition adjustment device 10B calculates the parameter adjustment amount Pa based on the updated updated machining result model so that the machining result according to the desired required specifications can be obtained, and adjusts the machining condition parameters.

[0142] The processing condition adjustment device 10B may update the processing result model by parallel translation of a simple linear regression model, or may update the reference processing result model using a complex AI (Artificial Intelligence) model such as supervised learning, unsupervised learning, reinforcement learning, etc., which will be described later.

[0143] When updating the reference machining result model using an AI model, the machining condition adjustment device 10B executes a process following the process performed when the reference machining result model Bm was derived. For example, when the reference machining result model is generated using a neural network described later, the machining condition adjustment device 10B uses the neural network used when the reference machining result model was generated as is when generating the updated machining result model. As a result, the machining condition adjustment device 10B can generate the updated machining result model Um by executing re-learning from the process following the process when the reference machining result model Bm was derived.

[0144] Thus, according to embodiment 7, the processing condition adjustment device 10B can adjust the processing condition parameters for obtaining processing results that satisfy the required specifications based on the updated processing result model that is updated to suit the usage pattern of the worker 5.

[0145] Embodiment 8 Next, an eighth embodiment will be described with reference to Fig. 12 to Fig. 15. In the eighth embodiment, a processing result model learning device 50, which will be described later, generates a reference processing result model.

[0146] 12 is a diagram showing a configuration of a processing result model learning device according to the eighth embodiment. The processing result model learning device 50 is a device that generates a reference processing result model by learning. The processing result model learning device 50 includes a learning data acquisition unit 51, a learning data evaluation unit 52, and a processing result model derivation unit 53.

[0147] The learning data acquisition unit 51 acquires learning data for learning the reference machining result model. Specifically, the learning data acquisition unit 51 acquires the machining test item D1, the machining state D2, the machine characteristic D3, the trajectory analysis result D4 which is the analysis result of the machining trajectory, the machining environment D5, and the machining result D6 as the learning data. Here, the learning data is data in which the machining test item D1, the machining state D2, the machine characteristic D3, the trajectory analysis result D4, the machining environment D5, and the machining result D6 are associated with each other.

[0148] The learning data acquisition unit 51 simultaneously acquires learning data including, for example, machining test items D1, machining states D2, machine characteristics D3, trajectory analysis results D4, machining environment D5, and machining results D6.

[0149] The learning data acquiring unit 51 acquires learning data from the worker 5. The learning data acquiring unit 51 may acquire at least a part of the learning data from the wire electric discharge machine 1, or may acquire at least a part of the learning data from another external device. The learning data acquiring unit 51 sends the acquired learning data to the learning data evaluating unit 52.

[0150] The learning data evaluation unit 52 weights the machining test items D1, the machining state D2, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5, and sends them to the machining result model derivation unit 53. That is, the learning data evaluation unit 52 sends the data obtained by weighting each of the learning data and the machining result D6 to the machining result model derivation unit 53 as learning data.

[0151] The processing result model derivation unit 53 derives a reference processing result model for inferring the processing result based on the learning data sent from the learning data evaluation unit 52. The derived reference processing result model is stored in the reference processing result model storage unit 45.

[0152] The machining result model learning device 50 may be disposed inside the machining condition adjustment devices 10B to 10E or the wire electric discharge machine 1. Moreover, the machining result model learning device 50 may be a device connected to the machining condition adjustment devices 10B to 10E or the wire electric discharge machine 1 via a network and separate from the machining condition adjustment devices 10B to 10E or the wire electric discharge machine 1. Moreover, the machining result model learning device 50 may exist on a cloud server.

[0153] In the above description, the learning data acquisition unit 51 simultaneously acquires the learning data including the machining test item D1, the machining state D2, the machine characteristic D3, the trajectory analysis result D4, the machining environment D5, and the machining result D6, but the learning data may be acquired at different timings. That is, the learning data acquisition unit 51 only needs to be able to accept the machining test item D1, the machining state D2, the machine characteristic D3, the trajectory analysis result D4, the machining environment D5, and the machining result D6 in association with each other, and may acquire the machining test item D1, the machining state D2, the machine characteristic D3, the trajectory analysis result D4, the machining environment D5, and the machining result D6 at different timings.

