WO2023157308A1

WO2023157308A1 - Wire electrical discharge machining device, wire electrical discharge machining method, learning device, and inference device

Info

Publication number: WO2023157308A1
Application number: PCT/JP2022/006991
Authority: WO
Inventors: 克彦林; 信行太田
Original assignee: 三菱電機株式会社
Priority date: 2022-02-21
Filing date: 2022-02-21
Publication date: 2023-08-24
Also published as: JP7233615B1; JPWO2023157308A1

Abstract

A wire electrical discharge machining device according to the present invention comprises: a machining mechanism (13) for electrical discharge machining; a thickness estimating device (24) that estimates the thickness of a workpiece at a position where machining is being carried out during rough machining of the workpiece by electrical discharge machining; a thickness output device (25) that outputs, to outside of the wire electrical discharge machining device, thickness data in which position data is associated with thickness information indicating the estimated thickness; a thickness input device (26) into which thickness data is input from outside of the wire electrical discharge machining device; a step position estimating device (27) that estimates, on the basis of the inputted thickness data, the position of stepped sections where the thickness of the workpiece changes; an electrical condition control device (22) that controls application of a pulse voltage on the basis of the thickness data and the step information indicating the position of the stepped sections when finishing of a workpiece by electrical discharge machining is being carried out after the rough machining; and a control device that controls the machining mechanism on the basis of the thickness data and the step information when finishing is being carried out.

Description

WIRE EDM MACHINE, WIRE EDM METHOD, LEARNING DEVICE, AND REASONING DEVICE

The present disclosure relates to a wire electric discharge machine, a wire electric discharge machining method, a learning device, and an inference device for electric discharge machining of a workpiece.

For wire electric discharge machines, machining conditions corresponding to the plate thickness of the workpiece to be machined or machining conditions corresponding to the required specifications such as surface roughness are prepared in advance. A worker who performs work using a wire electric discharge machine can perform high-precision machining, for example, a tolerance of 1/100 mm to 1/1000 mm, by appropriately selecting machining conditions suitable for machining from among prepared machining conditions. processing can be realized. However, in the case of a workpiece whose plate thickness changes abruptly in the process of machining one machining shape, a dimensional error may occur at a stepped portion, which is a boundary portion between portions of the workpiece having different plate thicknesses. In addition, streaks may occur on the surface of the stepped portion. Conventionally, in order to prevent the deterioration of machining quality due to the occurrence of dimensional errors or the occurrence of streaks, a method of estimating the step position of the workpiece, that is, the position of the step part, and performing electric discharge machining under the machining conditions suitable for the plate thickness has been proposed. It is

In Patent Document 1, the plate thickness of the workpiece is detected during rough machining of the workpiece, the thickness information is stored, and the step position is estimated and processed based on the thickness information during finish machining of the workpiece. A wire electrical discharge machine with changing conditions is disclosed. According to Patent Literature 1, a wire electric discharge machine stores plate thickness information in a storage area of the wire electric discharge machine.

Japanese Patent No. 6808868

In the technique of Patent Document 1, since the plate thickness information is accumulated in the storage area of the wire electric discharge machine that performed the rough machining, the amount of data of the plate thickness information that can be stored is the capacity of the storage area in the wire electric discharge machine. is limited to Therefore, there is a problem that the amount of plate thickness information that can be used for estimating the step position is limited, which may make it difficult to perform control for improving the processing quality.

The present disclosure has been made in view of the above, and performs control for improving processing quality without limiting the data amount of plate thickness information that can be used for estimating the step position in the work. An object of the present invention is to obtain a wire electric discharge machine capable of

In order to solve the above-described problems and achieve the object, a wire electric discharge machine according to the present disclosure is a wire electric discharge machine that performs electric discharge machining of a work by applying a pulse voltage between a wire electrode and the work. . The wire electric discharge machining apparatus according to the present disclosure includes a machining mechanism that is a mechanism for electric discharge machining, and when rough machining of the workpiece by electric discharge machining is performed, the plate thickness of the workpiece at the position where machining is performed. A plate thickness estimator for estimating a plate thickness, a plate thickness output device for outputting plate thickness data in which position information is linked to plate thickness information indicating the estimated plate thickness to the outside of the wire electric discharge machine, and a wire electric discharge machine A plate thickness input device that inputs plate thickness data from the outside, a step position estimator that estimates the position of a step where the plate thickness changes based on the input plate thickness data, and a rough machining An electric condition controller that controls the application of pulse voltage based on plate thickness data and level difference information indicating the position of a level difference part when the workpiece is being finish processed by electric discharge machining afterward, and finish machining is performed. and a control device for controlling the processing mechanism based on the plate thickness data and the step information when the plate thickness data and the step information are being processed.

The wire electric discharge machine according to the present disclosure has the effect of being able to perform control for improving machining quality without limiting the data amount of plate thickness information that can be used for estimating the step position in the work. Play.

