JPH1011125A - Working robot device and method for controlling continuous feeding speed - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a continuous feeding speed controlling method capable of using a tool in the best efficient state. SOLUTION: In the continuous feeding speed controlling method for detecting the load of a tool 4 by a power converter 6 and changing a feeding speed so that the load becomes objective load in the case of executing working by using a teaching playback type industrial robot 1 and a tool/cutter, the acceleration of the feeding speed is operated from a deviation between the tool load and the objective load and the change rate of the deviation to obtain a speed command.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a processing robot apparatus for performing deburring, cutting and the like using an industrial robot and various tools and cutting tools, and a continuous feed speed control method.

[0002]

2. Description of the Related Art Usually, robots are used for grinding, cutting,
When cutting / deleting, in general, the workpiece (work)
Activate the cutting tool suitable for the purpose. The cutting tool is rotated or vibrated by a tool.

A generally used robot is a teaching playback system, which repeats a movement taught in advance. Usually, the operation position and the feed rate are set according to the material of the work and the processing conditions (grinding depth and cutting area).

There are various types of tools, such as tools manufactured for human use and dedicated robots (including machine tools), but their capabilities are limited, and the processing load is set at the time of teaching. If the tool capacity exceeds the tool capability, the operation of the tool stops. Further, as the cutting tool is used, the working capacity is reduced due to wear or the like, so that the load on the tool is constantly increased and may eventually exceed the working capacity of the tool. Even in such a state, the robot attempts to continue the operation at the taught feed speed, so that the tool is overloaded, and eventually the cutting tool and the tool are damaged.

[0005] As a countermeasure, a tool or cutting tool which detects the load of a tool and temporarily stops the feed when overloaded, and restarts when the load is reduced by cutting or grinding at the stop position. Protection method (JP-A-2-13916)
No. 7) has been proposed by the present applicant.

[0006]

[Problems to be solved by the invention]

(1) In order to temporarily stop the robot when overload is detected, it is necessary to perform signal processing with the CPU of the robot controller. During that time, that is, from when overload is detected until the robot actually stops, The increase will continue. Therefore, the stop or the vibration is stopped in time, and the rotation or the vibration is stopped largely exceeding the capability of the tool, and there is a danger of breakage. To avoid this, provide a maximum load setting (maximum set value) that is higher than the set value for pausing (pause setting value) and does not stop rotation or vibration,
It is necessary to stop the system as soon as the set value is reached. However, when the system is stopped, it is difficult to restore the robot and peripheral devices to their original positions, so that the upper limit is not reached during normal grinding.

For this reason, the larger the interval between the temporary stop set value and the maximum upper limit set value, the higher the efficiency but the higher it cannot be. Furthermore, when the size of the load to be ground is uncertain such as deburring, the feed speed during teaching is set assuming that the maximum load occurs. Is low and efficiency is low.

(2) In the conventional method, since there is only a state of stop or operation, if the load is large, the system is temporarily stopped again at restart and repeated. In this state, the finishing state is poor, and the adverse effect on the tools and cutting tools is great. Therefore, if the number of times is counted and grinding cannot be performed within the set number of times, it is determined that processing is impossible and the system is stopped. Therefore, when the processing load becomes larger than that at the time of teaching due to a change in the lot of the work or the like, the operation often stops.

(3) When the feed is stopped, a large reaction force is also applied to the robot arm, so that a slight deflection is caused, and the trajectory tends to remain more than a normal trajectory. In order to return to the load state, a step is generated in that portion.

(4) When the load state (burr amount, etc.) of the object to be processed changes from that at the time of teaching, the only adjustment is to correct the feed speed taught.

Accordingly, an object of the present invention is to solve the above-mentioned problems and to provide a continuous feed speed control method capable of using a tool in the most efficient state.

Another object of the present invention is to solve the above-mentioned problems and to provide a machining robot device which can use a tool in the most efficient state and can improve machining efficiency and accuracy.

[0013]

SUMMARY OF THE INVENTION According to the present invention, when a machining operation is performed using an industrial robot of a teaching playback system and a tool / blade, a tool load is detected and a feed speed is adjusted so that the load becomes a target load. In the continuous feed rate control method of changing the feed rate, it is necessary to calculate the acceleration of the feed rate by fuzzy inference from the deviation between the tool load and the target load and the rate of change of the deviation to obtain the speed command in order to determine the feed rate. Features.

The present invention provides a continuous feed rate for detecting a tool load and changing a feed rate so that the load reaches a target load when performing a machining operation using an industrial robot of a teaching playback system and a tool / blade. In the control method, in order to determine the feed rate, the deviation between the tool load and the target load,
From the rate of change of the deviation, the speed command is obtained by calculating the acceleration of the feed speed by fuzzy inference, while setting the upper limit and temporarily stopping the robot feed when the upper limit is detected,
After a certain period of time, perform a load check again.If the load has dropped, lower the feed rate and restart the machine.If a large load is expected from the teaching stage, feed it just before grinding the part. It is characterized in that a speed deceleration command is given to forcibly decelerate to reduce the load peak.