[0154] In addition, the machining result model learning device 50 may derive a reference machining result model that associates at least one of the machining test items D1, the machining state D2, the machine characteristics D3, the trajectory analysis results D4, and the machining environment D5 with the machining result D6.

[0155] The learning algorithm used by the processing result model derivation unit 53 may be a known learning algorithm such as supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. Also, the learning algorithm used by the processing result model derivation unit 53 may be deep learning that learns to extract the feature amount itself. Also, the processing result model derivation unit 53 may perform machine learning according to other known methods, such as genetic programming, functional logic programming, or support vector machine. Also, the processing result model derivation unit 53 may generate a reference processing result model by mathematical optimization.

[0156] Here, as an example, a case where a neural network is applied to the learning algorithm used by the processing result model derivation unit 53 will be described. The processing result model derivation unit 53 learns a reference processing result model by so-called supervised learning according to, for example, a neural network model. In this case, the learning data is the supervised data. Here, supervised learning refers to a method of providing a learning device with a set of data of input and result (label), learning the features in the learning data, and inferring the result from the input.

[0157] A neural network is composed of an input layer consisting of multiple neurons, an intermediate layer (hidden layer) consisting of multiple neurons, and an output layer consisting of multiple neurons. The intermediate layer may be one layer or two or more layers.

[0158] Fig. 13 is a diagram for explaining a neural network used by the processing result model learning device according to the eighth embodiment. For example, in the case of a three-layered neural network as shown in Fig. 18, when a plurality of inputs are input to the input layer (X1 to X3), the values ​​are multiplied by weights W1 (w11 to w16) and input to the intermediate layer (Y1 to Y2). The results are further multiplied by weights W2 (w21 to w26) and output from the output layer (Z1 to Z3). This output result varies depending on the values ​​of weights W1 and W2.

[0159] The neural network used by the processing result model learning device 50 in Figure 12 learns a reference processing result model by so-called supervised learning according to learning data created based on a combination of a target variable and an explanatory variable acquired by the learning data acquisition unit 51.

[0160] The neural network used by the machining result model learning device 50 learns the machining result D6 corresponding to the machining test item D1, the machining state D2, the machine characteristic D3, the trajectory analysis result D4, and the machining environment D5. In other words, the neural network learns the second input corresponding to the first input by so-called supervised learning according to the learning data created based on the combination of the first input and the second input (correct answer) acquired by the learning data acquisition unit 51.

[0161] That is, the neural network used by the machining result model learning device 50 learns by forward propagation by inputting the machining test items D1, machining states D2, machine characteristics D3, trajectory analysis results D4, and machining environment D5, and adjusting the weights W1 and W2 so that the results output from the output layer approach the correct machining result D6 (correct answer).

[0162] In addition, the neural network used by the machining result model learning device 50 learns by adjusting the weights W1 and W2 through back propagation so that the machining result D6 is input to the output layer and the result output from the input layer approaches the correct machining test item D1, machining state D2, machine characteristic D3, trajectory analysis result D4, and machining environment D5.

[0163] The neural network learns the correspondence between the machining test items D1, machining states D2, machine characteristics D3, trajectory analysis results D4, and machining environment D5, and the machining results D6, and generates a reference machining result model that can output appropriate inference results when inference data is input.

[0164] For example, the neural network outputs an appropriate machining result D6 when a machining test item D1, a machining state D2, a machine characteristic D3, a trajectory analysis result D4, and a machining environment D5 are input. Also, when a machining result D6 is input, the neural network outputs an appropriate machining test item D1, a machining state D2, a machine characteristic D3, a trajectory analysis result D4, and a machining environment D5.

[0165] When the parameter adjustment unit 41 uses a reference machining result model, the required specifications are input to the reference machining result model to calculate machining test items D1, machining state D2, machine characteristics D3, trajectory analysis result D4, and machining environment D5, and output them to the machining result estimation unit 42. When the machining result estimation unit 42 uses a reference machining result model, the machining test items D1, machining state D2, machine characteristics D3, trajectory analysis result D4, and machining environment D5 received from the parameter adjustment unit 41 are input to the reference machining result model to calculate an estimated machining result. The required machining result included in the required specifications is information indicating the range of the machining result, and the estimated machining result is information indicating a specific value of the machining result. The estimated machining result is within the range of the required machining result, and is within the range desired by the operator 5.

[0166] The processing result model derivation unit 53 generates and outputs a reference processing result model by executing the above-mentioned learning. The reference processing result model storage unit 45 stores the reference processing result model output from the processing result model derivation unit 53.

[0167] 14 is a flowchart showing the procedure of processing executed by the machining result model learning device according to the eighth embodiment. Here, the procedure of processing in which the machining result model learning device 50 learns a reference machining result model will be described.

[0168] The learning data acquisition unit 51 acquires the machining test items D1, the machining state D2, the machine characteristics D3, the trajectory analysis results D4, the machining environment D5, and the machining results D6 as learning data (step S410).

[0169] The learning data evaluation unit 52 evaluates the machining test items D1, machining states D2, machine characteristics D3, trajectory analysis results D4, and machining environment D5 sent from the learning data acquisition unit 51 (step S420). That is, the learning data evaluation unit 52 weights the machining test items D1, machining states D2, machine characteristics D3, trajectory analysis results D4, and machining environment D5. The learning data evaluation unit 52 sends the weighted data and the machining result D6 to the machining result model derivation unit 53 as learning data.

[0170] The processing result model derivation unit 53 derives a reference processing result model based on the learning data sent from the learning data evaluation unit 52 (step S430). The reference processing result model storage unit 45 stores the reference processing result model derived by the processing result model derivation unit 53 (step S440).

[0171] The wire electric discharge machine 1 performs machining that satisfies the final required specifications (surface roughness) by, for example, successively executing machining steps. That is, the wire electric discharge machine 1 performs machining that gradually approaches the desired required specifications while repeatedly going back and forth along a similar machining trajectory to obtain the desired machining result. For this reason, the machining result model learning device 50 may derive a reference machining result model for each machining step included in the multi-step machining.

[0172] FIG. 15 is a diagram for explaining a plurality of steps from which the machining result model learning device according to the eighth embodiment derives a reference machining result model. The horizontal axis of the graph shown in FIG. 15 is the machining step, and the vertical axis is the machining result. The machining result model learning device 50 generates a reference machining result model used by the machining condition adjustment devices 10B to 10E. Note that, although the following describes a case where the machining result is surface roughness, the machining result may be something other than surface roughness.

[0173] 15 shows a case where the wire electric discharge machine 1 machines a workpiece 4 by executing a first machining step (1st) to a fifth machining step (5th). When the wire electric discharge machine 1 executes machining steps successively, each machining step has a machineable range, which is the range of machining results that can be produced.

[0174] The possible machining range is the range of machining results that can be achieved when machining is performed with variously changed machining condition parameters. For example, if there are three types of machining condition parameters and 10 types of parameter values ​​for each machining condition parameter, there are 1000 possible combinations of the machining condition parameters. In this case, the possible machining range is the range of machining results when machining is performed with the 1000 different machining condition parameters.

[0175] There are machining steps with a wide machining range and machining steps with a narrow machining range. The wire electric discharge machine 1 changes the machining result at each machining step, and finally obtains a machining result that satisfies the required specifications. In this case, the path along which the machining result changes will be different when high-speed machining is performed within the machining range and when accurate machining is performed. That is, the path along which the machining result changes will differ depending on whether emphasis is placed on machining speed or dimensional accuracy. In the eighth embodiment, the machining result model learning device 50 includes the changes in these machining results in the modeling target of the reference machining result model.

[0176] In Fig. 15, the machinable range in the first machining step is illustrated as machinable range R1. The wire electric discharge machine 1 performs machining under machining conditions that result in machining within the machinable range R1. In this case, for example, when high-speed machining is performed, the surface roughness resulting from the machining is large, and when low-speed machining (high-precision machining) is performed, the surface roughness resulting from the machining is small. In Fig. 15, the machining result when high-speed machining is performed in the first machining step is shown as machining result Fa, and the machining result when low-speed machining is performed is shown as machining result Fb. In the wire electric discharge machine 1, the machinable range in the next (next) machining step differs depending on the machining result (surface roughness) in the previous (previous) machining step.

[0177] 15, the processable ranges in the second to fourth machining steps when high-speed machining are performed are shown as processable ranges R2a to R4a, and the processable ranges in the second to fourth machining steps when low-speed machining are performed are shown as processable ranges R2b to R4b.

[0178] The wire electric discharge machine 1 performs machining for each machining step so as to satisfy the final required specifications (surface roughness) regardless of the machining conditions applied. In Fig. 15, the machineable range for the fifth machining step, which is the final machining step, is illustrated as machineable range R5.

[0179] In this way, the machining result at each machining step is affected by the machining result at the previous step. In order for the workpiece 4 to meet the final required specifications, there is an ideal machining result (the shaded area in FIG. 15) that should be achieved at each machining step. This ideal machining result is determined according to the required machining speed, dimensional accuracy, etc. By machining each machining step so that the machining result changes to the ideal machining result, the wire EDM 1 can obtain a machining result that meets the required specifications at the final, fifth machining step.

[0180] In addition, the ideal processing result to which the processing result should progress differs depending on the desired processing result, such as when it is desired to obtain a highly accurate processing result by increasing the total number of processing steps, when it is desired to shorten the processing time by reducing the total number of processing steps as much as possible, etc. For this reason, the processing result model learning device 50 includes the relationship between the ideal processing result of the previous processing step and the ideal processing result of the next processing step in the information used when modeling the reference processing result model.

[0181] The machining result model learning device 50 learns the reference machining result model, for example, by learning the relationship between the ideal machining result of the previous machining step, the ideal machining result of the next machining step, and the machining test items D1, machining conditions D2, machine characteristics D3, trajectory analysis results D4, and machining environment D5 in the next machining step.

[0182] The machining result model learning device 50 learns the relationship between the ideal machining result of the previous machining step, the ideal machining result of the next machining step, and the machining condition parameters in the next machining step, for example, to learn the reference machining result model.

[0183] As described above, according to the eighth embodiment, the machining result model learning device 50 learns the machining result D6 corresponding to the machining test item D1, the machining state D2, the machine characteristic D3, the trajectory analysis result D4, and the machining environment D5, thereby learning the reference machining result model, and can generate a reference machining result model that can infer the machining result with high accuracy. In addition, the machining result model learning device 50 can generate a reference machining result model corresponding to a complex shape.

[0184] In addition, since the machining result model learning device 50 learns the relationship between the ideal machining result of the previous machining step, the ideal machining result of the next machining step, and the machining test items D1, machining state D2, machine characteristics D3, trajectory analysis results D4, and machining environment D5 in the next machining step, it can derive a reference machining result model that takes into account the influence of the previous machining step, and can improve the modeling accuracy of the reference machining result model.

[0185] Embodiment 9 Next, a ninth embodiment will be described with reference to Figures 16 and 17. In the ninth embodiment, a test item model learning device 60, which will be described later, generates a test item model.

[0186] 16 is a diagram showing a configuration of a test item model learning device according to the ninth embodiment. The test item model learning device 60 is a device that generates test item models by learning. The test item model learning device 60 includes a learning data acquisition unit 61, a learning data evaluation unit 62, and a test item model derivation unit 63.

[0187] The learning data acquisition unit 61 acquires learning data for learning the test item model. Specifically, the learning data acquisition unit 61 acquires the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, the machining environment D5, and the machining test items D1 as learning data. Here, the learning data is data in which the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, the machining environment D5, and the machining test items D1 are associated with each other.

[0188] The learning data acquiring unit 61 simultaneously acquires learning data including, for example, the required specifications D11, the machine characteristics D3, the trajectory analysis results D4, the machining environment D5, and the machining test items D1.

[0189] The learning data acquiring unit 61 may acquire at least a part of the learning data from the wire electric discharge machine 1, or may acquire at least a part of the learning data from another external device. The learning data acquiring unit 61 acquires the required specifications D11 from the operator 5. The learning data acquiring unit 61 sends the acquired learning data to the learning data evaluating unit 62.

[0190] The learning data evaluation unit 62 weights the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5, and sends them to the test item model derivation unit 63. That is, the learning data evaluation unit 62 sends the weighted data of the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5, and the machining test items D1, to the test item model derivation unit 63 as learning data.

[0191] The test item model derivation unit 63 derives a test item model for inferring the processing test item D1 based on the learning data sent from the learning data evaluation unit 62. The test item model is stored in the test item model storage unit 43.

[0192] The test item model learning device 60 may be disposed inside the machining condition adjustment devices 10B-10E or the wire electric discharge machine 1. Moreover, the test item model learning device 60 may be a device connected to the machining condition adjustment devices 10B-10E or the wire electric discharge machine 1 via a network, and separate from the machining condition adjustment devices 10B-10E or the wire electric discharge machine 1. Moreover, the test item model learning device 60 may exist on a cloud server.

[0193] Although the learning data acquisition unit 61 simultaneously acquires the learning data including the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, the machining environment D5, and the machining test item D1, the learning data may be acquired at different timings. That is, the learning data acquisition unit 61 only needs to be able to accept the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, the machining environment D5, and the machining test item D1 in association with each other, and may acquire the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, the machining environment D5, and the machining test item D1 at different timings.

[0194] The test item model learning device 60 may derive a test item model that associates at least one of the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5 with the machining test item D1.

[0195] The learning algorithm used by the test item model derivation unit 63 may be a known learning algorithm such as supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, etc. Also, the learning algorithm used by the test item model derivation unit 63 may be deep learning that learns to extract the feature amount itself. Also, the test item model derivation unit 63 may execute machine learning according to other known methods such as genetic programming, functional logic programming, support vector machine, etc. As an example, a case where a neural network is applied to the learning algorithm used by the test item model derivation unit 63 will be described.

[0196] The test item model derivation unit 63 learns the test item model by so-called supervised learning according to a neural network model, for example. In this case, the learning data is the teacher data. The configuration of the neural network used by the test item model derivation unit 63 is the same as the neural network described in FIG. 13 of the eighth embodiment.

[0197] The neural network used by the test item model learning device 60 in Fig. 16 learns a test item model for inferring the machining test item D1 by so-called supervised learning according to the learning data created based on a combination of the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, the machining environment D5, and the machining test item D1 acquired by the learning data acquisition unit 61. That is, the neural network used by the test item model learning device 60 learns the machining test item D1 corresponding to the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5. In other words, the neural network learns the second input corresponding to the first input by so-called supervised learning according to the learning data created based on a combination of the first input and the second input (correct answer) acquired by the learning data acquisition unit 61.

[0198] That is, the neural network used by the test item model learning device 60 learns by inputting the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5, and adjusting the weights W1 and W2 so that the result output from the output layer approaches the machining test item D1 (correct answer).

[0199] The neural network learns the correspondence between the required specifications D11, the machine characteristics D3, the trajectory analysis results D4, and the machining environment D5 and the machining test items D1, and generates a test item model that can output an appropriate machining test item D1 when the required specifications D11, the machine characteristics D3, the trajectory analysis results D4, and the machining environment D5 are input.

[0200] The test item model derivation unit 63 generates and outputs a test item model by executing the above-mentioned learning. The test item model storage unit 43 stores the test item model output from the test item model derivation unit 63.

[0201] 17 is a flowchart showing the procedure of processing executed by the test item model learning device according to the 9th embodiment. Here, the procedure of processing executed by the test item model learning device 60 to learn a test item model will be described.

[0202] The learning data acquiring unit 61 acquires the required specifications D11, the machine characteristics D3, the trajectory analysis results D4, the machining environment D5, and the machining test items D1 as learning data (step S510).

[0203] The learning data evaluation unit 62 evaluates the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5 sent from the learning data acquisition unit 61 (step S520). That is, the learning data evaluation unit 62 weights the required specifications D11, the machine characteristics D3, the trajectory analysis result D4, and the machining environment D5. The learning data evaluation unit 62 sends the weighted data and the machining test items D1 to the test item model derivation unit 63 as learning data.

[0204] The test item model derivation unit 63 derives a test item model based on the learning data sent from the learning data evaluation unit 62 (step S530). The test item model storage unit 43 stores the processing test item model derived by the test item model derivation unit 63 (step S540).

[0205] As described above, according to the ninth embodiment, the test item model is learned by learning the machining test items D1 corresponding to the required specifications D11, the machine characteristics D3, the trajectory analysis results D4, and the machining environment D5, so that the machining test item model capable of inferring the test items with high accuracy can be generated.

[0206] Next, we will explain the hardware configurations of the machining condition adjustment devices 10A to 10E, the machining result model learning device 50, and the test item model learning device 60. Since the machining condition adjustment devices 10A to 10E, the machining result model learning device 50, and the test item model learning device 60 have similar hardware configurations, we will explain the hardware configuration of the machining condition adjustment device 10A here.

[0207] The machining condition adjustment device 10A is realized by a processing circuit. This processing circuit may be a processor and memory that executes a program stored in a memory, or may be dedicated hardware. The processing circuit is also called a control circuit.

[0208] FIG. 18 is a diagram showing a configuration example of a processing circuit in the case where the processing circuit included in the machining condition adjustment device according to the first embodiment is realized by a processor and a memory. The processing circuit 90 shown in FIG. 18 is a control circuit and includes a processor 91 and a memory 92. When the processing circuit 90 is configured using the processor 91 and the memory 92, each function of the processing circuit 90 is realized by software, firmware, or a combination of software and firmware. The software or firmware is described as a program and stored in the memory 92. In the processing circuit 90, each function is realized by the processor 91 reading and executing the program stored in the memory 92. That is, the processing circuit 90 includes a memory 92 for storing a machining condition adjustment program that results in the processing of the machining condition adjustment device 10A being executed. This machining condition adjustment program can also be said to be a program for causing the machining condition adjustment device 10A to execute each function realized by the processing circuit 90. This machining condition adjustment program may be provided by a storage medium in which the program is stored, or may be provided by other means such as a communication medium.

[0209] The test item inference device 30 and the machining result model update device 20 are realized by the processor 91 executing the machining condition adjustment program stored in the memory 92. That is, the machining condition adjustment program executed by the machining condition adjustment device 10A has a modular configuration including the test item inference device 30 and the machining result model update device 20, which are loaded onto the main storage device and generated on the main storage device.

[0210] The processing result model learning program used by the processing result model learning device 50 has a modular configuration including a learning data acquisition unit 51, a learning data evaluation unit 52, and a processing result model derivation unit 53, which are loaded onto the main memory device and generated on the main memory device.

[0211] In addition, the test item model learning program used by the test item model learning device 60 has a modular configuration including a learning data acquisition unit 61, a learning data evaluation unit 62, and a test item model derivation unit 63, which are loaded onto the main memory device and generated on the main memory device.

[0212] Here, the processor 91 is, for example, a CPU (Central Processing Unit), a processing device, an arithmetic device, a microprocessor, a microcomputer, or a DSP (Digital Signal Processor), etc. Also, the memory 92 is, for example, a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable ROM), an EEPROM (Electrically EPROM), a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, or a DVD (Digital Versatile Disc), etc.

[0213] Fig. 19 is a diagram showing an example of a processing circuit in the machining condition adjustment device according to the first embodiment, which is configured with dedicated hardware. The processing circuit 93 shown in Fig. 19 corresponds to, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination of these.

[0214] The processing circuits 90 and 93 may be partially realized by dedicated hardware and partially realized by software or firmware. In this way, the processing circuits 90 and 93 can realize the above-mentioned functions by dedicated hardware, software, firmware, or a combination of these. The machining condition adjustment device 10A may be realized by one processing circuit or multiple processing circuits.

[0215] The configurations shown in the above embodiments are merely examples, and may be combined with other known technologies, or the embodiments may be combined with each other. Also, parts of the configurations may be omitted or modified without departing from the spirit of the invention. [Explanation of symbols]

[0216] 1 Wire electric discharge machine, 2 Control device, 3 Power supply device, 4 Workpiece, 5 Operator, 6 Wire electrode, 7 Input unit, 8a Upper power supply, 8b Lower power supply, 10A to 10E Machining condition adjustment device, 11 Required specification input unit, 12 Machining result input unit, 13 Machining state acquisition unit, 14 Machine characteristic acquisition unit, 15 Machining trajectory analysis unit, 16 Machining environment acquisition unit, 17 Machining environment input unit, 18 Machining result measurement device, 20 Machining result model update device, 21 Update data acquisition unit, 22 Update data evaluation unit, 23 Update unit, 24 Update model storage unit, 30 Test item inference device, 31 Derivation data acquisition unit, 32 Derivation data evaluation unit, 33 Test item derivation unit, 34 Test item storage unit, 41 Parameter adjustment unit, 42 Machining result estimation unit, 43 Test item model storage unit, 44 Display unit, 45 Reference machining result model storage unit, 46 Update necessity judgment unit, 50 Machining result model learning device, 51, 61 Learning data acquisition unit, 52, 62 Learning data evaluation unit, 53 Machining result model derivation unit, 60 Test item model learning device, 63 Test item model derivation unit, 81 Wire electrode bobbin, 82 Wire electrode collection box, 83 Wire electrode transport roller, 84 Lower roller, 85 Wire running speed control motor, 86 Upper guide, 87 Lower guide, 88 Machining unit, 89X X-axis motor, 89Y Y-axis motor, 90, 93 Processing circuit, 91 Processor, 92 Memory, Bm Reference machining result model, D1 Machining test item, D2 Machining state, D3 Machine characteristic, D4 Trajectory analysis result, D5 Machining environment, D6, Fa, Fb Machining result, D11 Required specification, Pa Parameter adjustment amount, R1, R2a~R4a, R2b~R4b, R5 machining range, T surface plate, Um updated machining result model.

Claims

1. a test item inference device that acquires required specifications desired by an operator for electric discharge machining, and derives a minimum number of machining test items from the acquired required specifications using a trained machining test item model for inferring machining test items that are items of machining tests from the required specifications; a machining result model updating device that acquires machining results corresponding to the machining test as model updating data, infers machining condition parameters from the machining results, and updates the reference machining result model based on the acquired model updating data to generate an updated machining result model, which is an updated reference machining result model; a parameter adjusting unit that calculates, as a calculated machining condition parameter, a machining condition parameter for obtaining a machining result that satisfies the required specification from the required specification using the updated machining result model; a processing result estimation unit that estimates an estimated processing result corresponding to the calculation processing condition parameters based on the updated processing result model and the calculation processing condition parameters; Equipped with A processing condition adjusting device characterized by:

2. Further, a display unit is provided that displays the machining test items, the machining condition parameters, and the estimated machining results.

2. The processing condition adjusting device according to claim 1.

3. The processing result model update device acquire the model updating data including the machining state when the electric discharge machine executed the machining test, machine characteristics which are information on differences in machine configuration between the electric discharge machine and other electric discharge machines or differences between models, a trajectory analysis result which is an analysis result of a machining trajectory input to the electric discharge machine in order to obtain a desired workpiece when the electric discharge machine executed the machining test, the machining environment when the electric discharge machine executed the machining test, and at least one of the derived machining test items, and the machining result, and generate the updated machining result model based on the reference machining result model and the acquired model updating data; 3. The machining condition adjusting device according to claim 1 or 2.

4. The processing result model update device generating the updated processed model by weighting each element included in the model updating data; 4. The processing condition adjusting device according to claim 3.

5. The test item inference device acquiring derivation data including at least one of the machine characteristics, the trajectory analysis result, and the machining environment, and the required specifications, and deriving the machining test items from the acquired derivation data using the machining test item model; 4. The processing condition adjusting device according to claim 3.

6. The test item inference device deriving the processing test items by weighting each element included in the derivation data; 6. The processing condition adjusting device according to claim 5.

7. further comprising a machining result measuring device that acquires the machining result from the electric discharge machine and inputs the machining result to the machining result model updating device; 4. The processing condition adjusting device according to claim 3.

8. further comprising an update necessity determination unit that determines whether or not the reference machining result model needs to be updated based on the required specifications and the machining result corresponding to the machining test.

3. The machining condition adjusting device according to claim 1 or 2.

9. The electric discharge machining is a machining including a plurality of steps, and the reference machining result model is a learned model learned based on a machining result of a previous machining step, a machining result of a next machining step, and machining condition parameters in the next machining step.

3. The machining condition adjusting device according to claim 1 or 2.

10. an electric discharge machine for performing electric discharge machining; a machining condition adjusting device that adjusts machining condition parameters used by the electric discharge machine; and The processing condition adjusting device is a test item inference device that acquires required specifications desired by an operator for the electric discharge machining, and derives a minimum number of machining test items from the acquired required specifications using a trained machining test item model for inferring machining test items that are items of the machining test from the required specifications; a machining result model updating device that acquires machining results corresponding to the machining test executed by the electric discharge machine as model updating data, infers the machining condition parameters from the machining results, and updates the reference machining result model based on the acquired model updating data to generate an updated machining result model, which is an updated reference machining result model; a parameter adjusting unit that calculates, as a calculated machining condition parameter, a machining condition parameter for obtaining a machining result that satisfies the required specification from the required specification using the updated machining result model; a processing result estimation unit that estimates an estimated processing result corresponding to the calculation processing condition parameters based on the updated processing result model and the calculation processing condition parameters; Equipped with A processing system characterized by:

11. a test item inference step in which a machining condition adjustment device that adjusts machining condition parameters used by an electric discharge machine acquires required specifications desired by an operator for electric discharge machining, and derives a minimum number of machining test items from the acquired required specifications using a trained machining test item model for inferring machining test items that are items for machining tests from the required specifications; a machining result model updating step in which the machining condition adjusting device acquires machining results corresponding to the machining test as model updating data, and updates the reference machining result model based on a learned reference machining result model for inferring the machining condition parameters from the machining results and inferring the machining results from the machining condition parameters and the acquired model updating data, thereby generating an updated machining result model which is an updated reference machining result model; a parameter adjustment step in which the machining condition adjustment device calculates, as calculated machining condition parameters, machining condition parameters for obtaining a machining result that satisfies the required specifications from the required specifications using the updated machining result model; a machining result estimating step in which the machining condition adjusting device estimates an estimated machining result corresponding to the calculated machining condition parameters based on the updated machining result model and the calculated machining condition parameters; Including, A method for adjusting processing conditions.

12. a test item inference step in which a machining condition adjustment device that adjusts machining condition parameters used by an electric discharge machine acquires required specifications desired by an operator for electric discharge machining, and derives a minimum number of machining test items from the acquired required specifications using a trained machining test item model for inferring machining test items that are items for machining tests from the required specifications; a machining test execution step in which the electric discharge machine executes the machining test; a machining result model updating step in which the machining condition adjusting device acquires machining results corresponding to the machining test as model updating data, and updates the reference machining result model based on a learned reference machining result model for inferring the machining condition parameters from the machining results and inferring the machining results from the machining condition parameters and the acquired model updating data, thereby generating an updated machining result model which is an updated reference machining result model; a parameter adjustment step in which the machining condition adjustment device calculates, as calculated machining condition parameters, machining condition parameters for obtaining a machining result that satisfies the required specifications from the required specifications using the updated machining result model; a machining result estimating step in which the machining condition adjusting device estimates an estimated machining result corresponding to the calculated machining condition parameters based on the updated machining result model and the calculated machining condition parameters; Including, A processing method characterized by:

13. a test item inference step of acquiring required specifications desired by an operator for electric discharge machining, and deriving a minimum number of machining test items from the acquired required specifications using a trained machining test item model for inferring machining test items that are items of a machining test from the required specifications; a machining result model updating step of acquiring machining results corresponding to the machining test as model updating data, inferring machining condition parameters used by the electric discharge machine from the machining results, and updating the reference machining result model based on the acquired model updating data and a learned reference machining result model for inferring the machining results from the machining condition parameters, thereby generating an updated machining result model which is an updated reference machining result model; a parameter adjustment step of calculating machining condition parameters for obtaining a machining result that satisfies the required specifications from the required specifications using the updated machining result model; a processing result estimating step of estimating an estimated processing result corresponding to the calculation processing condition parameters based on the updated processing result model and the calculation processing condition parameters; A processing condition adjustment program that causes a computer to execute the above.

Images