1 is a diagram showing a configuration example of a wire electric discharge machining apparatus according to a first embodiment; FIG. FIG. 2 is a diagram showing functional configurations of a power supply unit and a control unit provided in the wire electric discharge machining apparatus according to the first embodiment; 4 is a flow chart showing the operation procedure of the wire electric discharge machine according to the first embodiment; FIG. 2 is a block diagram for explaining control during rough machining by the wire electric discharge machine according to the first embodiment; FIG. FIG. 2 is a block diagram for explaining control during finish machining by the wire electric discharge machine according to the first embodiment; FIG. 4 is a diagram showing a control unit to which a coordinate corrector, which is a position information correction unit, is added in Embodiment 1; FIG. 7 is a block diagram for explaining control during finish machining by the wire electric discharge machine to which the coordinate corrector shown in FIG. 6 is added; FIG. 2 is a diagram for explaining a machine coordinate system of the wire electric discharge machine according to the first embodiment; FIG. 4 is a diagram for explaining a relative coordinate system applicable to coordinates associated with plate thickness information in Embodiment 1. FIG. 1 is a diagram showing a configuration example of a control circuit according to a first embodiment; FIG. FIG. 11 shows a learning device according to a second embodiment; FIG. 8 is a diagram showing a configuration example of a neural network used for learning in the learning device according to the second embodiment; 4 is a flowchart showing the procedure of learning processing by the learning device according to the second embodiment; FIG. 11 shows an inference apparatus according to a third embodiment; 10 is a flowchart showing the procedure of inference processing by the inference device according to the third embodiment;

A wire electric discharge machining apparatus, a wire electric discharge machining method, a learning apparatus, and an inference apparatus according to embodiments will be described in detail below with reference to the drawings.

Embodiment 1.
FIG. 1 is a diagram showing a configuration example of a wire electric discharge machining apparatus 100 according to a first embodiment. A wire electric discharge machine 100 is a machine tool that processes a work 18 by generating electric discharge in a gap between the work 18 and a wire electrode 2, which is a machining electrode. The X-axis, Y-axis and Z-axis are the three axes of the machine coordinate system of the wire electric discharge machine 100 . For example, the XY plane is the horizontal plane, and the Z-axis direction is the vertical direction. In the following description, the plus Z direction is the upward direction, and the minus Z direction is the downward direction.

The wire electric discharge machine 100 includes a machining mechanism 13 that is a mechanism for electric discharge machining, a power supply section 14 that includes a machining power supply, and a control section 15 that includes a numerical control (NC) device that is a controller. .

The processing mechanism 13 includes the wire electrode bobbin 1 , the conveying roller 3 that conveys the wire electrode 2 drawn out from the wire electrode bobbin 1 , the upper feeder 4 arranged above the work 18 , and the work 18 below the work 18 . upper and lower guides 6 and 7 for supporting the wire electrode 2 during machining of the work 18; and a work table 8 on which the work 18 is placed. The processing mechanism 13 includes a lower roller 9 for transporting the wire electrode 2 used for processing, a recovery roller 10 for generating a driving force for transporting the wire electrode 2, and a wire electrode for recovering the wire electrode 2 after use. It has a collection box 11 and an X-axis drive motor 12X and a Y-axis drive motor 12Y that drive the work table 8 .

The NC device sends position commands to each of the upper guide 6 and lower guide 7. The upper guide 6 and the lower guide 7 support the wire electrode 2 at a position according to the position command and with an inclination according to the position command. Each of the upper feeder 4 and the lower feeder 5 is connected to a machining power supply. The NC device sends axis commands to each of the X-axis drive motor 12X and the Y-axis drive motor 12Y. The X-axis drive motor 12X drives the worktable 8 in the X-axis direction according to the axis command. The Y-axis drive motor 12Y drives the worktable 8 in the Y-axis direction according to the axis command. The processed portion 16 is the wire electrode 2 between the upper guide 6 and the lower guide 7 .

FIG. 2 is a diagram showing the functional configurations of the power supply section 14 and the control section 15 provided in the wire electric discharge machine 100 according to the first embodiment. FIG. 2 shows a power supply unit 14 and a control unit 15, an upper feeder 4 and a lower feeder 5 connected to a machining power supply 20, an upper guide 6 and a lower guide operating according to commands sent from the NC device 23. 7, an X-axis drive motor 12X and a Y-axis drive motor 12Y.

The power supply unit 14 includes a machining power supply 20 , a machining voltage detector 21 and an electrical condition controller 22 . The machining power supply 20 applies a pulse voltage between the wire electrode 2 and the work 18 according to the voltage command output by the NC device 23 . A machining voltage detector 21 detects a machining voltage. The machining voltage is the inter-electrode voltage applied between the wire electrode 2 and the workpiece 18 . The electrical condition controller 22 controls application of the pulse voltage.

The control unit 15 includes an NC device 23 , a plate thickness estimator 24 , a plate thickness output device 25 , a plate thickness input device 26 and a step position estimator 27 . The NC device 23 generates various commands according to machining conditions according to a machining program for electrical discharge machining. The NC device 23 controls the processing mechanism 13 and the power supply section 14 by outputting various commands.

The plate thickness estimator 24 estimates the plate thickness of the work 18 at the position where the work 18 is being rough-machined by electrical discharge machining. The plate thickness output device 25 outputs plate thickness data to the outside of the wire electric discharge machining apparatus 100 . The plate thickness data is data in which position information is linked to plate thickness information indicating the plate thickness estimated by the plate thickness estimator 24 . When the workpiece 18 is finished by electric discharge machining after rough machining, the thickness data is input to the thickness input device 26 from the outside of the wire electric discharge machine 100 . The step position estimator 27 estimates the position of the step based on the plate thickness data input to the plate thickness input device 26 . The stepped portion is a portion of the work 18 where the plate thickness changes, that is, a portion that hits the boundary between portions with different plate thicknesses.

