WO2022185431A1

WO2022185431A1 - Wire electrical discharge machining device, shape dimension compensator, wire electrical discharge machining method, learning device, and inference device

Info

Publication number: WO2022185431A1
Application number: PCT/JP2021/008094
Authority: WO
Inventors: 大介関本; 信行太田
Original assignee: 三菱電機株式会社
Priority date: 2021-03-03
Filing date: 2021-03-03
Publication date: 2022-09-09
Also published as: CN116348232A; JPWO2022185431A1; CN116348232B; JP6972443B1

Abstract

A wire electrical discharge machining device (100) provided with: a machining mechanism (30) that uses pulsed voltage from a wire electrode to perform wire electrical discharge machining on a workpiece having a plurality of thickness regions mutually different in terms of the thickness on the machining path; a thickness estimator (48) that estimates a thickness of the workpiece during wire electrical discharge machining; and a shape dimension compensator (35) that uses a machining voltage during machining, machining energy during machining, a machining speed during machining, the distance between a workpiece and a nozzle for supplying machining fluid to the wire electrode, and the thickness to calculate a voltage correction value, which is a correction value for machining voltage, an off-time correction value, which is a correction value for the off-time of a pulsed voltage, and a wire tension command, which is a command for the wire electrode, such that the machining dimensional difference between the thickness regions is reduced and the straightness in the workpiece in the wire electrode length direction is enhanced in each thickness region, wherein the machining mechanism (30) is controlled using the voltage correction value, off-time correction value, and wire tension command.

Description

Wire electric discharge machine, geometry compensator, wire electric discharge machining method, learning device, and reasoning device

The present disclosure relates to a wire electric discharge machine, a geometry compensator, a wire electric discharge machining method, a learning device, and an inference device that compensate for the dimension and shape of a workpiece after machining.

For wire electric discharge machines, the appropriate machining conditions differ depending on the thickness of the workpiece to be machined. Therefore, it is desired that the wire electric discharge machine selects appropriate machining conditions according to the plate thickness and performs wire electric discharge machining.

The wire electric discharge machining apparatus described in Patent Document 1 selects the electrical condition strength from the relationship between the plate thickness and the machining energy, and switches to the electrical condition corresponding to the electrical condition strength, thereby preventing disconnection of the wire electrode. .

JP-A-9-290328

However, in the technique of Patent Document 1, when the workpiece is a thin plate, the volume removed in the direction of progress of processing is small and the processing speed is high, so the discharge is less likely to spread to the side surface in the direction of progress. On the other hand, when the workpiece is a thick plate, the removal volume in the traveling direction is large and the machining speed is slow, so the discharge tends to fly to the side surface in the traveling direction.

For this reason, in the technique of Patent Document 1, when a large electrical discharge machining energy is input, in machining in which the plate thickness of the workpiece changes during machining, the machining groove width becomes narrower with a thinner plate, and a larger plate thickness , the groove width becomes thicker. As a result, for a workpiece whose plate thickness changes during processing, the processing dimensions differ for each plate thickness, and the precision of the processing dimension deteriorates. Also, the wire electrode bends during wire electric discharge machining. Therefore, even within a region of the same plate thickness of the workpiece, the width of the groove to be machined differs depending on the position of the wire electrode where the workpiece is machined. Due to the variation in dimensions, the accuracy of the processed shape deteriorates.

The present disclosure has been made in view of the above, and provides a wire electric discharge machine that can improve the accuracy of machining dimensions and machining shape even for a workpiece whose plate thickness changes during machining. with the aim of obtaining

In order to solve the above-described problems and achieve the object, the wire electric discharge machine of the present disclosure provides a workpiece having a plurality of plate thickness regions with different plate thicknesses on a machining path. A machining mechanism that performs wire electric discharge machining using pulses, and a thickness estimator that estimates the thickness of a workpiece during wire electric discharge machining. In addition, the wire electric discharge machine of the present disclosure provides machining voltage during machining, machining energy during machining, machining speed during machining, and the distance between the nozzle that supplies the machining fluid to the wire electrode and the workpiece Based on the distance and the plate thickness, the difference in processing dimensions between the plate thickness regions is small, and the straightness accuracy in the length direction of the wire electrode of the workpiece is high within each plate thickness region. A shape and size compensator is provided for calculating a voltage correction value that is a correction value for the machining voltage, a rest time correction value that is a correction value for the rest time of the voltage pulse, and a wire tension command that is a tension command for the wire electrode. The machining mechanism is controlled using voltage corrections, dwell time corrections, and wire tension commands.

The wire electric discharge machine according to the present disclosure has the effect of being able to improve the precision of machining dimensions and the precision of machining shape even for workpieces whose plate thickness changes during machining.

1 is a perspective view showing a configuration example of a wire electric discharge machine according to an embodiment; FIG. FIG. 4 is a perspective view showing another configuration example of the wire electric discharge machine according to the embodiment; FIG. 2 is a diagram for explaining the structure of a workpiece machined by the wire electric discharge machine according to the embodiment; A diagram for explaining the shape of the workpiece when the workpiece is machined without correcting the machining speed with respect to the plate thickness. FIG. 5 is a diagram for explaining the shape of a workpiece that has been machined without correcting the deflection of the wire electrode with respect to the plate thickness; 1 is a block diagram showing a functional configuration example of a wire electric discharge machine according to an embodiment; FIG. FIG. 4 is a diagram for explaining calculation processing of a voltage correction value by the geometry compensator according to the embodiment; FIG. 4 is a diagram for explaining voltage correction value information used by the geometry compensator according to the embodiment; FIG. 4 is a diagram for explaining a calculation process of a pause time correction value by the shape and dimension compensator according to the embodiment; A diagram for explaining the relationship between the amount of separation of the nozzle and the amount of bending of the wire electrode. FIG. 4 is a diagram for explaining the relationship between the wire tension and the deflection amount of the wire electrode; 1 is a flow chart showing a processing procedure of wire electric discharge machining by a wire electric discharge machine according to an embodiment; 1 is a block diagram showing a configuration example of a learning device according to an embodiment; FIG. 4 is a flow chart showing a processing procedure of learning processing by the learning device according to the embodiment; 1 is a block diagram showing a configuration example of an inference device according to an embodiment; FIG. 4 is a flowchart showing a processing procedure of inference processing by the inference device according to the embodiment; FIG. 2 is a diagram showing a hardware configuration example that implements the NC control device according to the embodiment;

A wire electric discharge machine, a geometry compensator, a wire electric discharge machining method, a learning device, and an inference device according to embodiments of the present disclosure will be described below in detail with reference to the drawings.

Embodiment.
FIG. 1 is a perspective view showing a configuration example of a wire electric discharge machine according to an embodiment. The wire electric discharge machine 100 includes a machining mechanism 30, a wire tension controller 31, a machining power source 32, and an NC (Numerical Control) controller 33, which is a numerical controller.

The processing mechanism 30 includes a wire electrode bobbin 1, a wire electrode 2, a tension load device 3, an upper feeder 4, a lower feeder 5, an upper guide 6, a lower guide 12, a surface plate 8, A lower roller 13 is provided. The processing mechanism 30 also includes a wire electrode collection box 10, a wire travel speed control motor 9, an X-axis drive motor 11X, and a Y-axis drive motor 11Y.

Also, the wire tension control device 31 is connected to a machining power source 32 and an NC control device 33 , and the machining power source 32 is connected to the NC control device 33 . The machining mechanism 30 is also connected to a wire tension controller 31 , a machining power source 32 and an NC controller 33 . In the following description, the two axes in the plane parallel to the upper surface of the plate-shaped surface plate 8 and perpendicular to each other are defined as the X-axis and the Y-axis. Also, the axis orthogonal to the X-axis and the Y-axis is defined as the Z-axis. 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 may be referred to as the upward direction, and the minus Z direction may be referred to as the downward direction.

The wire electrode bobbin 1 is wound with the wire electrode 2 and supplies the wire electrode 2 to the tension load device 3 . A wire electrode 2 is pulled out from a wire electrode bobbin 1 and sent to a tension load device 3 . The tension load device 3 conveys the wire electrode 2 and applies a tension load to the wire electrode 2 . The tension load device 3 feeds the wire electrode 2 to the lower roller 13 via the upper feeder 4 , upper guide 6 , lower feeder 5 and lower guide 12 . The wire electrode 2 passed through the lower roller 13 is sent to the wire electrode collection box 10 through the wire travel speed control motor 9 .

An upper guide 6 is arranged below the upper feeder 4 , a lower feeder 5 is arranged below the upper guide 6 , and a lower guide 12 is arranged below the lower feeder 5 . The upper power feeder 4 and the lower power feeder 5 are connected to a machining power source 32 to apply a voltage between the wire electrode 2 and the workpiece 7 .

The upper guide 6 and lower guide 12 support the position and inclination of the wire electrode 2 during machining of the workpiece 7, which is a workpiece. An upper nozzle 81 , which will be described later, is arranged below the upper guide 6 , and a lower nozzle 82 , which will be described later, is arranged above the lower feeder 5 . The upper nozzle 81 supplies machining fluid downward to the wire electrode 2 , and the lower nozzle 82 supplies machining fluid upward to the wire electrode 2 . The workpiece 7 is processed between the upper nozzle 81 and the lower nozzle 82 .

The wire electric discharge machine 100 of the present embodiment performs wire electric discharge machining on the workpiece 7 having a step. That is, the workpiece 7 has various plate thicknesses for each plate thickness region. In other words, the workpiece 7 has a plurality of thickness regions with different thicknesses on the machining path. For example, the workpiece 7 has a first thickness region of thickness regions to be processed, and a second thickness region adjacent to the first thickness region. is the second plate thickness, and the first plate thickness region and the second plate thickness region are continuously wire electric discharge machined. A workpiece 7 is placed on the surface plate 8 . The surface plate 8 is provided with a hole through which the wire electrode 2 is passed.

The lower roller 13 conveys the wire electrode 2 after machining the workpiece 7 on the surface plate 8 . The wire traveling speed control motor 9 is a collecting roller and generates driving force for conveying the wire electrode 2 . The wire electrode collection box 10 is a box for collecting the wire electrodes 2 sent from the wire travel speed control motor 9 . The X-axis drive motor 11X drives the surface plate 8 in the X-axis direction, and the Y-axis drive motor 11Y drives the surface plate 8 in the Y-axis direction.

The wire tension control device 31 is connected to the tension load device 3 and controls the wire tension that is the tension of the wire electrode 2 by controlling the tension load device 3 . The machining power supply 32 is connected to the upper power feeder 4 and the lower power feeder 5 , and controls the upper power feeder 4 and the lower power feeder 5 to generate electric discharge between the workpiece 7 and the wire electrode 2 . generate

The machining power supply 32 has a machining voltage detector 45 and a machining energy detector 46, which will be described later. The machining power supply 32 sends the machining voltage detected by the machining voltage detector 45 and the machining energy detected by the machining energy detector 46 to the NC controller 33 . The machining power supply 32 uses a voltage correction value (hereinafter referred to as a voltage correction value) and a voltage pulse pause time correction value (hereinafter referred to as a pause time correction value) sent from the NC control device 33. It controls the upper feeder 4 and the lower feeder 5 .

The NC controller 33 controls the machining mechanism 30, the machining power supply 32, and the wire tension controller 31. The NC control device 33 is connected to, for example, the X-axis drive motor 11X and the Y-axis drive motor 11Y. The NC control device 33 controls the position of the surface plate 8 in the X-axis direction and the Y-axis direction by sending axial movement commands to the X-axis drive motor 11X and the Y-axis drive motor 11Y. Thereby, the NC control device 33 controls the distance between the workpiece 7 placed on the surface plate 8 and the wire electrode 2, and the voltage between the workpiece 7 and the wire electrode 2 is controlled. to control.

The NC control device 33 is also connected to the wire travel speed control motor 9 and controls the wire travel speed control motor 9 . In FIG. 1, a connection line between the NC control device 33 and the wire travel speed control motor 9 is omitted.

The NC control device 33 calculates a wire tension command based on the machining voltage detected by the machining voltage detector 45 and the machining energy detected by the machining energy detector 46. A wire tension command is a command for controlling the tension of the wire electrode 2 . The NC control device 33 sends the calculated wire tension command to the wire tension control device 31 .

The NC control device 33 also calculates a voltage correction value and a rest time correction value based on the machining voltage detected by the machining voltage detector 45 and the machining energy detected by the machining energy detector 46. The NC control device 33 sends the calculated voltage correction value and idle time correction value to the machining power source 32 .

The wire electric discharge machining apparatus 100 moves the surface plate 8 in the X-axis direction and the Y-axis direction by the X-axis drive motor 11X and the Y-axis drive motor 11Y, and moves the workpiece 7 placed on the surface plate 8 and the wire electrode. 2 to a specific distance at which wire discharge is possible. Thereby, the wire electric discharge machine 100 performs wire electric discharge machining on the workpiece 7 with the wire electrode 2 . A case will be described below in which the workpiece 7 is moved in the X-axis direction and wire electric discharge machining is performed on the workpiece 7 .

Note that the wire electric discharge machine 100 may move the wire electrode 2 instead of the surface plate 8 . FIG. 2 is a perspective view showing another configuration example of the wire electric discharge machine according to the embodiment.

A wire electric discharge machine 101 shown in FIG. 2 includes a machining mechanism 34 instead of the machining mechanism 30 compared to the wire electric discharge machine 100 shown in FIG. Unlike the machining mechanism 30, the machining mechanism 34 does not have the X-axis drive motor 11X and the Y-axis drive motor 11Y. The wire electric discharge machine 101 sends an axis movement command to the upper guide 6 and the lower guide 12 . Thereby, in the wire electric discharge machine 100, the upper guide 6 and the lower guide 12 move in the X-axis direction and the Y-axis direction.

As described above, the wire electric discharge machine 100 shown in FIG. 1 is of a type that moves the surface plate 8 according to the axis movement command from the NC control device 33, and the wire electric discharge machine 101 shown in FIG. In this system, the upper guide 6 and the lower guide 12 are moved by an axial movement command from 33 . In the following description, the wire electric discharge machine 100 shown in FIG. 1 will be described.

Here, the reason why the processed dimensions and processed shapes vary when processing the workpiece 7 having various thicknesses for each thickness region will be described. The machining dimension is the dimension of the workpiece 7 after machining, and the machining shape is the shape of the workpiece 7 after machining. In this embodiment, the machining dimension is the dimension of the workpiece 7 in the Y-axis direction, that is, the dimension when viewed from the Z-axis direction, and the machining shape is the dimension when the workpiece 7 is viewed from the X-axis direction. is the shape of Since the workpiece 7 has a height in the Z-axis direction, machining dimensions differ for each height. The shape of the workpiece 7 is determined by the machining dimensions for each height.

FIG. 3 is a diagram for explaining the structure of the workpiece machined by the wire electric discharge machine according to the embodiment. FIG. 3 shows a region of the workpiece 7 in the vicinity of a portion to be machined by the wire electrode 2. As shown in FIG. The workpiece 7 is machined in the X-axis direction so that grooves parallel to the X-axis direction are formed in the workpiece 7 .

The workpiece 7 is repeatedly processed multiple times in the X-axis direction. For example, the workpiece 7 is rough-machined in the first machining, semi-finishing in the second machining, and finishing in the third machining.

The workpiece 7 has, for example, a first thickness region 21 having a first thickness, a second thickness region 22 having a second thickness, and a third thickness. It is composed of a third plate thickness region 23 and a fourth plate thickness region 24 having a fourth plate thickness. The first plate thickness is, for example, 200 mm, the second plate thickness is, for example, 150 mm, the third plate thickness is, for example, 100 mm, and the fourth plate thickness is, for example, 50 mm. Hereinafter, any one of the first thickness region 21, the second thickness region 22, the third thickness region 23, and the fourth thickness region 24 may be referred to as a thickness region.

When the wire electric discharge machine 100 processes the workpiece 7 in the order of the first thickness region 21, the second thickness region 22, the third thickness region 23, and the fourth thickness region 24. , the plate thickness to be processed changes in order of the first plate thickness, the second plate thickness, the third plate thickness, and the fourth plate thickness. The workpiece 7 is machined by the wire electrode 2 between the upper nozzle 81 and the lower nozzle 82 . The distance between the upper nozzle 81 and the lower nozzle 82 is, for example, 310 mm.

FIG. 4 is a diagram for explaining the shape of the workpiece when the workpiece is machined without correcting the machining speed with respect to the plate thickness. FIG. 4 shows the processed shape, which is the shape after processing of the first thickness region 21 and the fourth thickness region 24 when the workpiece 7 is viewed from above.

In the machining in which the thickness changes, if the machining is performed without considering the thickness, the removed volume of the workpiece 7 in the direction of progress of machining is small in the fourth thickness region 24, which is the region where the thickness is thin. Since the machining speed is fast, it is difficult for electric discharge to fly to the side in the direction of machining progress. On the other hand, in the thick first plate thickness region 21, the removal volume of the workpiece 7 in the machining progress direction is large and the machining speed is slow, so the discharge tends to fly to the side face in the machining progress direction.

Therefore, the amount of the workpiece 7 cut by machining is small in the fourth thickness region 24 having a small thickness, and the amount of the workpiece 7 cut by machining is reduced in the first thickness region 21 having a large thickness. will increase. As a result, the width of the machined groove is narrow in the region where the plate thickness is thin, and the width of the machined groove is wide in the region where the plate thickness is thick. The width of the machining groove, that is, the machining dimension, includes the vibration caused by running of the wire electrode 2, the amount of separation from the upper nozzle 81 to the workpiece 7, the amount of separation from the lower nozzle 82 to the workpiece 7, the machining voltage, and the input voltage. It changes depending on the discharge energy, etc.

Further, even within the same plate thickness region, the amount of bending of the wire electrode 2 differs depending on the height of the workpiece 7, so the progress speed of machining differs depending on the height of the workpiece 7. That is, the progress speed of machining differs depending on the amount of separation from the upper nozzle 81 to the workpiece 7 and the separation amount from the lower nozzle 82 to the workpiece 7 . The nozzle separation amount is the separation distance between the upper nozzle 81 and the workpiece 7 and the separation distance between the lower nozzle 82 and the workpiece 7 . Since the workpiece 7 has various plate thickness regions, the nozzle separation amount differs for each plate thickness region. Therefore, if the workpiece 7 is machined without correcting the machining conditions such as the machining speed, the machined shape will vary.

The wire electric discharge machine 100 of the present embodiment adjusts the machining conditions such as the machining voltage so that the machining dimensional and machining shape differences in each region in the first machining can be corrected in the second and subsequent machining. adjust. That is, if the machining dimensional difference or machining shape difference that occurs in the first machining is large, the machining dimensions and machining shape may not be corrected completely. , machining is performed in accordance with changes in plate thickness so that the machining groove width is kept constant to some extent. In other words, the wire electric discharge machining apparatus 100 is designed so that the difference in the machined groove width between the plate thickness regions and the variation in the machined shape within the plate thickness region, which are caused by the plate thickness change, approach a constant value at the time of the first machining. processed into

The NC control device 33 of the embodiment calculates a wire tension command, a voltage correction value, and a pause time correction value based on the machining voltage, electrical discharge machining energy per unit time, and nozzle separation amount. The NC control device 33 controls the wire tension command and the voltage so as to reduce the processing dimensional difference and the processing shape difference between the different thickness regions for the workpiece 7 having various thicknesses for each thickness region. A correction value and a pause time correction value are calculated.

The workpiece 7 has various plate thickness regions, and the amount of machining differs at each position from the bottom surface to the top surface of the workpiece 7 . Therefore, the dimension of the workpiece 7 after machining differs for each height from the bottom surface of the workpiece 7 . In the present embodiment, the average value of post-machining dimensions for each height from the bottom surface of the workpiece 7 in one plate thickness region is referred to as a machining dimension. The processed dimension may be the median value of post-processed dimensions for each height from the bottom surface of the workpiece 7 in one plate thickness region. The NC control device 33 calculates a wire tension command, a voltage correction value, and an idle time correction value so that the machining dimensions in each thickness region of the workpiece 7 do not vary among the thickness regions.

In the present embodiment, the machining shape of the workpiece 7 is indicated by the straightness accuracy of the workpiece 7 in the Z-axis direction. The straightness accuracy corresponds to the variation in dimensional accuracy of the machined dimensions of the workpiece 7 corresponding to the deflection amount of the wire electrode 2 during wire electric discharge machining. Since the bending of the wire electrode 2 is bending in a direction perpendicular to the Z-axis direction, it includes a bending component in the X-axis direction and a bending component in the Y-axis direction. Since the bending component in the Y-axis direction affects the machining dimension of the workpiece 7 in the Y-axis direction, the bending component in the Y-axis direction will be described below.

FIG. 5 is a diagram for explaining the shape of the workpiece when the workpiece is machined without correcting the bending of the wire electrode with respect to the plate thickness. The horizontal axis of FIG. 5 is the machining dimension of the workpiece 7 in the Y-axis direction, and the vertical axis is the height of the workpiece 7 . As shown in FIG. 5 , the machining dimension of the workpiece 7 in the Y-axis direction differs for each height of the workpiece 7 .

A dimension curve 65 is the processing dimension in the first plate thickness region 21, and a dimension curve 66 is the processing dimension in the second plate thickness region 22. A dimension curve 67 is the machining dimension in the third thickness region 23 and a dimension curve 68 is the machining dimension in the fourth thickness region 24 . For example, in the first thickness region 21, the height of the workpiece 7 is from 0 to 200 mm. In the first plate thickness region 21, the machining dimension is small in the central region in the Z-axis direction of the workpiece 7 where the amount of bending of the wire electrode 2 is large. In addition, the machining dimension is small in the end region of the workpiece 7 in the Z-axis direction where the amount of bending of the wire electrode 2 is small.

That is, even within one plate thickness region, the wire electrode 2 bends so that the workpiece 7 is processed at upper and lower end regions of the wire electrode 2 (hereinafter referred to as wire end processing regions). ) and a portion machined in the central region of the wire electrode 2 (hereinafter referred to as a wire central machining portion).

As a result, the bending of the wire electrode 2 is greater at the wire center machining location than at the wire end machining location, so the area to be machined becomes wider. As a result, the machined portion at the center of the wire has a larger amount of machining than the portion machined at the wire end portion, so that the post-machining dimension becomes smaller. As described above, even within one plate thickness region, the post-machining dimensions of the workpiece 7 vary for each height from the bottom surface of the workpiece 7 due to the bending of the wire electrode 2 . Variations in machining dimensions within one plate thickness region correspond to variations in machining shape.

The NC control device 33 stores dimensional curve information, which is information on the dimensional curves 65 to 68, and adjusts the machining voltage, rest time, and wire tension based on the dimensional curve information when machining the workpiece 7. Control.

Ideally, the machining dimension in a plane parallel to the XY plane (the Y-axis direction in this embodiment) within one plate thickness region of the workpiece 7 is to be the same. Therefore, the NC control device 33 provides a wire tension command, a wire tension command, a A voltage correction value and a rest time correction value are calculated. Specifically, the NC control device 33 provides a wire tension command, a voltage correction value, and a pause value that reduce the amount of bending of the wire electrode 2 in the Y-axis direction, which is the direction perpendicular to the X-axis direction, which is the processing progress direction. Calculate the time correction value. In other words, the NC control device 33 calculates the wire tension command, the voltage correction value, and the pause time correction value that increase the straightness accuracy in each plate thickness region of the workpiece 7 .

The smaller the bending amount of the wire electrode 2 in the Y-axis direction, the higher the straightness accuracy of the wire electrode 2 , the smaller the error in the machining dimension between the wire end machined part and the wire center machined part, and the workpiece 7 . The straightness accuracy of is improved. As a result, errors in the machined shape are reduced, and the NC control device 33 can improve the accuracy of the machined shape.

The higher the precision of the machining dimensions, the more the dimension curves 65-68 overlap. Also, the higher the precision of the machining dimensions, the less the dimensional curves 65 to 68 curve, and the dimensional curves 65 to 68 approach straight lines parallel to the vertical axis in FIG.

The wire electric discharge machine 100 uses the following four parameters that determine the machining shape during machining.
(A) Machining voltage (B) EDM energy per unit time (C) Nozzle distance (D) Wire tension

The above four parameters have the following effects on the machining of the workpiece 7, respectively.
・The machining voltage affects the machining groove width (machining dimensions and straightness accuracy).
・Electrical discharge machining energy per unit time affects straightness accuracy and machining speed.
・The amount of nozzle separation affects the electric discharge machining energy.
・The wire tension affects the straightness accuracy due to the deflection of the wire electrode 2 .

A low machining voltage corresponds to a short distance between the wire electrode 2 and the workpiece 7 . When adjusting the machining voltage, the wire electric discharge machine 100 adjusts the distance between the wire electrode 2 and the workpiece 7 by adjusting the feed speed of the wire electrode 2 in the direction in which machining progresses. For example, when lowering the machining voltage, the wire electric discharge machine 100 shortens the distance between the wire electrode 2 and the workpiece 7 by increasing the feed speed of the wire electrode 2 in the machining progress direction. In this case, the size of the workpiece 7 to be machined becomes large because the volume removed from the side face in the direction of progress of machining is small. On the other hand, when increasing the machining voltage, the wire electric discharge machine 100 increases the distance between the wire electrode 2 and the workpiece 7 by decreasing the feed speed of the wire electrode 2 in the direction of progress of machining. In this case, the size of the workpiece 7 to be machined becomes small because the volume removed from the side face in the direction of progress of the machining increases. The feed speed of the wire electrode 2 in the machining advancing direction corresponds to the machining speed.

When the feed speed in the direction of machining progress is increased, the force that separates the wire electrode 2 from the work piece 7 due to the explosive force generated by the electric discharge causes the wire electrode 2 to move due to the electrostatic attraction caused by the current flowing through the wire electrode 2. Since the force is larger than the force to bring the workpiece closer to the workpiece 7, the machined surface has a shape in a swelling direction. On the other hand, when the feed speed in the machining progress direction is lowered, the electrostatic attraction becomes superior to the explosive force generated by electric discharge, so that the machined surface shape becomes concave.

The wire electric discharge machining apparatus 100 calculates the electric discharge machining energy based on the energy per electric discharge pulse and the number of pulses. In the wire electric discharge machining apparatus 100, when the electric discharge machining energy is lowered, the bending amount of the wire electrode 2 is reduced, so the straightness accuracy can be improved. In addition, the wire electric discharge machine 100 can increase the machining speed by increasing the electric discharge machining energy, because the amount of machining can be increased.

The wire electric discharge machining apparatus 100 detects the nozzle separation amount by a nozzle separation amount detector 49, which will be described later, or receives it from the user through a setting input IF (Interface, interface) 20, which will be described later.

As the nozzle separation increases, the amount of machining fluid supplied to the gap between the wire electrode 2 and the workpiece 7 decreases, so the electric discharge machining energy that can be applied to the wire electrode 2 disconnection limit decreases. Also, the higher the wire tension, the less the bending of the wire electrode 2, so the straightness accuracy is improved.

The wire electric discharge machine 100 controls machining so that the machining dimensions for each plate thickness region to be machined and the machining shape within each plate thickness region approach a constant value. For this reason, the manufacturer of the wire electric discharge machine 100 acquires in advance the results of each machining dimension and each machining shape when machining is executed with various combinations of the parameters (A) to (D). . The manufacturer of the wire electric discharge machining apparatus 100 formulates the relationship between the above parameters (A) to (D), the machining dimensions and the machining shape, thereby producing the geometry compensator 35, which will be described later. A geometry compensator 35 is arranged in the NC controller 33 and calculates a wire tension command, a voltage correction value and a rest time correction value.

That is, the manufacturer of the wire electric discharge machining apparatus 100 uses the formulated function to minimize the machining dimensional difference between the plate thickness regions, and the straightness accuracy difference, which is the machining shape, to the minimum within each plate thickness region. A control model is constructed and set in the geometry compensator 35 . This control model corresponds to control by the geometry compensator 35 . As a result, the shape and dimension compensator 35 is based on the machining dimension and straightness accuracy of the workpiece 7 when wire electric discharge machining is performed with a plurality of combinations of machining voltage, electric discharge machining energy, nozzle separation amount, and wire tension. A voltage correction value, a pause time correction value, and a wire tension command are calculated using the control model set in the above.

Voltage correction value information 77, energy correction value information, and first to third correspondence relationship information, which will be described later, are set in the shape and dimension compensator 35 based on the machining result of the prior machining processing by the wire electric discharge machining apparatus 100. be. In addition to the above parameters (A) to (D), in the pre-processing by the wire electric discharge machine 100, the diameter of the wire electrode 2, the material of the workpiece 7, and other parameters that affect the machining shape. may be combined in various ways. In this case, the manufacturer of the wire electric discharge machine 100 constructs a control model for each diameter of the wire electrode 2 and each material of the workpiece 7 and sets it in the geometry compensator 35 .

The geometry compensator 35 uses a control model corresponding to at least one of the diameter of the wire electrode 2 and the material of the workpiece 7 specified by the user. The geometry compensator 35 calculates a wire tension command, a voltage correction value, and an idle time correction value based on the machining voltage, machining energy, machining speed, nozzle separation amount, etc. during machining.

FIG. 6 is a block diagram showing a functional configuration example of the wire electric discharge machine according to the embodiment. The machining power supply 32 has a machining voltage detector 45 , a machining energy detector 46 , a feedback controller 43 , and

calculators

41 and 42 . The NC control device 33 has a plate thickness estimator 48 , a nozzle separation amount detector 49 , a setting input IF 20 and a geometry compensator 35 .

In the machining power supply 32 , the calculator 41 is connected to the calculator 42 , and the calculator 42 is connected to the feedback controller 43 . A feedback controller 43 is also connected to the processing mechanism 30 . Specifically, feedback controller 43 is connected to upper feeder 4 and lower feeder 5 .

The machining mechanism 30 is also connected to a machining voltage detector 45 , a machining energy detector 46 , a wire tension controller 31 , a nozzle separation amount detector 49 and a plate thickness estimator 48 . Machining voltage detector 45 is connected to plate thickness estimator 48 and geometry compensator 35 . Machining energy detector 46 is connected to plate thickness estimator 48 and geometry compensator 35 . The geometry compensator 35 is connected to the plate thickness estimator 48 , nozzle separation amount detector 49 , setting input IF 20 , calculator 41 and wire tension controller 31 .

The wire tension control device 31 controls the tension load device 3 of the processing mechanism 30. The NC control device 33 also controls the X-axis drive motor 11X, the Y-axis drive motor 11Y, and the like of the processing mechanism 30 .

The machining voltage detector 45 is connected to the wire electrode 2 via the upper feeder 4 or the lower feeder 5 and to the workpiece 7 . The machining voltage detector 45 detects the machining voltage between the wire electrode 2 and the workpiece 7 during machining. The machining voltage detected by the machining voltage detector 45 corresponds to the distance between the wire electrode 2 and the workpiece 7 . The machining voltage detector 45 sends the detected machining voltage to the calculator 42 , plate thickness estimator 48 , and geometry compensator 35 .

The machining energy detector 46 is connected to the wire electrode 2 via the upper feeder 4 or the lower feeder 5 and to the workpiece 7 . The machining energy detector 46 detects discharge pulses generated between the wire electrode 2 and the workpiece 7 during machining. The machining energy detector 46 calculates the electric discharge machining energy based on the energy per one discharge pulse and the number of pulses. Machining energy detector 46 sends electrical discharge machining energy to calculator 42 , plate thickness estimator 48 , and geometry compensator 35 .

A computing unit 41 receives a command voltage and a pause time from the NC control device 33 . The command voltage is a command value of voltage used for wire electric discharge machining. Arithmetic unit 41 receives the voltage correction value and the quiescent time correction value sent from geometry compensator 35 . The voltage correction value is a correction value for correcting the command voltage received by the calculator 41 from the NC control device 33 . The pause time correction value is a correction value for correcting the pause time received by the calculator 41 from the NC control device 33 . The voltage correction value and the rest time correction value are correction values for improving the accuracy of the machining dimension and the accuracy of the machining shape for the workpiece 7 having a plurality of plate thickness regions with different plate thicknesses on the machining path. is.

The calculator 41 subtracts the voltage correction value and the pause time correction value from the received command voltage and pause time, and sends the result to the calculator 42 . Note that the calculator 41 may be arranged in the NC control device 33 .

The calculator 42 subtracts the current machining voltage sent from the machining voltage detector 45 from the command voltage sent from the calculator 41 and sends the result to the feedback controller 43 . Further, the calculator 42 subtracts the current electric discharge machining energy sent from the machining energy detector 46 from the pause time sent from the calculator 41 and sends the result to the feedback controller 43 .

A feedback controller 43 controls the machining mechanism 30 using the results calculated by the calculator 42 . Specifically, the feedback controller 43 corrects the position of the surface plate 8 in the X-axis direction and the Y-axis direction by sending an axis movement command to the X-axis drive motor 11X and the Y-axis drive motor 11Y. Thereby, the feedback controller 43 controls the machining voltage, electric discharge machining energy, etc. for the machining mechanism 30 .

The X-axis drive motor 11X and the Y-axis drive motor 11Y of the machining mechanism 30 are each connected to an encoder, which detects the machining speed and sends it to the plate thickness estimator 48.

The plate thickness estimator 48 is based on the machining speed sent from the machining mechanism 30, the machining voltage sent from the machining voltage detector 45, and the electric discharge machining energy sent from the machining energy detector 46. , the plate thickness of the workpiece 7 is estimated. The plate thickness estimator 48 estimates the plate thickness of a portion of the workpiece 7 that is being machined during machining in which the plate thickness changes. The plate thickness estimator 48 sends the estimated plate thickness to the geometry compensator 35 as a plate thickness estimated value.

The nozzle separation amount detector 49 detects the nozzle separation amount from the machining mechanism 30 during machining and sends it to the geometry compensator 35 . The setting input IF 20 accepts the reference plate thickness input by the user and sends it to the geometry compensator 35 . The reference plate thickness is a plate thickness used as a reference for processing dimensions. The wire electric discharge machine 100 processes a plate thickness region other than the reference plate thickness so that the workpiece 7 has the machining dimensions of the reference plate thickness.

For example, if the geometry compensator 35 stores information on dimension curves 65 to 68 as dimension curve information, and 200 mm is specified as the reference plate thickness, the geometry compensator 35 stores the dimension curve 65 of 200 mm. Controls the machining so that is nearly parallel to the vertical axis. That is, the geometry compensator 35 calculates a voltage correction value, a pause time correction value, and a wire tension command that make the 200 mm dimension curve 65 parallel to the vertical axis.

Further, the geometry compensator 35 calculates a voltage correction value, a pause time correction value, and a wire tension command so that the machining dimension of 50 mm to 150 mm approaches the dimension curve 65 that is parallel to the vertical axis.

When the reference plate thickness is not specified, the geometry compensator 35 controls the machining so as to approach the machining dimension (dimension curve) of a specific plate thickness. The geometry compensator 35 controls the machining so as to approach the machining dimension of the thinnest plate, for example.

The setting input IF 20 may receive the nozzle separation amount from the user and send it to the geometry compensator 35 . In this case, the NC control device 33 may not have the nozzle separation amount detector 49 .

The shape and dimension compensator 35 sets a voltage correction value, a pause time correction value, and a wire tension command based on the machining voltage, machining energy, machining speed, plate thickness estimate, nozzle separation amount, and reference plate thickness during machining. calculate.

The geometry compensator 35 sends the voltage correction value and the pause time correction value to the calculator 41 and sends the wire tension command to the wire tension control device 31 . The wire tension controller 31 controls the machining mechanism 30 according to the wire tension command. Specifically, the wire tension control device 31 controls the tension load device 3 according to the wire tension command.

In this way, the shape and dimension compensator 35 is based on the machining voltage during machining, the machining energy during machining, the machining speed during machining, the estimated plate thickness during machining, the nozzle separation amount during machining, and the reference plate thickness. Since the voltage correction value, the pause time correction value, and the wire tension command are calculated in this way, it is possible to improve the accuracy of machining dimensions and the accuracy of the machining shape in rough machining, which is the first machining.

Here, the details of the parameters (A) to (D) described above will be described. First, the correction of the machining voltage according to the plate thickness estimated value will be described. A geometry compensator 35 calculates a voltage correction value of the machining voltage based on the plate thickness estimated value.

FIG. 7 is a diagram for explaining calculation processing of a voltage correction value by the geometry compensator according to the embodiment. FIG. 7 shows the configuration of the voltage correction value calculator 85 included in the geometry compensator 35. As shown in FIG.

The voltage correction value calculation unit 85 includes

calculators

75 and 76 . The computing unit 75 calculates the plate thickness corresponding voltage correction value based on the plate thickness estimated value and sends it to the computing unit 76 . The plate thickness corresponding voltage correction value is a correction value of the machining voltage corresponding to the plate thickness estimated value. The arithmetic unit 75 calculates the plate thickness corresponding voltage correction value using the voltage correction value information indicating the correspondence relationship between the plate thickness estimated value and the plate thickness corresponding voltage correction value.

FIG. 8 is a diagram for explaining voltage correction value information used by the geometry compensator according to the embodiment. In the graph of the voltage correction value information 77 shown in FIG. 8, the horizontal axis is the plate thickness estimated value, and the vertical axis is the plate thickness corresponding voltage correction value. In the voltage correction value information 77, the plate thickness corresponding voltage correction value is 0 at a low plate thickness, and from a specific plate thickness, the plate thickness corresponding voltage correction value increases in proportion to the plate thickness. Above the thickness, the plate thickness corresponding voltage correction value is not changed. In addition, when the wire electrode 2 is unlikely to be disconnected, the plate thickness corresponding voltage correction value may be increased even if the plate thickness is equal to or greater than a specific plate thickness.

The arithmetic unit 75 calculates the plate thickness corresponding voltage correction value based on the voltage correction value information 77 and the plate thickness estimated value. The voltage correction value information 77 may be a mathematical expression indicating the correspondence relationship between the plate thickness estimated value and the plate thickness corresponding voltage correction value, or may be a data table.

The calculator 76 calculates the voltage correction value by adding the plate thickness corresponding voltage correction value to the measured machining voltage, which is the measured machining voltage. The calculator 76 sends the voltage correction value to the calculator 41 . In this manner, the voltage correction value calculator 85 corrects the machining voltage for the plate thickness estimated value. The machining speed slows down in thicker plate thickness regions, and the bending of the wire electrode 2 widens the gap between the wire electrode 2 and the side surface of the workpiece 7, lowering the straightness accuracy. For this reason, the voltage correction value calculation unit 85 calculates a voltage correction value for correcting the machining voltage to be higher in the thick plate thickness region, thereby correcting the machining speed so as to increase. As a result, the voltage correction value calculator 85 suppresses the processing dimensional difference between the thin plate thickness region and the thick plate thickness region.

Next, the correction of the electric discharge machining energy according to the plate thickness estimated value will be explained. The geometry compensator 35 calculates a rest time correction value for correcting the electric discharge machining energy based on the plate thickness estimate.

FIG. 9 is a diagram for explaining calculation processing of the downtime correction value by the geometry compensator according to the embodiment. FIG. 9 shows the configuration of the pause time correction value calculator 86 provided in the geometry compensator 35. As shown in FIG.

The pause time correction value calculator 86 has calculators 63 , 64 and 80 . The calculator 63 calculates the target electric discharge machining energy based on the plate thickness estimated value and sends it to the calculator 64 . The target electric discharge machining energy is a target value of the electric discharge machining energy according to the plate thickness estimated value. The calculator 63 calculates the target electrical discharge machining energy using the energy correction value information indicating the correspondence relationship between the plate thickness estimated value and the target electrical discharge machining energy.

The calculator 64 calculates electric discharge machining energy by subtracting the target electric discharge machining energy from the current electric discharge machining energy, and sends it to the calculator 80 . The computing unit 80 calculates a rest time correction value based on the electric discharge machining energy received from the computing unit 64 . A computing unit 80 calculates a pause time correction value by a combination of proportional control and integral control. In this way, the calculator 80 sets the target electrical discharge machining energy based on the calculated plate thickness estimated value, and calculates the pause time correction value such that the target electrical discharge machining energy matches the current electrical discharge machining energy. to control the pause time.

Next, the wire tension command or voltage correction value according to the nozzle separation amount will be explained. The geometry compensator 35 calculates a wire tension command or a voltage correction value for correcting the machining voltage based on the nozzle separation amount.

Since the wire electrode 2 bends under the influence of electrical discharge reaction force, electrostatic attraction, etc. during machining, the straightness accuracy of the workpiece 7, that is, the shape accuracy, varies depending on the installation height of the workpiece 7. . FIG. 10 is a diagram for explaining the relationship between the amount of separation of the nozzle and the amount of deflection of the wire electrode. In FIG. 10, the position of the lower nozzle 82 is set to a height of 0, and the position of the upper nozzle 81 is set to a height of T5.

A workpiece 7D is a workpiece that is machined by the wire electrode 2 in a region from height 0 to height T1. A workpiece 7C is a workpiece to be processed by the wire electrode 2 in a region from height T2 (>T1) to height T3 (>T2). The workpiece 7B is a workpiece that is machined by the wire electrode 2 in a region from height T4 (>T3) to height T5 (>T4).

The nozzle separation amount of the workpiece 7D is a distance R3 from the

upper nozzle

81 and 0 from the lower nozzle 82. The workpiece 7C is separated from the nozzle by a distance R2a from the upper nozzle 81 and by a distance R2b from the lower nozzle . The nozzle separation amount of the workpiece 7B is 0 from the upper nozzle 81 and the distance R1 from the lower nozzle 82 . Both R2a and R2b are smaller values than R1 and R3.

As shown in FIG. 10, during machining, the wire electrode 2 bends in the Y-axis direction, which is the direction perpendicular to the machining progress direction. When the wire electrode 2 bends, the amount of bending is greatest at the central portion between the upper nozzle 81 and the lower nozzle 82, and the closer to the upper nozzle 81 or the lower nozzle 82, the smaller the bending amount.

Thus, the shorter the distance between the nozzle closer to the workpiece 7 and the workpiece 7, the smaller the deflection amount. In other words, the smaller the absolute value of the difference between the distance of the workpiece 7 from the lower nozzle 82 and the distance from the upper nozzle 81, the larger the deflection amount. In the workpiece 7B shown in FIG. 10, the absolute value of the difference between the distance from the lower nozzle 82 and the distance from the upper nozzle 81 is R1. The absolute value of the difference from 81 to the distance is R3. 10, the absolute value of the difference between the distance from the lower nozzle 82 and the distance from the upper nozzle 81 is |R2a-R2b|, which is smaller than R1 and R3.

In this way, the workpiece 7C to be machined at the central portion of the wire electrode 2 has a high straightness accuracy because the wire electrode 2 has approximately the same deflection amount on both the upper and lower surfaces of the workpiece 7C. On the other hand, the

workpieces

7B and 7D machined by the ends of the wire electrode 2 have lower straightness accuracy because the wire electrode 2 bends differently between the upper and lower surfaces of the

workpieces

7B and 7C.

Therefore, if the shape and dimension compensator 35 can acquire the nozzle separation amount via the nozzle separation amount detector 49 or the setting input IF 20, the shape and dimension compensator 35 can perform processing control according to the nozzle separation amount, thereby making the workpiece 7 Straightness accuracy can be improved.

The shape and dimension compensator 35 calculates the wire tension for improving the straightness accuracy of the workpiece 7 based on the first correspondence information indicating the correspondence between the nozzle separation amount and the wire tension. The geometry compensator 35 can reduce the deflection amount of the wire electrode 2 by, for example, increasing the wire tension, thereby improving the straightness accuracy.

Further, the shape and dimension compensator 35 calculates a voltage correction value for improving the straightness accuracy of the workpiece 7 based on the second correspondence information indicating the correspondence between the nozzle distance and the voltage correction value. do. The shape and dimension compensator 35 can increase the machining speed by lowering the machining voltage by the voltage correction value, so it is possible to reduce the amount of machining at locations where the nozzle separation amount is large, and control the straightness accuracy and machining dimensional accuracy. can do.

Next, the wire tension command according to the plate thickness estimated value will be explained. A geometry compensator 35 calculates a wire tension command based on the plate thickness estimated value.

When the plate thickness of the workpiece 7 is thick, the processing speed becomes slow, so the amount of processing increases. In particular, at the central portion of the wire electrode 2, the bending of the wire electrode 2 increases the machining amount. In this case, since the wire electric discharge machine 100 can reduce the bending of the wire electrode 2 by increasing the wire tension, it is possible to reduce the machining amount in the region where the amount of bending of the wire electrode 2 is large, and improve the straightness accuracy. can be made

FIG. 11 is a diagram for explaining the relationship between the wire tension and the deflection amount of the wire electrode. In FIG. 11, the wire electrode 2B shows a wire electrode with a large amount of deflection, and the wire electrode 2A shows a wire electrode with an increased wire tension to reduce the amount of deflection.

As shown in FIG. 11 , the wire electric discharge machine 100 can suppress the bending of the wire electrode 2 by increasing the tension of the wire electrode 2 , thereby improving the straightness accuracy of the workpiece 7 . When increasing the wire tension, the wire electric discharge machine 100 increases the wire tension only to such an extent that the probability of the wire electrode 2 breaking is smaller than a specific value. The geometry compensator 35 calculates the wire tension for improving the straightness accuracy of the workpiece 7 based on the third correspondence information indicating the correspondence between the plate thickness estimated value and the wire tension.

Next, a processing procedure for wire electric discharge machining by the wire electric discharge machining apparatus 100 will be described. FIG. 12 is a flowchart showing a processing procedure of wire electric discharge machining by the wire electric discharge machine according to the embodiment.

When the wire electric discharge machine 100 starts wire electric discharge machining (step S10), the NC controller 33 collects data (step S20). Specifically, the plate thickness estimator 48 receives the machining voltage, electrical discharge machining energy, and machining speed. Further, the nozzle separation amount detector 49 detects the nozzle separation amount, and the setting input IF 20 receives the reference plate thickness.

The plate thickness estimator 48 estimates the plate thickness of the workpiece 7 based on the machining voltage, electric discharge machining energy, and machining speed (step S30). The plate thickness estimator 48 sends the estimated plate thickness to the geometry compensator 35 as a plate thickness estimated value.

The shape and dimension compensator 35 calculates a wire tension command, a voltage correction value, and an idle time correction value based on the plate thickness estimated value, the nozzle separation amount, and the reference plate thickness (step S40). The wire electric discharge machine 100 controls machining voltage, machining energy, and wire tension using the wire tension command, voltage correction value, and rest time correction value (step S50). Specifically, the feedback controller 43 feedback-controls the machining mechanism 30 with the voltage value and the rest time corresponding to the voltage correction value and the rest time correction value, and the wire tension controller 31 adjusts the wire tension of the wire electrode 2. Control.

It should be noted that the shape and dimension compensator 35 may estimate and store the machined groove width at the time of the first machining. In this case, the geometry compensator 35 estimates the machined groove width based on the first machining conditions used in the first machining and the dimension curve information. Moreover, the wire electric discharge machining apparatus 100 may store an estimated plate thickness value at the time of the first machining. The shape and dimension compensator 35 associates the estimated machining groove width and plate thickness estimated values with coordinate information indicating the machining position of the workpiece 7, and stores them as machining result information.

The shape and dimension compensator 35 uses, for example, the correspondence relationship between the machining groove width and the coordinate information included in the machining result information to determine the second machining conditions, which are the machining conditions used in the second and subsequent machining, and the offset Adjust at least one of the amounts. The offset amount is the amount of shift of the machining position (the position of the wire electrode 2 in the Y-axis direction) used in the second and subsequent machining operations toward the workpiece 7 side.

In the second and subsequent machining, the machining amount changes depending on the offset amount, making it difficult to estimate the plate thickness. Processing control can be executed based on the processing result information for the second and subsequent processing as well.

The geometry compensator 35 calculates the voltage correction value, the pause time correction value, and the wire tension command, for example, using the correspondence relationship between the plate thickness estimated value and the coordinate information included in the machining result information.

In this way, the wire electric discharge machining apparatus 100 uses the geometry compensator 35 to calculate the voltage correction value, the pause time correction value, and the wire tension command. From the machining of , the machining dimension and straightness accuracy can be improved regardless of the plate thickness region of the workpiece 7 .

Further, the wire electric discharge machine 100 controls the axis movement command for maintaining the machining voltage at a specific value and the time to start applying the machining voltage next time, so that the wire electric discharge machine 100 can continuously perform machining without changing the machining conditions. The control can control the machining shape and machining dimensions of the workpiece 7 .

In the present embodiment, a case where the shape and dimension compensator 35 calculates the voltage correction value, the rest time correction value, the wire tension, etc. has been described. , wire tension, etc. may be calculated. That is, the functions modeling the machining dimensions and straightness accuracy may be derived experimentally or may be derived by a learning device. When derived by the learning device, the learning device calculates the voltage correction value, the pause time correction value, and the wire tension command so that the machining dimensions of the plurality of plate thickness regions approach the machining dimensions of the specific plate thickness region. .

In the case of experimental derivation, the manufacturer of the wire electric discharge machining apparatus 100 creates a function modeling machining dimensions and straightness accuracy based on dimension information, which is information on machining dimensions contained in past machining results. Set in the compensator 35 .

When derived by a learning device, the learning device calculates the machining dimensions based on information obtained in the machining process (hereinafter referred to as process information) such as plate thickness estimated value, machining voltage, electric discharge machining energy, machining speed, and nozzle separation amount. Information such as a voltage correction value, a pause time correction value, wire tension, etc. (hereinafter referred to as precision improvement information) that can improve the precision of .

The learning device generates a learned model that derives accuracy improvement information that can improve the accuracy of machining dimensions and straightness accuracy from the process information. In other words, the learning device generates a learned model, which is a function that models the correspondence relationship between the process information and the accuracy improvement information that can improve the machining dimension accuracy and the straightness accuracy. The inference device uses the learned model to derive accuracy improvement information capable of improving machining dimensional accuracy and straightness accuracy from the process information.

<Learning phase>
13 is a block diagram of a configuration example of a learning device according to an embodiment; FIG. The learning device 50 includes a data acquisition section 51 and a model generation section 52 . The data acquisition unit 51 acquires processing results (behavior) and processing parameters (state) as learning data.

The machining results are the machining dimensions and machining shape (straightness accuracy). The machining parameters are a combination of parameters that affect the machining shape, such as plate thickness, wire diameter of the wire electrode 2, material of the workpiece 7, machining voltage, electric discharge machining energy, nozzle separation amount, and wire tension.

The model generation unit 52 learns the voltage correction value, the pause time correction value, and the wire tension command based on the learning data including the action that is the machining result and the state that is the machining parameter. That is, the model generator 52 generates the learned model 71 for inferring the voltage correction value, the pause time correction value, and the wire tension command from the machining parameters of the wire electric discharge machine 100 .

The model generation unit 52 can use known learning algorithms such as supervised learning, unsupervised learning, and reinforcement learning. As an example, a case where reinforcement learning is applied to the model generation unit 52 will be described. In reinforcement learning, an agent (action subject) in an environment observes the current state (environmental parameters) and decides what action to take. The environment dynamically changes according to the actions of the agent, and the agent is rewarded according to the change in the environment. The agent repeats this and learns the course of action that yields the most rewards through a series of actions. As representative methods of reinforcement learning, Q-learning, TD-learning, and the like are known. For example, in the case of Q-learning, a general update formula for the action-value function Q(s, a) is represented by formula (1) below.

In equation (1), s _t represents the state of the environment at time t, and a _t represents the action at time t. Action a _t changes the state to s _t+1 . r _t+1 represents the reward obtained by changing the state, γ represents the discount rate, and α represents the learning coefficient. γ is in the range of 0<γ≦1, and α is in the range of 0<α≦1. The action that is the result of processing becomes the action at, the state that is the processing parameter becomes the state _st , and the model generator 52 _learns the best action at in the state _st at time _t .

The update formula represented by formula (1) increases the action value Q if the action value Q of action a with the highest Q value at time t+1 is greater than the action value Q of action a executed at time t. On the contrary, the action value Q is decreased. In other words, the action value function Q(s, a) is updated so that the action value Q of action a at time t approaches the best action value at time t+1. As a result, the best behavioral value in a certain environment will be propagated to the behavioral value in the previous environment.

As described above, when the learned model 71 is generated by reinforcement learning, the model generation unit 52 includes the reward calculation unit 53 and the function update unit 54.

The reward calculation unit 53 calculates rewards based on the processing results and processing parameters. The remuneration calculator 53 calculates a remuneration r based on the machining accuracy, that is, the machining dimension accuracy and the machining shape accuracy. For example, if the machining accuracy is improved, the reward r is increased (for example, a reward of "1" is given.) On the other hand, if the machining accuracy is deteriorated, the reward r is reduced (for example, a reward of "-1" is given .).

The function updating unit 54 updates the functions for determining the voltage correction value, the pause time correction value, and the wire tension command according to the reward calculated by the reward calculation unit 53 and outputs them to the learned model storage unit 70 . For example, in the case of Q-learning, the action value function Q(s _t , a _t ) represented by Equation (1) is used as a function for calculating the voltage correction value, rest time correction value, and wire tension command.

The function updating unit 54 repeatedly executes the learning as described above. The learned model storage unit 70 stores the action-value function Q(s _t , a _t ) updated by the function updating unit 54 , that is, the learned model 71 .

Next, a processing procedure of learning processing by the learning device 50 will be described with reference to FIG. 14 is a flowchart of a procedure of learning processing by the learning device according to the embodiment; FIG. The data acquisition unit 51 acquires the processing result and processing parameters as learning data (step S110).

The model generation unit 52 calculates a reward based on the processing result and processing parameters (step S120). Specifically, the remuneration calculation unit 53 of the model generation unit 52 acquires the processing result and the processing parameters, and determines whether to increase the remuneration (step S130) or decrease the remuneration based on the predetermined processing accuracy. (step S140). Reward criteria are whether the precision of machined dimensions and the precision of machined shapes have improved or deteriorated. The model generator 52 determines to increase the reward when the accuracy of the machining dimension and the accuracy of the machining shape improve, and determines to decrease the reward when the accuracy of the machining dimension and the accuracy of the machining shape deteriorate.

The model generation unit 52 may determine that the reward should be increased or decreased if either one of the precision of the machining dimension and the precision of the machining shape improves and the other deteriorates. Further, the model generation unit 52 does not have to increase or decrease the reward when one of the precision of the machining dimension and the precision of the machining shape improves and the other deteriorates.

When the remuneration calculation unit 53 determines to increase the remuneration, it increases the remuneration in step S130. On the other hand, if the remuneration calculator 53 determines to reduce the remuneration, it reduces the remuneration in step S140.

The function updating unit 54 updates the action value function Q(s _t , a _t ) represented by Equation (1) stored in the learned model storage unit 70 based on the reward calculated by the reward calculation unit 53. (Step S150).

The learning device 50 repeatedly executes steps S110 to S150 described above, and stores the generated action-value function Q(s _t , a _t ) as a learned model 71 in the learned model storage unit 70 .

The learning device 50 according to the present embodiment stores the learned model 71 in the learned model storage unit 70 provided outside the learning device 50. It may be provided inside.

<Utilization phase>
15 is a block diagram of a configuration example of an inference apparatus according to an embodiment; FIG. The inference device 60 includes a data acquisition unit 61 and an inference unit 62 . The data acquisition unit 61 acquires processing parameters.

The inference unit 62 infers the processed information 79 using the learned model 71 stored in the learned model storage unit 70 . The machining information 79 is a voltage correction value, a pause time correction value, and a wire tension command. That is, the inference unit 62 inputs the machining parameters acquired by the data acquisition unit 61 to the learned model 71 to infer the voltage correction value, the pause time correction value, and the wire tension command suitable for the machining parameters. can be done.

In the present embodiment, a case has been described in which inference device 60 uses learned model 71 learned by model generation unit 52 of learning device 50. However, using learned model 71 acquired from another learning device, good too. Also in this case, the reasoning device 60 outputs the voltage correction value, the pause time correction value, and the wire tension command based on the learned model 71 acquired from another learning device.

Next, the procedure of inference processing by the inference device 60 will be described with reference to FIG. 16 is a flowchart of a procedure of inference processing by the inference apparatus according to the embodiment; FIG. The data acquisition unit 61 acquires inference data that is data for inferring the voltage correction value, the pause time correction value, and the wire tension command (step S210). Specifically, the data acquisition unit 61 acquires processing parameters.

The inference unit 62 inputs machining parameters to the learned model 71 stored in the learned model storage unit 70 (step S220), and obtains a voltage correction value, a pause time correction value, and a wire tension command. The inference unit 62 outputs the obtained data, that is, the voltage correction value, the pause time correction value, and the wire tension command to the wire electric discharge machine 100 (step S230).

The wire electric discharge machine 100 corrects the machining voltage and the rest time using the voltage correction value and the rest time correction value output from the inference unit 62 (step S240), and adjusts the wire tension using the wire tension command output from the inference unit 62. Control the tension of the electrode 2; As a result, the wire electric discharge machine 100 can improve the precision of the machined dimensions and the machined shape of the workpiece 7 .

In addition, in the present embodiment, the case where reinforcement learning is applied to the learning algorithm used by the inference unit 62 has been described, but it is not limited to this. As for the learning algorithm, supervised learning, unsupervised learning, or semi-supervised learning can be applied in addition to reinforcement learning.

In addition, as the learning algorithm used in the model generation unit 52, deep learning that learns to extract the feature amount itself can also be used, and other known methods such as neural networks, genetic programming, function Machine learning may be performed according to logic programming, support vector machines, and the like.

Note that the learning device 50 and the reasoning device 60 may be connected to the wire electric discharge machine 100 via a network, for example, and may be separate devices from the wire electric discharge machine 100 . At least one of the learning device 50 and the reasoning device 60 may be built in the wire electric discharge machine 100 . Furthermore, learning device 50 and reasoning device 60 may reside on a cloud server.

Also, the model generation unit 52 may learn the voltage correction value, the pause time correction value, and the wire tension command using learning data acquired from a plurality of wire electric discharge machines. Note that the model generation unit 52 may acquire learning data from a plurality of wire electric discharge machines used in the same area, or may acquire learning data from a plurality of wire electric discharge machines operating independently in different areas. The voltage correction value, the rest time correction value, and the wire tension command may be learned using the learning data. Also, it is possible to add or remove the wire electric discharge machine for collecting the learning data from the target on the way. Furthermore, the learning device 50 that has learned the voltage correction value, the pause time correction value, and the wire tension command for a certain wire electric discharge machine is applied to a different wire electric discharge machine, and for the other wire electric discharge machine, The voltage correction value, pause time correction value, and wire tension command may be re-learned and updated.

Here, the hardware configuration of the NC control device 33 will be explained. FIG. 17 is a diagram of a hardware configuration example that implements the NC control device according to the embodiment. The NC controller 33 can be implemented by a processor 91 , memory 92 , output device 93 and input device 94 .

An example of the processor 91 is a CPU (Central Processing Unit, central processing unit, processing unit, arithmetic unit, microprocessor, microcomputer, DSP (Digital Signal Processor)) or system LSI (Large Scale Integration). Examples of the memory 92 are RAM (Random Access Memory) and ROM (Read Only Memory).

The NC control device 33 is implemented by the processor 91 reading and executing a computer-executable control program for executing the operation of the NC control device 33 stored in the memory 92 . It can also be said that the control program for executing the operation of the NC control device 33 causes the computer to execute the procedure or method of the NC control device 33 . The control program for executing the operation of the NC control device 33 includes a program for machining the workpiece 7, a program for executing the operation of the geometry compensator 35, and the like.

The control program executed by the NC controller 33 has a module configuration including a plate thickness estimator 48, a geometry compensator 35, and a nozzle separation amount detector 49, which are loaded onto the main memory. and these are generated on the main memory.

The input device 94 accepts the reference plate thickness and the like and sends them to the processor 91 . The memory 92 stores voltage correction value information 77, energy correction value information, first to third correspondence information, dimension curve information, and the like. Also, the memory 92 is used as a temporary memory when the processor 91 executes various processes.

The output device 93 outputs the voltage correction value and the pause time correction value generated by the processor 91 to the machining power supply 32 . The output device 93 also outputs the wire tension command generated by the processor 91 to the wire tension control device 31 .

The control program may be stored in a computer-readable storage medium in an installable or executable format and provided as a computer program product. Also, the control program may be provided to the NC control device 33 via a network such as the Internet. It should be noted that the functions of the NC control device 33 may be partly realized by dedicated hardware such as a dedicated circuit, and partly realized by software or firmware.

The feedback controller 43, the wire tension control device 31, the learning device 50, and the inference device 60 have the same hardware configuration as the wire electric discharge machine 100, so description thereof will be omitted.

As described above, in the embodiment, the shape and dimension compensator 35 reduces the difference in machining dimension between the plate thickness regions based on the machining voltage, machining energy, machining speed, nozzle separation amount, and plate thickness estimated value. Also, the voltage correction value, the pause time correction value, and the wire tension command are calculated so that the straightness accuracy of the workpiece is high within the plate thickness region. Then, the machining mechanism 30 uses the voltage correction value, the rest time correction value, and the wire tension command to perform wire electric discharge machining on the workpiece 7 whose plate thickness changes during machining. As a result, the wire electric discharge machine 100 can improve the precision of machining dimensions and the precision of the machined shape even for the workpiece 7 whose plate thickness changes during machining.

The configuration shown in the above embodiment is an example, and can be combined with another known technique, and part of the configuration can be omitted or changed without departing from the scope of the invention. It is possible.

1 wire electrode bobbin, 2, 2A, 2B wire electrode, 3 tension load device, 4 upper feeder, 5 lower feeder, 6 upper guide, 7, 7B, 7C, 7D workpiece, 8 surface plate, 9 wire Travel speed control motor, 10 wire electrode collection box, 11X X-axis drive motor, 11Y Y-axis drive motor, 12 lower guide, 13 lower roller, 20 setting input IF, 21 first plate thickness area, 22 second plate thickness Region, 23 Third plate thickness region, 24 Fourth plate thickness region, 30, 34 Machining mechanism, 31 Wire tension control device, 32 Machining power supply, 33 NC control device, 35 Geometry compensator, 41, 42, 63 , 64, 75, 76, 80 calculator, 43 feedback controller, 45 machining voltage detector, 46 machining energy detector, 48 plate thickness estimator, 49 nozzle separation amount detector, 50 learning device, 51 data acquisition unit, 52 model generation unit, 53 reward calculation unit, 54 function update unit, 60 inference device, 61 data acquisition unit, 62 inference unit, 70 learned model storage unit, 71 learned model, 77 voltage correction value information, 79 processing information, 81 upper nozzle, 82 lower nozzle, 85 voltage correction value calculation unit, 86 rest time correction value calculation unit, 91 processor, 92 memory, 93 output device, 94 input device, 100, 101 wire electric discharge machine.

Claims

a machining mechanism that performs wire electric discharge machining on a workpiece having a plurality of thickness regions with different thicknesses on a machining path using voltage pulses from a wire electrode;
a thickness estimator for estimating the thickness of the workpiece during wire electric discharge machining;
Based on the machining voltage during machining, the machining energy during machining, the machining speed during machining, the clearance that is the distance between the nozzle that supplies the machining fluid to the wire electrode and the workpiece, and the plate thickness and correcting the machining voltage so that the difference in machining dimension between the plate thickness regions becomes small and the straightness accuracy of the wire electrode of the workpiece in the length direction becomes high within each of the plate thickness regions. a geometry compensator that calculates a voltage correction value that is a value, a rest time correction value that is a correction value for the rest time of the voltage pulse, and a wire tension command that is a tension command to the wire electrode;
with
The machining mechanism is controlled using the voltage correction value, the dwell time correction value, and the wire tension command,
A wire electric discharge machine characterized by:
The geometry compensator performs the machining of the workpiece when wire electric discharge machining is performed with a plurality of combinations of the machining voltage, the electric discharge machining energy, the separation distance, and the wire tension of the wire electrode. calculating the voltage correction value, the pause time correction value, and the wire tension command using a control model set based on the dimensions and the straightness accuracy;
The wire electric discharge machine according to claim 1, characterized in that:
At least one of the wire diameter of the wire electrode and the material of the workpiece is set in the control model.
The wire electric discharge machine according to claim 2, characterized in that:
The geometry compensator is
Based on the dimension information, which is the information of the machining dimension contained in the past machining result, the voltage correction value and the pause are adjusted so that the machining dimension of the plurality of plate thickness regions approaches the machining dimension of the specific plate thickness region. calculating a time correction value and the wire tension command;
The wire electric discharge machine according to any one of claims 1 to 3, characterized in that:
The shape and dimension compensator estimates a machining groove width of the workpiece based on the first machining conditions and the dimensional information used in the first wire electric discharge machining of the workpiece, and Based on the plate thickness estimated in the first wire electric discharge machining of the workpiece and the machined groove width, the wire electrode used in the second and subsequent wire electric discharge machining is applied to the workpiece. Adjust the offset amount and the second processing condition, which is the amount of shift of
The wire electric discharge machine according to claim 4, characterized in that:
The specific thickness region is the thinnest thickness region among the plurality of thickness regions,
The wire electric discharge machine according to claim 4, characterized in that:
The plate thickness of each of the plate thickness regions estimated when performing wire electric discharge machining using a voltage pulse from a wire electrode on a workpiece having a plurality of plate thickness regions with different plate thicknesses on the machining path, Based on the machining voltage, machining energy, machining speed, and the distance between the workpiece and the nozzle that supplies machining fluid to the wire electrode when machining each of the plate thickness regions, the The correction value of the machining voltage is such that the difference in machining dimensions between the plate thickness regions becomes small and the straightness accuracy of the wire electrode of the workpiece in the length direction becomes high within each of the plate thickness regions. calculating a voltage correction value, a pause time correction value that is a correction value for the pause time of the voltage pulse, and a wire tension command that is a tension command to the wire electrode;
A geometry compensator characterized by:
A machining step in which a wire electric discharge machine performs wire electric discharge machining on a workpiece having a plurality of thickness regions each having a different thickness on a machining path, using voltage pulses from a wire electrode,
The processing step includes
an estimation step in which the wire electric discharge machine estimates the plate thickness of the workpiece during wire electric discharge machining;
In the wire electric discharge machine, the machining voltage during machining, the machining energy during machining, the machining speed during machining, the distance between the nozzle that supplies machining fluid to the wire electrode and the workpiece, and based on the plate thickness, the difference in machining dimension between the plate thickness regions becomes small, and the straightness accuracy in the length direction of the wire electrode of the workpiece becomes high within each of the plate thickness regions. a calculation step of calculating a voltage correction value that is a correction value of the machining voltage, a rest time correction value that is a correction value of the rest time of the voltage pulse, and a wire tension command that is a tension command to the wire electrode;
including
The wire electric discharge machine controls the wire electric discharge machining using the voltage correction value, the pause time correction value, and the wire tension command.
A wire electric discharge machining method characterized by:
Machining parameters of a wire electric discharge machine that performs wire electric discharge machining on a workpiece having a plurality of thickness regions with different thicknesses on a machining path using voltage pulses from a wire electrode; a data acquisition unit that acquires learning data including machining results of the wire electric discharge machine;
Using the learning data, from the machining parameters of the wire electric discharge machine, a voltage correction value that is a correction value of the machining voltage, a rest time correction value that is a correction value of the rest time of the voltage pulse, and the wire electrode A model generation unit that generates a trained model for inferring a wire tension command that is a tension command of
A learning device comprising:
The learned model is
Based on the machining voltage during machining, the machining energy during machining, the machining speed during machining, the clearance that is the distance between the nozzle that supplies the machining fluid to the wire electrode and the workpiece, and the plate thickness , the voltage correction value is such that the difference in machining dimension between the plate thickness regions becomes small and the straightness accuracy of the wire electrode of the workpiece in the length direction becomes high within each of the plate thickness regions. , inferring the dwell time correction value and the wire tension command;
10. The learning device according to claim 9, characterized by:
a data acquisition unit that acquires machining parameters of a wire electric discharge machine that performs wire electric discharge machining on a workpiece having a plurality of thickness regions with different thicknesses on a machining path using voltage pulses from a wire electrode; ,
Using a learned model for inferring the machining result of the wire electric discharge machine from the machining parameters, the voltage correction value that is the correction value of the machining voltage, the voltage pulse from the machining parameters acquired by the data acquisition unit an inference unit that infers and outputs a rest time correction value that is a correction value for the rest time of and a wire tension command that is a tension command to the wire electrode;
An inference device characterized by comprising:
The learned model is
Based on the machining voltage during machining, the machining energy during machining, the machining speed during machining, the clearance that is the distance between the nozzle that supplies the machining fluid to the wire electrode and the workpiece, and the plate thickness , the voltage correction value is such that the difference in machining dimension between the plate thickness regions becomes small and the straightness accuracy in the length direction of the wire electrode of the workpiece becomes high within each of the plate thickness regions. , inferring the dwell time correction value and the wire tension command;
12. The reasoning apparatus according to claim 11, characterized by: