Jump Control System for an Electric Discharge Machine

1. Technical Field

The present invention relates to a jump control system for an electric discharge machine, and more particularly, to a jump control system for an electric discharge machine in which a jump period and jump distance corresponding to the discharge state are automatically decided by a fuzzy control.

2. Background Art

In an electrode feed control in an electric discharge machine such as an electric discharge engraving machine, an average voltage of a discharge voltage waveform and a delay time before the start of an electric discharge, etc., are detected, and the distance between the electrode and a workpiece is controlled by servomotors so that these detected values are adjusted to a reference target voltage or a reference target delay time.

The control of the distance between the electrode and the workpiece is generally effected in the following manner.

(a) The average voltage of the discharge voltage waveform or the delay time before the start of the electric discharge is detected.

(b) The difference between this detected value and the reference target voltage or the reference target delay time is obtained.

(c) A value obtained by multiplying this difference by a certain gain is used as a move command for the servomotors.

(d) The above items (a) to (c) are repeated at predetermined intervals.

Accordingly, the electrodes are controlled so as to always obtain a reference target distance.

In such an electrical discharge machining, sludge (machining tailings) accumulates between the electrode and the workpiece, and therefore, a jump control is effected in which the electrodes are moved far apart, to discharge the sludge, and are then returned to their respective original positions.

Although this jump control is essential to an electrical discharge machining, it makes no contribution to the electrical discharge machining itself. Therefore, if the jump period of the jump control is too short, or if the jump distance is too long, the electrical discharge machining time is prolonged. On the other hand, if the jump period is too long, or if the jump distance is too short, an ineffective electric discharge, such as an arc or a short, frequently occurs, and thus the discharge state becomes unstable, the machining speed is lowered, and the accuracy of the machined surfaces is poor.

Further, the discharge state is momentarily changed by various factors, for example, the amount of sludge between the electrode and the workpiece, the machining depth, the ionic conditions of machined portions, and the machining shape, etc., and thus it is difficult to accurately detect the discharge state due to the existence of many ambiguous factors. In the conventional electrical discharge machining, therefore, the jump conditions are decided before the machining, on the basis of experience and experiments, and assuming the existence of the worst case.

For a high-speed stable machining, however, the jump control conditions must be optimally controlled during the machining operation, depending on the discharge state, and thus the method of controlling the jump period and jump distance during the machining operation is one of the most important problems now confronting the existing electrical discharge machining technique.

The present invention has been contrived in consideration of these circumstances, and an object thereof is to provide a jump control system for an electric discharge machine in which a jump period and jump distance corresponding to the discharge state are automatically decided by a fuzzy control.

To solve the above problem, according to the present invention, there is provided a jump control system for an electrical discharge machine for controlling a jump of the electrical discharge machine, in which machining is effected while automatically setting the period and distance of the jump to optimum values by a fuzzy control, using the discharge state of the electrical discharge machine as an input.

The optimum jump period and distance are controlled by the fuzzy control based on the discharge state, and accordingly, the machining speed and the electrical discharge machining state can be improved, thereby ensuring an improved accuracy of the machined surfaces.

FIG. 1 is a diagram showing an outline of a CNC electric discharge engraving machine according to one embodiment of the present invention;

FIG. 2 is a diagram showing a time-based change of the discharge stability;

FIG. 3 illustrates an example of a fuzzy rule for inferring an increment of a jump distance;

FIG. 4 illustrates an example of a fuzzy rule for inferring an increment of a jump period;

FIG. 5 is a diagram showing another embodiment of the present invention; and

FIG. 6 illustrates an example of a fuzzy rule for inferring the discharge stability.

One embodiment of the present invention will now be described with reference to the drawings.

FIG. 1 is a diagram showing an outline of a CNC electric discharge engraving machine according to one embodiment of the present invention.

The electric discharge machine comprises an electrode 34, an XY movement table 36, a power supply 37 for machining, and X-, Y-, and Z-axis servomotors 31, 32, and 33. The XY movement table 36, which is driven by the X- and Y-axis servomotors 31 and 32, moves a workpiece 35 in the X and Y directions, and the electrode 34, which is driven by the Z-axis servomotor 33, moves in the Z direction. The machining power supply 37 applies a high-frequency discharge voltage between the electrode 34 and the workpiece 35, and the workpiece 35 is machined to a desired shape by an electric discharge between the electrode 34 and the workpiece 35 caused by a relative movement of the XY movement table 36 and the electrode 34.

The X-, Y-, and Z-axis servomotors 31, 32, and 33, and the machining power supply 37, are controlled by a numerical control device. The following is a description of the arrangement of the numerical control device.

A machining condition setting means 13 outputs various control parameters for controlling the whole electric discharge machine. The various data for the machining condition setting means 13 is input through a CRT/MDI 11. The machining condition setting means 13 outputs a machining power supply parameter to the machining power supply 37, to thereby control the output of the machining power supply 37, and further, the machining condition setting means 13 outputs control parameters to the electrode swing control means 14 and electrode servo feed control means 15. The respective outputs of the electrode swing control means 14 and the electrode servo control feed means 15 are applied to the input of servo control means 16, and based on these outputs, the servo control means 16 controls the servomotors 31, 32, and 33.

A gap voltage waveform detecting means 17a detects the waveform of a gap voltage between the electrode 34 and the workpiece 35, and the output thereof is sent to an A/D converter 18. Further, a discharge current detecting means 17b detects a discharge current, and the output thereof is sent to the A/D converter 18. The A/D converter 18 digitizes the respective output waveforms of the gap voltage waveform detecting means 17a and the discharge current detecting means 17b, and the outputs thereof are sent to a waveform feature extracting means 19. The waveform feature extracting means 19 extracts an average voltage Vm, discharge delay time Tn, and discharge current value In, etc., from the digitized voltage waveforms, and outputs same to a discharge state estimating means 20. Based on this data, the discharge state estimating means 20 estimates and digitizes the discharge state, and the digitized data Si and .delta.Si are fetched as inputs to a fuzzy inference section 12a of a fuzzy control unit 12.

The fuzzy inference section 12a executes a fuzzy inference on the basis of the input data Si and .delta.Si, and sends output data .delta.t and .delta.d to an interpretive section 12b. The symbols .delta.t and .delta.d represent the increments of a jump period and a jump distance, respectively. The interpretive section 12b delivers data .delta.T and .delta.D, which is non-fuzzy data corresponding to the data .delta.t and .delta.d, respectively, to a jump period/distance modifying means 23, and the jump period/distance modifying means 23 modifies the jump period and the jump distance, on the basis of the data .delta.T and .delta.D, and outputs a modified jump period T and a jump distance D to a jump control means 24. The jump control means 24 outputs a control value based on the jump period T and the jump distance D, to the servo control means 16, and thus a jump control of the electric discharge machine is executed. The fuzzy inference in the fuzzy control unit 12 will be described in detail later.

The distance estimating means 21 estimates the distance between the electrode and the workpiece from the output of the average voltage Vm from the waveform feature extracting means 19, and sends it to a computing element 22. The computing element 22 calculates the difference between a target reference value from the machining condition setting means 13 and the estimated distance estimated by the distance estimating means 21, and sends the displacement to the electrode servo feed control means 15. Based on the control parameters from the machining condition setting means 13 and the displacement from the computing element 22, the electrode servo feed control means 15 sends a servo move command to the servo control means 16. The servo control means 16 drives the servomotors 31, 32, and 33 in accordance with move commands output from the electrode swing control means 14, the electrode servo feed control means 15 and the jump control means 24, so that the distance between the electrode 34 and the workpiece 35 is adjusted to a target value set in the machining condition setting means 13.

Before explaining the operation of the fuzzy control unit 12, the operation of the distance estimating means 21 will be described. Various waveforms of the electric discharge between the electrode 34 and the workpiece 35 are shown on the right-hand side of FIG. 6, wherein waveform C1 represents a short state; C2, a state of an arc; C3, a state of a normal spark I; C4, a state of a spark II subject to a substantial discharge delay time; and C5, an open state. As seen from these waveforms, the open and short states are established when the average voltage Vm of the discharge is high and low, respectively. Accordingly, the distance estimating means 21 standardizes the open and short states of the electric discharge as 0 and 1 and outputs these values to the computing element 22. The machining condition setting means 13 sends the electric discharge in the spark-I state as the target reference value to the computing element 22, and thus the displacement of the distance is obtained.

The discharge state estimating means 20, like the distance estimating means 21, receives the average voltage Vm and the discharge delay time Tn from the waveform feature extracting means 19. Based on the input average voltage Vm and discharge delay time Tn, the spark-I, open, and short states of the electric discharge are standardized as 0, +1 and -1, respectively. The value Si of each of these states is given as a discharge stability. FIG. 2 shows the time-based change of the discharge stability Si. The rate of this change, that is, the differential .delta.Si of the discharge stability, is given as a discharge stability change .delta.Si. The discharge state estimating means 20 calculates the discharge stability Si and the rate of change .delta.Si thereof, and outputs same to the fuzzy control unit 12. A method also exists by which the discharge stability can be more accurately estimated by using the discharge current In in addition to the average voltage Vm and the discharge delay time Tn.

Based on the discharge stability Si and the rate of change .delta.Si thereof, the fuzzy control unit 12 executes a fuzzy inference. FIGS. 3 and 4 show the rule of this fuzzy inference; wherein FIG. 3 illustrates an example of the fuzzy rule for inferring the increment .delta.d of the jump distance, and FIG. 4 illustrates an example of the fuzzy rule for inferring the increment .delta.t of the jump period. Each rule consists of four rules.

First, the rule for inferring the increment .delta.d of the jump distance shown in FIG. 3 will be described. In FIG. 3, A11, A12, A13, and A14 represent membership functions related to the discharge stability Si; A21, A22, A23 and A24 represent membership functions related to the rate of change .delta.Si of the discharge stability; and B1, B2, B3, and B4 represent membership functions related to the increment .delta.d of the jump distance. The axis of ordinate of these membership functions represents the adaptation of each inference.

The first rule R1 implies that, if Si and .delta.Si are Big and Positive Big, respectively, then .delta.d is Negative Big.

The second rule R2 implies that, if Si and .delta.Si are Medium and Positive Medium, respectively, then .delta.d is Negative Small.

The third rule R3 implies that, if Si and .delta.Si are Small and Negative Medium, respectively, then .delta.d is Positive Small.

The fourth rule R4 implies that, if Si and .delta.Si are Small and Negative Big, respectively, then .delta.d is Positive Big.

The following is a description of the rule for inferring the .delta.t of the jump period shown in FIG. 4. In FIG. 4, A11, A12, A13, and A14 represent membership functions related to the discharge stability Si; A21, A22, A23 and A24 represent membership functions related to the rate of change .delta.Si of the discharge stability; and B1, B2, B3, and B4 represent membership functions related to the increment .delta.t of the jump distance. The axis of ordinate of these membership functions represents the adaptation of each inference.

The fifth rule R5 implies that, if Si and .delta.Si are Big and Positive, respectively, then .delta.t is Positive Big.

The sixth rule R6 implies that, if Si and .delta.Si are Big and Zero, respectively, then .delta.d is Positive Small.

The seventh rule R7 implies that, if Si and .delta.Si are Medium and Negative Small, respectively, then .delta.d is Negative Small.

The eighth rule R8 implies that, if Si and .delta.Si are Small and Negative Big, respectively, then .delta.d is Negative Big.

The fuzzy inference section 12a executes inference on the basis of these eight rules, and delivers the values of the respective increments .delta.d and .delta.t of the jump distance and the jump period to the interpretive section 12b. The values of the respective increments .delta.d and .delta.t of the jump distance and the jump period are obtained by the gravity center method or the like, in accordance with the aforementioned fuzzy rules. The obtained increments .delta.d and .delta.t of the jump distance and the jump period are digitized in the interpretive section 12b, and are sent to the jump period/distance modifying means 23, whereupon a jump control is effected by means of the jump control means 24.

These fuzzy rules and membership functions are stored as a data base, and the fuzzy inference also may be made by using this data base.

Alternatively, a plurality of data bases may be provided, depending on the material of the workpiece and the machining method, etc., and the fuzzy inference is effected by selecting the required data bases.

FIG. 5 is a diagram showing another embodiment of the present invention. Since like reference numerals are used to designate the same components as shown in FIG. 1, a description of those components is omitted. In the present embodiment, a discharge state fuzzy estimating means outputs the discharge stability Si and the rate of change .delta.Si thereof by fuzzy inference. More specifically, in the present embodiment, the discharge state is estimated by fuzzy inference in accordance with the maximum value Vn of the discharge voltage and discharge delay time Tn output from the waveform feature extracting means 41.

Accordingly, the present embodiment differs from that shown in FIG. 1 in that the waveform feature extracting means 41 delivers the maximum value Vn of the discharge voltage and the discharge delay time Tn to the discharge state fuzzy estimating means 40, and the discharge state fuzzy estimating means 40 executes a fuzzy inference in accordance with these outputs.

A fuzzy control unit in the discharge state fuzzy estimating means 40 executes a fuzzy inference in accordance with the maximum value Vn of the discharge voltage and the discharge delay time Tn, and FIG. 6 shows the rule of this fuzzy inference. Namely, FIG. 6 illustrates an example of the fuzzy rule for inferring the discharge stability Si, which fuzzy rule consists of five rules.

In FIG. 6, A11, A12, A13, and A14 represent membership functions related to the maximum value Vn of the discharge voltage; A21, A22, A23, A24, and A25 represent membership functions related to the discharge delay time Tn; and B1, B2, B3, B4, and B5 represent membership functions related to the discharge stability Si. The axis of ordinate of these membership functions represents the adaptation of each inference. A discharge state waveform corresponding to each rule is indicated on the right of the rule.

The first rule R1 implies that, if Vn and Tn are Small and Zero, respectively, then Si is Negative Big (SHORT STATE).

The second rule R2 implies that, if Vn and Tn are Medium and Small, respectively, then Si is Negative Small (ARC STATE).

The third rule R3 implies that, if Vn and Tn are Big and Medium, respectively, then Si is Zero (SPARK I STATE).

The fourth rule R4 implies that, if Vn and Tn are both Big, then Si is Positive Small (SPARK II STATE).

The fifth rule R5 implies that, if Tn is Very Big, then Si is Positive Big (OPEN STATE).

The fuzzy control unit of the discharge state fuzzy estimating means 40 executes the inference on the basis of these five rules, digitizes the discharge stability Si and the discharge stability rate of change .delta.Si obtained on the basis of the discharge stability Si at the interpretive section thereof, and outputs these values to a fuzzy control unit 12. The fuzzy control unit 12 executes a fuzzy inference in the same manner as described above, to thereby effect a jump control.

According to the present embodiment, as described above, the fuzzy inference is executed with the discharge state as an input, the optimum jump distance and jump period are obtained, and the jump control is executed on the basis of these factors, and accordingly, the optimum jump control can be effected corresponding to the momentarily changing discharge state. Therefore, since wasteful jump controls can be reduced, in contrast to the conventional case, the machining speed is increased and the discharge state stabilized, to thereby provide an outstanding effect such as an improvement in the accuracy of machined surfaces. Moreover, since the jump distance and the jump period can be automatically determined and controlled, data on the jump need not be set as a machining condition.

According to the present invention, as described above, an optimum and high-level control can be effected depending on the discharge state, thus ensuring a considerable improvement in the machining accuracy and machining speed, in contrast with the conventional jump control in which the jump conditions are determined on the basis of experience and experiments.