The electrical condition controller 22 controls the application of the pulse voltage based on the plate thickness data and the step information during the finishing process. The step information is information indicating the position of the step. The NC device 23 controls the processing mechanism 13 based on the plate thickness data and the step information during finishing processing.

The plate thickness output device 25 outputs plate thickness data by writing the plate thickness data to a file 17 stored outside the wire electric discharge machining apparatus 100 . The file 17 is stored in storage means such as a storage device or a storage medium. The plate thickness output device 25 may output the plate thickness data by writing the plate thickness data to a machining program stored outside the wire electric discharge machining apparatus 100 . The plate thickness data is input to the plate thickness input device 26 by the plate thickness input device 26 reading the plate thickness data from the file 17 or from the processing program.

Next, the operation of the wire electric discharge machine 100 will be described. FIG. 3 is a flow chart showing the operation procedure of the wire electric discharge machine 100 according to the first embodiment. The wire electric discharge machine 100 performs machining a plurality of times until obtaining a machined shape that meets the required specifications. Rough machining is machining performed first among a plurality of times of machining, and is machining in which machining speed is emphasized rather than shape accuracy. Finish machining is machining that is performed after rough machining among a plurality of machining operations. The number of finishing processes is arbitrary. FIG. 3 shows a sequence of operations performed by the wire electric discharge machine 100 to adjust the machining of the stepped portion of the workpiece 18 during rough machining and finish machining.

The wire electric discharge machine 100 performs the operations of steps S1 and S2 in rough machining. In step S1 , the wire electric discharge machine 100 estimates the plate thickness of the workpiece 18 using the plate thickness estimator 24 . In step S2, the wire electric discharge machining apparatus 100 writes the thickness data to the file 17 or machining program by the thickness output device 25. FIG. Thereby, the wire electric discharge machining apparatus 100 outputs the plate thickness data to the outside of the wire electric discharge machining apparatus 100 .

The wire electric discharge machine 100 performs operations from step S3 to step S5 in finish machining. In step S3 , the wire electric discharge machine 100 uses the plate thickness input device 26 to read the plate thickness data from the file 17 or the machining program. That is, the wire electric discharge machine 100 reads plate thickness data from the outside of the wire electric discharge machine 100 .

In step S4, the wire electric discharge machining apparatus 100 estimates the position of the stepped portion based on the read plate thickness data. In step S5, the wire electric discharge machine 100 adjusts at least one of the electrical conditions and the axis command based on the plate thickness data and the step information. The electrical condition is a condition for applying a pulse voltage, such as a voltage value or a pulse rest time. The wire electric discharge machine 100 controls at least one of the application of the pulse voltage and the machining mechanism 13 based on the plate thickness data and step information by adjusting at least one of the electrical conditions and the axis command. As described above, the wire electric discharge machine 100 completes the operation according to the procedure shown in FIG.

FIG. 4 is a block diagram for explaining control during rough machining by the wire electric discharge machine 100 according to the first embodiment. The machining power supply 20 applies a pulse voltage between the wire electrode 2 and the workpiece 18 according to the voltage command from the NC device 23 . The X-axis drive motor 12X and the Y-axis drive motor 12Y move the work table 8 in the X-axis direction and the Y-axis direction in accordance with axis commands from the NC device 23 . The wire electric discharge machine 100 adjusts the distance between the wire electrode 2 and the work 18 by moving the work table 8 . The wire electric discharge machine 100 controls the electric discharge energy in the machining section 16 based on the voltage command and the axis command. Thus, the wire electric discharge machine 100 controls electric discharge machining.

The machining voltage detector 21 detects the machining voltage in the machining section 16 . The wire electric discharge machine 100 adjusts the electrical conditions indicated by the voltage command by feedback of the machining voltage detected by the machining voltage detector 21 . The machining power source 20 applies a pulse voltage according to the voltage command with the adjusted electrical conditions. The wire electric discharge machine 100 adjusts the axis command by feedback of the machining voltage detected by the machining voltage detector 21 . The X-axis drive motor 12X and the Y-axis drive motor 12Y move the work table 8 according to the adjusted axis commands. Adjustment by feedback of machining voltage is not limited to adjustment of both electrical conditions and axis commands. The wire electric discharge machine 100 should adjust at least one of the electrical condition and the axis command by feedback of the machining voltage.

The plate thickness estimator 24 estimates the plate thickness at the position where rough processing is being performed based on the processing data. The machining data is data such as machining voltage, machining current, the number of discharge pulses, or machining speed, and is data representing the state of the machining unit 16 . The plate thickness estimator 24 generates plate thickness data in which position information is linked to the plate thickness estimation result. The thickness estimator 24 outputs thickness data to the thickness output device 25 . The plate thickness output device 25 writes the plate thickness data to the file 17 or processing program.

FIG. 5 is a block diagram for explaining control during finish machining by the wire electric discharge machine 100 according to the first embodiment. The wire electric discharge machine 100 controls electric discharge machining based on a voltage command and an axis command. The wire electric discharge machine 100 adjusts at least one of the electrical conditions and the axis command by feedback of the machining voltage detected by the machining voltage detector 21 .