Further, the present invention relates to a load transducer for detecting a tool load in a machining robot apparatus for performing a machining operation by controlling an industrial robot with a robot controller using an industrial robot of a teaching playback system and a tool / blade. A continuous feed rate in which a FUZZY calculation CPU, an A / D converter, a D / A converter, D / O, and D / I are connected via a data bus to a CPU that controls various connected devices and performs signal processing. A control unit and, in order to determine the feed speed, to obtain a speed command by calculating the acceleration of the feed speed by fuzzy inference from the deviation between the tool load and the target load, and the rate of change of the deviation; and FU
The function of the ZZY calculation CPU is provided to the CPU for the continuous feed speed control unit or the management CPU of the robot controller.

[0016]

FIG. 1 shows a first embodiment of the present invention. In the figure, 1 is a robot main body, 2 is a robot controller of a teaching playback system, 3 is a continuous feed speed control unit, 4 is a high-frequency electric tool attached to an arm of the robot main body 1, 5 is a power supply of the high-frequency electric tool 4 It is an inverter device used as. A drive motor of the high-frequency electric tool 4, for example, an induction motor 4M is connected to an output terminal of the inverter device 5, and a load current and a voltage thereof are detected and used as an input of a load transducer (power converter) 6.

The robot controller 2 has a management CP
It comprises a U2-1, a data memory 2-2, a control CPU 2-3, a position detection interface 2-4, a servo interface 2-5, a servo driver 2-6, and an I / O interface unit 2-7. The I / O interface unit 2-7 includes a D / I interface 2-7A, a D / O interface 2-7B, and an A / D
It has a converter 2-7C.

The management CPU 2-1 has a function of performing an inching operation for moving the robot to an arbitrary position, a teaching operation (man-machine interface), and storing and reproducing teaching data. Also, adaptive control for stopping or changing the operation speed is processed according to the condition of each input / output signal. The data memory 2-2 stores teaching data and control parameters. The control CPU 2-3 controls each axis of the arm based on the teaching position data provided from the management CPU 2-1. The position detection interface 2-4 gives a pulse from an absolute (ABS) encoder attached to the servomotor of each axis to the control CPU 2-3. The servo interface 2-5 converts each axis operation angle command from the control CPU 2-3 into pulse data to be given to the servo driver 2-6. The servo driver 2-6 drives the motor of each axis. The D / I interface 2-7A inputs a digital signal. D / O interface 2-7B
Outputs a digital signal. A / D converter 2-7C
Converts an analog signal into a digital signal.

The robot controller 2 is an I / O of a general teaching playback type controller.
Add A / D converter 2-7C to the interface,
The voltage signal can be handled by converting it to internal data of 0 to 100% with respect to 0 V to full scale.

The continuous feed speed control unit 3 is a CPU 3-1 for controlling various connected devices and processing signals.
And a fuzzy (FUZZY) calculation CPU 3-2, A / D
Converter 3-3, D / A converter 3-4, D / O (digital output device) 3-5, D / I (digital input device) 3-6
Are connected by a data bus.

The load transducer 6 measures the input power of the tool 4 and outputs a voltage proportional thereto. This output voltage is input to the A / D converter 3-3 of the continuous feed speed control unit 3.

In the above configuration, it is possible to set a load that exerts the maximum performance on the tool 4 regardless of rotation or vibration. In the present invention, in order to always follow the load, the load deviation and the change rate of the deviation are constantly monitored, the acceleration is calculated by FUZZY inference according to the change, and the acceleration is calculated. The load of the tool 4 is adjusted by changing the feed rate.

FIG. 2 shows a basic signal processing flow (flow diagram). The continuous feed speed control unit 3 stores in advance an initial set value and knowledge required for FUZZY control. When the power is turned on, first, the target load value A _s and other parameters are set (step S1). Subsequently, the output voltage signal of the load transducer 6 is converted into internal data by the A / D converter 3-3 of the continuous feed rate control unit 3 (S2). The converted data is CP
Preprocessing is performed by U3-1.

[0024] That is, for calculating the load deviation .DELTA.A _i from the load measured values A _i by subtracting the target load value A _s (S3), to match the input range of the membership function, the gain G ₁ in .DELTA.A _i
Is multiplied to calculate ΔA _INP (S4), and ΔA _{i −1} calculated in the previous process is subtracted from ΔA _i to calculate a deviation change rate Δa _i (S5), which is adjusted to the input range of the membership function. Therefore, processing such as multiplying Δa _i by the gain G ₂ and calculating Δa _INP (S6) is sequentially performed. Thereafter, the FUZZY inference is executed by the FUZZY calculation CPU 3-2 (S7). F used in this FUZZY inference
Knowledge of UZZY control (rules and membership functions)
Is shown in Table 1 and FIGS. The details will be described later. A method of obtaining an inference result using this knowledge is described in FUZZ.
The centroid method, which is a general method of Y inference, is adopted.

