JPH09262671A - Arc depression correcting method of automatic pipe circumference welding equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the arc performance of an automatic pipe circumference welding equipment to perform the circumferential automatic welding of a steel pipe while performing the high-speed weaving. SOLUTION: When the whole circumference of end faces of two pipes which are fixed in a butted condition is welded by the pulse MAG welding method, the two-dimensional weaving of a torch is performed in the right-to-left direction orthogonal to the tangential line of the pipe circumference and in the vertical direction along a grooves of the pipes. The current pulse width 123 is obtained based on the measured pulse welding current 13, and by detecting either of the case where the current pulse width is larger than the predetermined upper limit value, or the case where the ratio of the current pulse width to the commanded pulse width is larger than the predetermined upper limit value, the dullness of the current pulse waveform is detected (133), and when the dullness is detected, the torch is controlled to be lowered in the direction of the pipe surface.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pipe circumference automatic welding apparatus, and more particularly, to a pipe circumference automatic welding apparatus for measuring and analyzing a pulse welding current waveform and a pulse welding voltage waveform and performing arc tracing-related correction. The present invention relates to an arc dullness correction method.

[0002]

2. Description of the Related Art At the time of laying a gas pipeline, the end faces of two pipes fixed in abutting manner are connected by full circumference welding. When this welding is performed by automatic welding, a belt-like guide rail is wound around a pipe, and a welding head is run on the guide rail to perform welding by a MAG (metal active gas) method or the like. The arc copying is to monitor the current and voltage during welding and position the torch relative to the work.

In conventional butt-arc welding of pipes, driving such as weaving is generally performed by a pulse motor for a relatively low speed operation and by a servomotor for a high speed operation. In order to keep the torch at the optimum position for the workpiece during welding, the current and voltage during welding are monitored, and feedback control of the torch position is performed so that those values become appropriate. Is being done. A current / voltage measurement method for this purpose is to smooth a pulse welding current / voltage signal through a low-pass filter and measure it as a digital current / voltage signal by an A / D converter. However, the cutoff frequency of the low-pass filter was set much lower than the repetition frequency of the pulse welding current / voltage (for example,
(Cutoff frequency 10 Hz for pulse frequency 100 to 400 Hz).

[0004] Conventional techniques relating to pipe circumferential welding are disclosed in Japanese Patent Application Laid-Open Nos. 63-183776 and 4-2.
There is one described in Japanese Patent No. 84973. JP 63
The technique described in JP-183776A relates to an arc-rotation type all-position welding apparatus. In this method, the groove shape is narrow and high precision machining is required, and it is difficult to apply the method to a work having a wide groove width. JP-A-4-284973
The technique described in Japanese Patent Application Laid-Open No. H11-209,897 relates to an automatic arc welding method in which the welding of a groove is controlled in real time, but does not solve the problem of measuring individual pulse waveforms of pulse welding current / voltage.

[0005]

In order to perform optimum welding in pipe circumferential welding, high-speed weaving, etc., arc output voltage control, pulse accent (pulse width and pulse current waveform collapse) prevention control, etc. are performed. Will be required.

In conventional pulse welding current / voltage measurement,
It was difficult to measure the current / voltage waveform with high speed and high accuracy, and the measurement was at a response time of at most 10 times the repetition cycle of the pulse welding current / voltage. Therefore, for example, even if a pulse accent occurs and an attempt is made to correct it, there is a drawback in that the timing of current, voltage, weaving command, etc. for correction is delayed.

The present invention provides an arc blunting correction method which makes it possible to improve the arc performance of a pipe circumference automatic welding apparatus for automatically welding the circumference of a steel pipe while performing high-speed weaving.

[0008]

The arc blunting correction method for a pipe circumference automatic welding apparatus according to the present invention is a method for welding the entire circumferences of the end faces of two pipes fixed by butting by pulse MAG welding. Is a blunting correction method for an automatic welding device that performs two-dimensional weaving of a torch in the left-right direction and the up-down direction orthogonal to the tangent of the pipe circumference along the groove of, and the current pulse width is based on the measured pulse welding current. If the current pulse width is larger than a predetermined upper limit value, and by detecting any one of the case where the ratio of the current pulse width to the command pulse width is larger than a predetermined upper limit value, The present invention is characterized in that bluntness of the current pulse waveform is detected, and when the bluntness is detected, the torch is controlled so as to be lowered toward the pipe tube surface.

According to this structure, since the arc blunting is detected by monitoring the pulse width itself of the pulse welding current, the occurrence of the arc blunting can be detected more reliably and with good responsiveness.

In the above method, preferably, the torch is lowered at a predetermined speed until the current pulse width reaches an allowable value.

Further, preferably, based on an average current of the pulse welding current, a vertical arc tracing correction is performed in which the torch is moved in the vertical direction so that the average current matches a command value, and the arc blunting correction is being performed. Invalidates the vertical arc scanning correction. As a result, the arc blunting can be reliably eliminated by prioritizing the arc blunting correction with respect to the left and right arc copying correction.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below in detail with reference to the drawings. Here, an example in which the present invention is applied to high-frequency pulse MAG welding will be described. The pulse MAG welding is described, for example, in “Welding Technique”, February 1989, pp. 67-76.

A schematic configuration of the welding apparatus of the present invention will be described with reference to FIG. The personal computer 1 is for creating CAD data (design values such as the outer diameter, thickness, material, and groove shape of a pipe) for automatic welding control, and the created data is a flexible disk (floppy disk). Disc 1a. The control device 2 performs a substantial control of the automatic welding (control of a power supply and a torch position, etc.).
And reads the CAD data from the flexible disk 1a in which the CAD data is stored. Of course, data transfer may be performed by well-known data communication without using a flexible disk. The control device 2 sets actual welding conditions using the read data, and sets various welding parameters in accordance with the set conditions by using data previously recorded in a logic table described later. For details of such logic table and creation of welding data using this,
It is disclosed in Japanese Patent Application No. 7-173921 previously proposed by the present applicant.

Further, in the control device 2, the set welding parameters are converted into NC language (numerical control language) used for actual welding control, and the converted language is internally used as control data in the form of a control data table. To the memory of. Using the control data recorded in this memory, the drive of the power supply device 3 and the welding head 4 is controlled,
Measurement and measurement of pulse voltage waveform and pulse current waveform
It is configured to perform arc copying related correction using the analysis value. The main control items of the control device 2 are the welding voltage and current of the welding head 4, the weaving of the welding torch 8 mounted on the head 4 with respect to the pipe groove 93, the moving speed of the welding head 4, and the like. The head 4 is mounted on a guide rail 12 wound on the outer periphery of the pipe so as to be movable in the circumferential direction. The head 4 is provided with a wire supply unit 6 for supplying a welding wire to the welding torch 8.

In FIG. 2, the control device 2, the power source 3, the head 4,
The electrical connection between the pipes 5 is shown. As can be seen from the figure, the power source 3 applies a pulse welding voltage between the wire 7 of the head 4 and the pipe 5 via the power transmission cable 9. Thereby, an arc 91 (FIG. 12) is generated between the wire 7 and the surface of the pipe 5.
The welding wire 7 is fed at a constant speed and melted by an arc 91 to become a weld metal 92 (FIG. 12).
It solidifies in (FIG. 12) and the base material 5 is joined.

As shown in FIG. 3, the position of the welding head 4 on the outer circumference of the pipe (expressed in time and minutes) X axis, the weaving groove width direction of the torch 8 (pipe axis direction) Y axis, and the welding height. A servo mechanism is provided for each of the Z axis of the direction (pipe radial direction) and the torch swivel axis A, and a total of four axes are controlled by the controller 2. The two-dimensional weaving refers to an operation of moving the torch 8 along the groove in the Y-axis direction and the Z-axis direction orthogonal to the welding progress direction (X-axis).

Components attached to the welding head 4 in the present embodiment will be described with reference to FIG. Welding head 4
Include encoders 42x, 42y, and 42z (generally designated as 42) for detecting a change in the position of the head for each of the X, Y, and Z axes, and a mechanical origin (not shown) provided for each of the axes. ) Detecting the origin sensors 41x, 41y, 41z
(Generally indicated by reference numeral 41). Each axis encoder 42 is a device that generates a number of pulses corresponding to the head movement amount in the axial direction, and the movement amount can be obtained by counting the number of pulses. However, the X-axis encoder 42x is configured by an absolute value encoder so that the absolute position can be maintained even during the power-off period of the motor. Y
Regarding the axis and Z axis, even if the current position information is lost, the movement stroke is small, and the trouble and time for detecting and setting a new mechanical origin are not a problem, so that a simpler incremental (relative positioning) encoder is used. Is adopted. The origin sensor 41 of each axis is a sensor for detecting a mechanical origin indicating a mechanical origin of each axis, and for example, an optical sensor can be used. X
The mechanical origin of the shaft is an index provided at one predetermined position on the circumference of the guide rail 12. In this example, the mechanical origin is a light blocking member (not shown) provided at the top of the guide rail 12. The mechanical origin of the Y axis is a light blocking member (not shown) provided at a retreat position of the torch in the Y axis direction. Similarly,
The mechanical origin of the Z-axis is a light blocking member (not shown) provided at the uppermost position in the Z-axis direction of the torch. Of course, the sensor is not limited to an optical sensor, and for example, a sensor such as a proximity sensor can be used. The target position of the torch can be designated with reference to the origin of each of these axes. The detection and setting of these origins are performed as an initial work of the welding work.

The internal structure of the controller 2 will be described with reference to FIG.

The control device 2 has a control unit 223, a storage unit (memory) 221, a hard disk 222, a welding head control unit 230, a CRT / keyboard 225, and a pulse waveform measuring unit 211.

The computer 1 may be built in as shown by a broken line in the control device 2 of FIG. The control unit 223 can be configured by a microprocessor, and the storage unit 221 provides a storage area and a work area for the control program and various data. The hard disk 222 is a large-capacity storage device that stores control programs, various data, and the like in a nonvolatile manner. Specifically, the hard disk 222 stores a predetermined welding condition and welding parameters in a logic table (described later).
Are stored in a table format as. The control unit 223 reads the CAD data stored in the storage unit 221, determines the welding conditions, and reads the welding parameters that match the welding conditions from the logic table. Further, the read welding parameters are converted into a language for controlling the welding head 4 and the power supply device 3 and stored in the storage unit 221 as control data in the form of a control data table (not shown). This control data is transmitted to a welding head controller 230 described later.
Sent to The CRT / keyboard 225 includes the control unit 22
A display unit and an input unit for providing a user interface between the operator 3 and an operator; Welding head controller 23
0 is a transmission / reception unit 2 for transmitting / receiving a signal from the control unit 223 and transmitting / receiving a signal to / from the remote operation device 234.
31, a head drive unit 2 for outputting a drive command of a servomotor for controlling the operations of the welding head 4 and the torch 8
33, and a power control unit 232 that controls the power supply device 3. The pulse waveform measuring unit 211 is a hardware circuit for measuring a pulse welding voltage and a pulse welding current from the power supply device 3. An example of the configuration of this hardware circuit will be described later with reference to FIG.

FIG. 6 shows a flowchart of processing of the pipe circumference automatic welding apparatus in the present embodiment.

In FIG. 6, first, the control device 2 reads the pipe diameter, plate thickness, material and groove shape from the CAD data (61). Welding conditions (laminating plan, stacking plan,
Welding parameters corresponding to the current, voltage, rotation speed, etc. for each layer and address are determined (62, 63). The determined welding parameters are converted into NC language (64). The NC data obtained as a result is stored in a control table (not shown) as control data. Further, using this NC data, the power supply unit and the welding head are operated (65
~ 68). During this operation, the position interval and the like of the head are detected using pulse welding current / voltage detection means, and it is determined whether or not the operation state is set to the optimum state. The trajectory data of the torch is corrected, welding conditions and the like due to changes in the width of the root gap are corrected, and welding parameters are reset according to the data in the logic table. After the resetting, the operation is continued, the state of the arc is detected and compared with the design value, and the automatic welding is executed as needed while controlling the arc voltage and the like (69 to 72).

That is, as welding condition parameters stored in advance in the logic table 62, for each welding layer, welding current, welding voltage, head moving speed, torch rotation speed, arc start and end positions for each position. Etc., and the torch trajectory correction 6 based on the result of the groove shape measurement 66 with the wire for these welding condition parameters.
7 and correction of welding conditions by changing root gap width 68
I do. Thereafter, automatic operation is started (69), and arc sensing (welding line scanning), current / voltage sampling, and torch position (up / down, left / right) control are performed (70). In addition, pulse accent prevention control and arc voltage control are performed (71). The details of the arc copying related correction performed based on the arc sense will be described later. Various parameters such as current welding current, voltage, each axis position, and various correction amounts of automatic welding can be monitored as a current state (status) by real-time display on a CRT (7).
2).

FIG. 7 shows an example of the logic table.
This is an example of a file that defines the setting of the weaving stop time. There are various other files, but only one example is shown here. In this logic table, “welding posture”, “interlayer”,
About four conditions of "rotation speed" and "weave width",
Equal: “=”, not equal: “! =”, Greater:
“>”, Small: “<”, above: “> =”, below: “<
The data is stored so as to meet each condition of “=”.

For example, suppose the following conditions are given. That is, the welding conditions are upward, the layers are finished, the rotation speed is 100, and the weave width is 10.0. The table contents matching this condition are searched, and the item No. of the table is searched. Select 5. As a result, data of a rotation speed of 120 or less, a weave width of 11.0 or less, left stop 0.15, middle stop 0.00, and right stop 0.15 are selected.
When these welding parameters are selected, the data is then passed to the actual NC execution control program.

FIG. 8 shows the correspondence between CAD program items and NC control program items. At the time of initial value setting, bore diameter, groove information, servo parameters, arc copying correction parameters, and rework welding parameters are set as CAD information. These parameters are converted into the number of guide rail rotation pulses (the number of encoder pulses) for the aperture, and the groove information is converted into touch sense information for NC control. Other items are also converted for NC control according to the correspondence shown in FIG. This system is
Provisional operations are executed based on the NC control data obtained in this manner, and for each operation, those requiring correction are corrected based on the detection result, and the process shifts to the actual welding operation.

With this configuration, a person can
When D data is created and the welding head is set on the pipe, the system automatically determines the welding conditions and executes the welding automatically, which eliminates the need for a person to teach the welding conditions and the like. Variation is reduced. Also,
Since most conditions are stored in the form of a logic table, the I
Compared to the case of using F to THEN, the conditions can be changed and the reason for selection can be easily corrected.

An example of two-dimensional weaving of the torch performed by the present welding apparatus will be described with reference to FIGS. 9 and 10. FIG.
FIG. 10 shows changes in the Y-axis position and the Z-axis position with the horizontal axis as time, and FIG. 10 shows the movement path of the weaving operation along the X-axis direction. In the weaving operation, as shown in the figure, the torch is reciprocated with a predetermined width and period in the Y-axis and Z-axis directions to make the trajectory of the torch wavy. In this example, as described above, weaving is temporarily stopped at the left end (L), the center (C), and the right end (R). The speed at the time of oblique movement on the left and right in FIG. 10 and the speed at the time of horizontal movement can be separately specified. The speed is specified by a speed (tangential speed) along the movement trajectory of the wire tip.

FIG. 11 shows the weaving command locus and the actual torch movement locus (servo locus) in comparison. The horizontal axis of the figure shows time, and the vertical axis shows the position in the Y-axis direction. As can be seen from the figure, due to the delay in the response of the servo mechanism to the movement command of the torch, the servo track follows the command track later. In the example of the figure, the weaving width (moving range in the Y-axis direction) is set to the right end, right, center, left,
Sampling of various data described later is performed in the five sections at the left end. The determination of each section is performed based on the servo trajectory indicating the actual mechanical torch position according to the output of the Y-axis encoder. A dead zone may be provided at the boundary of each section to reduce errors.

In the arc copying-related correction described later, the following data is obtained according to the weaving position.

Left / Middle / Right smoothed current value Left / Middle / Right smoothed voltage value Left end / Left / Middle / Right / Right end arc short time Left / Middle / Right current pulse period Arc short means welding wire (electrode) ) And the base material or the molten metal are too short, resulting in a short circuit between them.

FIG. 12 shows a hardware configuration example of the measuring circuit 211.

The welding power supply device 3 supplies a large pulse current to the welding torch 8. The pulse welding current / voltage measuring circuit 211 receives the pulse welding current signal 1 from the welding power source device 3.
3 and pulse welding voltage signal 32 are obtained and input to filters 14 and 33, respectively. It should be noted that “voltage signal” and “current signal” referred to in this specification mean that the signal represents a welding voltage and a signal representing a welding current, respectively, and the signal is represented by a voltage level and a current level. It does not mean that. Usually, these voltage and current signals are represented by voltage levels.

FIG. 13 shows an example (filter input) of the pulse welding current / voltage signal. In the figure, (a) shows the voltage signal 32 and (b) shows the current signal 13. Voltage signal 32
Is a pulse waveform of a predetermined cycle, and a high-frequency noise component 15 is superimposed. A base voltage portion 51 of the pulse waveform includes a voltage drop waveform portion 50 indicating occurrence of an arc short. The current signal 13 corresponding to the voltage signal 32 also includes the high-frequency noise component 15 like the voltage signal 32. A high-frequency noise component 15 is superimposed on the pulse welding current signal 13, and a waveform portion 52 corresponding to the waveform portion 50 of the voltage signal 32 is shown. The filters 14 and 33 have a cut-off frequency (for example, 4 KHz) higher than the pulse frequency of the pulse power supply as described above, remove the high-frequency noise component 15, and perform the pulse welding current signal 16 and the pulse welding voltage signal 34. Is output. Thereby, it is possible to prevent a malfunction of the electrical characteristic measurement due to the high frequency noise. Further, by using a filter having such a cutoff frequency, the pulse waveforms of the current signal 13 and the voltage signal 32 can be maintained, and the electrical characteristics can be measured in pulse units by a method described later.

FIG. 14 shows an example of the pulse welding current / voltage signal output by the filters 14 and 33.
3A shows a voltage signal, and FIG. 3B shows a current signal. The pulse welding current signal 16 and the pulse welding voltage signal 34 are obtained by removing the respective high frequency noise components 15 from the voltage signal 32 and the current signal 13 by the filters 14 and 33. In the figure, I1 is the peak value of the current signal, E2 is the base voltage of the voltage signal, T1 is the pulse width of the current signal (and voltage signal), T2 is the base width of the current signal (and voltage signal), and T3 is the voltage signal (And current signal) arc short width (arc short time), T4
Is the repetition period of the current signal (and the voltage signal), H1 is the level of the threshold level signal 22 for measuring the pulse width and the base width with respect to the current signal 16, and H2 is the voltage signal 34
Is the level of the threshold level signal 38 for measuring the arc short width with respect to.

The arc short width is obtained from the voltage waveform because the arc short waveform tends to appear clearly in the voltage signal. The reason why the pulse width and the like are determined from the current waveform is that the difference between the base level and the peak level is larger and clearer in the voltage waveform.

The peak hold circuit 17 in the measuring circuit 211 of FIG. 12 detects and holds the peak value of the pulse welding current signal 16 output from the filter 14 at each time interval determined by the timing signal 20. This peak value is
The digital signal 19 is converted by the A / D conversion circuit 18 according to the timing signal 20. It should be noted that the timing signal 20 is generated by the control device 2 by an interrupt circuit 24 described later.
The control device 2 outputs in response to the interruption to.

The comparator 21 compares the threshold level signal 22 given from the control device 2 with the pulse welding current signal 16 (see FIG. 14B), and when the pulse welding current signal 16 exceeds the threshold level signal 22. The output signal 23 is set to the low level, and when the pulse welding current signal 16 becomes smaller than the threshold level signal 22, the output signal 23 is set to the high level. Interrupt circuit 24
Outputs an interrupt signal to the control device 2 at the falling edge of the output signal 23. Thereby, the control device 2 recognizes the rise of the pulse waveform of the welding current signal 13 and generates the timing signal 20 described above. As will be described later, this can be used as timing for processing such as welding line scanning control during arcing at the time of high-speed weaving. The output signal 23 is input to the counter 25, and the inverted signal of the output signal 23 by the inverting circuit 27 is input to the counter 30. The counter 25 indicates that the output signal 23 is at a high level (that is, the current signal 16 is at the threshold signal 2 level).
2 (lower level period), enabled and the controller 2
Is counted. As a result, the time T2 of the current signal 16 is obtained and output as the base width 26. On the other hand, the counter 25 is enabled while the output signal 23 is at a low level (that is, while the current signal 16 is at a higher level than the threshold signal 22), and counts the clock signal 44 from the control device 2. As a result, the time T1 of the current signal 16 is obtained, and the pulse width 31
Is output as Based on the pulse width 31, a pulse accent can be prevented. The pulse period T4 is obtained by adding time T1 and time T2. Alternatively, the pulse period can be obtained from the command value. As will be described later, the pulse period, along with the arc short width,
It can be used for welding line copying control during arcing.

The comparator 37 compares the arc shorting threshold level signal 38 provided from the control device 2 with the voltage signal 34 (see FIG. 14A), and the voltage signal 34 is the arc shorting threshold level signal 38.
When the voltage becomes smaller, the output signal 39 is set to a high level, and when the voltage signal 34 exceeds the threshold level signal 38, the output signal 39 is set to a low level. Counter 41
Counts the clock signal 45 from the controller 2 while the output signal 39 is at a high level. Thereby, the voltage signal 3
4 is obtained and output as the arc short width 42. It should be noted that the threshold level signal 22 and the arc short threshold level signal 38 are transmitted to the control device 2 in accordance with changes in welding operation conditions (pipe diameter, steel type, etc.).
Can be variably controlled by. Reference numeral 46 denotes a signal between the pulse current / voltage measurement circuit 211 and the control device 2, including the various signals described above.

In order to perform highly accurate measurement, of course, the clock signals 43, 44 and 45 having a cycle sufficiently shorter than the time width of the measurement object are used.

FIG. 15 shows a measuring circuit 21 shown in FIG.
1 shows an outline of timing for each signal. Reference numerals 16 and 34 denote outputs of the filters 14 and 33, respectively.
4 is the same as that shown in FIG. The output signal 23 of the comparator 21 falls at the rising edge of the current signal 16 output from the filter 14, whereby the interrupt circuit 24 interrupts the control device 2. Controller 2 generates timing signal 20 in response.
The timing signal 20 causes the peak hold circuit 1
7 is reset, and subsequently, the peak value of the signal 16 is detected and held until the next timing signal 20 is generated. In this example, the peak value is the high level value of the pulse of signal 16 and is held until the next occurrence of timing pulse signal 20. Output signal 23 of comparator 21
Is high during the base width time T2 of the current signal 16, and this period is measured by the counter 25 as described above. The output signal 28 of the inverting circuit 27 has a waveform obtained by inverting the output signal 23, and the high-level period of
Corresponds to the pulse width time T1. This time T1 is measured by the counter 30 as described above. The output signal 39 of the comparator 37 is at a high level during the arc short time T3 of the voltage signal 34, and this time T3 is measured by the counter 41 described above.

In the illustrated example, the arc short width and the pulse width / pulse period are measured with the same low-pass filter output, but the pulse width / period is slightly lower than the cutoff frequency (for example, 1 kHz). ), Another filter may be used to reduce malfunction due to noise.

Further, although not shown, the smoothing current and the smoothing voltage can be measured by using a low pass filter having a low cutoff frequency (for example, 10 Hz) as in the conventional case. There is a difference between the actual arc output current / voltage value according to the arc output command and the current / voltage value obtained from the waveform measurement due to the characteristics of the output waveform of the welding power source. Therefore, as shown in the following equation, a value obtained by correcting data calculated by various sampling processes in consideration of a current / voltage difference obtained in advance is used as a current / voltage feedback value.

Ifb = Iad + IOFSmod Vfb = Vad + VOFSmod where Ifb is a current feedback value, Vfb is a voltage feedback value, Iad is a measured current value, Vad is a measured voltage value, IOFSmod is a current error that varies depending on the power supply mode, and VOFSmod is a power supply mode. It is a voltage error that changes depending on

Controls such as arc tracing correction and arc voltage adjustment, which will be described later, are performed using the corrected current / voltage feedback values. The current value and the voltage value in the following description are those after this correction, without special mention.

Next, various arc scanning corrections according to the present invention will be described below. Sampling of various data based on the measurement of the welding current and the welding voltage is performed in two stages of an average current / voltage value measurement process / pulse waveform measurement process and an interval timer process. The average current / voltage measurement process is a process that is started every fixed period (for example, 1 ms), and reads the current / voltage feedback value smoothed at the low cutoff frequency described above to obtain a value in which the error is added. . The pulse waveform measurement process is started for each feedback pulse cycle (measured pulse cycle), and calculates the peak current / voltage value, pulse cycle, peak pulse width, and arc short time detected based on the welding current / voltage waveform described above. read out. The interval timer process is started at regular intervals (for example, 4 ms), and averages the current, voltage, and arc short time values output from the pulse waveform measurement process. This interval timer cycle is also called a servo control cycle. In addition, during the weaving operation, various average values in the right end, right, middle, left, and left end sections are obtained according to the weaving position. If weaving is not in progress, 32
Averaging is performed on data for 8 ms.

FIG. 16 shows a procedure for obtaining the left and right arc scanning correction amounts and the vertical arc scanning correction amount based on the measurement result of the pulse welding current 13 and the signals used therefor.

<Left and right arc tracing> In the left and right arc tracing, the weaving center of the torch accurately passes through the groove center in a single pass (welding one layer in one pass) within the groove. Therefore, the current pulse cycle at both ends of the weaving (the left end and the right end) is detected, and both are controlled to be the same. If the center of the weaving is displaced from the groove, the distance between the torch and the groove wall at the left and right ends is different, and the current values of both are different. By confirming this current value, the deviation of the weaving center can be detected. But,
The smoothing current 122 is smoothed at a low cutoff frequency as described above, and the current values at the left and right ends cannot be accurately obtained from this signal. Therefore, in the present embodiment, attention is paid to the fact that the current pulse period is proportional to the current value, and the current value is estimated based on the pulse period. In this example, as the preconditions, a weaving cycle of 50 ms or more (frequency of 20 Hz or less), a weaving Y-axis speed of 5 m / min or less, and a current pulse cycle of 1 ms or less are assumed.

Specifically, in order to obtain the right and left current values with higher accuracy, first, the smoothing current values of all sampling points of the horizontal moving portion (excluding the left end and the right end) within one weaving cycle are calculated from the smoothing current 122. Averaged average current value 1
Find 35. The horizontal part is used because the arc is relatively stable in this part. On the other hand, the peak current value, pulse period, pulse width, and arc short time (123) are obtained from the current waveform of the pulse current 13. Further, an average pulse period 150 is obtained by averaging all the pulse periods of the horizontal movement portion within one weaving period. From the average pulse period 150 and the average current value 135, the pulse period −
The current value conversion coefficient 126 is obtained by the following equation.

Cycle-current conversion coefficient TAR = Iav / T
av (A / ms) Here, Iav represents an average current value and Tav represents an average period.

From the current waveform 123, according to the weaving section, the left current pulse period and the right current pulse period 127
Ask for. Since control is performed according to the overall tendency, the current pulse periods of several points including the leftmost end and the rightmost end are averaged on the left and right. The number of points to be averaged is variable depending on the weaving cycle. For example, weaving period (Hz)
Is less than 1 Hz, 32 points, less than 2 Hz is 16 points, less than 4 Hz is 8 points, less than 16 Hz is 2 points, and more than 16 Hz is 1 point.

Therefore, the left current value and the right current value 128 are obtained by the following equation.

Left current value El = Tl.TAR Right current value Er = Tr.TAR Here, Tl and Tr represent the right current period and the left current period (ms), respectively. The left and right current values 128 are displayed as the status information described above.

Using the right and left current values 128 of the previous weaving thus obtained, the position of the torch is controlled so as to shift the center position of the current weaving so that both current values are equal (129, 130). The movement for this correction amount is performed gradually in each servo control cycle over one weaving cycle. That is, instead of moving the correction amount at a time, the movement is dispersed in one weaving cycle and moved in small increments. The dispersed individual movement amount is
Since the number of servo control periods within one weaving period can be known, the total correction amount is divided by the number of servo control periods. The reason why the movement for the entire correction amount is not performed collectively at once is that if the movement in the horizontal direction is large at once, the arc is greatly affected, and the balance of the left and right arcs is caused by the left and right arc copying correction. This is because a situation occurs in which the object collapses. For example, if weaving is performed at the start of weaving, which starts moving from the center to the left, only the left is affected and the right is imbalanced.

The left and right current values are 128 depending on the welding conditions.
In consideration of non-uniformity, an offset Ao is added when evaluating the left-right current difference as described later. In addition, a dead zone is provided that does not react to a minute left-right current difference in consideration of the reliability of the current waveform. This dead zone width is
Preferably, it is changed according to the weaving width. Also, in order to prevent the arc copying correction from operating indefinitely in the state where the malfunction of the left and right arc copying correction has occurred,
Left and right correction limits LMT1, L of the left and right arc copying correction amounts
MTr (μm) is provided. The left / right current difference Ad and the current correction amount ΔYn are obtained by the following equations.

When Ad = El-Er | Ad + Ao |> = NSBlr, ΔYn = (Ad + Ao) · Glr ΔYn> 0 and | Σ (ΔY) + ΔYn |> = LMTr, ΔYn = LMTr−Σ (ΔY) ΔYn <0 and | Σ (ΔY) + ΔYn |> = LMTl, ΔYn = LMT1-Σ (ΔY) where NSBlr is the dead zone width (A), Glr is the left and right arc scanning gain (μm / A), and Σ (ΔY) ) Represents the integrated value (μm) of the correction amount up to the previous time. Left and right current offset,
Each value of the dead zone width, the left and right arc copying gain, and the left and right correction limit is a value specified in advance.

Although not related to the left-right arc tracing, the arc breakage is detected when the arc short-circuit time of the current waveform 123 exceeds a predetermined allowable time (13).
1). This arc break occurrence position (X-axis position) is recorded in the arc break occurrence position table (132).

<Upper and Lower Arc Trajectory> The upper and lower arc trace is to correct the vertical position of the torch so that the command current value and the measured current value match. As aforementioned,
If the distance between the torch and the groove wall is different, the current value changes. In principle, when the current value becomes higher than the command current value, it is determined that the torch is too low and the wire protrusion length is short, and the torch is raised. Conversely, when the current value is lower than the command current value, it is determined that the torch is too high and the wire protrusion length is long, and the torch is lowered.

Specifically, the smoothing current 122 similar to that described above is used.
The upper and lower arc copying corrections are instructed from the average current 135 obtained from (136). Also in this case, a dead zone for the command current difference is provided. Further, in order to prevent the arc scanning correction from operating indefinitely in a state where the malfunction of the vertical arc scanning correction has occurred, a correction limit of the vertical arc scanning correction amount is provided.

The current correction amount ΔZn is obtained as follows.

When | Ad |> = NSBud, ΔZn = Ad · Gud ΔZn> 0 and | Σ (ΔZ) + ΔZ |> = LMTu, ΔZn = LMTu−Σ (ΔZ) ΔZn <0 and | Σ (ΔZ ) + ΔZ |> = LMTd, ΔZn = LMTd−Σ (ΔZ) where Ad is the command current difference (A), NSBud is the dead zone width (A), and Gud is the vertical arc scanning gain (μm /
A) and Σ (ΔZ) are integrated values (μ
m), LMTu and LMdd are upper and lower correction amount limits (μ
m). The dead zone width, the vertical arc copying gain, and the vertical correction amount limit are values specified in advance.

In this way, the vertical arc scanning correction amount ΔZn is obtained during a certain weaving cycle period, and the torch is moved up and down by the correction amount at once at the starting point of the next weaving (137). Performing the correction at once like this is
This is because, if the correction of the torch in the vertical direction is dispersed like a right and left arc and corrected at different Y-axis positions, a right and left imbalance will occur.

On the other hand, arc dullness detection 133 is performed from the current waveform 123. The arc dulling means that the pulse waveform of the current becomes dull due to the influence of the resistance value or the like (the rising and falling speed decreases and the pulse width increases). One cause of the arc dulling is that the protruding length of the wire is large and the arc length is long. This makes it impossible to perform normal pulse welding because the shield gas does not reach the molten pool. Therefore, it is checked whether or not arc dulling has occurred based on the pulse width or the pulse peak value for each servo control cycle. That is, the current pulse width this time is
It is determined that arc blunting has occurred when it exceeds a predetermined allowable value. The allowable value can also be specified as a ratio (%) to the command pulse width.

More specifically, the current pulse width PWn is detected by satisfying the following equation.

PWn> = PWlim or PWn / tp> = PRat Here, PWlim is a preset pulse width upper limit value,
tp indicates a command pulse width, and PRat indicates a preset pulse width ratio upper limit value. The unit of pulse width is 0.01 ms.

When the arc blunting is detected, an instruction is given to lower the torch at a speed Vs designated in advance (134).
As a result, the protruding length of the wire is reduced, the resistance is reduced, and the bluntness is reduced. The arc blunt correction ends when the current pulse width is equal to or less than the blunt detection value for a certain period of time.

When the arc blunting correction is being performed, the vertical arc tracing correction 136 is invalid (14
9).

Further, in order to avoid an adverse effect due to the measurement of the abnormal current value at the start of the arc, the arc blunting correction is invalidated for a designated time at the start of the arc. The torch correction width is specified by a program / parameter.

The average current 135 is displayed as status data (148).

Next, FIG. 17 shows procedures for adjusting (correcting) the arc voltage based on the measurement result of the pulse welding voltage 32, correcting the weaving width, and correcting the weaving locus, and the signals used therefor.

<Arc voltage adjustment> The arc voltage adjustment is to automatically adjust the arc voltage in order to absorb the fluctuation of the arc voltage due to the environmental conditions during the welding work. Specifically, a smoothed voltage 124 obtained by filtering the pulse voltage 32 at a low cutoff frequency is sampled, and all data sampled during a half cycle of weaving is averaged to obtain an average voltage 138.

On the other hand, the arc short time is obtained from the voltage waveform 125 of the pulse voltage 32. Further, the average arc short time 139 is obtained by averaging all the arc short times detected in the horizontal movement portion of one weaving cycle. The voltage command is adjusted so that the average arc short-circuit time has an appropriate value (140). That is, AS
In the case of n> ASVup, ASd = ASn−ASVup In the case of ASn <ASVlo, ASd = ASn−ASVlo where ASn is the current arc short time, ASVup is the allowable arc short upper limit value, ASd is the current arc short time difference, and ASVlo is the allowable arc. This is the short-circuit lower limit value. These units are 0.001 ms.

From these, the current voltage adjustment amount ΔVn (unit is 0.1 V) is obtained by the following equation.

ΔVn = ASdGas where Gas is a voltage calculation coefficient (unit is 0.1 V / m
s).

In the arc voltage adjustment (141), the voltage command is adjusted so that the command voltage (including the arc length correction amount ΔVn) and the measured average voltage 138 match. Here, in order to prevent the arc voltage adjustment from operating indefinitely in the state where the malfunction of the arc voltage adjustment has occurred, a limit of the voltage adjustment amount is provided.

ΔVn = (Vc + Val) -Vfb Here, Vc is the command voltage value, Val is the arc correction amount, and V is the arc correction amount.
fb is an output voltage (feedback) value.

However, when ΔVn> ΔVmax, ΔVn = ΔVmax, where ΔVmax is the allowable maximum voltage adjustment amount once.

When Vlo <= ΔVn <= Vup, ΔVn = 0.

ΔVn> 0 and | Σ (ΔV) + ΔVn |>
= VLMu, ΔVn = VLMu−Σ (ΔV) ΔVn <0 and | Σ (ΔV) + ΔVn |> = VLMd, ΔVn = VLMd−Σ (ΔV) where Σ (ΔV) is the previous voltage Integrated value of adjustment amount,
VLMu indicates the maximum voltage adjustment amount (increase) and VLMd maximum voltage adjustment amount (decrease).

In this way, the arc voltage adjustment amount ΔVn is obtained, and the voltage is corrected at the center position of the weaving every half period of the weaving (142).

<Weaving Width Correction> The weaving width correction is to make the weaving width appropriate for the groove width, and changes the width of the horizontally moving portion of the weaving independently. The left-right arc tracing also corrects the amount of movement in the left-right direction by weaving, but the left-right arc tracing is performed according to the left-right current value 128 based on the pulse current 13 as described above. On the other hand, the weaving width correction is performed according to the left and right end arc short-circuit time 143 based on the pulse voltage 32, and both corrections function independently. Further, the left and right arc copying works for the entire weaving, whereas the weaving width correction differs in that the left and right weaving widths can be adjusted separately.

The weaving width correction basically compares the left end arc short time and the right end arc short time with the upper limit allowable value and the lower limit allowable value, and increases or decreases the left and right weaving widths when they are out of the allowable values. . That is, when the left end arc short time is larger than the upper limit allowable value, the left weaving width is narrowed, and when it is smaller than the lower limit allowable value, it is widened. On the other hand, when the right end arc short time is larger than the upper limit allowable value, the right weaving width is narrowed, and when it is smaller than the lower limit allowable value, it is widened. When the left and right weaving widths are asymmetrical after this weaving width change, the logical weaving center points are shifted so that the left and right weaving widths are symmetrical. The weaving speed is also adjusted so that the weaving period does not change after the weaving width is changed.

Specifically, from the voltage waveform 125, the arc short times at the points corresponding to the weaving sections at the left end and the right end are extracted and averaged to obtain the left end arc short time and the right end arc short time 143. The number of points for averaging is 127
Similarly, it can be made variable according to the weaving period.

The arc shorting time 143 thus obtained is individually compared with a predetermined upper / lower limit allowable value, and the weaving width is corrected as follows (1
44). The change of the left and right weaving width is not a correction by the shift amount like the above-mentioned left and right arc copying, but is a change of the weaving target point positions from the weaving center point to the left and right ends.

I) When the left end arc short time is larger than the upper limit allowable value, it is judged that the torch is too close to the left and the left weaving width is narrowed, and ii) the left end arc short time is the lower limit allowable value and the upper limit. If it is in the middle of the allowable value, do nothing, iii) If the left end arc short time is smaller than the lower limit allowable value, determine that the torch is not left enough and widen the left weaving width. iv) If the right end arc short time is greater than the upper limit, narrow the right weaving width, v) Do nothing if the right end arc short time is between the lower limit and the upper limit, vi) Right end If the arc short time is less than the lower limit allowable value, increase the right weaving width.

In order to prevent the weaving width correction from operating indefinitely while the weaving width correction malfunction occurs, a correction amount limit is set.

Change amount ΔW at the time of changing the weaving width
n (μm) is as follows.

I) Arc short time ASn> upper limit value A
In the case of Sup, ΔWn = (ASup−ASn) · Gwm, where | Σ (ΔW) + ΔWn) |> = WLMe, ΔWn = WLMe−Σ (ΔW) where ASn is the arc short time this time and Gwm is weaving. The width change coefficient (μm / ms), Σ (ΔW) is the integrated value of the weaving width change amounts up to the previous time, and WLMe is the maximum weaving width change amount (on the widening side).

Ii) Arc short time ASn> lower limit value A
In the case of Sdn, ΔWn = (ASdn−ASn) · Gwm, where | Σ (ΔW) + ΔWn) |> = WLMs, ΔWn = WLMs−Σ (ΔW) where WLMs is the maximum weaving width change amount (narrowing side) ).

Note that ASup, ASdn, Gwm, WL
Me and WLMs are values designated in advance.

As a result of adjusting the weaving width in this way, when the left and right weaving widths are asymmetric, the weaving center point is shifted according to the following equation to make the weaving symmetrical. However, it should be noted that this is a logical center point recognized by the control device, and not a direct control with respect to the movement of the torch.

ΔWc = (ΔWr−ΔWl) / 2 where ΔWc is the weaving center position change amount (μ
m), ΔWr is the right weaving width change amount, and ΔWl is the left weaving width change amount.

Furthermore, when the weaving width is adjusted and the overall weaving width is changed, the weaving speed is changed according to the following equation in order to keep the weaving period constant.

KF = (W + ΔW) / W NFw = KF · Fw where KF is the weaving speed change coefficient, W is the width of the horizontal portion of the weaving (μm), ΔW is the amount of change of the weaving width, and Fw is the horizontal portion of the weaving. Speed, NFw
Is the speed (μm / sec) of the horizontal portion of the weaving after the change.
c).

<Weaving locus correction> Weaving locus correction is applied to the welding surface, especially to an irregular bead shape.
The weaving locus of the torch (Z-axis position in this case) is set appropriately, and the weaving locus is corrected independently of the left, middle, and right Z-axis heights of the weaving horizontal movement part. Do as much as possible.
The above-mentioned vertical arc tracing also controls the Z-axis position of the torch, but since the signal used by the vertical arc tracing is the average current value 135 of the smoothing current 122, the correction is for the entire weaving width. On the other hand, weaving trajectory correction is left, center, and
It enables fine adjustment in each section on the right.

In principle, the weaving locus correction calculates the arc short time of each of left, middle and right (horizontal moving part of the weaving) from the voltage waveform 125 based on the pulse voltage 32 (145), and the upper and lower limits thereof are determined. Each weaving height for the left, middle, and right is adjusted according to the result of comparison with the allowable value. To see the arc short time,
This is to determine whether the torch is too close to the bead surface. In addition, the current waveform 1 based on the pulse current 13
The respective current period ratios of the left, middle, and right are obtained from 23 (146, 1
50, 151), which also adjusts the left, middle, and right weaving heights. To see the pulse current period ratio,
This is to determine whether the torch is too far from the bead surface (this cannot be determined by the arc short time).

Left / center / right arc short time 145
As, the arc short time of several points in each section is averaged. The averaging number is variable according to the weaving cycle as described above. If the arc short time is too long, it is determined that the torch is too close to the bead surface, and control is performed to raise the torch. On the other hand, the average pulse period 150 is an average of all pulse periods of the horizontal moving portion within one weaving period, as in the case of the left-right arc scanning described above. The left / middle / right current pulse period 146 is obtained by averaging the pulse periods of several points in each section of the current pulse period detected in the horizontal moving portion of the weaving. Also in this case, the averaging number is variable according to the weaving cycle. Left / middle / right current pulse period 14 with respect to average pulse period 150
The ratio of 6 is obtained as the left / middle / right pulse period ratio 151. If the pulse cycle ratio is too large, it is determined that the torch is too far from the bead surface, and the torch is controlled to be lowered.

Specifically, weaving locus correction 147
Then, the weaving height is corrected as follows according to each of the left / middle / right arc short times 145 and each pulse period ratio 151.

I) When the left arc short-circuit time exceeds the upper limit value, the weaving left height is increased, and ii) when the left current pulse period ratio is the upper limit value or less and the left arc short circuit time is the upper limit value or less. Iii) If the left current pulse period ratio exceeds the upper limit, lower the weaving left height, and iv) if the middle arc short time exceeds the upper limit, weaving center height V) If the medium current pulse period ratio is less than the upper limit value and the medium arc short time is less than the upper limit value, the height of the weaving center is not changed, and vi) The medium current pulse period ratio does not exceed the upper limit value. When it exceeds, the weaving center height is lowered, vii) When the right arc short time exceeds the upper limit value, the weaving right height is increased, viii) The right current pulse period ratio is less than or equal to the upper limit value, and the right arc short If the applied time is more than the upper limit, without changing the weaving right height, ix) right current pulse period ratio may exceed the upper limit, to reduce the weaving right height.

The current change amount ΔZ when changing the left, middle, and right heights of the weaving is determined as follows. It should be noted that a limit of the correction amount is provided to prevent the weaving locus correction from operating indefinitely while the weaving locus correction malfunctions.

I) When increasing the weaving height, the correction amount is calculated from the arc short time as follows.

ΔZn = (ASup−ASn) · Gup However, in the case of | Σ (ΔZn) + ΔZn |> = ZLMu, ΔZn = ZLMu−Σ (ΔZn) where ASup is the upper limit value of the arc short time, ASn.
Is the left / center / right arc short time this time, Gup is the upward weaving height change coefficient (μm / ms), Σ (ΔZ
n) is the weaving height change amount (μm) up to the previous time, Z
LMu is an upward maximum weaving height change amount.

Ii) When lowering the weaving height, the correction amount is calculated from the current pulse period.

ΔZn = (APup−APn) · Gdn However, in the case of | Σ (ΔZn) + ΔZn |> = ZLMd, ΔZn = ZLMd−Σ (ΔZn), where APup is the current pulse cycle upper limit value and APn is the current left. Middle / right current pulse period, Gdn is a downward weaving height change coefficient (μm / ms), and ZLMd is a downward maximum weaving center height change amount.

When the whole weaving locus is changed as a result of such weaving locus correction, the weaving speed is changed as follows in order to keep the weaving period constant.

WLn = ((WL / 2). (WL / 2) +
(ΔZl) (ΔZl)) + ((WL / 2) · (WL /
2) + (ΔZr) (ΔZr)) LF = WLn / WL NFw = KF · Fw where WLn is the moving distance of the weaving horizontal portion after trajectory correction, and WL is the commanded moving distance of the horizontal portion of weaving, ΔZl Is the left weaving difference height, ΔZr is the right weaving difference height, KF is the weaving speed change coefficient, F
w is the speed of the horizontal portion of the weaving, and NWFw is the speed of the horizontal portion of the weaving after the change.

The status of the average voltage 138, the average arc short time 139, the left / middle / right arc short time 145, and the left / middle / right pulse cycle 146 are displayed.

As described above, by performing the vertical and horizontal arc scanning correction, the weaving width / orbit correction, etc. for each layer, the groove width variation, especially the arc with a wide groove width, can be sufficiently traced and the welding amount And the shape of the bead can be made flat, and blow holes and undercuts can be prevented.

Although the preferred embodiment of the present invention has been described above, the specific contents of the system configuration and processing are merely examples, and various modifications and changes can be made without departing from the scope of the claims.
Changes are possible. For example, the specific various numerical values given in the above description are merely examples, and do not limit the present invention.

[0110]

According to the present invention, it is possible to improve the arc performance of a pipe circumference automatic welding apparatus for automatically welding the circumference of a steel pipe while performing high-speed weaving. In particular, since the arc blunting is detected by monitoring the pulse width itself of the pulse welding current, the occurrence of the arc blunting can be more reliably detected with good responsiveness, and this can be eliminated.

[0111]

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a schematic configuration of a pipe circumference automatic welding apparatus according to the present invention.

FIG. 2 is an explanatory diagram of an electrical connection relation of each part of the welding apparatus of FIG.

FIG. 3 is an explanatory view of the operation of the head and the torch in the welding apparatus of FIG.

FIG. 4 is an explanatory diagram of accessory parts of a head in the welding apparatus of FIG.

5 is a block diagram showing an internal configuration of a control device of the welding device of FIG. 1. FIG.

6 shows a flowchart of processing of the welding apparatus of FIG.

7 is an explanatory diagram of an example of a logic table used in the processing shown in FIG.

8 is a CAD and NC used in the process shown in FIG.
5 is an explanatory diagram of a correspondence relationship of control items of FIG.

FIG. 9 is a timing chart showing a temporal change of the torch position in the weaving operation performed by the welding apparatus of FIG.

FIG. 10 is an explanatory diagram of a torch locus in the weaving operation of FIG. 9.

FIG. 11 is a timing chart showing the relationship between the command locus and the servo locus in the weaving operation.

12 is a block diagram showing a configuration example of a pulse waveform measuring section 211 shown in FIG.

13 is a waveform diagram showing the voltage signal and the current signal of FIG.

FIG. 14 is a waveform diagram showing a voltage signal and a current signal after filtering in FIG.

FIG. 15 is a timing chart for explaining the operation of the pulse waveform measuring section in FIG.

FIG. 16 is an explanatory diagram of an arc tracing-related correction according to the present invention.

FIG. 17 is an explanatory diagram of another arc-related correction according to the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Computer, 2 ... Control device, 3 ... Power supply device, 4 ...
Head, 5: Pipe, 6: Wire supply unit, 7: Wire,
8 welding torch, 12 guide rail, 16 current signal, 32 voltage signal, 91 arc, 92 weld metal,
93 ... groove, 41 ... sensor, 42 ... encoder, 211
... Pulse waveform instrumentation unit, 221 ... Storage unit, 222 ... Hard disk, 223 ... Control unit, 225 ... CRT, keyboard, 230 ... Welding head control unit, 231 ... Transceiving unit, 2
32 ... Power control unit, 233 ... Head drive unit, 234 ... Remote operation device,

 ─────────────────────────────────────────────────── ─── Continued front page (72) Inventor Ikuo Mibu 2982-3 Sugaya, Naka-machi, Naka-gun, Ibaraki Prefecture (72) Kenichi Maeda 1225-47 Tenjinbayashi, Hitachiota-shi, Ibaraki Prefecture

Claims (3)

[Claims] 1. When the end faces of two pipes that are fixed to each other by butt welding are welded all around by a pulse MAG welding method, the pipes are aligned in the left-right direction and in the up-down direction orthogonal to the tangent line of the pipe circumference along the groove of the pipe. A dullness correction method for an automatic welding device that performs two-dimensional weaving of a torch, wherein a current pulse width is obtained based on a measured pulse welding current, and the current pulse width is larger than a predetermined upper limit value, and By detecting any one of the case where the ratio of the current pulse width to the command pulse width is larger than a predetermined upper limit value, the bluntness of the current pulse waveform is detected, and when the bluntness is detected, the torch is moved to the pipe pipe surface. A method for correcting arc blunting in a pipe circumference automatic welding device, characterized by controlling so as to lower the direction. 2. The arc blunting correction method for a pipe circumference automatic welding apparatus according to claim 1, wherein the torch is lowered at a predetermined speed until the current pulse width reaches an allowable value. 3. A vertical arc tracing correction for moving the torch in the vertical direction so that the average current matches the command value is performed based on the average current of the pulse welding current, and during the execution of the arc blunt correction, The arc blunting correction method for a pipe circumference automatic welding apparatus according to claim 1 or 2, wherein the vertical arc copying correction is disabled.