The plate thickness input device 26 reads plate thickness data from the file 17 or the processing program. The plate thickness input device 26 outputs plate thickness data to each of the step position estimator 27 and the electrical condition controller 22 . The step position estimator 27 detects changes in the plate thickness at each position of the workpiece 18 from the plate thickness data. The step position estimator 27 estimates the position where the plate thickness abruptly changes as the position of the step. The step position estimator 27 generates step information indicating the estimated position. The bump position estimator 27 outputs the bump information to the electrical condition controller 22 .

The electric condition controller 22 determines the timing at which the position where finishing is being performed reaches the step portion based on the step information. The electrical condition controller 22 adjusts the electrical conditions according to the plate thickness indicated by the plate thickness data. Thereby, the electric condition controller 22 controls the application of the pulse voltage based on the plate thickness data and the step information during the finishing process. The electric condition controller 22 adjusts the position of the work table 8 indicated by the axis command according to the plate thickness indicated by the plate thickness data. Thereby, the NC unit 23 controls the electric discharge machining according to the machining conditions adjusted based on the plate thickness data and the step information during finishing machining. Note that the adjustment based on the plate thickness data and the step information is not limited to the adjustment of both the electrical conditions and the axis command. The wire electric discharge machine 100 should adjust at least one of the electrical conditions and the axis command based on the plate thickness data and the step information. That is, the wire electric discharge machine 100 controls at least one of the application of the pulse voltage and the machining mechanism 13 based on the plate thickness data and the step data.

The wire electric discharge machine 100 outputs the plate thickness data generated in the rough machining to the outside of the wire electric discharge machine 100. In addition, the wire electric discharge machining apparatus 100 generates step information based on the plate thickness data input from the outside of the wire electric discharge machine 100 during finish machining, and generates a pulse based on the plate thickness data and the step information. It controls the application of voltage or the processing mechanism 13 . The wire electric discharge machining apparatus 100 can achieve high machining quality by enabling machining suitable for the plate thickness at the stepped portion.

According to Embodiment 1, since it is not necessary to accumulate plate thickness data inside the wire electric discharge machining apparatus 100, it is necessary to set a limit on the data amount of plate thickness information that can be used for estimating the position of the stepped portion. disappear. By storing the plate thickness data in the file 17 or machining program stored outside the wire electric discharge machine 100, it is possible to ensure the traceability of the machining process by the wire electric discharge machine 100.

The wire electric discharge machining apparatus 100 may include a position information correction unit that corrects the position information included in the plate thickness data. Here, an example in which a position information correction unit is added to the control unit 15 will be described. FIG. 6 shows the controller 15 to which the coordinate corrector 28, which is the position information corrector, is added in the first embodiment.

The coordinate corrector 28 corrects the coordinates, which are the position information included in the thickness data, when the position or inclination of the workpiece 18 on the work table 8 has changed since the thickness data was generated. The coordinate corrector 28 corrects the coordinates so that the change in the position or tilt of the workpiece 18 is offset. The electrical condition controller 22 controls the application of the pulse voltage based on the plate thickness data including the corrected positional information and the step information. The NC device 23 controls the processing mechanism 13 based on the plate thickness data including the corrected position information and the step information. The wire electric discharge machine 100 can control the pulse voltage or the machining mechanism 13 based on the plate thickness data by correcting the coordinates with the coordinate corrector 28 even when the position or inclination of the work 18 is changed. .

The position or inclination of the work 18 may be changed by removing the work 18 from the work table 8 after rough machining and placing the work 18 on the work table 8 again. Further, when the work 18 is machined by changing the model of the wire electric discharge machine 100 for each machining step such as rough machining and finish machining, the position or inclination of the work 18 in the finish machining is the same as the position of the work 18 in the rough machining. Or it can be changed from tilt. As an example of changing the model of the wire electric discharge machine 100, the wire electric discharge machine 100 using water as the machining fluid is used for rough machining, and the wire electric discharge machine 100 using oil as the machining fluid is used for finish machining. are mentioned. By using the wire electric discharge machining apparatus 100 that uses water as a machining fluid for rough machining, machining can be performed at a high machining speed. By using the wire electric discharge machining apparatus 100 that uses oil as a machining fluid for finish machining, high-precision machining can be performed at a low machining speed.

Also, when changing the model of the wire electric discharge machine 100 for each machining step, the size of the work table 8 may differ for each model. In this case as well, the wire electric discharge machine 100 can control the pulse voltage or the machining mechanism 13 based on the plate thickness data in finishing machining by correcting the coordinates with the coordinate corrector 28 .

FIG. 7 is a block diagram for explaining control during finish machining by the wire electric discharge machine 100 to which the coordinate corrector 28 shown in FIG. 6 is added. The plate thickness input device 26 outputs the input plate thickness data to the coordinate corrector 28 . The coordinate corrector 28 corrects the coordinates included in the thickness data when the position or inclination of the workpiece 18 has changed since the thickness data was generated. The coordinate corrector 28 outputs the plate thickness data whose coordinates have been corrected to the step position estimator 27 and the electrical condition controller 22, respectively. As a result, the wire electric discharge machine 100 adjusts the pulse voltage or the machining mechanism 13 based on the thickness data when the position or inclination of the workpiece 18 on the work table 8 is changed after the thickness data is generated. can be controlled.