The calculation result is output as a speed command acceleration ΔV _OUT . Thereafter, ΔV _i is calculated by multiplying ΔV _OUT by a gain G ₃ so as to obtain an optimum increase / decrease amount (S
8), which added to the integrated value V _i-1 output in the previous process (integration) and a velocity command value V _i to (S9). Velocity command value V _i is output to the robot controller 2 through the D / A converter 3-4 (S10). In the robot controller 2, the analog input voltage is converted to 0 to 100% by the A / D converter 2-7C (S11). By multiplying this output value by the designated speed at the time of teaching, the feed speed at the time of machining is changed.

The continuous feed speed control unit 3 monitors the operation state of the tool 4 by the load transducer 6, and outputs a "grinder abnormality" signal to the robot controller 2 when it is determined that the tool 4 has failed. Further, when the load increases so rapidly that even the above control cannot complete the processing, a signal for stopping the robot 1 is output to protect the tools / cutting tools and the robot.

Table 1 shows the rules used for FUZZY inference.

[0028]

[Table 1]

This table shows the deviation ΔA _i and the rate of change Δa _{i of the} deviation.
Are shown as a matrix. For example, “PM” in the second row and third column of the table represents the following combination in the fuzzy rule.

"If ΔA _i = NS and Δa _i = N
M then ΔV _i = PM ”Symbols PL, PM, PS, ZR, NS, NM, NL in the table
Are the membership functions described below as “positive and large”, “positive and medium”, “positive and small”, “zero”, “negative and small”, “negative and medium”, and “negative and large”, respectively. Indicates the label.

The features of the rule, the larger the deviation .DELTA.A _i, also the absolute value of the higher acceleration [Delta] V _i change rate .DELTA.a _i deviation increases increases, the degree of change is one that basic to be higher, Non-linear control can be realized in which the larger the change in the load is, the more sensitive the response is, and the closer to the target, the less the response is. In this embodiment, the range of “ZR” is widened near the center of the matrix in order to secure control stability especially near the target load.

The combination is not limited to the above example, and the range of each label is increased or decreased according to the controllability of the target work, and an optimum combination is selected.

Next, the membership function will be described. The condition part is two of the load deviation ΔA _i and the rate of change of the deviation Δa _i .
Input. 3 and 4 show the membership functions. The input value is input as a ratio of the load transducer 6 to the full scale.

The feature of the membership function of the load deviation ΔA _i of the input (1) is that the interval between each fuzzy set becomes narrower toward the outside, so that the response becomes sensitive when the deviation is large, In the vicinity, a characteristic close to a dead zone is provided to achieve both improvement of responsiveness and stability of control.

The shape of the fuzzy set includes a bell shape, a trapezoid shape and the like in addition to the triangular shape in the illustrated example. In the embodiment,
A typical triangular type is illustrated.

The feature of the membership function of the load deviation change rate Δa _i of the input (2) is that the interval between each fuzzy set becomes narrower toward the outside so that the response is sensitive when the deviation is large. In the vicinity of 0, a characteristic close to a dead zone is provided to achieve both improvement of responsiveness and stability of control.

The format of the fuzzy set is 3 in the illustrated example.
In addition to the square shape, there are a bell shape, a trapezoid shape and the like. In the embodiment, a typical triangular type is illustrated.

The conclusion part is the first output feed speed increase ratio [Delta] V _i. FIG. 5 shows the membership function. The output value is finally multiplied and used.

The features of the membership functions of the feed rate increase rate [Delta] V _i output by setting as more intervals of the fuzzy set close to 0 is narrowed, [Delta] V _i increases output greater the label That is, such a nonlinear output value can be obtained.

The type of the fuzzy set of the embodiment is as follows.
This type is called a singleton, and there are other types such as a triangle type, a bell type, and a trapezoidal type. However, since the control characteristics are not significantly affected, the singleton type which is easy to set is exemplified.

In the above embodiment, the continuous feed speed control unit 3 is installed separately from the robot controller 2, and D / O, D / I , D / A interfaces, the operation of which can be summarized as follows: (1) When the "tool start" signal is received, the control unit 3
Outputs an operation signal to the inverter device 5, and after the start-up is completed,
Check no-load current, (2) "Control ON / OFF"
Upon receiving the signal, the activation of the tool 4 is confirmed, and if there is no abnormality, the robot 1 is sent to the D / A converter 3-4 according to the aforementioned knowledge.
(3) If the check after the start of the tool 4 is abnormal, the control unit 3
Outputs an "grinder abnormality" and a "robot stop" signal, and the robot controller 2 displays an abnormality and stops.

