JPH08187578A - Storing method of data for position correcting in multi-layer welding control - Google Patents

Abstract

PURPOSE: To greatly save buffer memory capacity to be used by improving data storing method in executing multi-layer welding by welding with use of a laser sensor mounted to a robot in tracking method. CONSTITUTION: When a program data for one point is read and welding of a first layer for multi-layer welding by welding is executed (T2, T3) and a traveling quantity of robot exceeds a setting value (d), a data for position correcting is read from a buffer memory region A (T), a traveling target value is corrected (T6). Successively, as data for correcting after second layer, a coefficient and constant member of the approximate linear expression expressing linear region is determined, a distributing pulse is calculated from the traveling target position obtained by T6 (T11). Further, a degree of program digestion is discriminated by T12, T13 (T13). In welding after second layer, in accordance with robot traveling quantity, the data for position correcting stored at first layer is read from a buffer memory region B (T9), the traveling target position is corrected (T10).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data storage method for controlling multi-pass welding, more specifically, data in multi-pass welding control for performing welding line tracking using a laser sensor mounted on a welding robot. Regarding storage method.

[0002]

2. Description of the Related Art In a welding line tracking control system using a laser sensor which is currently adopted, a laser sensor senses a region in front of a tool tip point of a robot to detect the position of the welding line, and the detection result is detected. The position correction data created based on the sensor data indicating the robot position and the data indicating the robot position at the time of detection are sequentially written in a predetermined buffer memory area, while the stored data are sequentially read to determine the robot movement target position. Meanwhile, control is performed to move the tool tip point of the robot along the welding line.

When performing multi-layer welding by this method, the position correction data created during the welding of the first layer is used not only for tracking during the welding of the first layer, but also for the robot position correction for the second and subsequent layers. It is accumulated for the purpose.

That is, at the time of welding the first layer, position correction data used for tracking control at the time of welding the second and subsequent layers is written at regular intervals in a buffer memory area secured in advance in the system. Saved. The writing of the data for position correction is executed every time the robot moves a certain distance. Therefore, when the buffer memory capacity that can be secured in the system is small, or when it is necessary to write data at a short writing interval in a long multi-layer welding section, the buffer memory capacity becomes insufficient. Was there.

[0005]

SUMMARY OF THE INVENTION An object of the present invention is to overcome the above problems. That is, the invention of the present application is to improve the method of storing position correction data in multilayer welding control using a laser sensor, and to greatly save the buffer memory capacity to be used. Further, through this, even when a system having no large buffer memory capacity is used, it is possible to perform multi-layer welding by a tracking control method using a laser sensor for a long section.

[0006]

The present invention has, as a basic configuration for solving the above technical problems, "a robot controller, a robot controlled by the robot controller, and a robot supported by the robot. A welding torch and a laser sensor for detecting the position of the welding line on the workpiece to be welded. For robot position correction in multi-pass welding control using a welding robot system that follows the welding line. A method of storing data,
A robot in the straight line section of the second and subsequent layers based on the sensor data obtained by using the laser sensor to move the robot to follow the welding line in the straight section during the welding of the first layer The position correction data for moving is created, and the created position correction data is written in the buffer memory area prepared in the welding system. A method for storing robot position correction data in the multi-layer welding control, wherein the position correction data is data of a parameter that defines a linear expression that approximately expresses the welding line extending position in the straight line section "(claim 1) was proposed.

According to the present invention, in the above basic structure, "the step of creating position correction data for moving the robot in the straight line section on and after the second layer is the execution of welding on the first layer. Based on the data of the weld line positions relating to at least two points detected therein, a parameter defining a linear expression that approximately expresses the weld line extending position of the straight line section is determined "(claim 2). , The rationale for creating the above parameters is clarified.

Further, in order to improve the approximation accuracy in that case, "the step of preparing position correction data for moving the robot in the straight line section on and after the second layer is the welding on the first layer. A parameter that defines a linear expression that approximately expresses the welding line extension position of the straight line section is determined by the least squares method using the data of the welding line positions regarding at least three points detected during the execution of " It is a proposal to impose the requirement of item 3).

[0009]

According to the method of the present invention, with respect to the straight line section in which the multi-layer welding is performed, the second and subsequent layers are formed in the form of parameters that define a linear expression that approximately expresses the welding line extending position of the straight line section. The robot position correction data for welding is stored. Since the number of such parameters is very small, a large saving of the used memory capacity can be achieved in the straight section included in the multi-pass welding section as compared with the conventional method. Typical parameters that define a linear expression that approximately expresses the welding line extension position in a straight line section are a coefficient a representing the slope of the straight line and a constant term b. Therefore, even if a large number (N) of straight line sections are included in the process of one cycle of robot movement, the number of data to be finally stored only increases by about twice the number of straight line sections.

The position correction data for moving the robot in the straight line section on the second and subsequent layers is based on the welding line position data on at least two points detected during the execution of the welding on the first layer. It is obtained by determining parameters that define a linear expression that approximately represents the welding line extension position of the straight line section.

In order to improve the approximation accuracy in that case,
A parameter that defines a linear expression that approximately expresses the welding line extension position of the straight line section by the least squares method using the data of the welding line positions of three or more points detected during the execution of the first layer welding. Should be decided.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the present invention, a laser sensor is used that senses a region near the robot position on the traveling direction side. As is well known, a laser sensor forms a light spot array on a surface to be inspected (hereinafter referred to as a work surface) by deflecting and scanning a laser beam, and forms a light spot image on a light detecting means. Then, three-dimensional position information regarding the light spot or the light spot array is obtained.

FIG. 1 exemplifies a schematic configuration of a laser sensor used in the present invention. In FIG. 1, reference numeral 10 is a detection unit, which is a laser oscillator 11, an oscillating mirror (galvanometer) 12 for scanning a laser beam, It has an optical system 13 which captures reflected light and forms an image on the light receiving element 14.

On the other hand, the control unit 20 constituting the sensor board
Has a CPU 21 composed of a microprocessor,
Input / output device 28 and ROM to U21 via bus 29
And a memory 25 including a RAM are connected. Then, the input / output device 28 detects the position from the laser drive unit 22 that drives the laser oscillator 11 to generate a laser beam, the mirror operation unit 23 that swings the swing mirror 12, and the position received by the light receiving element 14. The signal detection unit 24 is connected. The input / output device 28 is also connected to a line 28 for exchanging various commands and data with a robot control device (not shown).

When a laser sensor start command from the robot controller is received, the laser sensor drive program stored in the memory 25 is started, and the CPU 21 sends a laser drive command to the laser drive unit 22 to drive the laser oscillator 11. , Generate a laser beam. In parallel with this, a mirror scanning command is sent to the mirror scanning unit 23, and the oscillating mirror 12 oscillates, so that the laser beam generated from the laser oscillator 11 scans the object 30.

The laser beam diffused and reflected on the object 30 forms an image on the light receiving element 14 by the optical system 13 according to the reflection position on the object. The light receiving element includes
Split-type CCD (Charge Coupled)
Device), PSD (Po) of non-division type / integration type element
position Sensitive Detecto
r) and the like are used.

Here, it is assumed that a one-dimensional CCD array of a laser sensor is used as the light receiving element 14. Light striking the light-receiving surface of the light-receiving element 14 (image of reflected light)
Are converted into photoelectrons and stored in the cell. The charges accumulated in the cells are sequentially output from the first end every predetermined period in accordance with the CCD scanning signal from the signal detection unit 24, and are subjected to processing such as AD conversion via the signal detection unit 24 and the input / output device 28 to be updated. Data is stored in the memory 25.

The scanning cycle of the CCD is set sufficiently shorter than the scanning cycle of the swing mirror 12 (for example, several hundredths).
The change of the swing angle of the swing mirror 12 and the CCD
The transition of the element output state can be grasped at any time. C
CD element output status is the maximum output cell position (cell number)
The cell position where reflected light hits is detected.
From this position, the position of the object 30 is calculated from the sensor.

FIG. 2 is an explanatory diagram for obtaining the coordinate position (X, Y) of the object 30 from the sensor based on the position xa detected by the light receiving element 14. Assuming that the sensor origin (0, 0) is on the line connecting the center of the optical system and the center of the light receiving element 14,
This line is the Y axis, and the axis orthogonal to this Y axis is the X axis. The distance from the origin to the center of the optical system is L1, the distance from the center of the optical system to the center point of the light receiving element L2 is from the sensor origin. The distance in the X-axis direction to the swing center of the swing mirror 14 is D, the Y-axis distance from the sensor origin to the swing center of the swing mirror is L0, and the Y-axis of the laser beam reflected by the swing mirror 12 is the Y-axis. The angle with respect to the direction is θ, and the light receiving position on the light receiving element 14 is xa. The coordinate position (X, Y) at which the laser beam hits the object and is reflected can be obtained by performing the following formulas.

X = xa · [(L1−L0) · tan θ + D] / (xa + L2 · tan θ) (1) Y = [L1 · xa + L2 · (L0 · tan θ−D)] / (xa + L2 · tan θ) (2) The CPU 21 of the control unit 20 activates the position calculation program stored in the memory 25 according to a command from the robot controller, and calculates the above formulas (1) and (2) at a predetermined cycle. The process corresponding to is executed. The calculation result is transferred to the robot controller. The data transferred to the robot controller is used for calculation of the three-dimensional position of the reflection position together with the position data of the robot, as will be described later.

In the present embodiment, the sensing according to the above principle is carried out only when the welding of the first layer is executed. When performing the welding of the first layer, the position of the welding line is calculated from the data collected by the laser sensor and the data of the robot position at the time when the data was obtained. Then, using the result, the position correction amount of the robot (corresponding to the deviation amount between the planned welding line position and the detected welding line position) is obtained,
The robot position is corrected as in the case of performing normal tracking control.

On the other hand, when the welding of the first layer is executed, position correction data for tracking control at the time of welding of the second and subsequent layers is created and written in the buffer memory area in the system. The processing contents including creation / writing of position correction data used in this embodiment will be described later.

FIG. 3 is a sketch diagram for explaining the arrangement at the time of executing the multi-pass welding by taking one straight fillet multi-pass welding section as an example. In the figure, W1 and W2 are workpieces to be welded, and here it is assumed that multi-layer welding is performed along the corner lines indicated by reference numeral 4. Reference numeral 5 represents a weld layer that has already been formed. Corner line is from position P to P '
It extends to and forms one straight section.

Welding is performed by a welding torch 2 attached to the tip of the robot arm designated by the reference numeral 1 by omitting most of the robot main body through an appropriate attachment mechanism. Reference numeral 3 represents the tool tip point of the robot set at the welding torch tip position. Further, reference numeral 5 represents the weld bead of the first layer which has already been formed.

A laser sensor LS is mounted on the robot arm tip 1 alongside the welding torch 2, and the laser beam LB is deflected and scanned so as to straddle the corner line 4. Reference numeral 6 represents a light spot locus drawn by the laser beam LB.

As shown by the arrow, the welding robot repeats a predetermined number of movements of moving the welding torch 2 along the corner line 4 in the direction of the arrow to the corner end position P'with the welding torch 2 ignited. Multi-layer welding is applied to the entire corner. However, after reaching the position P ′, there may be a case where the welding of one or more other welding sections (not shown) is performed and then the process returns to the position P and the welding of the next layer is performed. The number of path movements between PP 'depends on the number of weld layers to be formed.

Next, FIG. 4 is a block diagram showing an essential part of the entire system including the robot controller used when performing the multi-layer welding in the above arrangement. To explain this, a robot controller 40 has a central processing unit (hereinafter referred to as CPU) 41. C
The PU 41 includes a memory 42 including a ROM, a memory 43 including a RAM, a non-volatile memory 44, and a liquid crystal display unit 45.
A teaching operation panel 46 including a robot axis control section 47 connected to the robot body 1 via a servo circuit 48, a control section 20 of the laser sensor LS (see FIG. 1), and a welding power source section 2 '.
The general-purpose interfaces 49 connected to each are connected to each other via the bus line BL.

The ROM 42 stores various programs for the CPU 41 to control the robot body 1, the laser sensor controller 20, the welding power source 2'and the robot controller 40 itself. The RAM 43 is a memory used for temporary storage of data and calculation. Non-volatile memory 4
Various parameter setting values and robot operation programs are input / stored in 4. In the present embodiment, the buffer memory areas A, B, C, etc. used to store the position correction data in the processing described later are preset in the non-volatile memory 44. Data of the robot movement amount d, which is an index for determining the writing / reading timing of the correction data, is also stored in advance in the non-volatile memory 44.

As the teaching path designated by the operation program, it is assumed that points P and P'in FIG. 3 are taught by the playback method as the motion form "straight line". Shapes of workpieces W1 and W2 during work execution,
If the dimensions and the positioning state match the work used for the teaching by the playback method, the teaching path substantially matches the corner line 4 in each work, but if not (for example, the positioning is incorrect) , The teaching path has a considerable deviation from the corner line 4 in each work.

In the following, on the premise of the above matters, data to be used for robot position correction at the time of welding the second layer and thereafter at the time of welding the first layer and to the buffer memory area in the non-volatile memory 44. The processing procedure in this embodiment including the writing of will be described with reference to FIGS.

FIG. 5 is a flow chart outlining the processing contents related to writing to the buffer memory area A.
First, the laser sensor 3 is activated and collection of detection data is started (step S1). The collected data is analyzed at an appropriate cycle set in the laser sensor, and sensor data which is a source for calculating the position of the welding line 4 is created and stored (step S2).

Then, it is confirmed that the welding of the first layer is not completed (step S3), and the movement amount of the robot is further checked (step S4). Unless the amount of movement of the robot exceeds the set value d, the process returns to step S2 and the sensor data is repeatedly created and stored.

If it is determined in step S4 that the movement amount of the robot exceeds the set value d, the positional deviation amount of the welding line (specified by the program) is calculated based on the data of the robot position corresponding to the sensor data created during that time. Position correction data that represents the amount of deviation from the current position)
The data is written in the buffer memory area A in the nonvolatile memory 14 (step S5).

The steps S2 to S5 will be described below.
It is repeated until the welding of the first layer is completed. Since the laser sensor 3 does not perform the welding during the welding of the second and subsequent layers, if the welding of the first layer is completed, step S
From 3 to step S6, the laser sensor 3 is deenergized and the process ends.

Next, the flowchart of FIG. 6 is the same as that of FIG.
7 is an outline of a process executed in parallel with the process described in the flowchart of FIG. First, in step T1, one point of program data is read, and it is judged whether or not it is related to the movement of the multi-layer welding execution section (step T2). Further, if it relates to the movement of the multi-layer welding execution section, it is judged whether or not it relates to the welding of the first layer (step T3).

When YES is determined in both steps T2 and T3, the movement amount of the robot at that time is checked (step T4). When it is determined that the movement amount of the robot exceeds the set value d, the position correction data written in the buffer memory area A in step S5 in the flowchart of FIG. 4 is read (step T5), and the reading is performed. The movement target position of the robot is corrected using the corrected position correction data (step T
6).

If it is determined in step T4 that the amount of movement of the robot does not exceed the set value d, the process directly proceeds to step T6 and the position correction data read from the buffer memory area A last time is used. Correct the movement target position of the robot.

As a result, the preparation for the tracking movement at the time of welding the first layer is completed. Next, the processing related to the storage of the correction data for the second and subsequent layers is executed (step T
7). The processing content will be described later.

Then, from the movement target position obtained in step T6, the amount of movement of each axis is obtained by the calculation of inverse transformation, and the distribution pulse is calculated and sent to the servo control section of each axis (step T11). In the following step T12, it is judged whether or not the program execution for one point can be ended. If the determination in step T12 is YES, the end of the program itself is further confirmed (step T13). If not completed, the process returns to step T1 and the above processing cycle is executed again.

When the welding of the first layer is completed, step T3
Then, a negative determination is made, and the control shifts to tracking movement control using the position correction data. First, in step T8, it is determined whether the movement amount of the robot exceeds the set value d.
If yes, the position correction data created in the first layer is read from the buffer memory area B (step T
9). Since this position correction data is created and stored in step T7, it is expressed by a parameter value that defines a linear expression that approximately represents the welding line extending position in the straight line section, as will be described later. There is. Therefore, the robot position correction amount is obtained through a calculation in which the movement amount of the robot from the movement start point is substituted into a linear expression defined by its parameters (coefficients and constant terms).

The data of the calculated position correction amount is used to correct the movement target position of the robot in the subsequent step T10. If the determination in step T8 is no, the process directly proceeds to step T10, and the movement target position of the robot is corrected using the position correction amount data calculated last time.

As a result, the preparation for the tracking movement at the time of welding the second layer is completed. Then, from the movement target position corrected in step T10, the amount of movement of each axis is obtained by the calculation of inverse transformation, the distribution pulse is calculated, and the pulse is sent to the servo control section of each axis (step T11).

Thereafter, the same processing is repeated, and when it is judged that all the instructions of the program to be executed are completed (step T13), all the processing is ended. In addition,
Program commands other than movement of the multi-layer execution section (for example,
When each axis movement for approaching the work from the air cut point) is read, the process goes directly from step T2 to step T11 to calculate the distribution pulse and send it to the servo control section of each axis. Be done). Therefore, in this case, tracking control is not performed. Then, when the program data of the multi-pass welding section is read again, a yes determination is made in step T2, the tracking control is restarted, and the welding of the first layer or the second and subsequent layers is performed based on the processing described above. Is executed.

Next, referring to the flowchart of FIG.
An outline of the process shown in step T7 in the flowchart of FIG. 6 will be described.

First, it is confirmed whether or not the robot has moved a predetermined distance d (or time) since the previous data acquisition (step V1). If it has not elapsed, the subsequent processing is not performed. If so, step V2
Proceed to and confirm that the section is taught in the linear movement format. If not, the subsequent processing is not performed.

If the section is taught in the linear movement format, then in step V3, the position correction amount of the robot (used for tracking) is stored in the buffer memory together with the data of the movement distance from the starting point of the section. Store in area C. Then. In step V4, the index i indicating the number of data in the buffer memory area (the number of data for one welding point position is a unit) is incremented.

Then, in step V5, it is determined whether or not it is the final point of one block (the detection point of the final point P'in the straight line section in FIG. 3). If it is not the final point of the straight line section, it is further confirmed that the data size limit value n is equal to or less than the data size limit value n determined according to the capacity of the buffer memory area (step V6), and the process proceeds to step V7.

In step V7, the data in the buffer memory area is used to calculate the data required for the least squares method. Next, recalculation is performed using the above calculation result and the data stored in the data storage buffer memory area B ′, and the result is stored in the data storage buffer B ′. That is, the data for the least-squares calculation that has been calculated up to that point is updated by using the latest welding line position data.

Then, the buffer memory area for calculation is cleared, the processing is terminated, and the process returns to step T8 in the flowchart of FIG. When the sensing point by the laser sensor reaches the final point (point P ′ in FIG. 3) of the straight line section,
A yes decision is made in step V5, and step V10
Go to. In step V10, data required for the least-squares method is calculated using the data (limit value n or less) in the buffer memory area. Next, recalculation is performed using the above calculation result and the data stored in the data storage buffer memory area B ′ to determine the data for the least squares calculation. Then, the calculation by the least square method is executed to determine the coefficient and the constant term of the linear expression that approximately expresses the extending position (position as a straight line) of the welding line in the straight line section.

The calculation result is stored in the buffer memory area B for storing the position correction data for welding the second and subsequent layers (step V12). This completes the storage of the position correction data for one straight line section.

When the linear approximation is performed by the least squares method and the approximation formula is Y = aX + b .. (1), the coefficients a and b are as follows.

A = {ΣY · ΣX ² −ΣX · ΣXY} / {N · ΣX ² − (ΣX) ² } ·· (2) b = {N · ΣXY−ΣX · ΣY} / {N · ΣX ² − (ΣX) ² } ... (3) Therefore, as the data (see step V7) necessary for performing the calculation by the least square method, the data of ΣX, ΣY, ΣXY, ΣX ² , N (4) You should prepare.