The coordinates linked to the thickness information in the thickness data may be the coordinates of a coordinate system based on the work 18 instead of the coordinates of the machine coordinate system unique to the wire electric discharge machining apparatus 100 . A coordinate system based on the work 18 is a coordinate system having a preset position on the work 18 as an origin. In the following description, a coordinate system based on the workpiece 18 is called a relative coordinate system.

FIG. 8 is a diagram for explaining the machine coordinate system of the wire electric discharge machining device 100 according to the first embodiment. In the example shown in FIG. 8, one point on the work table 8 is assumed to be the origin of the machine coordinate system.

If the coordinates linked to the plate thickness information are the coordinates of the machine coordinate system, the work 18 is removed from the work table 8 after the plate thickness data is generated, and the work 18 is placed on the work table 8 again. may change the position of the workpiece 18 with respect to the origin. Alternatively, the tilt of the workpiece 18 with respect to the machine coordinate system may be changed. During finish machining, if the position or inclination of the workpiece 18 is changed from that during rough machining, the wire electric discharge machine 100 cannot control the pulse voltage or the machining mechanism 13 based on the plate thickness data. .

When the workpiece 18 is machined by changing the model of the wire electric discharge machine 100 for each machining step such as rough machining and finish machining, the position or inclination of the workpiece 18 during finish machining is the same as the position or inclination of the workpiece 18 during rough machining. can be changed from Further, when the model of the wire electric discharge machine 100 is changed for each machining step, the size of the work table 8 may differ for each model. In such a case as well, the wire electric discharge machine 100 cannot control the pulse voltage or the machining mechanism 13 based on the plate thickness data.

FIG. 9 is a diagram for explaining a relative coordinate system that can be applied to coordinates linked to plate thickness information in the first embodiment. The x-axis, y-axis and z-axis shown in FIG. 9 are the three axes of the relative coordinate system. In the example shown in FIG. 9, the origin of the relative coordinate system is the starting point of the machining locus 19, that is, the machining start position. The origin of the relative coordinate system may be a position other than the machining start position. The origin of the relative coordinate system may be a position other than the position on the machining locus 19 . The wire electric discharge machine 100 can set any position of the workpiece 18 as the origin of the relative coordinate system.

By using the coordinates included in the plate thickness data as the coordinates of the relative coordinate system, the wire electric discharge machine 100 can apply the pulse voltage or the machining mechanism 13 based on the plate thickness data even if the position or inclination of the work 18 is changed. can be controlled. The wire electric discharge machine 100 can also control the pulse voltage or the machining mechanism 13 based on the plate thickness data even when the model of the wire electric discharge machine 100 is changed.

Next, a machining voltage detector 21, an electrical condition controller 22, an NC device 23, a plate thickness estimator 24, a plate thickness output device 25, a plate thickness input device 26, a step position estimator 27, and a coordinate corrector 28. A hardware configuration that implements the components will be described. Each component described above is implemented by a processing circuit, which is a circuit in which a processor executes software. The processing circuit executing the software is, for example, the control circuit shown in FIG. FIG. 10 is a diagram showing a configuration example of the control circuit 30 according to the first embodiment. The control circuit 30 comprises an input section 31 , a processor 32 , a memory 33 and an output section 34 .

The input unit 31 is an interface circuit that receives data input from outside the control circuit 30 and provides it to the processor 32 . The output unit 34 is an interface circuit that sends data from the processor 32 or memory 33 to the outside of the control circuit 30 . When the processing circuit is the control circuit 30 shown in FIG. 10, the above components are implemented by the processor 32 reading out and executing programs corresponding to the respective components stored in the memory 33 . Memory 33 is also used as temporary memory in each process performed by processor 32 . The processor 32 may output data such as calculation results to the memory 33 for storage, or may store data such as calculation results in an auxiliary storage device via the volatile memory of the memory 33 .

The processor 32 is a CPU (Central Processing Unit, also referred to as a central processing unit, processing unit, arithmetic unit, microprocessor, microcomputer, processor, or DSP (Digital Signal Processor)). The memory 33 is non-volatile such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (registered trademark) (Electrically Erasable Programmable Read Only Memory), etc. Alternatively, a volatile semiconductor memory, a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, a DVD (Digital Versatile Disc), or the like.

Although FIG. 10 is an example of hardware in which the above components are implemented by a general-purpose processor 32 and memory 33, the above components may be implemented by dedicated hardware circuits. The processing circuit, which is a dedicated hardware circuit, can be 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 circuit. Each component described above may be realized by a combination of the control circuit 30 and a dedicated hardware circuit.

According to Embodiment 1, the wire electric discharge machining apparatus 100 outputs the plate thickness data to the outside of the wire electric discharge machining apparatus 100, and determines the position of the stepped portion based on the plate thickness data input from the outside of the wire electric discharge machining apparatus 100. to estimate As a result, the wire electric discharge machine 100 can perform control for improving machining quality without limiting the data amount of plate thickness information that can be used for estimating the step position in the workpiece 18. play.

Embodiment 2.
Embodiment 2 describes a learning device for using machine learning for at least one of plate thickness estimation and step position estimation. FIG. 11 shows a learning device 40 according to the second embodiment. In the second embodiment, the same reference numerals are given to the same components as in the first embodiment, and the configuration different from the first embodiment will be mainly described.

