Automatic welding method for thick-walled pipe complex groove

Abstract

The application provides an automatic welding method for a complex groove of a thick-walled pipe and relates to the technical field of robot automatic welding. The method comprises the following steps: establishing a welding bead submodel according to a wire feeding speed, a welding speed deposition coefficient and a welding wire diameter, combining a groove model determined according to a size parameter to calculate a layer number of a filled groove and a welding bead number, so as to complete welding bead model establishment; then, a coordinate system is established according to the welding bead model, the relationship between the stroke of a motor and the pulse value is measured and determined, and multiple data related to the groove, the sub-welding bead and the welding bead are correlated with each other based on the coordinate system, so as to establish an automatic welding motion model, and welding is performed according to the automatic welding motion model. The method has the technical advantages of high welding precision, convenient welding process and good welding result quality.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot automatic welding, and particularly relates to an automatic welding method for a thick-walled pipe complex groove. BACKGROUND

[0002] With the continuous development of China's industrial production mode, the welding technology as the "industrial tailor" is gradually transforming towards flexibility, refinement and intelligence. In the fields of water supply engineering, petrochemical industry, chemical industry, power industry, agricultural irrigation and urban construction, the welding quality of thick-walled pipes which bear the role of liquid and gas transportation directly determines the service life and stability of the pipes, and plays a decisive role in the cost of a long time span.

[0003] At present, the position of the welding torch before welding mostly needs to rely on the experience of manual judgment, and the welding process is relatively complex. During welding, real-time manual fine adjustment is needed through visual correction to ensure that the position of the welding torch and the arc starting point have a suitable distance, so that the welding quality is very dependent on the technical level of the operator, and the precision is limited. Even if the size of the current during welding is used to judge the arc related information, there will be problems such as large noise interference and unstable data affecting the quality of the welding results. SUMMARY

[0004] In view of the technical problems existing in the prior art, the purpose of the present application is to provide an automatic welding method with higher precision, more convenient welding process and better welding result quality.

[0005] To achieve the above-mentioned purpose, the present application provides a welding method for a thick-walled pipe complex groove, which is applied to a welding robot comprising a welding torch, a swing shaft, a crawling shaft, a wire feeding shaft and a welding torch shaft; the swing shaft is used to control the displacement of the welding torch parallel to the pipe wall; the crawling shaft is used to control the longitudinal displacement and transverse displacement of the welding robot parallel to the pipe wall; the wire feeding shaft is used to deliver the welding wire; and the welding torch shaft is used to control the displacement of the welding torch perpendicular to the pipe wall.

[0006] The welding method comprises the following steps:

[0007] establishing a groove model according to the shape and size parameters of the groove cross section;

[0008] establishing each sub-welding bead model according to the shape and size parameters of each sub-welding bead cross section, in combination with the wire feeding speed, the welding speed, the deposition coefficient and the welding wire diameter;

[0009] calculating the number of welding layers, the layer height of each welding layer and the number of sub-welding beads for filling the groove in combination with the groove model and each sub-welding bead model, and correcting the layer height of each welding layer and the number of sub-welding beads according to the calculation results to establish a welding bead model;

[0010] A coordinate system is established based on a welding bead model to determine the starting point coordinates of the welding torch of each sub-welding bead;

[0011] The welding robot is installed at the welding position, and a pulse is inputted. The controller controls the motor to drive the swing shaft, the crawling shaft, the wire feeding shaft and the welding torch shaft to displace, measures the corresponding stroke of the displacement in the coordinate system, and records the pulse value of the pulse and the stroke data of the stroke;

[0012] The stroke data is polynomial fitted to establish a relationship between the stroke and the pulse value;

[0013] The relationship is associated with the shape and size parameters of the groove cross section, the shape and size parameters of the cross section of each sub-welding bead, the number of welding layers of the filled groove, the layer height of each modified welding layer and the number of sub-welding beads, based on the coordinate system, to obtain an automatic welding motion model;

[0014] The welding robot is welded according to the automatic welding motion model.

[0015] Preferably, for any welding layer, a plurality of sub-welding beads with parallelogram cross sections and one sub-welding bead with trapezoidal cross section can be formed. Among them, the parallelogram is equal in height to the trapezoid.

[0016] Optionally, the shape of the cross section of the sub-welding bead can also be rectangular and triangular.

[0017] Preferably, the coordinate system can adopt a three-dimensional vertical coordinate system. The rectangular coordinate system is the most commonly used coordinate system type, which is convenient for understanding and remembering. In addition, a cylindrical coordinate system or a polar coordinate system can also be used.

[0018] Preferably, before the polynomial fitting of the stroke data, preprocessing can be performed: removing the intercept term and transposing the two columns of data composed of the stroke data and the pulse value.

[0019] When the programmed process after the automatic welding motion model is established, QT platform can be used, or other programming languages such as C language can be used for programming.

[0020] It can be seen that in the technical scheme, firstly, the object to be welded, i.e., the complex groove of the thick-walled pipeline, is measured to determine the shape and size parameters of the groove cross section, and a model thereof is established; then the wire feeding speed and the welding speed of the current motor are measured and confirmed, the shape and size parameters of each sub-welding bead are confirmed in combination with the deposition coefficient and the welding wire diameter, the sub-welding bead model is established, and the number of layers and the number of welding beads required for filling the groove are further confirmed and can be corrected, and the welding bead model is established. Further, based on the welding bead model, a coordinate system is established, the motor is driven to run, the relationship between the stroke and the pulse value of the swing shaft, the crawling shaft, the wire feeding shaft and the welding gun shaft is detected and recorded, and the confirmed data are associated to construct an automatic welding motion model, thereby providing a guarantee for realizing high-precision automatic welding. In summary, compared with the traditional manual welding, the technical scheme provided by the application has the technical effects of higher welding precision, more convenient welding process and better welding result quality. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 FIG. 1 is a schematic diagram of the establishment process of an automatic welding motion model in an embodiment of the application.

[0022] Figure 2 FIG. 2 is a schematic diagram of the shape of a complex groove type in an embodiment of the application.

[0023] Figure 3 FIG. 3 is a schematic diagram of the arc motion trajectory of a welding gun in an embodiment of the application.

[0024] Figure 4 FIG. 4 is a schematic diagram of the welding gun station posture and the arc starting point in an embodiment of the application.

[0025] Figure 5 FIG. 5 is a schematic diagram of the filling of a groove cross section in an embodiment of the application.

[0026] Figure 6 FIG. 6 is a schematic diagram of the structure of a specific welding layer in an embodiment of the application.

[0027] Figure 7 FIG. 7 is a schematic diagram of the welding starting points of two welding beads in an embodiment of the application.

[0028] Figure 8 FIG. 8 is a schematic diagram of the groove cross section coordinate system involved in an embodiment of the application.

[0029] Figure 9 FIG. 9 is a schematic diagram of the data acquisition and recording software involved in an embodiment of the application.

[0030] Figure 10 FIG. 10 is a schematic diagram of the data acquisition and processing process in an embodiment of the application.

[0031] Figure 11is a fitted curve graph of the stroke of the wire feeding shaft in an embodiment of the present application.

[0032] Figure 12 is a quadratic equation fitted curve graph of the first segment of the stroke of the welding gun shaft in an embodiment of the present application.

[0033] Figure 13 is a quintic equation fitted curve graph of the second segment of the stroke of the welding gun shaft in an embodiment of the present application.

[0034] Figure 14 is a fitted curve graph of the simplified processing of the second segment of the stroke of the welding gun shaft in an embodiment of the present application. DETAILED DESCRIPTION

[0035] In the prior art, the judgment of the position of the welding gun before welding is mostly relied on artificial experience, the welding process is relatively complex, the precision is limited and unstable, and the quality of the welding result cannot be well guaranteed. To solve the above problems, the present application provides an automatic welding method for thick-walled pipe complex groove.

[0036] Next, the inventor will describe the implementation process and details of the technical solution in an embodiment, the process of which is as shown in Figure 1 .

[0037] Embodiment 1

[0038] Generally speaking, because the thick-walled pipe is subjected to a large amount of trace elements in the soil, the corrosive medium thereof will cause certain erosion to the plate and cause problems such as cracking, so the pipe is mostly medium plate; and in order to improve the butt joint efficiency and ensure the butt joint quality of the pipe, the pipe in different environments is processed by adopting a composite groove form, which can effectively solve the defects such as slag inclusion and incomplete penetration in the butt joint construction of the large wall thickness pipe on site. As shown in Figure 2 , the composite groove type generally needs multi-pass and multi-layer welding. Figure 2 Among them, C is the width of the weld cover beyond the edge of the groove; e1 is the distance between the highest point of the cover and the outer wall of the pipe; e2 is the distance between the lowest point of the root weld cover and the inner wall of the pipe. Both C and e1 exist in the pipe welding, but generally satisfy C≤2mm and e1≤3mm, so they can be ignored in the process of establishing the model.

[0039] The automatic welding method of thick-walled pipes is to set a track on the pipe to facilitate the circumferential movement of the automatic welding equipment along the track, and the welding torch is fixed on the equipment and follows the circumferential movement of the pipe. In this process, the pose of the welding torch can be regarded as being in a static state with the outer side of the pipe and being perpendicular to the plane of each layer. In pipe welding, the welding of a single welding bead is mostly the reciprocating movement of the welding torch, that is, the swing welding method, which can ensure the flatness of each layer of the welding surface, increase the welding torch swing, and also increase the width of each welding bead, which is beneficial to the welding of the groove with a large cross section similar to the groove, and can generally avoid uneven weld forming. The welding torch swing can reduce the weld height to a certain extent, greatly improve the mechanical properties of the weld metal, and is more conducive to the realization of the multi-layer and multi-pass welding model in the model. In the embodiment, the motion trajectory of the arc is as shown in Figure 3 , and the pose of the welding torch and the position of the arc starting point in the trapezoidal welding bead are as shown in Figure 4 .

[0040] A simplified filling diagram of the cross section of the groove is as shown in Figure 5 . The groove is composed of four parts, from top to bottom, they are the cover welding layer (I layer), there will be a little residual welding excess on the top of the layer; the filling welding upper layer (II layer) with a groove angle of a1; the filling welding lower layer (III layer) with a groove angle of a2 and the root welding (backing welding bead) layer (IV layer). Let the pipe wall thickness be T, the distance from the top of the filling welding lower layer to the inner wall of the pipe be H, the root face be b, the root welding layer height be L, and the gap between the two pipes be G. The initial welding bead and the intermediate welding bead cross sections of the filling welding upper and lower layers adopt parallelogram fitting, and the last welding bead cross section adopts isosceles trapezoidal fitting; the cover welding layer also includes the cover part of the outer wall, which can be regarded as a welding bead with the welding layer itself, and its cross section is fitted with a rectangle. Among them, the junction between each part of the model is regarded as an ideal plane, and they do not affect each other.

[0041] The welding filler filling in the groove adopts the equal-height filling strategy to establish the welding bead planning model. The core idea of the strategy is that the height of each layer in the filling welding upper and lower layers is equal, that is, the welding layer height of the upper layer is t1, and the welding layer height of the lower layer is t2. Taking the filling welding lower layer as an example, first, the width b2 of each sub-welding bead in a single layer needs to be selected according to engineering experience and actual project environment, that is, the height of all parallelograms in this part is t2, the bottom is b2, and the area is S a ; the last sub-welding bead cross section adopts trapezoidal fitting, and the area is S b . The number of layers needs to be determined by the layer height, and the layer height can be inversely deduced from the area of the parallelogram in a single layer and the sub-welding bead width. The cross-sectional area of a single sub-welding bead is determined by the welding wire filling amount per unit time, and the larger the wire feed rate, the larger the cross-sectional area that can be welded. Therefore, only the wire feed speed V s (m·min-1 ) and welding speed V h (mm.s -1 ) of the process parameters, S a can be obtained from equation (1) after converting and combining the two units, and the height of the parallelogram is obtained from equation (2) by back calculation, which is the layer height t2 of the layer.

[0042]

[0043]

[0044] In the formula, d is the diameter of the welding wire; a is the deposition coefficient. After the layer height is determined, the total number of layers filled by the filler welding under layer is

[0045]

[0046] The value of n is generally not an integer. If only the integer part n t is taken, the layer will not be completely filled at this time; if the integer is further taken, the welding wire filling will exceed the layer, resulting in a filling protrusion, which will affect the subsequent welding of the upper layer of the filler welding (the upper layer will directly affect the appearance of a large covering excess material). Therefore, the layer height t2 is modified

[0047]

[0048] The modified layer height is realized by adjusting the wire feeding speed V s or the welding speed V h .

[0049] After the layer height is determined, the layer is set as the i-th layer, where i∈[1, n t ], the cross-sectional view of the layer is shown in Figure 6 , and the upper base of the isosceles trapezoid is set as f i1 ; the lower base is set as f i2 . The total cross-sectional area of the i-th layer is

[0050]

[0051] After the cross-sectional area of the i-th layer is obtained, the ratio N of the cross-sectional area of the i-th layer to the area S ax of the modified parallelogram is the number of sub-welding beads in the i-th layer. Similarly, in general, N is also a non-integer. The width of the last trapezoidal sub-welding bead will be determined by the decimal part N a , and it is difficult to form an appropriate fusion width if it is too large or too small, so N still needs to be modified. According to the existing technology, it can be judged that when N a is about 0.4, a good weld can be formed. Therefore, when N a ≥ 0.4, the number of parallelogram sub-welding beads P of the i-th layer is the integer part Nb of N; when Na When <0.4, the number of parallelogram sub-welds P of the ith layer is Nb-1.

[0052] Similarly, the number of layers and the number of sub-welds of the upper layer of the filler weld can also be obtained by this method, but it should be noted that because the surface welding layer is the top layer of the groove, its layer height t0 cannot be too small or too large. Through experiments, the ratio of t0 to t1 is preferably controlled between 0.8-1.2 to obtain a better filling effect of the weld. If the ratio is too large or too small, it will cause problems such as incomplete filling of the groove or excessive excess height. Therefore, when correcting the layer height of the upper layer, the range of the ratio of t0 to t1 should be considered. In terms of weld calculation, the simplified models of the upper layer and the lower layer of the weld are fitted into the same structure, so the sub-weld model of the lower layer is directly applied. The root layer is relatively independent compared with the upper layer and the lower layer of the weld and the surface welding layer, and it is only a single layer and a single pass, so the number of layers and the number of sub-welds do not need to be calculated, and the corresponding process parameters are set to directly weld.

[0053] After determining the number of layers and the number of sub-welds of the weld, the welding starting point of the welding torch head is determined. The model in this embodiment is for the welding of thick-walled pipes, and the cross section of the groove around the center of the pipe is actually a hollow cylinder. Therefore, for a single sub-weld, after starting welding at the starting point and making a swing motion for one cycle, it can still return to the starting point plane position. In other words, after completing the welding of a sub-weld, it can be moved to another sub-weld in the same layer to continue welding.

[0054] For the two sub-welds used in this embodiment, in the parallelogram sub-weld, the welding starting point is the intersection of the two diagonals; in the trapezoidal sub-weld, the welding starting point is the intersection of the vertical bottom edge of the two diagonal intersection points and the line parallel to the starting point and the bottom edge of the parallelogram. The welding starting point of the two sub-welds is shown in Figure 7 .

[0055] In some embodiments, multiple sub-welds with rectangular and two triangular cross sections can also be used to fill a welding layer.

[0056] In order to prevent the welding torch head from colliding with the groove sidewall due to too large swing amplitude or causing incomplete fusion of the sidewall after welding due to too small swing amplitude, this embodiment sets a correction factor m1, m2 from the edge of the fitted sub-weld. Generally, m1, m2 are valued at 2-3 mm, and at the same time, it must satisfy that the single-side swing width is less than half of the sub-weld width. Taking the lower layer of the filler weld as an example, it needs to satisfy

[0057]

[0058]

[0059] If not, increase the correction factor appropriately to prevent wall collision.

[0060] A coordinate system as shown in FIG. 1 is established with the center of the gap between the inner walls of the two pipes (the center of the lower bottom of the bevel) as the coordinate origin, the straight line on which the bottom edge of the bevel lies as the y-axis, and the straight line perpendicular to the bottom edge as the z-axis. The track of the motor device climbing around the pipe is the x-axis, because the value of the x-axis is only affected by the climbing shaft in the motor and is relatively independent, so it is not included in the coordinate system. Figure 8

[0061] Alternatively, the coordinate system can also be established with another origin and axis direction, or a polar coordinate system or a cylindrical coordinate system can be used. Whether a different position and direction is used as the origin and axis direction of the coordinate system, or another coordinate form is directly used, the essence is only a difference in mathematical calculation, and it will not have an essential impact on the technical solution of the present application.

[0062] For example, the starting point coordinates (y, z) of the welding torch filling the first sub-welding pass in the first layer of the lower layer of the weld are

[0063]

[0064] As can be seen, the coordinate position of the welding torch on the y and z axes of the welding pass is obtained by adding or subtracting the swing size to the starting point.

[0065] After the coordinate system is established, a motor motion model is constructed based on the coordinate system.

[0066] The motor driver drives the motor to move by giving a pulse value. Therefore, after fixing a pulse equivalent (1000 in this embodiment), the relationship between the stroke of the motor (i.e. the actual distance of movement) and the given pulse value must be known. The motion trajectory of the motor itself has a certain error that can be ignored, and after assembling the welding torch head and other equipment, the error will increase. Therefore, in this embodiment, the relationship between the two is calculated to obtain more accurate results.

[0067] Specifically, after determining the direction, pulse equivalent, acceleration and other parameters through the test program written in the controller (Ubuntu version of Linux), a pulse value is input to drive the motor to move a certain component in a certain direction. In this embodiment, the parameter settings are as shown in FIG. 3. Subsequently, the actual distance of movement is measured using a high-precision vernier caliper, and the pulse value and actual distance are recorded and stored in a table in.csv format for subsequent data processing. Figure 9

[0068] Alternatively, the actual distance of movement can also be achieved using methods such as trajectory monitoring measurement.

[0069] ​​In the embodiment, the stroke of the motor includes the displacement stroke of five mechanical components in the welding process. Among them, the stroke of the swing shaft refers to the actual distance of the arc movement when the welding gun is welding; the stroke of the creep shaft is divided into two parts, which respectively refer to the actual distance of the motor creeping in the x-axis and y-axis directions; the stroke of the wire feeding shaft refers to the length of the welding wire consumed in welding, that is, the length of the welding wire fed in; the stroke of the welding gun shaft refers to the stroke of the movement accumulated by the vertical displacement of the welding gun when changing the welding layer, which mainly includes the movement distance in the y-axis and z-axis directions. It is obvious that the relationship between the stroke of the welding gun shaft and the pulse value is nonlinear, while the relationship between the stroke of the other four shafts and the pulse value is linear.

[0070] In the embodiment, the curve fitting toolbox (CurveFitting) of Matlab is used to fit the curve.

[0071] In some embodiments, to intuitively display the relationship between the stroke of the five mechanical components and the pulse value, Python is used to preprocess the data: subtract the minimum value in the stroke and the pulse value, that is, remove the intercept term, and transpose the two columns of data. Matlab reads the transposed data and imports the processed data into a new table. The above data acquisition and processing process is shown in Figure 10 .

[0072] After importing the data of the table into Matlab, input “cftool” in the command line window to open the fitting toolbox. In the selection of fitting type, the five shafts use polynomial fitting (Polynomial), among which the four shafts with linear change use a first-degree equation with Degree 1. The fitting result is taken as an example for the stroke of the wire feeding shaft, and its fitting result is shown in Figure 11 .

[0073] Alternatively, in addition to using Matlab, Origin, Mathmatica, Maple and other tools can also be used for data fitting. Further, the preprocessing of the data can also be realized by other various ways, such as R language, C language and C++.

[0074] In the embodiment, the stroke of the welding gun shaft is fitted using a multiple equation with Degree n. The data cannot be directly expressed by an expression when fitting, so the data is divided and fitted in segments, so that both segments have good fitting results.