The CPU of the control unit 3 of the first embodiment
If 3-1 is upgraded and fuzzy inference processing is performed by software, the FUZZY calculation CPU 3-2 can be omitted. In such a case, if the A / D converter is built-in or if it can be added, the robot controller only needs to be equipped with a continuous feed speed control unit without greatly modifying the existing controller. Get higher. Although the load on the CPU of the control unit 3 increases, the cost can be reduced by omitting the FUZZY calculation CPU.

FIG. 6 shows a second embodiment of the present invention. In the second embodiment, the main part of the continuous feed rate control unit 3 (the CPU 3-1 and the FUZZY calculation CP
U3-2, a control unit 3 'composed of an A / D converter 3-3) is incorporated in the I / O interface unit 2-7' of the robot controller 2 'to send and receive signals and output feed speed commands. I / of the robot controller 2 '
O data bus.

In the above configuration, the robot controller 2 'controls the control unit 3' to provide the robot A / D
Recognize in the same way as a converter. Processing of each signal is performed according to the first embodiment.
Is the same as

In the second embodiment, the D / O, D / I and D / A converters on the feed speed control unit side and the A / D converter of the robot controller are omitted, and the control unit 3 'is replaced with the robot controller. Since it is incorporated in 2 ', it is not necessary to add a panel or the like, so that the cost is greatly reduced. Further, since the A / D and D / A conversion processing of the speed command is eliminated, the processing can be speeded up. Furthermore, even if the robot controller is built in, the FUZZY calculation CPU
Because of having 3-2, the control cycle is quick and the response is good.

In the second embodiment, if the CPU 3-1 of the control unit 3 'is upgraded and the fuzzy inference processing is performed by software, the FUZZY calculation CPU 3-2
Can be omitted. Thereby, the cost can be further reduced.

FIG. 7 shows a third embodiment of the present invention. In the third embodiment, the management CPU of the robot controller 2 ″
The FUZZY calculation CPU 3-2 is provided on the board, and the control is performed by the management CPU 2-1. That is, as in the second embodiment, the continuous feed speed control unit is incorporated in the robot controller.

With such a configuration, the management CPU 2-
Since the control of the program can be performed in the step 1, the setting change or the like can be made into a dedicated command, or a man-machine interface on the robot side such as a teaching box can be used, thereby improving usability. Also, the input / output interface on the feed speed control unit side can be omitted, or the FUZ
Since the ZY calculation CPU is incorporated in the robot controller, it is not necessary to increase the number of boards, so that the cost is greatly reduced. Further, since the A / D and D / A conversion processing of the speed command is eliminated, the processing can be sped up. In addition, even if the robot controller is built in, the FUZZY operation CPU 3
Due to having -2, the control cycle is quick and the response is good.

In the third embodiment, the management CPU 3-
If the fuzzy inference processing is also performed by upgrading 1 to 1, the FUZZY calculation CPU 3-2 can be omitted.
Thereby, the cost can be further reduced.

Features of each of the first to third embodiments (1) Set a target load and control so as to change the feed rate so that machining can always be performed with that load (the tool can be used in the most efficient state) And). (2) From the deviation to the target load and the rate of change of the deviation, FUZZ
Obtaining acceleration by Y inference. Reasons for Using FUZZY Inference for Load Calculation a. Improve target following performance To converge to the target load quickly, make the response faster when the deviation is large, and slow it when the deviation is small, to achieve both control stability and high response (deviation of deviation). The same applies to the rate of change), and FUZZY control is adopted, and the construction of nonlinear control is facilitated by selecting how to compose the knowledge (rules and membership functions). b. Improvement of adaptability to changes in control conditions Since control performance changes continuously, control can be performed under optimal conditions. c. Improvement of adaptability to sudden changes When sudden changes occur, read-ahead control can be performed according to the magnitude of the deviation at that time and the rate of change of the deviation. (3) Since the calculated acceleration is integrated into the feed rate, it can be controlled without any problem even if the final feed rate command value changes due to differences in the material of the work and the processing conditions. FIG. 8 shows a fourth embodiment of the present invention. The fourth embodiment aims at performing processing without stopping the line as much as possible by utilizing the features of the continuous feed control even when there is a sudden change in load exceeding the control response limit due to processing delay. In continuous feed speed control, a speed change command voltage is output to a robot controller in accordance with a tool load, and the controller internally converts the voltage to a numerical value of 0 to 100% to adjust a feed speed specified in a program. Therefore, when the load is low, 100% of the speed change command is maintained, and the feed speed is exactly specified by the program. If the speed specified in this program is increased, the rate of change when the load suddenly increases will increase according to the speed, so the efficiency will be improved when the load variation is large,
Since the rate of change when the load suddenly increases increases according to the speed, the load peak due to the processing time lag increases. If this load peak greatly exceeds the capability of the tool, the operation of the tool will stop, so the system must be stopped for protection, and if this occurs frequently, the overall efficiency will decrease, resulting in a decrease in efficiency. Increase of the maximum speed that can be specified by the program cannot be expected.