The learning device 40 learns the relationship between the machining data about the state of electric discharge machining and at least one of the thickness of the workpiece 18 and the position of the step portion of the workpiece 18 where the thickness changes. . A case in which the learning device 40 learns the relationship between the processing data and both the plate thickness and the position of the step will be described below as an example.

The learning device 40 includes a data acquisition unit 41 and a model generation unit 42. Processing data, plate thickness data, and step information are input to the data acquisition unit 41 . The data acquisition unit 41 creates learning data using the input processing data, plate thickness data, and step information. In Embodiment 2, learning data is data in which processing data, plate thickness information, and step information are associated with each other. In this manner, the data acquisition unit 41 acquires learning data including processing data, plate thickness information, and step information.

The model generation unit 42 uses the learning data to generate a learned model 43 for inferring the plate thickness and the position of the step from the processing data. The trained model storage unit 44 stores the generated trained model 43 . A learned model storage unit 44 shown in FIG. 11 is an external storage unit of the learning device 40 . The trained model storage unit 44 may be provided inside the learning device 40 .

A known algorithm such as supervised learning, unsupervised learning, or reinforcement learning can be used as the learning algorithm used by the model generation unit 42 . As an example, a case of applying a neural network will be described.

For example, the model generation unit 42 learns the plate thickness and the position of the step by so-called supervised learning according to the neural network model. Here, supervised learning is a method of learning a feature in the learning data by giving a set of input and result data to the learning device 40 and inferring the result from the input.

The learning data includes inputs and labels that are the results corresponding to the inputs. Processing data corresponds to input, and plate thickness information and step information correspond to labels. A neural network is composed of an input layer consisting of a plurality of neurons, a hidden layer which is an intermediate layer consisting of a plurality of neurons, and an output layer consisting of a plurality of neurons. The intermediate layer may be one layer, or two or more layers.

FIG. 12 is a diagram showing a configuration example of a neural network used for learning in the learning device 40 according to the second embodiment. The neural network shown in FIG. 12 is a three-layer neural network. The input layer includes neurons X1, X2, X3. The middle layer contains neurons Y1 and Y2. The output layer contains neurons Z1, Z2, Z3. Note that the number of neurons in each layer is arbitrary. A plurality of values input to the input layer are multiplied by w11, w12, w13, w14, w15, and w16, which are weights W1, and input to the intermediate layer. A plurality of values input to the intermediate layer are multiplied by w21, w22, w23, w24, w25, and w26, which are weights W2, and output from the output layer. The output result output from the output layer changes according to the values of weights W1 and W2.

In Embodiment 2, the neural network learns the plate thickness and the position of the step by so-called supervised learning according to learning data acquired by the data acquisition unit 41 . That is, the neural network adjusts the weights W1 and W2 so that the processed data is input to the input layer and the result output from the output layer approaches the thickness information and step information. Learn location. The model generation unit 42 generates the learned model 43 by executing the learning as described above, and outputs the learned model 43 . The learned model storage unit 44 stores the learned model 43 output from the model generation unit 42 . The model generation unit 42 may read the already generated learned model 43 from the learned model storage unit 44 and update the learned model 43 by re-learning according to the learning data.

Next, learning processing by the learning device 40 will be described. FIG. 13 is a flow chart showing the procedure of learning processing by the learning device 40 according to the second embodiment. In step S11 , the learning device 40 acquires learning data including processing data, plate thickness information, and step information using the data acquisition unit 41 . The data acquisition unit 41 creates learning data using the processing data, the plate thickness data, and the step information, all of which are acquired at the same time. Note that the data acquisition unit 41 only needs to create learning data in which the processing data, the thickness information, and the step information are associated with each other, and does not necessarily acquire the processing data, the plate thickness data, and the step information at the same time.

In step S12, the learning device 40 uses the model generation unit 42 to learn the plate thickness and the position of the stepped portion by so-called supervised learning according to the learning data, and generates or updates the learned model 43. In step S13, the model generating unit 42 outputs the generated or updated learned model 43. FIG. With the above, the learning device 40 ends the learning process according to the procedure shown in FIG. 13 . The learned model storage unit 44 stores the learned model 43 obtained by the learning process.

In Embodiment 2, the case where supervised learning is applied to the learning algorithm used by the model generation unit 42 has been described, but learning other than supervised learning may be applied to the learning algorithm. The model generator 42 may perform machine learning using learning algorithms such as reinforcement learning, unsupervised learning, or semi-supervised learning. The model generation unit 42 may perform machine learning using learning algorithms such as deep learning, genetic programming, inductive logic programming, or support vector machines.

In Embodiment 2, the learning device 40 is a device external to the wire electric discharge machining device 100 . The learning device 40 may be a device connected to the wire electric discharge machine 100 via a network. The learning device 40 may be a device existing on a cloud server. The learning device 40 is not limited to a device external to the wire electric discharge machine 100 , and may be a device built into the wire electric discharge machine 100 .

The learning device 40 is not limited to learning the plate thickness and the position of the stepped portion according to learning data created for one wire electric discharge machining device 100 . The learning device 40 may learn the plate thickness and the position of the stepped portion according to learning data created for a plurality of wire electric discharge machines 100 . The learning device 40 may acquire learning data from a plurality of wire electric discharge machines 100 used at the same place, or may acquire learning data from a plurality of wire electric discharge machines 100 used at different places. may be obtained. The learning data may be obtained from a plurality of wire electric discharge machines 100 that operate independently of each other at a plurality of locations. After starting acquisition of learning data from a plurality of wire electric discharge machines 100, a new wire electric discharge machine 100 may be added as a target for acquiring learning data. Also, after starting acquisition of learning data from a plurality of wire electric discharge machines 100, some of the plurality of wire electric discharge machines 100 may be excluded from targets for which learning data is acquired.

The learning device 40 that has learned about one wire electric discharge machine 100 may learn about other wire electric discharge machines 100 other than the wire electric discharge machine 100 concerned. The learning device 40 can update the learned model 43 by re-learning the other wire electric discharge machining device 100 .

The learning device 40 should just learn the relationship between the processing data and at least one of the plate thickness and the position of the stepped portion. When the learning device 40 learns the relationship between the processing data and the plate thickness, the processing data and the plate thickness data are input to the data acquisition unit 41 . The data acquisition unit 41 uses the input processing data and plate thickness data to create learning data, which is data in which the processing data and plate thickness information are associated with each other. The data acquisition unit 41 acquires learning data including processing data and plate thickness information. The model generator 42 uses the learning data to generate a learned model 43 for inferring the plate thickness from the processed data.

In addition, when the learning device 40 learns the relationship between the processing data and the position of the step, the processing data and the step information are input to the data acquisition unit 41 . The data acquisition unit 41 uses the input processing data and step information to create learning data, which is data in which the processing data and the step information are associated with each other. The data acquisition unit 41 acquires learning data including processed data and step information. The model generator 42 uses the learning data to generate a trained model 43 for inferring the position of the step from the processed data.

The learning device 40 is realized by hardware similar to the hardware shown in FIG. Each part of the learning device 40 is realized by a processing circuit, which is a circuit in which a processor executes software. Each part of the learning device 40 may be realized by a dedicated processing circuit, or by a combination of a processing circuit in which a processor executes software and a dedicated processing circuit.

According to the second embodiment, the learning device 40 generates the learned model 43 using learning data including processing data and at least one of plate thickness information and step information. The learning device 40 can generate a learned model 43 for inferring at least one of the plate thickness and the position of the step from the processed data.

Embodiment 3.
Embodiment 3 describes an inference device that infers at least one of the plate thickness and the position of the step using the learned model 43 . FIG. 14 shows an inference device 50 according to the third embodiment. In the third embodiment, the same components as those in the first or second embodiment are denoted by the same reference numerals, and the configuration different from that in the first or second embodiment will be mainly described.

The inference device 50 infers at least one of the thickness of the workpiece 18 and the position of the step portion of the workpiece 18 where the thickness changes, based on the machining data about the state of electric discharge machining for the wire electric discharge machine 100 . In the following description, an example will be described in which the inference device 50 infers both the plate thickness and the position of the stepped portion based on the processing data.

The inference device 50 includes a data acquisition unit 51 and an inference unit 52 . Processed data is input to the data acquisition unit 51 . That is, the data acquisition unit 51 acquires processed data. The inference unit 52 reads the learned model 43 for inferring the plate thickness and the position of the step from the processed data from the learned model storage unit 44 . The inference unit 52 outputs thickness information and step information by inputting processing data, which is data for inference, to the learned model 43 . A trained model storage unit 44 shown in FIG. 14 is an external storage unit of the inference device 50 . The learned model storage unit 44 may be provided inside the inference device 50 . In the third embodiment, the inference device 50 uses the learned model 43 generated by the learning device 40 according to the second embodiment to infer the plate thickness and the position of the step from the processed data.

Next, the inference processing by the inference device 50 will be described. FIG. 15 is a flowchart showing the procedure of inference processing by the inference device 50 according to the third embodiment. In step S21 , the inference device 50 acquires processed data using the data acquisition unit 51 . In step S22 , the inference unit 52 of the inference device 50 inputs processed data to the learned model 43 . In step S23, the inference unit 52 outputs plate thickness information and step information. With the above, the inference device 50 ends the inference processing according to the procedure shown in FIG. 15 . The inference device 50 sends the plate thickness information and the step information to the wire electric discharge machining device 100 .

In the third embodiment, the inference device 50 is a device external to the wire electric discharge machining device 100 . The inference device 50 may be a device connected to the wire electric discharge machine 100 via a network. The inference device 50 may be a device existing on a cloud server. The reasoning device 50 is not limited to a device external to the wire electric discharge machining device 100 , and may be a device built into the wire electric discharge machining device 100 .

The inference device 50 may use the learned model 43 to infer at least one of the plate thickness and the position of the step. When the inference device 50 infers the plate thickness from the processed data, the inference unit 52 outputs plate thickness information by inputting the processed data to the learned model 43 for inferring the plate thickness from the processed data. Further, when the inference device 50 infers the position of the stepped portion from the processed data, the inferred portion 52 inputs the processed data to the learned model 43 for inferring the position of the stepped portion from the processed data. to output

The inference device 50 is realized by hardware similar to the hardware shown in FIG. Each part of the inference device 50 is implemented by a processing circuit, which is a circuit in which a processor executes software. Each part of the inference device 50 may be realized by a dedicated processing circuit, or by a combination of a processing circuit in which a processor executes software and a dedicated processing circuit.

According to the third embodiment, the inference device 50 inputs the machining data to the learned model 43 for inferring at least one of the thickness and the position of the stepped portion from the machining data, thereby obtaining the thickness information and the step information. Output at least one. The inference device 50 can infer at least one of the plate thickness and the position of the step from the processing data.

The wire electric discharge machine 100 controls machining based on at least one of the plate thickness and the position of the step portion inferred by the inference device 50 . The wire electric discharge machining apparatus 100 can achieve high machining quality by enabling machining suitable for the plate thickness at the stepped portion.

The configuration shown in each embodiment above is an example of the content of the present disclosure. The configuration of each embodiment can be combined with another known technique. Configurations of respective embodiments may be combined as appropriate. A part of the configuration of each embodiment can be omitted or changed without departing from the gist of the present disclosure.

1 wire electrode bobbin, 2 wire electrode, 3 transport roller, 4 upper feeder, 5 lower feeder, 6 upper guide, 7 lower guide, 8 work table, 9 lower roller, 10 recovery roller, 11 wire electrode recovery box, 12X X-axis drive motor, 12Y Y-axis drive motor, 13 Machining mechanism, 14 Power supply unit, 15 Control unit, 16 Machining unit, 17 File, 18 Workpiece, 19 Machining trajectory, 20 Machining power supply, 21 Machining voltage detector, 22 Electricity Condition controller, 23 NC unit, 24 plate thickness estimator, 25 plate thickness output device, 26 plate thickness input device, 27 step position estimator, 28 coordinate corrector, 30 control circuit, 31 input unit, 32 processor, 33 memory , 34 output unit, 40 learning device, 41, 51 data acquisition unit, 42 model generation unit, 43 learned model, 44 learned model storage unit, 50 reasoning device, 52 reasoning unit, 100 wire electric discharge machining device.

Claims

A wire electric discharge machine for performing electric discharge machining of a work by applying a pulse voltage between a wire electrode and the work,
a machining mechanism that is a mechanism for electrical discharge machining;
a plate thickness estimator for estimating the plate thickness of the work at a position where the work is being rough-machined by the electrical discharge machining;
a plate thickness output device that outputs plate thickness data in which position information is linked to plate thickness information indicating the estimated plate thickness to the outside of the wire electric discharge machining apparatus;
a plate thickness input device for inputting the plate thickness data from the outside of the wire electric discharge machine;
a step position estimator for estimating the position of a stepped portion of the workpiece where the plate thickness changes, based on the input plate thickness data;
Controlling the application of the pulse voltage based on the plate thickness data and step information indicating the position of the stepped portion when the workpiece is being finished by the electrical discharge machining after the rough machining. an electrical condition controller;
a control device that controls the processing mechanism based on the plate thickness data and the step information when the finish processing is being performed;
A wire electric discharge machine comprising:
The wire electric discharge machining apparatus according to claim 1, wherein the position information included in the plate thickness data is coordinates of a coordinate system having a preset position in the workpiece as an origin.
A position information correction unit that corrects the position information included in the plate thickness data input to the plate thickness input device,
The electrical condition controller controls application of the pulse voltage based on the plate thickness data including the corrected position information and the step information,
2. The wire electric discharge machining apparatus according to claim 1, wherein the control device controls the machining mechanism based on the plate thickness data including the corrected position information and the step information.
4. Any one of claims 1 to 3, wherein the plate thickness output device writes the plate thickness data to a file stored outside the wire electric discharge machining apparatus or to a machining program for the electric discharge machining. The wire electric discharge machine according to 1.
A wire electric discharge machining method in which a wire electric discharge machine performs electric discharge machining of a work by applying a pulse voltage between a wire electrode and the work,
a step of estimating the plate thickness of the work at the position where the work is being rough-machined by the electrical discharge machining;
A step of outputting plate thickness data in which position information is linked to plate thickness information indicating the estimated plate thickness to the outside of the wire electric discharge machining apparatus;
a step of reading the plate thickness data from outside the wire electric discharge machine;
a step of estimating the position of a stepped portion of the workpiece where the plate thickness changes, based on the read plate thickness data;
When finish machining of the workpiece by the electrical discharge machining after the rough machining is being performed, the application of the pulse voltage and the a step of controlling at least one of a machining mechanism that is a mechanism for electric discharge machining;
A wire electric discharge machining method comprising:
For a wire electric discharge machine that performs electric discharge machining of the work by applying a pulse voltage between a wire electrode and the work, machining data on the state of the electric discharge machining, the plate thickness of the work, and the plate of the work A learning device for learning a relationship with at least one of the positions of a stepped portion whose thickness changes,
a data acquisition unit that acquires learning data including the processed data and at least one of plate thickness information indicating the plate thickness and step information indicating the position of the stepped portion;
a model generation unit that uses the learning data to generate a trained model for inferring at least one of the plate thickness and the position of the stepped portion from the processing data;
A learning device comprising:
For a wire electric discharge machine that performs electric discharge machining of the work by applying a pulse voltage between the wire electrode and the work, the thickness of the work and the thickness of the work are determined based on machining data on the state of the electric discharge machining. An inference device for inferring at least one of the positions of the stepped portion where the plate thickness changes,
a data acquisition unit that acquires the processed data;
By inputting the processed data into a trained model for inferring at least one of the thickness and the position of the stepped portion from the processed data, thickness information indicating the plate thickness and the position of the stepped portion are indicated. an inference unit that outputs at least one of step information;
An inference device characterized by comprising: