JP2946333B2 - Adaptive control device for machining machine - Google Patents

Abstract

PURPOSE:To easily realize complex adaptive control which facilitates the addition and change of techniques and considers various factors in an adaptive control device for use on an electrical discharge machine by describing techniques such as know-how of machining and the like in a memory unit independently of an inference unit. CONSTITUTION:The control device is provided with a knowledge memory unit 42 in which plural techniques for changing machining state are described, a state memory unit 41 storing machining states and machining conditions, and an inference unit 43. When a voltage is applied to between a tool electrode 1 and a work 2 by a machining power supply 22, the work 2 is machined. An electrode-position control unit 21 controls the electrode 1 in response to the voltage-difference between an average voltage between electrodes obtained by a state recognition unit 23 and a reference voltage in order to keep a correct gap distance between the electrode 1 and the work 2. Further, the inference unit 43 synthesizes the amount of operations obtained by plural techniques such as techniques ambiguously expressed by skilled operators, and realizes an adaptive machining based on complex know-how of machining. The change of technique is simplified by describing the technique in the memory unit 42 independently of the inference unit 43.

Description

Description: TECHNICAL FIELD The present invention relates to an adaptive control device for various types of processing machines, and in particular, independently describes, as a knowledge base, a plurality of methods for realizing a desired processing state, such as processing know-how. Further, the present invention relates to a processing machine adaptive control device capable of easily maintaining a desired processing state at all times by controlling based on this.

Also, the present invention relates to a processing machine adaptive control device capable of easily adding, changing, and sharing methods.

Further, the present invention relates to a machining machine adaptive control device which always maintains an electric discharge machining state in electric discharge machining optimally and realizes a jump operation of a machining electrode so as to maximize electric discharge machining efficiency.

The present invention also relates to a processing machine adaptive control device that collects a processing method of an operator during processing and performs automatic processing based on the processing method.

[Related Art] Hereinafter, an electric discharge machine will be described as an example of various kinds of machining machines.

FIG. 20 is a block diagram of a conventional adaptive control device of an electric discharge machine disclosed in Japanese Patent Publication No. 62-10769, in which (1) is a machining electrode, (2) is a workpiece, (3) ) Is a processing tank, (4) is a processing liquid, (5) is a spindle, (6) is a drive motor, (7) is a speed or position detector, (20) is a processing machine, and (21) is an electrode position control unit. , (22) is the machining power supply, (2)
3) is a state recognition unit, and (31) is an adaptive control unit. In addition,
In this specification, (1), (2), (3), (4),
The part including (5), (6), (7), (21), and (22) will be referred to as a processing machine.

Next, the operation will be described. First, the operator selects the initial processing conditions in consideration of the material and size of the workpiece, the processing amount, the required finishing accuracy, and the like for the processing to be performed. Set this to 20). For example, the setting of the peak value of the current pulse, the pulse width and the pause time, the electrode pulling cycle, the electrode pulling amount, the electrode servo parameters, and the like are performed.

After the completion of the initial setting, machining is started, and a pulse voltage is applied between the machining electrode (1) and the workpiece (2) by the machining power supply (22). 2) is processed along with the movement of the processing electrode (1). The electrode position control unit (21) compares the average electrode voltage obtained from the state recognition unit (23) with a reference voltage in order to keep the machining electrode (1) and the workpiece (2) at an appropriate distance for discharge. The position or speed of the processing electrode (1) is controlled according to the difference voltage.

In electrical discharge machining, machining electrode (1) and workpiece (2)
Is generally as small as about ten microns to several tens of microns, and particularly when the processing area is large, processing waste generated by the processing is difficult to be discharged through the clearance. For this reason, abnormal discharge is likely to occur, for example, processing dust stays in the processing gap and discharge concentrates on that portion. This situation occurs because the amount of processing debris exceeds the discharge capacity.To prevent this, detect or predict abnormal conditions to reduce the amount of processing debris, Actions such as enhancing the ability may be taken.

FIG. 21 shows the movement of the position of the machining electrode (1). FIG. 21 (a) shows the case where normal machining is performed.
(B) is a case where an abnormality occurs in the machining gap. The machining electrode (1) vibrates about 10 to 100 microns during machining, but when normal machining is performed, the extreme lower point that moves from when the machining electrode (1) drops to when it rises (Ten
1) is gradually descending as processing progresses. However, when an abnormality occurs in the machining gap, the extreme lower point (101) tends to rise as seen in FIG. 21 (b). Therefore, the pulse width of the current pulse supplied from the machining power supply (22) is narrowed in order to detect the rise of the extreme lower point (101) and reduce the amount of machining scrap generated, and to increase the machining scrap discharging ability. It can be seen that measures such as increasing the regular lifting amount of the processing electrode (1) should be taken.

In FIG. 20, the state recognition unit (23) is a position detector (7).
The lower point (101) is detected from the resulting movement of the position of the processing electrode (1), and the adaptive controller (31) is notified of the rise or fall of the lower point (101). The adaptive control unit (31) determines that an abnormality has occurred in the machining gap when this rise exceeds a certain threshold value, and for example, reduces the pulse width of the current pulse to suppress the amount of machining waste generated, or A command, such as increasing the amount of lifting of the electrode, in order to increase the processing chip discharge capacity, is sent to the electrode position control unit (21) and the processing power supply (22).

FIG. 22 is a circuit diagram of the adaptive control unit.
When the rise value (110) of the extreme lower point obtained in 3) exceeds a certain threshold value, a command (111) for increasing the electrode pull-up amount is output to the electrode position control unit (21).

Next, another example of the conventional processing machine adaptive control device will be described.

FIG. 23 is a block diagram showing a conventional electric discharge machine adaptive control device disclosed in, for example, JP-A-61-297017.
In the figure, the electric discharge machining process including the electric discharge phenomenon to (51),
(52) is a state quantity of the electric discharge machining process (51), (53) is an electrode control system, (54) is a machining gap distance between a machining electrode and a workpiece adjusted by the electrode control system (53), (55) ) Is a state detector that detects the state quantity (52), (56) is the detected value, (5)
7) is a command value setting device for setting the state of the electric discharge machining process (51), (58) is its command value, (59) is the deviation obtained from the command value (58) and the detection value (56), 60) is a jump controller for controlling a jump operation, (61) is a jump operation amount, and (62) is whether to perform machining gap control according to the deviation (59) or to perform a jump operation according to the jump operation amount (61). A switch for switching, (63) is an operation amount to the electrode control system which is an output of the switch (62), (64)
Is a position signal of the machining electrode, (65) is a jump setting device for setting a jump amount and a jump cycle based on a machining depth in advance to perform an appropriate jump operation, (66)
Is a jump command value given by the jump setting device (65) to the jump controller in accordance with the position signal (64) of the machining electrode.

Fig. 20 shows the mechanical parts of the processing machine as mechanical elements (1).
In contrast to the expressions (7) to (7), in FIG. 23, the distance between the poles is represented as an input, and the state of machining is represented as a control object as an output.

The operation in the case of performing the processing gap control will be described with reference to FIG. In the figure, a command value (58) of a command value setting device (57) for setting a desirable state of the electric discharge machining process and a detection value (5) of a state detector (55) for detecting the state of the electric discharge machining process are shown.
The deviation (59) is obtained from 6) and the deviation (5
9) is given as the manipulated variable (63) of the electrode control system (53), and the electrode control system (53) adjusts the machining gap distance (54) so that the deviation (59) becomes zero, and It works to always maintain.

However, as the machining progresses, machining waste generated by electric discharge machining becomes stationary between the machining electrode and the workpiece, and short circuits frequently occur in the machining gap. The machining state cannot be maintained.

Therefore, in general, processing waste that is standing between the processing electrode and the workpiece is eliminated by using a pumping action by a jump operation of the processing electrode.

In addition, the jump operation is to ignore the command value (58) and the detection value (56) during the machining gap control, forcibly raise the machining electrode from the machining position by the set amount, and return to the original machining position. The return operation is performed periodically.

As shown in the figure, the jump operation of the machining electrode sets the jump conditions such as the jump amount, which is the electrode pulling amount, and the jump period, which is the electrode pulling period, according to the processing depth in the jump setting unit (65) in advance. Aside
The machining depth is obtained from the position signal (64) of the machining electrode during machining, and the jump command value (66) is sent to the jump controller (60) by referring to the jump conditions set in the jump setting device (65). The feed / jump controller (60) is controlled by giving a jump operation amount (61) as an operation amount (63) to an electrode control system (53) via a switch (62).

As is clear from the above description, the jump operation of the machining electrode is indispensable to always maintain a stable electric discharge machining state, but from the viewpoint of machining efficiency, it wastes time that does not contribute to machining, and is optimal. It is important to perform a machining electrode jump operation in order to improve machining efficiency.

In order to realize this optimal machining electrode jump operation,
The jump conditions such as the jump amount and the jump cycle set in the jump setting unit (65) can be adjusted not only for the machining depth but also for the combination of machining power pulse conditions, machining electrode shape, material of machining electrode and workpiece, etc. It is necessary to determine the value, and generally, it depends very much on the technique of a skilled worker.
Particularly, a skilled worker monitors the degree of instability of the electric discharge machining state and adaptively changes the jump operation.

[Problem to be Solved by the Invention] Since the conventional processing machine adaptive control device is configured as described above, for example, when changing the electrode lifting amount, the rise of the extreme lower point simply exceeds a certain threshold value. In order to determine the change of the electrode lifting amount only by the result obtained by using the method of increasing the electrode lifting amount at the time of using, a complicated method such as a method expressed vaguely by skilled workers was used There is a first problem that control is difficult to realize.

Further, since the conventional processing machine adaptive control device is configured as described above, for example, when adding or changing another method of determining the increase or decrease amount of the electrode lifting, changing the hardware that realizes the method, Alternatively, if the method is implemented by software, it is necessary to make changes to the entire software for testing the amount of electrode pulling up using the method, and the know-how of the manufacturer or user cannot be easily added or changed, and various know-how must be added. When sharing with other processing machines, it is necessary to share not only the technique but also hardware or software for realizing the technique, and there is a second problem that it takes time and effort.

Further, since the conventional electric discharge machining control device is configured as described above, if an attempt is made to set jump conditions in accordance with a technique of a skilled worker in order to perform an optimum jump operation, the qualitative method included in the technique is not included. It is difficult to properly describe the ambiguous expression as a jump condition, and when the jump operation is automatically changed in accordance with the instability of the electric discharge machining state on behalf of the skilled worker, There is a third problem in that it is difficult to accurately describe a criterion for determining the degree of instability of the metal, and it is difficult to improve the EDM efficiency.

Also, since the conventional processing machine adaptive control device is configured as described above, it is difficult to add or modify the processing method of the worker. There is a fourth problem that the operator has to clarify what is seen and how he is operating in the processing state.

The first invention of the present invention has been made to solve the first problem as described above. Since the operation amount can be determined based on a plurality of methods, and the methods can be easily added or changed, An object of the present invention is to provide a processing machine adaptive control device capable of realizing automatic processing or the like by incorporating complicated processing know-how and the like possessed by an operator.

The second invention of the present invention has been made to solve the second problem as described above, and it is possible to perform automatic processing and the like based on a complicated method such as the processing know-how of an operator, and easily. It is an object of the present invention to obtain a processing machine adaptive control device capable of adding, changing, and sharing a method.

A third invention of the present invention has been made in order to solve the third problem as described above, and includes a jump condition effective for performing an optimum jump operation and a criterion for determining the degree of instability of an electric discharge machining state. It is an object of the present invention to obtain a processing machine adaptive control device that can appropriately and easily describe techniques of a skilled worker, and execute a jump operation optimally and adaptively automatically by these techniques. .

A fourth invention of the present invention has been made to solve the above-mentioned problems, and can perform automatic machining based on a machining technique of an operator, and can easily collect and correct the machining technique. It is an object of the present invention to obtain a processing machine adaptive control device.

[Means for Solving the Problems] A processing machine adaptive control device according to a first invention of the present invention includes:
In a processing machine adaptive control device that performs control by changing processing conditions during processing, a status storage unit that describes a plurality of types of processing states or processing states and processing conditions, and the plurality of types of processing states or processing states and processing conditions A knowledge storage unit that describes a method of determining each processing condition in accordance with each degree, and the plurality of types of processing states or the processing states and the respective degrees of the processing conditions described in the situation storage unit and the knowledge. An inference unit for obtaining a plurality of processing conditions based on each of the methods described in the storage unit, synthesizing the plurality of processing conditions, and obtaining a value of the processing condition to be changed.

A processing machine adaptive control device according to a second invention of the present invention includes:
A time-series data recording unit that records, in a time-series, the processing state obtained by the state recognition unit, the processing condition changed by the operator, and the value of the processing condition to be changed obtained by the inference unit; A knowledge updating unit that updates the method described in the knowledge recording unit when there is a difference between the value of the processing condition changed by the worker and the value of the processing condition to be changed, which is recorded in a recording unit. It is a thing.

A processing machine adaptive control device according to a third aspect of the present invention includes:
The control is performed by changing the processing conditions during the processing.
In a processing machine adaptive control device, a jump controller that controls a jump operation of a processing electrode, a plurality of types of processing states or a state detector that detects processing states and processing conditions, and the plurality of types detected by the state detector. A state storage unit that describes the machining state or the machining state and the machining condition, and a knowledge storage that describes a method of determining each jump operation condition according to each degree of the plurality of machining states or the machining state and the machining condition. And a plurality of jump operation conditions obtained based on the plurality of types of machining states or machining states and machining conditions described in the situation storage unit and the respective methods described in the knowledge storage unit, An inference unit that synthesizes a plurality of processing conditions and outputs a jump operation command value to the jump controller.

A processing machine adaptive control device according to a fourth invention of the present invention is:
The method of determining processing conditions described in the knowledge storage unit is described in a rule format consisting of an antecedent part and a THEN consequent part.

A processing machine adaptive control device according to a fifth invention of the present invention is:
The method of determining processing conditions described in the knowledge storage unit is described in the form of a rule using a fuzzy set consisting of an antecedent part and a THEN consequent part, and synthesis of the processing condition values in the inference unit is performed by fuzzy inference. is there.

[Operation] In the first invention of the present invention, in the first invention of the present invention, the inference unit is based on a method in which a plurality of types of machining states or machining states and machining conditions are stored in the state storage unit. Adaptive machining based on complex machining know-how by combining multiple machining conditions and finding the value of the machining condition to be changed, and describing the method in the knowledge storage unit independently of the inference unit. Add and change easily.

In the second invention of the present invention, the time-series data recording unit records the processing state, the processing condition changed by the operator, and the value of the processing condition to be changed obtained by the inference unit in a time-series manner. After the end, if there is a difference between the value of the processing condition changed by the operator and the value of the processing condition to be changed, which is recorded in the time-series data recording unit, the method described in the knowledge recording unit is updated, and the knowledge storage unit is updated. The worker's machining method stored in the storage device becomes more substantial and performs optimal control.

In the third invention of the present invention, the inference unit calculates a plurality of types of machining states or machining states described in the situation storage unit and a plurality of jump operation conditions obtained based on the machining conditions and the method described in the knowledge storage unit. Then, the plurality of processing conditions are combined, a command value of the jump operation is output to the jump controller, and the optimum jump operation is executed and adaptively changed.

In the fourth aspect of the present invention, the method of determining the processing conditions described in the knowledge storage unit is described in a rule format including the IF antecedent part and the THEN consequent part. Are easily added, changed, and shared, and the inference unit determines the processing conditions based on the complicated processing know-how according to this rule.

According to a fifth aspect of the present invention, a method for determining processing conditions described in a knowledge storage unit is described in a rule format based on a fuzzy set including an antecedent part and a THEN consequent part. The fuzzy reasoning is used to synthesize the values of the above, so that various processing know-how of the operator can be easily added, changed,
In addition, the inference unit determines the processing conditions based on the complicated processing know-how according to this rule.

Hereinafter, a first embodiment of the present invention will be described with reference to the drawings.
In FIG. 1, (1) to (23) are the same as or correspond to those of the conventional device, (31a) is an adaptive control unit, (41) is a situation storage unit, (42) is a knowledge storage unit, and (43) ) Is an inference unit.

Next, the operation will be described. The knowledge storage unit (42) describes a plurality of methods for determining the change in the amount of electrode pull-up as shown in FIG. In the present invention, the presence or absence of an increase is simply determined by a threshold value. In contrast, in the present invention, the knowledge storage unit (42) is described by software independently of the inference unit (43). It is possible to describe more complicated methods. Method 2 shows a method of increasing / decreasing the amount of electrode withdrawal based on the distribution density change rate from the time of application of a pulse voltage applied to the electrode and the workpiece to the start of discharge (hereinafter referred to as no-load time). The description of the method in the knowledge storage unit (42) can be expressed by using hardware such as an op-amp and a switch in addition to the description by software.

FIG. 3 shows a method and a situation storage unit which are described in the knowledge storage unit (42) by the inference unit (43), and determine each processing condition in accordance with a plurality of types of processing state or each processing state and the degree of each processing condition. It is the flowchart which showed the procedure which calculates | requires the increase / decrease amount of an electrode pulling up using several types of processing conditions or processing conditions and processing conditions memorized by (41). First, the inference unit (43) reads the method 1 from the knowledge storage unit (42) and calculates the amount of increase / decrease of the electrode pull-up by the method 1 and Z ₁ according to the degree of the rise of the extreme lower point described in the situation storage unit (41). (Steps S31 to S34). Decrease amount of electrode pull-up by Method 2 by the value of the distribution density changing rate approaches 2 and unloading time are obtained analogously to Z ₂ (step S35, S32 to S34). Next, in this embodiment, since N = 2, step S35
There YES, decrease amount Z _t of the electrode pulled by synthesizing the two results obtained by the two methods is determined,
It is output to the electrode position controller (21) and the machining power supply (22) (steps S36 to S37). In the above synthesis, for example, it is conceivable to take the average of each result as in the following equation (1).

Here, when the method 2 is used, the distribution density change rate of the no-load time is required, and this is measured by the state recognition unit (23) for measuring the no-load time of a certain section, which is shown in FIG. What is necessary is just to calculate the distribution density change rate in accordance with the shown equation (5) and to describe it in the situation storage unit (41). In addition, the value of the rise of the extreme lower point is required in the method 1, which may be obtained by the state recognition unit (23) and described in the situation storage unit (41).

As described above, by combining a plurality of results to obtain a processing condition (the amount of increase / decrease of the electrode lifting), it is possible to realize complicated adaptive control based on a plurality of methods.

In the above embodiment, the knowledge storage unit (42) describes two techniques for determining the increase / decrease amount of the electrode pull-up using the rise of the extreme lower point and the distribution density change rate during the no-load time.
The pulse width of the current pulse, the pause time of the current pulse, the peak value of the current pulse, the sound generated during processing, the vibration state of the processing electrode (1), the appearance of bubbles in the processing liquid (4), etc. Two or more methods for determining the processing conditions such as the cycle of raising the electrode are described in the knowledge storage unit (42), and if the processing conditions are determined by this, it is possible to realize more advanced and complicated adaptive control. is there.

In the above embodiment, the formula (1) is used as a method of synthesizing the result in the inference unit (43).
Of course, it is also possible to combine using various methods such as a maximum value or a minimum value.

Further, in the above embodiment, the method described in the knowledge storage unit (42) is described in a free format as shown in FIG. It can also be described using. For example, when Method 2 is described using this format, it is as shown in FIG. Similarly, method 1
Is also expressed using the rule format of “if, then, please”, the description format is unified, the description in the knowledge storage unit (42) becomes easy, and the processing in the inference unit (43) is performed. This has the effect of being simple.

Further, in the above embodiment, the method described in the knowledge storage unit (42) is, for example, "if 0.8
Although it is shown as a quantitative value such as `` if below, '' the processing know-how of the worker is often expressed in qualitative words such as `` large '' and `` small '', and such a qualitative value It is also possible to describe the method by rules using fuzzy sets that can be expressed. FIG. 5 shows the method shown in FIG. 4 described using this fuzzy set. At this time, a membership function shown in FIG. 6 is used to handle qualitative expressions such as “large” and “small”.

For example, in the method 2A, "if the change rate of the distribution density during the no-load time is small" is expressed, and the fuzzy set corresponding to "small" is represented by "figure 6" in FIG. Is represented by the membership function for "small". For example, if the value of the change rate is 0.7
If so, the value of the membership function is 1, and if the value of the rate of change is 0.9, the value of the membership function is 1/3.
Here, the value 1 of the membership function means that it completely belongs to the set, and the value 0 indicates that it does not belong to the set at all.

When the method is described by a rule using a fuzzy set as shown in FIG. 5, the processing conditions are determined by fuzzy synthesis (also called fuzzy inference).
Various methods have been proposed for fuzzy synthesis. For example, the following methods are available.

 Now, suppose that the rules were expressed as follows.

Here, x _i and y _i are the situation storage unit (41)
U is the electrode position control unit (21) or the processing power source (2).
Processing conditions output to 2), Is a fuzzy set such as "large" or "small", and the subscript i indicates that it corresponds to the i-th rule. further, _Denote by f _Ai , f _Bi , f _{Ci the} membership function for
Processing conditions u _t is determined is obtained by (2) to (4) below. _Ci (u) = f _Ci (u) * (f _Ai (x _i ) ∧f _Bi (y _i )) (2) Here, ∧ and ∨ are operators that take minimum and maximum values, respectively, and * is a product or α-cut operator. Thus the fuzzy synthesis, (2) determine the result of each rule in the formula, the plurality of result (3), by determining the processing condition u _t synthesized by (4), based on a plurality of methods Complicated adaptive control can be easily realized.

Furthermore, in the above embodiment, the machining machine is an electric discharge machine, but a laser machining machine, a beam machining machine, an electrolytic machining machine, which can realize adaptive control by adjusting machining conditions according to a machining state during machining. It is clear that complex adaptive control can be easily realized in NC lathes, NC grinders, etc. by combining the results obtained by multiple methods and determining the processing conditions. For example, in a laser processing machine, a method of adjusting the output of the light source to prevent sagging at the corner portion, and a means for adjusting the output of the light source with respect to a change in the thickness of the plate, combine the results obtained to combine the output of the light source with the output of the light source. An adaptive control device to be determined can be realized.

As described in the above embodiment, in the present invention, the inference unit (4
Since the method is described “independently” of 3) in the knowledge storage unit (42), when the method is added or changed, the data of the method stored in the knowledge storage unit (42) is added.・ It only needs to be changed. The procedure for actually executing a certain technique or the procedure for combining a plurality of techniques is shown in FIG. 2 or (2).
Fuzzy inference described in Equations (4) (in the inference unit (43)), and need not be changed.

For example, considering a case in which a method of determining the electrode pull-up amount based on the vibration of the electrode is added as Method 3, when "independently described", this method is simply referred to as the knowledge storage unit (4
You only need to add the description in 2) (database change only), but if you do not write the information independently, how to execute this method, synthesize it with other methods, and finally Think about how to decide the amount to raise,
Changes need to be made, including the program of the adaptive control unit (31).

Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. In the first embodiment, among the first and second methods shown in FIG. 2, an example using a method in the form of a rule “if, if,” is used for the second method. For both methods 1 and 2
This is the format of the rule.

In the second embodiment, a method of determining a change in the amount of electrode pull-up as shown in FIG. 7 is described in the knowledge storage unit (42), and the rules 1a to 1c are the same as those in the related art. Is used to determine the increase in the amount of lifting of the electrode. Conventionally, the presence or absence of the increase is simply determined by a threshold value. In the present invention, however, the knowledge storage unit (42) is independent of the inference unit (43). Therefore, more complicated methods can be described. Rules 2a to 2c show a method of increasing or decreasing the amount of electrode pulling up according to the distribution density change rate of the time from the application of a pulse voltage applied to the electrode and the workpiece to the start of discharge (hereinafter referred to as no-load time). I have.

Next, the operation will be described with reference to the flowchart shown in FIG. In FIG. 3, the method i of steps S32, 33, and 34 is replaced with the rules ia, ib, and ic. First, the inference unit (4
3) to obtain the electrode raising decrease amount due to rule 1 a to 1 c, the Z ₁ in accordance with the degree of increase of the electrode under point described in status storage unit reads the rule 1 a to 1 c from the knowledge storage section (42) (41) (Steps S31 to S34). Decrease amount of electrode pull-up by rule 2a~2c the value of the distribution density changing rate rules 2a~2c and unloading time are obtained analogously to Z ₂ (step S35, S
32 to S34). Then electrode pull-up decrease amount Z _t is determined by combining the results obtained by this method, is the output control unit in the electrode position (21) and processing power (22) (step S36～S37). The above synthesis is performed, for example, in the aforementioned (1)
It is conceivable to take the average of each result as in the formula.

Here, when using the rules 2a to 2c, the distribution density change rate of the no-load time is required. This is because the no-load time of a certain section is measured by the detection value processing unit (23), What is necessary is just to calculate the distribution density change rate according to the equation (5) shown in the figure and to describe it in the situation storage unit (41). Also, rule 1a
In the case of ~ 1c, the value of the rise of the extreme lower point is required, but this value may also be obtained in the detection value processing unit (23) and described in the situation storage unit (41).

As described above, by combining a plurality of results to obtain a processing condition (the amount of increase / decrease of the electrode lifting), it is possible to realize complicated adaptive control based on a plurality of methods.

In the second embodiment described above, the knowledge storage unit (42) describes a method of determining the increase / decrease amount of the electrode pull-up using the rise of the extreme lower point and the change rate of the distribution density during the no-load time.
The pulse width of the current pulse, the pause time of the current pulse, the peak value of the current pulse, the electrode, using the sound generated during processing, the vibration state of the processing electrode (1), the appearance of bubbles in the processing liquid (4), and the like. If a technique for determining the processing conditions such as the pulling cycle is added to the knowledge storage unit (42) and the number of processings is determined by this, it is a matter of course that more advanced and complicated adaptive control can be realized.

In the second embodiment, the formula (1) is used as a method of synthesizing the result in the inference unit (43).
Of course, it is also possible to combine using various methods such as sum, maximum value or minimum value.

Furthermore, some or all of the data of the rules described in the knowledge storage unit (42) can be shared with other processing machine adaptive control devices using communication means or the like. As a result, in a factory or the like having a plurality of similar processing machines, common processing know-how can be centrally managed, and only processing know-how unique to each processing machine can be managed by each processing machine control device. Become.

Furthermore, if some or all of the rules described in the knowledge storage unit (42) are stored on a magnetic disk and exchanged as necessary, the same processing machine can be used based on different processing know-how. Processing can be easily realized. Here, instead of magnetic disk, optical disk, IC
Similar effects can be obtained by using other storage media such as a card, an IC cartridge, a magnetic bubble memory, and a magnetic tape.

Furthermore, in the above-described second embodiment, the machining machine is an electric discharge machine, but a laser machining machine, a boom machining machine, and an electrolysis machine capable of realizing adaptive control by adjusting the number of machining according to the machining state during machining. Even for processing machines, NC lathes, NC grinding machines, etc., complex adaptive control can be easily performed by describing the processing know-how possessed by the operator using rules and combining the results obtained to determine the processing conditions. It is clear that it will be possible. For example, in a laser beam machine, the output of the light source is determined by synthesizing the results obtained from the method of adjusting the output of the light source to prevent sagging at a corner portion and the method of adjusting the output of the light source with respect to a change in plate thickness. An adaptive control device can be realized.

The rule shown in FIG. 7 is composed of rules 1a to 1c and rules 2a to 2c.
It is also possible to describe one rule that summarizes 1c and use it to implement it.

In the second embodiment, since the knowledge storage unit (42) is described in the form of rules, the description format is unified,
Only the main part can be rewritten, such as only the antecedent part and only the consequent part. Therefore, it is easier to add or change the method than that shown in FIG. In the case of FIG. 2, it is time-consuming to change it because it is necessary to understand the text of the entire method.

Hereinafter, a third embodiment of the present invention will be described. In FIG. 8, (51) to (63) are the same as the conventional one, but (42)
a) is a knowledge storage unit, (68) is a method taken out of the knowledge storage unit (42a), and (69) is a plurality of types of processing conditions required at least for the method stored in the knowledge storage unit (42a). A quantity, (70) is a situation detector for detecting the machining situation quantity (69), (71) is a detection value thereof, and (41a) is a value for storing at least one of a current value and a past value of the detection value (71). The situation storage unit, (73) is a situation amount necessary for the above method (68) extracted from the situation storage unit (41a), (43a) is stored in the knowledge storage unit (42a), and a plurality of types of machining states are stored. Method of determining each jump motion according to each degree (68)
And an inference unit that comprehensively determines an optimal jump operation and an adaptive change from the situation amount (73) stored in the situation storage unit (41a).
0).

FIG. 1 shows mechanical elements of the processing machine by (1) to (7), as in FIGS. 20 and 23, whereas FIG. As a control target, which is a distance between poles and a processing state quantity as an output. Therefore, it is difficult to clearly indicate the direct correspondence between FIG. 1 and FIG. 8, but the corresponding parts are as follows.

The electrode position control unit (21) corresponds to one including an electrode control system (53), a jump controller (60), and a switch (62).

The state recognition unit (23) corresponds to one including the state detector (55) and the situation detector (70).

The situation storage section (41) corresponds to the one including the situation storage section (41a) and the command value setting device (57).

 The knowledge storage unit (42) corresponds to the knowledge storage unit (42a).

 The inference unit (43) corresponds to the inference unit (43a).

Next, how to generate the optimal command value (75) for the jumping operation of the machining electrode in the embodiment of FIG. 8 will be described. FIG. 9 is an explanatory diagram showing an example of an effective method for causing the machining electrode to perform a jump operation. In the conventional electric discharge machining control device, such a method could not be described appropriately and easily. However, here, for example, a method shown in FIG. It can be described in the above-mentioned knowledge storage unit (42a) in a rule format including the antecedent part THEN consequent part. That is, “the machining current is large” as included in the method of FIG. 9,
A qualitative and ambiguous expression such as “small clearance”, “deep machining depth”, and “small amount of jump operation” is expressed in the membership function. For example, the feature of “Machining current is large” in Method 1 is that the machining current is large.
The value of the membership function is set to “0” in the sense that it is not satisfied at all if it is less than 15 amps, and the value of the membership function is set to “1” in the sense that it is completely satisfied if the machining current is 35 amps or more. If the value is 15 to 35 amps, the value of the membership function is set to “0 to 1” in the sense that the value is satisfied between 0 and 1. Similarly, other qualitative and ambiguous expressions can be appropriately and easily described by membership functions.

On the other hand, the situation storage unit (41a) detects, by the situation detector (70), a plurality of types of machining states required for the technique stored in the knowledge storage unit (42a) or machining states that are machining conditions. Stored. Known data such as machining current is input and stored as known data.

In the case of FIG. 9, the necessary machining conditions are machining current, clearance, and machining depth, and the machining current is known because it is a parameter that can be set as one of machining conditions by an operator.

The clearance is mainly determined by the gap servo voltage (command value of the machining gap control system), the gap applied voltage value, and the set value of the pressurizing current, and is known as machining condition data.

The machining depth is obtained from the electrode position detector based on the difference between the machining start position and the current position.

The inference unit (43a) performs fuzzy inference from the method stored in the knowledge storage unit (42a) and the situation stored in the situation storage unit (41a) according to the procedure shown in FIG. Determine the operation command value (75). In the figure, (73a), (73b), and (73c) are known data and detected values of the machining current, the clearance, and the machining depth stored in the situation storage unit (41a), respectively. In fuzzy inference, we investigate how each of these methods (73) satisfies the qualitative and ambiguous expression of the antecedent part described by the membership function in each method, and finds the most satisfactory in the antecedent part. Cut the upper limit of the membership function in the consequent part with the value of the membership function with a small degree, and combine the membership functions so that they always have the largest function value among the respective membership functions. The area centroid position CG of the obtained membership function is obtained. This value is the optimal command value (75) for the jump operation of the machining electrode.

In FIG. 11, the antecedent part of the method describes three machining situations, and the consequent part describes one jump operation amount, but there is nothing to limit these. In addition, it goes without saying that an optimum command value of the jumping operation of the machining electrode can be similarly obtained even when the number of methods is increased. further,
In the third embodiment, the jump operation is not adaptively changed according to the degree of instability of the electric discharge machining state. However, the jump operation can be realized in the same manner as described above.

In the third embodiment, the fuzzy set is used in the knowledge storage unit and the fuzzy inference is performed in the inference unit. However, it is naturally possible to use the knowledge expression and the inference method used in other general expert systems. And the third
An effect similar to that of the embodiment is obtained.

Hereinafter, a fourth embodiment of the present invention will be described with reference to the drawings. No.
FIG. 12 is a block diagram showing a configuration of a processing machine adaptive control device according to one embodiment of the present invention. In the figure, (10) is an operator and (20) is a processing machine. (42) is the worker (10)
(23) is a state recognition unit that recognizes a plurality of types of processing states, (41) is a processing machine (20)
A state storage unit that stores the processing conditions set in the processing conditions and the processing state recognized by the state recognition unit (23) as the processing state;
3) uses a processing method that determines each processing condition according to the degree of each of the plurality of types of processing states stored in the knowledge storage unit (42), and uses the processing state stored in the state storage unit (41). An inference unit for inferring the operation of the processing machine (20). (87) is a first inference unit composed of the above (23), (41), (42) and (43).
It is a configuration group.

(88) is a time-series data recording unit that records the processing state recognized by the state recognition unit (23), the processing conditions set in the processing machine (20), and the time-series data of the operation performed by the worker (10). , (89) extract the processing method from the contents recorded in the time-series data recording unit (88), generate and modify the processing method together with the output of the inference unit (43), and store the contents in the knowledge storage unit (42). (90) is a second configuration group including a time-series data storage unit (88) and a knowledge updating unit (89).

Next, the operation will be described. When processing the A1 material in a certain processing machine (20), a processing method for adjusting the processing condition A as shown in FIG. 42) is stored in a format as shown in FIG. 13 (b).

When the processing is started, as shown in the flowchart of FIG. 15, after the presence or absence of the end is confirmed (step S41),
The state recognition unit (23) processes the signal from the sensor attached to the processing machine (20), obtains the processing speed and the loudness of the processing sound, and recognizes the processing state (step S42). The processing speed and the loudness of the processing sound are stored in the status storage unit (41) as the processing status together with the currently set processing conditions (step S43). The inference unit (43) infers an operation from the processing state by using the method of the knowledge storage unit (42) (step S44). For example, when the processing sound 3/2, the processing speed 1, and the condition B4 are obtained in relation to a predetermined reference value, K ₁ = −2 / 3 and K ₂ = −1 from the table of FIG. , K ₃ =
Since 4, (Operation) = K ₁ · (processing speed) + K ₂ · (processed _{sound) + K 3 = (- 2/3} · 1) - (1 · 3/2) +4 = 11/6 , and the processing Condition A is set to “11/6”. This state is shown in FIG. 13 (c), and is represented by a three-dimensional function including the processing speed, the processing sound, and the operation amount.

As described above, the processing machine adaptive control device controls the processing machine (20) by the processing method of the worker using the first configuration group, and performs the above steps (S41) to (S44) until the processing is completed. repeat.

Next, an operation in the case of modifying the processing method of the worker stored in the knowledge storage unit (42) will be described.

As shown in the flowchart of FIG. 16, the operator changes some or all of the machining conditions during machining (step
S51), set processing conditions in the time-series data recording unit (88),
The processing state and time series data of the processing condition changing operation performed by the operator are recorded (step S52). When the processing conditions are changed by the operator, the output of the inference unit (43) is also recorded (step S52). Then, after the processing, the knowledge updating unit (89) sets the processing conditions selected by the operator and the inference unit (43).
If there is a difference from the processing conditions inferred by
The processing method stored in 2) is modified (step S53).

This specific example will be described with reference to FIG. It is assumed that as a result of processing the A1 material whose condition B is “3”, the contents of the time-series data recording unit (89) are as shown in FIG. 14 (a). At this time, the output of the inference unit (43) (inference value)
Since there is a difference between the operation and the operator's operation, the processing method is modified. This correction is performed by the least squares method. When the operation is y, the processing speed is u ₁ , and the processing sound volume is u ₂ , the processing method is represented by y = K ₁ u ₁ + K ₂ u ₂ + K ₃ (1). Now, by recording the operation of the worker, (y _i , u _1i , u _2i ) (i = 1, 2,..., 6) (2) is obtained. At this time, e _i = K ₁ · u _1i + K ₂ · u _2i + K ₃ −y _i (3) K ₁ , K ₂ , and K ₃ that minimize
Store in 2). Equation (4) is partially differentiated with respect to K ₁ , K ₂ , and K ₃ , and By simultaneous equations (5), (6) and (7), K ₁ = −1, K ₂ = −3 / 2, K ₃ = 7 (8) is obtained. Then, the knowledge storage unit (42) is obtained by the equation (8).
To update. As a result, the content of the knowledge storage unit (42) is
The content shown in FIG. 14B is replaced with the content shown in FIG. 14B, and the modification of the processing method of the worker is completed.

Next, the processing methods of the workers are collected, and the knowledge storage unit (42)
Will be described. Here, an example in which processing methods related to Cu material processing are collected will be described. As in the case of the modification of the processing method for the A1 material, the set processing conditions, the processing state, the operation to be performed by the operator, and the processing inferred by the inference unit (43), as in the case where the processing method of the A1 material was corrected earlier. The time series data of the condition is recorded, and the knowledge updating unit (89) obtains K ₁ and K ₂ K ₃ of the equation (1) by the least square method.

The processing methods collected in this way are stored in the knowledge storage unit (4
This is stored in 2). At this time, a big difference from the case of the previous correction is that while the value already stored in the case of the correction is updated, a new storage location is secured in this case. The point is to store the processing method. The processing method of the worker is collected as described above, and the addition to the knowledge storage unit (42) is completed.

In the above embodiment, the case where there is one worker is shown. However, when there are two or more workers, it may be provided in two or more knowledge storage units (42) and assigned to each worker. The knowledge storage unit (42) may be divided into two or more areas and assigned to each worker. In such a case, the worker sets his / her own processing method in the knowledge storage unit (42).
Can be stored.

To obtain data for each worker as described above,
For example, the name of the worker or personal information equivalent thereto is input to the knowledge storage unit (42) using a keyboard or an IC card, a magnetic card, a magnetic disk, or an optical disk.

Further, the processing machine in the present invention may be any of an electric discharge machine, a laser processing machine, a beam processing machine, an electrolytic processing machine, an NC lathe or an NC machine tool, and each processing machine corresponds to a processing state. The changing signal is sent to the state recognition unit (2
By inputting in 3), the same effect as in the fourth embodiment can be obtained.

17 to 19 are block diagrams of a machining machine adaptive control device according to fifth to seventh embodiments of the present invention, respectively.
The fifth embodiment shown in the figure is an example in which the second configuration group is provided with an example teaching unit. In this embodiment, the example teaching unit (91) generates a processing example, and the time series data recording unit (88) can record how the operator responds to the processing example. That is, when the operator collects or corrects the processing method, the example processing is performed using the first configuration group (87) and the second configuration group (90), and part or all of the processing conditions are inferred by the inference unit (43). The processing method stored in the knowledge storage unit (42) is updated by the knowledge updating unit (89) by the operator without using the output of (2). Since the processing method is updated in this manner, the processing method of the worker can be reliably collected.

In addition, a part or all of the processing methods stored in the knowledge storage unit (42) include an antecedent part describing conditions to be determined and contents to be executed when the conditions are satisfied or not satisfied. Rules consisting of the consequent part (IF-THEN
~).

In addition, some or all of the rules are fuzzy rules that express the antecedent and consequent parts using fuzzy sets that deal with human ambiguity such as "large amplitude" and "slightly lower." There may be. In this case, the membership function and the relationship between the membership functions for the fuzzy set used in expressing the antecedent part and the consequent part are stored in the knowledge storage part (42) and stored in the situation storage part (41). It is desirable that the inference unit (43) performs a fuzzy synthesis of synthesizing a plurality of results obtained by the obtained processing state and processing conditions and the fuzzy rules relating to those situations stored in the knowledge storage unit (42). Find the processing conditions to achieve the processing state.

In this case, the processing method of the knowledge storage unit (42) may be modified by changing the membership function.

In the sixth embodiment shown in FIG. 18, a rotary knob (92) that can be freely moved by an operator is provided in the second group of components so that analog information can be input to the time-series data recording unit (88). This is an example. In this embodiment, since the amount that the operator feels can be input by the rotary knob (92), for example, when the operator uses a method of “changing the processing condition when the processing sound is loud”, “large” is used. Can be recorded simultaneously, and more reliable processing methods can be collected.

Although an example of the rotary knob (92) is shown as an input device for inputting analog information, a slide knob or a joystick may be used, and these positions are measured by, for example, a potentiometer or a rotary encoder (not shown). The time-series data storage section (88) takes in the data via an A / D converter or an up / down counter (not shown).

In connection with the input device, a plurality of push buttons are provided on the input device, and by pressing one or more push buttons, at least one of the following items (a) to (e) is input. I do.

(A) Immediately before the operator changes the processing conditions, a change operation is to be performed.

(B) The change operation has been completed after the operator has changed the processing conditions.

(C) Self-scoring of the change operation after the operator changes the processing conditions.

(D) Canceling the change operation after the operator changes the processing conditions.

(E) The reason why the operator performed the change operation after changing the processing conditions.

Further, the seventh embodiment of FIG. 19 is an example in which a CRT display device (93) is provided. In this embodiment, the knowledge storage unit (4
Since the processing method stored in 2) can be displayed, it is possible to confirm whether collection or correction has been performed reliably. A liquid crystal panel, a plasma display, or the like may be used instead of the CRT display device (93), and in each case, at least one of the following (a) to (g) is displayed.

(A) Processing conditions.

(B) Output of the knowledge update unit (89).

(C) The method stored in the knowledge storage unit (42) has been updated.

(D) Inference situation in the inference unit (43).

(E) All or arbitrary methods stored in the knowledge storage unit (42).

(F) Contents stored in the time-series data storage unit (88).

(G) The name of the worker or personal information equivalent thereto.

Further, when the example teaching section (91) is generating an example, the fact is displayed on the above-mentioned CRT display device or the like so as to inform an operator or the like of the operation state of the processing machine.

Further, the knowledge storage unit (42) may be configured to be detachable as in the second embodiment. In this case, the knowledge storage unit (42) can be used between processing machines of the same model installed at different locations. ) Can be used in common, so that each processing machine can perform processing based on the same processing method. As the storage medium in that case, for example, a magnetic disk, an optical disk, an IC card, an IC cartridge, a magnetic bubble memory,
There are magnetic tapes and the like.

In the above embodiment, regarding the processing condition and the processing state used when the inference unit performs the inference, when the control is simple, the processing state such as the processing speed and the processing sound can be obtained even if the processing condition is not clearly understood. Alone, adaptive control is possible.

In this specification, the “processing state” is an amount representing the processing state obtained by the detector.

The “processing condition” is an amount that affects processing such as a set value and a target value.

“Processing status” or simply “status” indicates both “processing status” and “processing conditions”.

[Effects of the Invention] As described above, in the conventional method, it is necessary to realize processing based on a simple method or to describe a very complicated method for realizing a complicated method. According to a first aspect of the present invention, in a processing machine adaptive control device that performs control by changing processing conditions during processing, a status storage unit that describes a plurality of types of processing states or processing states and processing conditions; A knowledge storage unit that describes a method of determining each processing condition according to the degree of each type of processing state or processing state and processing condition, and the plurality of types of processing state or processing state described in the situation storage unit An inference unit for obtaining a plurality of processing conditions based on the respective degrees of the processing conditions and the respective techniques described in the knowledge storage unit, synthesizing the plurality of processing conditions, and obtaining a value of the processing condition to be changed; The method of processing know-how can be described in the knowledge storage unit independently of the inference unit.
The change is facilitated, and complicated adaptive control in consideration of various factors can be easily realized.

A time-series data recording unit that records, in chronological order, the machining state obtained by the state recognition unit, the machining condition changed by the operator, and the value of the machining condition to be changed obtained by the inference unit; Recorded in the sequence data recording unit, if there is a difference between the value of the processing condition changed by the worker and the value of the processing condition to be changed, a knowledge updating unit that updates the method described in the knowledge recording unit; , The processing technique can be easily updated.

Also, in a processing machine adaptive control device that controls by changing processing conditions during processing, a jump controller that controls a jump operation of a processing electrode, and a situation detection that detects a plurality of types of processing states or processing states and processing conditions. A state storage unit that describes the plurality of types of processing states or processing states and processing conditions detected by the state detector, and according to the degree of each of the plurality of types of processing states or processing states and processing conditions. A knowledge storage unit that describes a method of determining each jump operation condition; and the plurality of types of machining states or machining states and machining conditions described in the situation storage unit and the respective methods described in the knowledge storage unit A plurality of jump operation conditions obtained on the basis of the above, synthesize the plurality of processing conditions, and output a command value for a jump operation to the jump controller. Since, with the method having the skilled operator relating jumping operation can be properly and easily described, run and adaptive changes optimal jump operation by these techniques can be automatically performed.

In addition, the method of determining the processing conditions described in the knowledge storage unit is described in the rule format consisting of the antecedent part and the consequent part of the THEN. By sharing the information, management and sharing of processing know-how can be facilitated.

In addition, the method of determining processing conditions described in the knowledge storage unit is described in the form of a rule using a fuzzy set consisting of an antecedent part and a THEN consequent part, and synthesis of processing condition values in the inference unit is performed by fuzzy inference. Therefore, addition and change of the method can be further facilitated, and management and sharing of the processing know-how can be facilitated by sharing the knowledge storage unit.

[Brief description of the drawings]

FIG. 1 is a block diagram of a processing machine adaptive control device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram describing a method of changing the amount of increase / decrease of electrode lifting, and FIG. FIG. 4 is a flow chart, FIG. 4 is an explanatory diagram describing a method of changing the amount of increase / decrease of the electrode pulling using a rule,
FIG. 5 is an explanatory diagram describing a method of changing the amount of increase / decrease of the electrode lifting using a rule expressed by a fuzzy set,
FIG. 6 is an explanatory diagram describing a membership function for a fuzzy set, FIG. 7 is an explanatory diagram describing a method of changing the amount of increase / decrease of electrode pull-up by rules, and FIG. 8 is a discharge according to a third embodiment of the present invention. FIG. 9 is a block diagram of a machining control device, FIG. 9 is an explanatory diagram showing an example of an effective method for performing a jumping operation of a machining electrode, and FIG. 10 is an example describing the method of FIG. 9 using a fuzzy set. Illustrated explanatory drawing, No.
FIG. 11 is an explanatory diagram showing a process of fuzzy inference by the method of FIG. 9, FIG. 12 is a block diagram showing a configuration of a processing machine adaptive control device according to a fourth embodiment of the present invention, FIG. 13 and FIG. FIGS. 15 and 16 are flow charts showing the operation of the apparatus shown in FIG. 12, and FIGS. 17 to 19 are processing machine adaptive control devices according to the fifth to seventh embodiments of the present invention. FIG. 20 is a block diagram of a conventional processing machine adaptive control device, FIG. 21 is an explanatory diagram showing the movement of a processing electrode, FIG. 22 is a circuit diagram of an adaptive control unit, FIG. FIG. 2 is a block diagram showing another example of a conventional processing machine adaptive control device. In the figure, (10) is a worker, (20) is a processing machine, (21)
Is an electrode position control unit, (22) is a machining power supply, (23) is a detection value processing unit, (31a) is an adaptive control unit, (41) and (41a) are situation storage units, (42) and (42a) are Knowledge storage, (43), (43
a) is an inference unit, (60) is a jump controller, (70) is a situation detector, (87) is a first component group, (88) is a time-series data storage unit, (89) is a knowledge update unit, 90) is the second group, (9
1) is an example teaching unit, (92) is a rotary knob, and (93) is a CRT display device. In the drawings, the same reference numerals indicate the same or corresponding parts.

Continued on the front page (31) Priority claim number Japanese Patent Application No. 63-165081 (32) Priority date July 4, 1988 (July 4) (33) Priority claim country Japan (JP) (72) Inventor Juichi Maruyama Hyogo 8-1-1, Tsukaguchi-Honmachi, Amagasaki-shi, Japan Mitsubishi Electric Corporation Applied Equipment Research Laboratories (56) References JP-A-67-172751 (JP, U) JP-B 62-10769 (JP, B2) (58) Survey Field (Int.Cl. ⁶ , DB name) B23H 7/20 B23H 1/02 B23Q 15/12 G05B 13/02

Claims (5)

(57) [Claims] 1. A processing machine adaptive control device for performing control by changing processing conditions during processing, comprising: a state storage unit describing a plurality of types of processing states or processing states and processing conditions; A knowledge storage unit that describes a method of determining each processing condition according to each degree of the processing state and the processing condition; and each of the plurality of types of processing state or the processing state and the processing condition described in the situation storage unit A plurality of processing conditions based on the degree of and the respective methods described in the knowledge storage unit, and combining the plurality of processing conditions, and an inference unit for obtaining a value of the processing condition to be changed. Characteristic processing machine adaptive control device. 2. A time-series data recording section for recording in a time-series a processing state obtained by a state recognition section, a processing condition changed by an operator, and a value of a processing condition to be changed obtained by an inference section; If there is a difference between the value of the processing condition changed by the operator and the value of the changing processing condition recorded in the time-series data recording unit, the knowledge for updating the method described in the knowledge recording unit The processing machine adaptive control device according to claim 1, further comprising an updating unit. 3. A processing machine adaptive control device for controlling a processing condition by changing a processing condition during processing, a jump controller for controlling a jump operation of a processing electrode, and detecting a plurality of types of processing states or processing states and processing conditions. A state detector that describes the plurality of types of processing states or processing states and processing conditions detected by the state detector; and a degree of each of the plurality of types of processing states or processing states and processing conditions. A knowledge storage unit that describes a method of determining each jump operation condition according to the plurality of types of processing states or the processing states and processing conditions described in the situation storage unit and the knowledge storage unit described in the knowledge storage unit A plurality of jump operation conditions obtained based on the respective methods are obtained, the plurality of processing conditions are combined, and a command value of the jump operation is output to the jump controller. Machine adaptive control apparatus characterized by comprising a part. 4. The method according to claim 1, wherein the method for determining the processing condition described in the knowledge storage unit is described in a rule format including an IF antecedent part and a THEN consequent part. 2. The processing machine adaptive control device according to 1. 5. A method for determining processing conditions described in a knowledge storage unit is described in a rule format using a fuzzy set including an antecedent part and a THEN consequent part, and the synthesis of the processing condition values in the inference unit is fuzzy. 2. The method according to claim 1, wherein the calculation is performed by inference.
The processing machine adaptive control device according to claim 3.