In the fourth embodiment, by adding an automatic return operation when the system is stopped, the peak load is suppressed, the speed specified by the program is increased, and the line stop situation is avoided.

In the figure, 1 is a robot main body, 2 is a teaching playback type robot controller, 3 is a continuous feed speed control unit, 4G is a high frequency electric angle grinder attached to the arm of the robot main body 1, and 5 is this grinder 4G. This is an inverter device used as a power supply for the power supply. A drive motor of the grinder 4G, for example, an induction motor is connected to an output terminal of the inverter device 5, and a load transducer is detected by detecting a load current of the drive motor.
6 ′.

In the fourth embodiment, the set value at which the system is stopped is set to a level immediately before the operation stops due to exceeding the capability of the tool in view of safety. This set value is referred to as the “maximum value”. For example, in the case of a 3kw high-frequency electric angle grinder, the maximum output is 5k when the load current is 26A.
If w is exceeded, the output and the torque suddenly decrease when the w is exceeded, and the output shaft falls into a rotation stopped state. Therefore, the safety factor is set to 20A so as not to be exceeded even if a processing delay occurs.

For this reason, even if the system is stopped, there is a case where the robot does not actually reach the unrotatable state. If the "maximum limit" is detected, the automatic operation is not stopped as the system is stopped but the robot is fed. If the load is checked again after a certain period of time and the load is reduced again, the feed speed command is set to a low speed and restarted. Due to this control operation, even if the load is large, a sudden increase in the load does not occur, and after the restart, machining is performed at an optimum feed speed to operate with continuous feed speed control. FIG. 9 shows the processing flow. The process shown in this flowchart is performed by the control CPU 3-1 in the continuous feed speed control unit 3. The processing flow is as follows.

First, a load check is performed (S21). If the load is normal, continuous feed speed control is executed (S22), and then it is determined whether or not the grinding is completed (S23). If the "upper limit" is detected in the load check, the robot temporarily stops (S24) and waits for about 1 second (S2).
5). Thereafter, a reload check is performed (S26). If the load has decreased, the feed speed command is changed (10 to 20%).
(S27), and the robot is restarted (S28). If the load has not decreased, the system is stopped (S29). When grinding is not completed and when the robot is restarted, the process returns to the initial stage load check (S21).

With the addition of this processing, even if the tool is stopped due to an increase in the load peak, if the tool is not stopped yet, the tool is restarted to continue the machining, and the frequency of system stop is reduced. Therefore, the maximum speed at the time of grinding specified by the program is increased, and the processing efficiency when the load is small is improved. This is particularly effective when the load is not constant and fluctuates greatly.

If the load is expected to increase in advance,
If it is possible to predict in advance that a large load will be applied to the tool from the teaching stage, such as a casting gate or a hoist, for example, the teaching robot program sends a `` feed speed reduction command (D / O output) and the feed speed command
By forcibly decelerating to about% and starting grinding, it is possible to keep the load peak low, and this control operation is added. This forced deceleration is performed by starting grinding and removing the tool load from the no-load state (determining that grinding has started), or by elapse of a predetermined period of time from the start of deceleration (gating is performed more than assumed during teaching). And when the lift is small), the grinding is performed by returning to the normal continuous feed rate control, so that even if the load increase is not large, deterioration of the work efficiency due to forced deceleration is minimized. can do. FIG. 10 shows a processing flow when the "feed speed deceleration command" is input. The process shown in this flowchart is
The control is performed by the control CPU 3-1 in the continuous feed speed control unit 3. The processing flow is as follows.

First, a feed speed deceleration command check is performed (S31), and when it is OFF, continuous feed speed control (S3) is performed.
2), a routine of determining whether or not the grinding is completed (S33) is executed, and when ON, the feed speed deceleration command is held (10 to 20).
%) (S34), deceleration hold time check (S3
5) A routine for determining whether the load current is equal to or greater than the no-load current value upper limit (S36) is executed. Deceleration hold time (1 to
When the load current exceeds the upper limit (about 2 seconds) and when the load current exceeds the upper limit, the process of releasing the feed speed command fixation (S37) is performed, and the process returns to the continuous feed speed control process (S32).

With the addition of this processing, the suspension is not performed in the above-described state, and the grinding process is performed without stopping or suspending the system even when there is a lot of predictable frying and gates. That is, the grinding operation is efficient and stable. This is effective when a large load can be predicted in advance, but the magnitude fluctuates.

If the processing at the time of detecting the "maximum upper limit" and the processing at the time of inputting the "feed speed deceleration command" are performed, the applicable range can be expanded, which is useful for continuous feed speed control.

Next, an application example of the present invention will be described. One of application examples is an example in which the above-described high-frequency electric angle grinder is used as a tool (Embodiment 4 in FIG. 8), and is applied to burr grinding of a casting. High frequency electric angle grinder
Suitable for heavy grinding with offset whetstone as a tool. The drive motor is a 2-pole induction motor, three-phase two
A power supply (inverter device) of 00 V and 250 Hz is used. The use of the current converter 6 'for the load transducer is the same as in the embodiment (FIG. 8), and the operation is also the same.

As a power tool relatively frequently used for grinding or cutting of burrs, a high-frequency electric angle grinder, a high-frequency spindle (generally using a high-frequency power source,
Grinding wheels or end mills at high rotation speeds of 10,000 to tens of thousands, and internal grinders (High rotation speed grinders of 10,000 to tens of thousands, using internal whetstones as cutting tools, mainly grind the inner surface of castings etc. ), And straight grinders (at a lower speed than the spindle, mainly using a flat grindstone to grind a large area). The power source of the tool can be used even with a single-phase 100 V 50/60 Hz if a load transducer suitable for the use is used, and an angle grinder or a straight grinder for manual work can be used similarly. Further, in addition to deburring of castings, the present invention can be applied to grinding of welding beads and deburring of other materials such as FRP.

FIG. 11 shows a deburring robot device (Embodiment 5) using an air spindle. When cutting machined burrs using the air spindle,
The pulse pickup 7 is attached to the air spindle 4Sa so as not to cut too much. In the fifth embodiment, a tachometer is used as the load transducer 6 ", and the input from the pulse pickup 7 is converted into a rotation speed and output as a voltage proportional thereto. The output of the load transducer 6" is controlled as a continuous feed speed. Inputting to the unit 3, calculating by fuzzy inference, outputting a speed command to the robot controller 2, and controlling the robot 1 while changing the feed speed specified by the program based on the speed command is the same as described above. .

Other pneumatic tools include an air angle grinder (a cutting tool is an offset grindstone, which is more suitable for light grinding than an electric tool), an air internal grinder (a grindstone is mainly used to grind the inner surface of a work using an internal grindstone as a cutting tool), Air straight grinders (at lower speeds than spindles, mainly with flat grindstones for grinding large areas). In addition, the same configuration can be applied to burr grinding of FRP, grinding of weld beads, and the like.

FIGS. 12 and 13 show an FRP using a power tool.
An example (embodiment 6 and 7) of product trimming (cutting processing) is shown. This example is structurally the same as the above-described fourth embodiment. Trimming is performed by attaching a cutter (a saw blade on which diamond is electrodeposited) 8 as a cutting tool to the electric angle grinder 4G (FIG. 12), and the electric spindle 4S
This is a case where a router bit 8 'is attached as a cutting tool to e for trimming (FIG. 13).

The former is suitable for use in cutting a straight portion over a long distance. The latter is suitable for trimming small R portions and curves. The signal processing and operation are the same as in the fourth embodiment.

FIGS. 14 and 15 show an example (Eighth Embodiment) of sluice cutting of a cast product by a floor type cutting machine. In the eighth embodiment, an electric floor-standing cutting machine 11 is used, and a tip saw 11A is attached to the cutting machine 11 as a cutting tool, and the robot 1 grips the work 12 to cut the gate 12A. The tip saw 11A is rotationally driven by a drive motor 11B, a pulley 11C, a belt 11D, and the like. The tool load is a load transducer 6 'using a current converter.
Detected by Here, a current converter is used, but a power converter may be used. In general, in the case of such an application, since the capacity of the motor is large and the inertia increases with the capacity, the torque converter may be used as a means for capturing the load fluctuation more quickly. In that case, as shown in FIG. 15, a torque sensor 13 is inserted through a coupling 11E to a shaft to which the tip saw 11A is attached, and connected to an amplifier (load transducer) 6T. The signal processing and operation are the same as in the fourth embodiment.

FIGS. 16 and 17 show an example of the torque converter (manufactured by Kubota Corporation). FIG. 16 is an operation principle diagram, and FIG. 17 is a block diagram. When torque is applied to the shaft 31, tensile stress and compressive stress act on the grooved portions A and B, respectively, and the magnetic permeability of the grooved portions A and B increases and decreases, respectively. This is voltage / current converted and output. for that reason,
Excitation coils 32A and 32B of two systems and detection coil 33A,
33B is provided and connected to the amplifier 6T. Amplifier 6T
Oscillator circuit 34, power amplifier 35, rectifier circuits 36A and 36B, filters 37A and 37B, arithmetic circuit 38, V
/ I converter 39 and the like. In the A coil system,
The magnetic permeability of the grooved portion A increases due to the tensile stress + σ, and the output voltage increases. In the B coil system, the magnetic permeability of the grooved portion B decreases due to the compressive stress −σ, and the output voltage decreases. The torque is calculated based on the difference between the two output voltages.

FIGS. 18 and 19 show examples (embodiments 9 and 10) of drilling an FRP product. In the ninth embodiment (FIG. 18), a hole saw 9 is used as a cutting tool, and this is attached to the electric drill 4De. A current transducer is used as a load transducer. Signal processing and operation
This is the same as the fourth embodiment.