In the present embodiment, the distance from the starting point (P in FIG. 3) of the straight line section is X, and the x and y components of the correction amount.
With each z component as Y, three sets of data are calculated and stored for every n data. Then, from the last stored data of (4) above, using the above equations (2) and (3), the coefficient of the linear approximation equation and the value of the constant term are calculated, and the result is the welding of the second and subsequent layers. It is stored as data for robot position correction at the time.

It is obvious that the laser sensor used when applying the present invention may be of a CCD camera type, and it is sufficient to make a slight modification to the above embodiment. .

[0055]

According to the present invention, with respect to a straight line section for carrying out a multi-layer deposit, robot position correction data for welding the second and subsequent layers in the form of parameters that define a linear expression expressing the straight line section. Is stored, a significant saving of the used memory capacity is achieved as compared with the conventional method.

Therefore, even when a system having no large buffer memory capacity is used, it is possible to carry out multi-layer welding by a tracking control method using a laser sensor for a long section.

[Brief description of drawings]

FIG. 1 is a diagram illustrating a schematic configuration of a laser sensor.

FIG. 2 is an explanatory diagram for obtaining a coordinate position (X, Y) of an object from a sensor based on a position detected by a light receiving element of a laser sensor.

[FIG. 3] Taking one straight fillet multi-layer welding section as an example,
It is a sketch drawing explaining arrangement at the time of execution of multilayer welding in this example.

FIG. 4 is a block diagram showing a main part of a configuration of an entire system including a robot controller used when performing multi-layer welding in the arrangement shown in FIG.

FIG. 5 is a flowchart outlining processing contents related to writing to the buffer memory area A in the embodiment.

FIG. 6 is a flowchart showing an outline of processing executed in parallel with the processing described in the flowchart of FIG. 5 in the embodiment.

FIG. 7 is a flowchart showing an outline of processing executed in step T7 in the flowchart of FIG. 6 in the embodiment.

[Explanation of symbols]

 1 Welding robot body (arm tip) 2 Welding torch 2'Welding power source 3 Tool point 4 Corner line 5 Welding layer 6 Light spot locus 10 Laser sensor detector 11 Laser oscillator 12 Swing mirror 13 Lens 14 Light receiving Element (CCD array or PSD) 20 Laser sensor control unit 21 CPU (Laser sensor) 22 Laser drive unit 23 Mirror scanning unit 24 Signal detection unit 25 Memory (Laser sensor) 28 Input / output device 29 Bus (Laser sensor) 30 Object Surface 40 Robot controller 41 CPU (robot controller) 42 Memory (ROM) 43 Memory (RAM) 44 Non-volatile memory 45 Liquid crystal display 46 Teaching operation panel 47 Robot axis controller 48 Servo circuit 49 General-purpose interface BL bus (robot control) Equipment) LB Ray Beam LS laser sensor P, P 'teaching point position W1, W2 workpiece

Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display area G05D 3/12 T

Claims (3)

[Claims] 1. A welding robot including a robot controller, a robot controlled by the robot controller, a welding torch supported by the robot, and a laser sensor for detecting the position of a welding line on the workpiece to be welded. A method for storing robot position correction data in multi-layer welding control, which allows a robot to move along a welding line using a system, wherein the robot is moved to the section when performing welding on the first layer of a straight section. A step of creating position correction data for moving the robot in the linear section on and after the second layer based on the sensor data obtained by using the laser sensor to move following the welding line, Writing the created position correction data into a buffer memory area prepared in the welding system, A method for storing the robot position correction data in the multi-layer welding control, wherein the position correction data written in is data of a parameter that defines a linear expression that approximately expresses the welding line extending position of the straight line section. . 2. The step of creating position correction data for moving the robot in the straight line section of the second layer and thereafter,
Determining a parameter that defines a linear expression that approximately expresses the welding line extension position of the straight line section based on the data of the welding line position relating to at least two points detected during the execution of the first layer welding. A method for storing robot position correction data in multilayer welding control according to claim 1, including: 3. A step of creating position correction data for moving the robot in the straight line section on the second and subsequent layers,
A parameter that defines a linear expression that approximately expresses the welding line extension position of the straight line section by the least squares method using the data of the welding line positions regarding at least three points detected while executing the welding of the first layer. Including the step of determining
A method for storing robot position correction data in the multi-layer welding control according to claim 1.