JP2003230960A - Control device for automatic welding position copying - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a control device for automatic welding position copying, where the position of a tungsten electrode in a welding groove is easily and accurately determined to conduct the control of automatic copying. <P>SOLUTION: The data of the image of a welding part, which is photographed by an image processing device 3 by a charge-coupled-device (CCD) camera 1, is analyzed to obtain following values: an electrode shift amount which is the difference between a intermediate position among welding groove faces and the position of the tungsten electrode 9; an arc tilt angle difference that is the difference between angles, which a pair of tangential lines coming into contact with the upper contour of the image in the width direction of an arc (a) makes with a horizon; the length difference of welding groove face boundaries where the arc (a) and the image of the molten pool come into contact with the welding groove face; and a hanging down amount deviation which is the difference between the lower part vertical positions of a pair of hanging down parts formed at the bottom of the image of the molten pool. The proper position of the tungsten electrode 9 in the width direction is calculated according to either of the above values, and the electrode 9 is shifted to a proper position by a laterally shifting device 5. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention photographs the state of an arc and a molten pool formed between an electrode and a base material inserted in a groove formed between plate-shaped base materials, and takes an image. The present invention relates to a welding position automatic copying control device for controlling the movement of an electrode in the width direction based on the result of analyzing data and calculating an appropriate position in the width direction of the electrode.

[0002]

2. Description of the Related Art A main vessel of a power generation boiler and a pressure vessel such as a reactor pressure vessel (RPV) for nuclear power generation are structures made by welding a large number of thick plates having a thickness of about 30 mm to 150 mm. Is. For example, main pipes that form part of a boiler for power generation, that is, main steam pipes, high-temperature reheat steam pipes, low-temperature reheat steam pipes, main water supply pipes, etc., must withstand high temperatures and high pressures, It is composed of thick-walled steel pipe in diameter. The main piping guides high-temperature, high-pressure steam heated by the coil from the boiler to the turbine, and low-temperature in the turbine.
It plays the role of returning the low-pressure steam to the boiler coil again. In this way, the piping itself has a high temperature,
Since the boiler is exposed to high pressure, it expands and contracts significantly when the boiler is started and stopped, and stress concentrates especially on the welded joint. Therefore, high quality welding is required for the main piping from the boiler to the turbine.

As a welding method that meets such requirements, the amount of heat input to the base metal is suppressed, welding heat distortion is small, and work efficiency can be improved, for example, a groove width of 5 mm to 12 mm.
The narrow groove arc welding method can be used.

Further, when the base material is stainless steel or high alloy steel, JP-A-11-179545 and JP-A-20
Japanese Patent Laid-Open No. 11-77 discloses a non-consumable electrode type TIG (Tungsten Inert Gas Arc) welding method as disclosed in Japanese Patent Application Laid-Open No. 00-210771, and a base material of low alloy or mild steel.
Consumable electrode type GMA (Gas Metal Ar) as disclosed in Japanese Laid-Open Patent Publication No. 307 and Japanese Patent Laid-Open No. 11-104833.
c) Welding methods are often applied.

FIG. 7 is a schematic view showing an outline of narrow groove TIG welding according to the prior art. First, the opposing base materials 8a,
A narrow groove 25 having a width of 5 mm to 12 mm is processed between 8b, and a long welding torch 6 and a wire torch 13, which are thinner than the groove width, are arranged and inserted in the welding direction of the narrow groove 25. A tungsten electrode 9 is attached to the tip of the welding torch 6, and the weld is shielded from ambient air by an inert gas (blocking gas) consisting of argon gas blown from the tip of the welding torch 6. Then, the addition voltage 12 is fed to the welding portion through the wire torch 13 while applying an arc voltage with the tungsten electrode 9 as a negative electrode and the base materials 8a and 8b as a positive electrode to form the arc a. As a result, the base material 8
The base materials 3a and 3b are welded by filling the narrow groove 25 of a and 8b with welding metal.

A shield gas box 7 is integrally attached to the welding torch 6, and the blocking gas is caused to flow out from here in the direction of the narrow groove 25 to fill the entire inside of the narrow groove 25 with the blocking gas, whereby Complements the air barrier. The addition wire 12 is wound around a wire reel 18 and is supplied from a wire feeding device 17. The supplied addition wire 12 is guided to the weld through the conduit 16 and the wire torch 13. A welding device (not shown) equipped with the welding torch 6, the wire feeding device 17, etc. moves in the direction of the arrow.

[0007] In this example, the multi-layer welding system of one layer and one pass is used. However, since one layer is a laminated layer of 2 mm to 4 mm, when the plate thickness becomes large, the welding time will be several hours, which is a continuous work. Becomes Also, to ensure welding quality,
It is necessary to monitor the arc welding situation within a very narrow groove. Especially, it is very important to adjust the electrode position in the groove, and when the tungsten electrode 9 is biased to one of the left and right wall surfaces of the groove, the adjacent wall surfaces are excessively melted while the opposing wall surfaces are not fused well. Will occur. In order to reduce the work burden on the welder who carries out highly accurate welding work in this way, it is possible to perform remote control while monitoring images captured by a CCD (Charge Coupled Device) camera, and to use image processing equipment to process image data. The position of the groove and the position of the electrode are analyzed and analyzed, and automatic control is performed so that the electrode is positioned at the center of the groove in the width direction.

FIG. 8 is an explanatory view of a welded portion showing a state in which an arc is asymmetrically generated in the groove. As shown in the figure, the tip of the tungsten electrode 9 is usually used in a state of being polished into a conical shape, but the degree of wear of the tip may be different on the left and right. In such a case, the arc / molten pool a
The shape of m becomes asymmetrical. This phenomenon is particularly likely to occur when the arc length is long. Further, even when the tip of the tungsten electrode 9 is displaced from the center of the cylindrical shape, a left-right asymmetric arc / molten pool am may similarly occur.

In addition to this, when the lower bead of the front layer is formed unevenly, asymmetrical arc / molten pool am may occur due to the bead shape. Further, the addition wire 12 passes through the wire torch 13 and protrudes from the tip thereof to be inserted into a predetermined position of the molten pool.
When the addition wire 12 has a bending tendency, the insertion position may shift. Furthermore, a difference in temperature (heat capacity difference) occurs between the left and right base materials due to a significant difference in the shape, for example, the size of the material to be processed, and as a result, there is a difference in the degree of melting of the left and right wall surfaces of the groove. Similarly in these cases,
The shape of the molten pool am may be left-right asymmetric.

FIG. 9 is an explanatory view showing the state of the molten portion when the insertion position of the addition wire 12 into the molten pool is displaced, and FIG. 10 is the state where the position of the tungsten electrode 9 is displaced from the groove center. . In these cases as well, the shape of the arc / molten pool am becomes asymmetrical, which makes the degree of melting on the left and right wall surfaces of the groove uneven, and the bead shape is also affected, resulting in welding defects such as poor fusion and non-fusion. It was causing the outbreak.

FIG. 11 is a photograph of the melted portion. FIG. 11A shows the arc formed at the correct position, and FIG. 11B shows the arc forming position due to the asymmetry of the tip shape of the tungsten electrode 9. (C) shows that the formation position of the arc is deviated due to the deviation of the insertion position of the addition wire 12 into the molten pool. As shown in these figures, in the images taken by the CCD camera of the fusion zone,
The image of the arc and the molten pool can be seen as an integrated arc and molten pool am. Further, the image of the addition wire 12 appears as a black shadow in the center of the image of the arc / molten pool am.

By the way, the above-mentioned JP-A-11-10483.
In the invention disclosed in Japanese Patent Publication No. 3, the numerical values of the characteristic points on the left and right of the lower image of the molten pool are extracted from the image of the arc / molten pool am captured by a CCD camera, and the difference between these numerical values becomes zero. The control of moving the position of the tungsten electrode 9 is performed.

[0013]

The above-mentioned prior art is effective in the consumable electrode type GMA welding method in which the electrode also serves as the additive wire, but the arc is generated from the non-consumable tungsten electrode and the additive wire is used. Non-consumable electrode type TI that advances the welding equipment while inserting
In the G welding method, since the shape of the lower part of the weld pool changes depending on the insertion state of the additive wire into the weld pool, it is not possible to properly determine the bias of the arc and an effective method for determining the position within the groove of the electrode. It can not be said.

An object of the present invention is to provide a welding position automatic copying control device capable of easily and accurately determining the position of the tungsten electrode in the groove from the image data of the welded portion and performing automatic copying control of the electrode. To provide.

[0015]

In order to achieve the above-mentioned object, the present invention is to move an electrode inserted in a groove formed between a pair of plate-shaped base materials in a welding advancing direction and to form a groove wall surface. An image capturing the state of the melted portion in which the base material is melt-bonded by an arc formed by a direct current voltage applied between the electrode and the base material Means, image processing means for analyzing image data of the melted portion photographed by the imaging means to calculate an appropriate position in the width direction of the electrode, and according to data of the appropriate position of the electrode calculated by the image processing means. The present invention is directed to a welding position automatic copying control device including a movement control means for controlling the movement of the movement means in the width direction.

Then, the arithmetic means is an arc tilt angle difference which is a difference between an angle formed by a pair of tangents to an upper contour of the image in the width direction of the arc sandwiching the electrode and a reference line, and the arc and the molten pool. Droop amount deviation which is the difference in the boundary length of the groove wall where the pair of groove walls are in contact with the image in the width direction or the difference between the vertical positions of the lower ends of the pair of hanging parts formed at the bottom of the image in the width direction of the molten pool. According to any one of the above, the proper position of the electrode in the width direction is calculated.
It is achieved by means of.

A second means of the present invention is the first means according to the first means, wherein the arithmetic means is an electrode displacement amount which is a difference between a midpoint between a pair of groove walls and an electrode position, an arc inclination angle difference, and an opening angle. It is characterized in that an electrode position correction value, which is a linear combination of the front wall boundary length difference and the drooping deviation, is calculated and set as an appropriate position in the width direction.

[0018]

Embodiments of the present invention will be described in detail below with reference to the drawings. FIG. 1 and FIG. 2 are a front view and a side view seen from inside a groove showing a schematic configuration of a non-consumable electrode type TIG welding apparatus according to an embodiment of the present invention. Portions that are the same as or can be regarded as the same as those in the conventional example are denoted by the same reference numerals, and redundant description will be omitted.

In these figures, 1 is a CCD camera for photographing a welded portion in the narrow groove 25, 2 is a monitoring camera for displaying an image of the welded portion photographed by the CCD camera 1, and 3 is a CC.
An image processing device that analyzes image data captured by the D camera 1 to calculate an appropriate tip position of the tungsten electrode 9.
Is the narrow groove 2 of the tungsten electrode 9 based on the data of the tip position of the tungsten electrode 9 calculated by the image processing device 3.
A welding control device 5 for controlling the left and right position in 5 is a left and right moving device which is controlled by the welding control device 4 and moves a mounting base, which will be described later, in the left and right directions perpendicular to the welding proceeding direction. A weld bead formed on the inner bottom, 14
Is a mounting base for mounting the welding torch 6 and the wire torch 13 on the left and right moving device 5.

In many cases, the image of the welded portion is taken by the CCD camera 1 from diagonally forward in the welding advancing direction. In the case of narrow groove welding of a thick plate, the width of the narrow groove 25 formed by the groove processing of the base material 8 is about 5 mm to 12 mm. Therefore, insert the CCD camera 1 into the narrow groove 25. Can not
The CCD camera 1 is installed diagonally above the narrow groove 25 and monitors the welded portion. Since the arc a has a substantially conical shape, the image of the arc a captured by the CCD camera 1 is a combined image of an ellipse and a triangle.

The image data obtained by the CCD camera 1 is taken into the image processing device 3, the image is displayed on the monitor television 2, and the tip position of the tungsten electrode 9 is determined from the characteristics of the shape of the arc / molten pool am image. , Narrow groove 2
The positions and the like of the wall surfaces 20a and 20b of No. 5 are calculated, and an appropriate left and right position of the tungsten electrode 9 is calculated. The data of the proper left and right positions of the tungsten electrode 9 are transferred to the welding control device 4. The welding control device 4 drives the left / right moving device 5 by a predetermined amount based on the position data to move the tungsten electrode 9 to an appropriate left / right position within the narrow groove 25.

The tip position of the tungsten electrode 9 and the wall surfaces 20a, 2 of the narrow groove 25, which will be described below, are performed in the image processing apparatus 3.
A method of calculating the deviation amount from the position of 0b and the corrected position of the tungsten electrode 9 in the narrow groove 25 will be described.

FIG. 3 is a diagram for explaining a method of analyzing an arc / molten pool am image. Since the arc light has a very high light intensity in the ultraviolet region, the lens of the CCD camera 1 is covered with a special filter to suppress the luminance in the ultraviolet region and the weld is photographed. The image data taken by the CCD camera 1 is taken into the memory of the computer in a specifiable format.

As shown in FIG. 11, the arc a and the molten pool are heated to a high temperature by Joule heat and shine, and the tungsten electrode 9
Since the tip of each of them also receives the heat and is heated and glows, it becomes a bright image, and the surrounding area and the addition wire 12 are at a low temperature, so that it becomes a dark image. As described above, since the boundary between the arc a and the molten pool is not clear because the molten pool reflects the arc light, those images are integrated arc / molten pool am images.

FIG. 3 shows a situation in which the arc a is biased to the right side even though the tip of the tungsten electrode 9 is located at the center of the narrow groove 25. Figure 3
It is assumed that the origin of the XY coordinates is taken at the upper left corner of the area divided into squares and divided into 510 in each of the X and Y directions, and a pixel is located at the intersection of each coordinate. Then, the brightness of each pixel is set to 8
Bits, that is, data of 255 gradations will be represented.
First, the center position of the tungsten electrode 9 is obtained by the following procedure.

An electrode detection scanning line 1h is set above the above area, and pixel data on this electrode detection scanning line 1h is set to X = 0.
Starting from the beginning. Then, a point A where the pixel data sharply increases and a point B where the pixel data sharply decreases are detected. These boundary points A and B clearly coincide with the contour of the tungsten electrode 9. Therefore, the midpoint between the boundary point A and the boundary point B is obtained and set as the center position of the tungsten electrode 9.

Next, the image data is examined along the arc detection scanning line lla extending downward from the boundary point A, and the upper contour of the arc a is detected by detecting the sudden increase point. afterwards,
While shifting the arc detection scanning line lla to the left (X decreasing direction) from the boundary point A, the upper contour of the arc a where the image data rapidly increases is continuously detected, and the boundary point C immediately before the upper contour of the arc a cannot be detected is detected. To detect. Then, the boundary point C is assumed to coincide with the wall surface 20a of the narrow groove 25. Similarly, the arc detection scanning line llb directed downward from the boundary point B
The image data is examined along with, the upper contour of the arc a is detected by detecting a sudden increase point, and then the arc detection scanning line l
lb is shifted rightward to detect the boundary point D immediately before the upper contour of the arc a cannot be detected. The boundary point D is supposed to coincide with the wall surface 20b of the narrow groove 25.
The center coordinates X of the narrow groove 25 can be obtained from the X coordinates of the boundary points C and D thus obtained.

Next, the amount of deviation from the proper position of the tungsten electrode 9 is calculated from the X coordinates of these boundary points A to D. That is, when the X coordinates XA to XD of the boundary points A to D are set, the electrode displacement amount ΔX is expressed by the following equation.

ΔX = (XC + XD) / 2− (XA + XB) / 2 (1) Simply thinking, when ΔX> 0, it is considered that the tungsten electrode 9 is biased to the left, so the tungsten electrode 9 is moved to the right. When ΔX <0, the tungsten electrode 9 may be moved to the left.

However, in the case of narrow groove welding, the groove width is very narrow, and the distance between the peripheral edge of the tungsten electrode 9 and the wall surfaces 20a, 20b of the narrow groove 25 is close, so that the arc a is biased. Depends on more complex factors. For example, in the example shown in FIG. 3, the tungsten electrode 9 has a narrow groove 25.
Despite being located at the center of the arc, the arc a is generated biased to the right. In such a case, the deviation of the arc a cannot be prevented by the above-mentioned simple method.

FIG. 4 is a schematic diagram showing the state of the welded portion when the arcs are generated without deviation. As described above, the arc light is reflected on the surface of the molten pool, and since the CCD camera 1 captures the reflected light, it is not possible to capture an image of a portion that is a shadow of the arc light. Further, in TIG welding, an arc voltage is applied immediately below the tungsten electrode 9 to form a concave shape. Therefore, the lower line of the arc / molten pool am image in the foreground often does not indicate a boundary. When the arc a is generated without deviation as shown in FIG. 4, the upper contour portion has a left-right symmetrical conical shape, and the lower portion also has a substantially cylindrical rotationally symmetrical shape. On the other hand, as shown in FIGS.
In the case where the arc a is biased as shown in FIG. 5, the arc / molten pool am image is distorted and does not have a bilaterally symmetrical shape.

Therefore, in this embodiment, the arc / molten pool a
Characteristic numerical values representing the distortion of the m image are compared on the left and right, and when they are different, the tungsten electrode 9 is controlled so as to be moved in a direction in which the difference between them is reduced. First, the angle formed by the upper contour line of the arc / molten pool am image and the horizontal line, that is, the arc inclination angle θ, becomes smaller as the distance between the peripheral edge of the tungsten electrode 9 and the wall surfaces 20a, 20b of the narrow groove 25 decreases. Eventually it approaches zero. Therefore, the difference between the left and right arc tilt angles θL and θR, that is, the arc tilt angle difference Δθ is calculated.

Δθ = θR-θL (2) If the arc inclination angle difference Δθ> 0 (for example, θL3 and θR3 in FIG. 9), it means that the arc a is generated biased to the left. Move electrode 9 to the right,
If Δθ <0 (for example, θL2, θR2 in FIG. 8 and FIG.
.Theta.L4, .theta.R4) and the arc a is biased to the right, the tungsten electrode 9 may be controlled to move to the left. When the arc / molten pool am image is distorted, the arc / molten pool am image and the wall surface 20 of the narrow groove 25
Left and right groove wall boundary length L, which is the contact length with a and 20b
A difference occurs in R and LL. Therefore, the left and right groove wall boundary lengths L
The difference between R and LL, that is, the groove wall boundary length difference ΔL is calculated.

ΔL = LR-LL (3) The groove wall boundary lengths LR and LL are the wall surface 20 of the narrow groove 25.
It represents the melting length of a, 20b, and the longer this length is, the greater the influence of the arc a on the wall surfaces 20a, 20b. If ΔL <0 (for example, LL3 in FIG. 9)
LR3), since the arc a is generated biased to the left, the tungsten electrode 9 is moved to the right and ΔL> 0.
If so (for example, LL2, LR2 in FIG. 8, LL in FIG. 10)
4, LR4) and the arc a is generated biased to the right, so that the tungsten electrode 9 is controlled to move to the left.

As described above, the addition wire 12 is wound around the wire reel 18 and is supplied from the wire supply device 17 that accommodates the addition wire 12 and the conduit 16 and the wire torch 13.
Since it is guided to the welded portion through the wire, it is given a bending tendency at the time of being housed in the winding or at the time of feeding, and bends in the bending direction when protruding from the tip of the wire torch 13. For this reason, the position of the addition wire 12 inserted into the molten pool shifts to the left and right. When the insertion position of the addition wire 12 deviates from the regular position in this way, the shape of the arc / molten pool am is also affected.

For example, as shown in FIG. 9 described above, when the addition wire 12 is biasedly inserted to the left, molten metal overflows on the left side of the molten pool, the height of the molten pool increases, and the arc on the front side occurs. Since the light is blocked, the boundary of the molten pool cannot be identified in the arc / molten pool am image. The shape of the lower part of the arc / molten pool am image is such that the left side where the addition wire 12 is inserted rises and the right side falls. As described above, with respect to the shape of the lower portion of the arc / molten pool am image, the deviation amount of the arc a cannot be determined, and the deviation amount of the addition position of the addition wire 12 can be determined.

Although the deviation amount of the addition position of the addition wire 12 can be directly obtained from the insertion position of the addition wire 12,
It is possible to obtain the one in which the molten state is more reflected by obtaining from the peculiar amount of the shape of the arc / molten pool am. If the deviation of the addition wire 12 is large, the shape of the molten bead naturally also becomes uneven, so it is necessary to correct the addition position of the addition wire 12, but since the wire torch 13 is integrally fixed to the welding torch 6, In the present embodiment, the addition wire 12 is adjusted by adjusting the left and right positions of the tungsten electrode 9.
Correct the bias of the addition position of. Specifically, the maximum points of the left and right Y coordinates of the arc / molten pool am, that is, the drooping points YL,
The Y-coordinates of YR are compared, and the tungsten electrode 9 is moved toward the larger absolute value (in the above example, the absolute value of YR3 is larger than the absolute value of YL3, that is, on the lower side).