The tenth embodiment is an example in which an air drill is used when only drilling holes or when using a small-diameter hole saw, and a hole saw 9 'is attached to the air drill 4Da. In this case, the number of rotations is detected in the same manner as in the fifth embodiment, and is used for feed speed control. When using an air drill, the prior application (Japanese Patent Laid-Open No.
No. 50910) is attached, but in the present embodiment, since load control can be performed, the necessity is eliminated.

[0071]

【The invention's effect】

A. Overall (basic) effect (1) Even if the speed command at the time of teaching is rough, machining is performed while automatically changing to the optimum speed, so it is necessary to limit the feed speed at the maximum load expected to occur Gone
Shortening the cycle time during normal operation increases machining efficiency, (2) increases the ability to respond to load fluctuations, reduces line stoppage, and (3) eliminates the need to specify the teaching feed speed precisely. The work to adjust the speed to the optimum value is reduced, and the time required for cutting and trying to start the line can be shortened. (4) Since the speed change command continuously changes between 0 and 100%, The operation is smooth. (5) Even if the tool stops due to an increase in the load peak, if the tool has not stopped yet, it will be restarted and processing will continue, reducing the frequency of system stoppage and specifying it in the program. The maximum speed at the time of grinding is high, and the processing efficiency is improved when the load is small. (6) Grinding can be performed without stopping or pausing the system even when there are many It is now being, efficient, and becomes a stable grinding operation, the effect of such.

B. Application effects a. In the case of burr grinding or cutting (1) Since the burr is always ground while changing the feed rate so that the target load of the tool is achieved, even if there is variation in the burr amount between lots and individual products, it is efficient. Grinding can be done, cycle time can be shortened,
(2) Even if burrs that exceed the capability of the tool are generated, the feed rate is reduced to cope with this, so the range of burrs that can be grinded is widened, and the frequency of system stoppages is reduced. (3) Robots during overload Because the feed speed of the
It is not as extreme as restarting, and the finish quality is stable. (4) Processing continues without overload even if the load increases due to the wear of the cutting tool, extending the service life and shortening the use time. There are effects such as extension.

B. In the case of cutting (1) Objects with large fluctuations in tool load (Thickness is not constant, such as FRP parts laminated by hand, and the cutting area is large due to the cylindrical shape like the gate of castings) When it is applied to a material that changes, or a material whose starting and end cannot be cut at the same time, such as when cutting multiple gates at the same time), the cutting thickness is adjusted automatically to adjust the cutting speed according to the load. Even if it fluctuates, it can be cut efficiently. Also, when cutting the gate, the feed speed is automatically adjusted only by teaching the two points of the start and end of the cut, so the cutting time can be shortened. (2) The start and end positions of the cut may change In the case of the target product, the same running speed as the cutting speed is provided before and after that, so that it does not suddenly cut at the approaching speed and does not perform the separating operation before the cutting is completed Although the efficiency is inevitable due to the cutting, the load on the blade is detected and the feed rate is optimized for cutting.
The speed of the idle section can be increased several times as compared with the conventional one, and the idle time can be greatly reduced. (3)
During cutting, the blade can be cut under a certain load, preventing damage. (4) Even if a thick product exceeding the tool's ability appears, the cutting speed can be reduced by reducing the feed speed, so the range that can be cut is expanded, The frequency of system stop decreases. (5) The feed rate of the robot at the time of overload is not extreme like pause and restart, and the finish quality is stable because it changes continuously. Even if the load increases due to wear, processing can be continued without overloading, so the service life (use time)
Is prolonged.

C. In the case of drilling (1) If the object is an FRP part (thickness is not constant) that is manually laminated, the drilling speed according to the load is automatically adjusted even if the tool load fluctuates greatly. for,
Drilling can be performed efficiently even if the thickness varies. In addition, the feed speed is automatically adjusted only by teaching the two points at the start and end of machining, so that the machining time can be shortened. (2) Target products that may change the drilling start and drilling end positions In the case of, the same speed as the drilling speed is provided before and after that, so that the vehicle does not cut at the speed of approaching suddenly and does not perform the withdrawal operation before the drilling is completed, so the efficiency by the control operation Although it is inevitable to decrease, the speed of the idle running section can be increased several times compared to the conventional speed to detect the load of the blade and adjust the feed speed to the optimum value for drilling, greatly increasing the idle running time. Can be shortened,
(3) During drilling, the cutting tool can be cut with a certain load to prevent damage. (4) Even if a thick product exceeding the tool's ability appears, the feed rate can be reduced to cope with it. (5) The robot's feed speed during overload is not as extreme as that of pause and restart, and the finish quality is stable because it changes continuously. (6) Even if the load increases due to wear of the cutting tool, machining can be continued without overloading.
This has the effect of extending the life (use time).

[Brief description of the drawings]

FIG. 1 is a block circuit diagram showing a first embodiment of the present invention.

FIG. 2 is an operation flowchart of signal processing in the first embodiment.

FIG. 3 is an example of a membership function of an input (1) (load deviation ΔA _i ) of the first embodiment.

FIG. 4 shows the input (2) (change rate of load deviation Δ) of the first embodiment.
a _i ) Membership function example.

FIG. 5 is an example of a membership function of output (feed rate increase rate ΔV _i ) of the first embodiment.

FIG. 6 is a block circuit diagram showing a second embodiment of the present invention.

FIG. 7 is a block circuit diagram showing a third embodiment of the present invention.

FIG. 8 is a block circuit diagram showing a fourth embodiment of the present invention.

FIG. 9 is a flowchart at the time of detecting “upper limit” in the fourth embodiment.

FIG. 10 is a flowchart when a “feed speed deceleration command” is input in the fourth embodiment.

FIG. 11 is a block circuit diagram showing a fifth embodiment of the present invention.

FIG. 12 is a side view showing a tool and a cutting tool according to a sixth embodiment of the present invention.

FIG. 13 is a side view showing a tool / blade according to a seventh embodiment of the present invention.

FIG. 14 is a block circuit diagram showing an eighth embodiment of the present invention.

FIG. 15 is a layout and connection diagram showing an example of use of a torque converter according to an eighth embodiment.

FIG. 16 is an operation principle diagram of the torque converter.

FIG. 17 is a block diagram of a torque converter.

FIG. 18 is a side view showing a tool / blade according to a ninth embodiment of the present invention.

FIG. 19 is a side view showing a tool / blade according to a tenth embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Robot main body 2, 2 ', 2 "... Robot controller 2-1 ... Management CPU 2-3 ... Control CPU 2-4 ... Position detection interface 2-6 ... Servo driver 2-7 ... I / O interface unit 2- 7C A / D converter 3, 3 'Continuous feed speed control unit 3-1 CPU 3-2 FUZZY calculation CPU 3-3 A / D converter 3-4 D / A converter 4 High frequency Power tool 4M Drive motor 4Sa Air spindle 4Se Electric spindle 4G High frequency angle grinder 5 Inverter 6 Load transducer (power converter) 6 'Load transducer (current converter) 6 "Load transducer (rotometer) 6T: Load transducer (torque sensor amplifier) 7: Pulse pickup

Claims (7)

[Claims] 1. A continuous feed rate control method for detecting a tool load and changing a feed rate so that the load reaches a target load when performing a machining operation using an industrial robot of a teaching playback system and a tool / blade. A continuous feed speed control method comprising: calculating a feed speed acceleration by fuzzy inference from a deviation between a tool load and a target load and a change rate of the deviation to obtain a speed command in order to determine the feed speed. . 2. A continuous feed rate control method for detecting a tool load and changing a feed rate so that the load reaches a target load when performing a machining operation using an industrial robot of a teaching playback system and a tool / blade. In order to determine the feed speed, the speed command is obtained by calculating the acceleration of the feed speed by fuzzy inference from the deviation between the tool load and the target load and the rate of change of the deviation, and the upper limit value is set. Continuous feed speed control method characterized by temporarily stopping the feed of the robot when the upper limit value is detected, checking the load again after a certain period of time, and if the load has decreased, reducing the feed speed and restarting. . 3. A continuous feed rate control method for detecting a tool load and changing a feed rate so that the load reaches a target load when performing a machining operation using an industrial robot of a teaching playback system and a tool / blade. In order to determine the feed speed, the speed command is obtained by calculating the acceleration of the feed speed by fuzzy inference from the deviation between the tool load and the target load and the rate of change of the deviation, while a large load is predicted from the teaching stage. If so, a continuous feed speed control method characterized in that a feed speed deceleration command is given immediately before grinding of the part to forcibly decelerate the load and reduce the load peak. 4. In a machining robot device for performing machining work by controlling an industrial robot with a robot controller using an industrial robot of a teaching playback system and a tool / blade, a load transducer for detecting a tool load and various connections. CP that controls equipment and performs signal processing
U includes a FUZZY operation CPU, an A / D converter, a D / A converter, a continuous feed speed control unit having a configuration in which D / O and D / I are connected by a data bus. A machining robot device wherein a speed command is obtained by calculating an acceleration of a feed speed by fuzzy inference from a deviation between a tool load and a target load and a rate of change of the deviation. 5. The processing robot device according to claim 4, wherein a main part of the continuous feed speed control unit is incorporated in an I / O interface unit of the robot controller to reduce the number of boards. 6. A FUZZ for a continuous feed rate control unit.
5. The processing robot device according to claim 4, wherein the Y operation CPU is incorporated in a management CPU board of the robot controller, and the function is provided to the robot controller. 7. The method according to claim 4, wherein the FUZZY calculation CPU is omitted, and its function is replaced by a CPU for a continuous feed speed control unit or a management CPU of a robot controller. Processing robot device.