That is, the drooping amount deviation ΔY of the arc / molten pool am image is ΔY = YR-YL (4) After all, when the drooping amount deviation ΔY is positive, the tungsten electrode 9 is moved to the left and ΔY When is negative, tungsten electrode 9
Control to move to the right.

In the present embodiment, as described above, the electrode deviation amount ΔX, the arc inclination angle difference Δθ, the groove wall boundary length difference ΔL, and the drooping amount deviation Δ, which are obtained by the above equations (1) to (4).
Depending on whether Y is positive or negative, ΔX> 0, Δθ> 0, ΔL <0, Δ
If Y <0, move the tungsten electrode 9 to the right and
If <0, Δθ <0, ΔL> 0, ΔY> 0, the tungsten electrode 9 is controlled to move to the left, and the arc a
Even if the addition position of the or addition wire 12 is deviated, the shape of the arc a can be made into a normal symmetrical state. Therefore, in the present embodiment, the electrode position correction value ΔT is defined by a linear combination of the above deviations, as shown by the following equation. If the electrode position correction value ΔT is positive, the tungsten electrode 9 is moved to the right, and if it is negative. For example, the tungsten electrode 9 is controlled to move to the left.
That is, .DELTA.T = a.DELTA.X + b.DELTA..theta.-c.DELTA.L-d.DELTA.Y (5) where a, b, c, d are coefficients for adjusting the degree of influence of each deviation on the electrode position correction value .DELTA.T.

FIG. 5 is a flow chart of the electrode position correction process. First, when the welding operation is started, the welded portion is photographed by the CCD camera 1 (S1), and the image processing device 3 takes in the image data. And the arc due to the deviation of arc a
Each deviation ΔX, Δθ, which characterizes the distortion of the molten pool am image,
The values of ΔL and ΔY are calculated according to the equations (1) to (4) (S2). Next, according to the equation (5), weighting and subtraction of each of the biases ΔX, Δθ, ΔL, and ΔY is performed to calculate the electrode position correction value ΔT (S3). Then, the electrode position correction value ΔT is transmitted to the welding control device 4 (S4).
After that, it is judged whether or not the welding work is completed (S5),
If the welding work has been completed, this process is ended, and if the welding is continued, the process returns to S1.

On the other hand, on the welding control device 4 side, it is judged whether or not the electrode position correction value ΔT is received (S6), and if it is received, it is judged whether or not the electrode position correction value ΔT is positive ( S
7) If positive, the welding torch 6 (tungsten electrode 9) is moved to the right by the electrode position correction value ΔT (S8). S7
If it is N, it is determined whether or not the electrode position correction value ΔT is negative (S9), and if negative, the welding torch 6 (tungsten electrode 9) is moved to the left by the electrode position correction value ΔT. Thereafter, in S11, it is determined whether or not the welding work is completed. If the welding work is completed, this process is terminated, and if the welding is continued, the process returns to S6.

By executing the electrode position correction processing in this way, even if there is a deviation of the arc a or a deviation of the addition position of the addition wire 12 to the molten pool, the electrode left-right position automatic copying control for narrow groove TIG welding is performed. Can be adjusted to the optimum state,
The degree of melting on the left and right walls of the groove can be made uniform,
The occurrence of welding residue and welding defects can be reduced.

If the deviation of the arc a or the deviation of the addition position of the addition wire 12 is significant, welding may not be possible. Therefore, in order to keep the welding quality stable, this is detected and welding is automatically terminated. You may do it.

Next, another embodiment of the present invention will be described. As the groove width becomes wider, the tungsten electrode 9 shown in FIG.
The arc cannot reach the wall surfaces 20a, 20b of the groove 25 unless it is swung left and right within the groove 25. The horizontal scanning of the electrode position in this case will be described. The fact that the tungsten electrode 9 has reached the left and right ends of the swing during welding is detected by a detection means (not shown), the welding condition of the left and right ends of the swing of the tungsten electrode 9a or 9b is photographed by the CCD camera 1, and this is image processed. Then, the electrode position A at the swing left end, the groove wall boundary position C at the left end, the electrode position B at the swing right end, and the groove wall boundary position D at the right end are calculated by the above formula (1). The positional deviation ΔX of the electrode swing, the left arc / molten pool tilt angle θL at the swing left end calculated from the arc / molten pool shape, and the right arc / molten pool tilt angle θR at the swing right end from the above formula (2) The groove wall boundary length calculated by the above equation (3) from the arc tilt angle difference Δθ obtained in step 1, the left groove wall boundary length LL at the swing left end, and the right groove wall boundary length LR at the swing right end. Difference ΔL, left arc of the rocking left end / molten pool lower end position YL, right arc of the rocking right end The control parameters of the droop amount deviation ΔY calculated in-melt pool bottom end position YR the above equation (4), the formula (5)
The electrode position correction value ΔT is obtained, the optimum position of the electrode swing position is calculated, and the swing position is adjusted to perform left and right scanning.

Although an example of TIG welding has been described in the above embodiment, the present invention can be applied to GMA welding.

[0046]

As described above, according to the present invention, the proper position in the width direction of the electrode is determined according to the difference in arc inclination angle, the boundary length of the groove wall or the deviation of the drooping amount, and the moving means moves the electrode to the proper position. Since the electrode is moved to the position, the position in the groove of the electrode can be easily and accurately determined and the electrode can be automatically copied.

[Brief description of drawings]

FIG. 1 is a non-consumable electrode type TIG according to an embodiment of the present invention.
It is the front view seen from the inside of a groove which shows a schematic structure of a welding device.

FIG. 2 is a side view showing the schematic configuration of the non-consumable electrode type TIG welding apparatus as viewed from inside the groove.

FIG. 3 is a diagram for explaining a method of analyzing an arc / molten pool image.

FIG. 4 is a schematic diagram showing a state of a welded portion when an arc is generated without deviation.

FIG. 5 is a flowchart of an electrode position correction process.

FIG. 6 is a diagram for explaining an arc / molten pool image analysis method when the present invention is used for TIG welding in a wide groove.

FIG. 7 is an explanatory diagram showing an outline of a narrow groove TIG welding method.

FIG. 8 is a schematic view showing a state of a welded portion where an arc is asymmetrically generated in a groove.

FIG. 9 is a schematic diagram showing a state of a welded portion when the position of insertion of the addition wire into the molten pool is displaced.

FIG. 10 is a schematic diagram showing a state of a welded portion when the position of the tungsten electrode is deviated from the groove center.

FIG. 11 is a diagram showing a state of a welded portion.

[Explanation of symbols]

1 CCD camera 2 surveillance TV 3 Image processing device 4 Welding control device 5 Left and right moving device 6 welding torch 8,8a, 8b Base material 9 Tungsten electrode 10 Weld beads 12 Doped wire 13 wire torch 14 Mounting base 20a, 20b groove wall 25 narrow groove a arc am arc molten pool

Claims (2)

[Claims] 1. A moving means capable of moving an electrode inserted in a groove formed between a pair of plate-shaped base materials in a welding advancing direction and also in a width direction toward a groove wall surface, Image capturing means for capturing a state of a fused portion where the base material is fused and joined by an arc formed by a DC voltage applied between the electrode and the base material, and image data of the fused portion captured by the image capturing means. Image processing means for analyzing the appropriate position of the electrode in the width direction, and movement control for controlling the movement of the moving means in the width direction according to the data of the appropriate position of the electrode calculated by the image processing means. In the welding position automatic copying control device including a means, the computing means is an arc inclination in which a pair of tangents to an upper contour of the image of the arc in the width direction sandwiching the electrode is a difference between angles formed by a reference line and each of the tangent lines. Difference in angle, boundary width difference between the image in the width direction of the arc and the molten pool and a pair of the groove walls, or lower end portions of a pair of hanging parts formed at the bottom of the image in the width direction of the molten pool A welding position automatic copying control device, characterized in that an appropriate position in the width direction of the electrode is calculated according to any one of the drooping amount deviations which are differences in vertical position. 2. The calculation means is a linear combination of an electrode displacement amount which is a difference between a midpoint between a pair of groove walls and an electrode position, an arc inclination angle difference, a groove wall boundary length difference and a drooping deviation. 2. The welding position automatic copying control device according to claim 1, wherein the electrode position correction value is calculated to be an appropriate position in the width direction.