[0075] In the embodiment, through the fitting deduction of each value, it is finally determined that n of the two segments is 2 and 5, that is, the first segment of the stroke of the welding gun shaft is fitted by a quadratic equation curve, and the second segment is fitted by a quintic equation curve. The fitting results are shown in Figure 12 , Figure 13 .

[0076] The fitting error square sum (SEE) tends to 0 and the correlation coefficient square value (R-Square) tends to 1, and the fitting effect is better. After the Matlab processing, the function relationship between the actual value y and the set value x of the travel of the swing shaft (formula 9), the crawling shaft (formula 10), the wire feeding shaft (formula 11) and the welding torch shaft (formula 12, 13) is

[0077] y = 0.1945 * x (9)

[0078] y = 0.0366 * x (10)

[0079] y = 0.2706 * x (11)

[0080] y = 0.0016 * x 2 -0.3267 * x + 88.43 (x ∈ (0, 90)) (12)

[0081]

[0082] In the embodiment, although the crawling shaft is annularly moved on the pipe, the essence is still the linear motion of the gears driven by the two servo DC motors, and therefore the five shafts can be regarded as linear motion trajectories.

[0083] In the test of the measurement distance, the inching mode is adopted, that is, after the pulse value (units) of a fixed amount is input, the motor driver drives the motor to displace (unit: mm) and measures the travel. The specific pulse value is the product of the time value of inching and the pulse equivalent. Except for the welding torch shaft, the conversion between the travel and the pulse value of the motor is shown in Table 1.

[0084] Table 1: Correspondence table of pulse value and travel

[0085]

[0086] In the conversion process of the welding torch shaft, if the conversion is completely performed according to formula (10), it is too cumbersome. Therefore, the simplified processing is performed: the quintic equation of the second segment is fitted as a combination of three linear equations; the first segment is still fitted as a quadratic equation. The relationship image of the two after the simplification is as shown in Figure 14 .

[0087] As shown in formula (13), the relationship between the actual value y and the set value x of the travel of the welding torch shaft in the second segment is Figure 14

[0088] y = -0.2386 * x + 93.1 (x ∈ (100, 160)) (14)

[0089] y = -0.1096 * x + 71.66 (x ∈ (170, 490)) (15)​

[0090] y = -0.2484 * x + 139.6 (x e (500, 560)) (16)

[0091] It can be deduced from the formula that the stroke of the welding torch shaft in the four intervals divided by the pulse value and the conversion of the pulse value are shown in Table 2.

[0092] Table 2 Correspondence table of the stroke of the welding torch shaft and the pulse value

[0093]

[0094] According to the above processing mode, the establishment of the motor motion model is completed based on the relationship between the stroke of the motor and the pulse value.

[0095] Further, the relationship is associated with the shape and size parameters of the groove cross section, the shape and size parameters of each sub-weld cross section, the number of welding layers of the filled groove, the layer height of each modified welding layer and the number of sub-welds, and the automatic welding motion model is established. Further, the model is programmed and applied to the control of the thick-walled pipeline welding crawling welding robot to realize automatic welding of the pipeline.

[0096] In the embodiment, the programming of the model is specifically programmed by using the QT platform. In addition, those skilled in the art can also use other tools for programming to realize the programming of the model, such as C language.

[0097] It can be seen that the measurement and confirmation of the size parameters of the groove, the confirmation of the wire feeding speed and the welding speed, the measurement and confirmation of the relationship between the stroke of the motor and the pulse value, and the establishment of the three models to finally construct the automatic welding motion model are the key to guarantee that the welding can overcome the technical problems of insufficient precision and poor welding quality caused by manual welding, that is, the problems of inaccurate manual correction of the welding torch position by artificial vision and large noise interference and instability of the arc related information judged by the current size are solved. A set of practical automatic welding method is provided for the relatively complex composite groove in thick-walled pipeline welding, and the automatic welding process is more convenient and fast. In short, the technical scheme provided by the present application has the advantages of higher welding precision, more convenient welding process and better welding quality.

[0098] The above is only a preferred embodiment of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method of welding a complex groove of a thick-walled pipe, characterized by, The welding method is applied to a welding robot comprising a welding torch, a swing axis, a crawl axis, a wire feeding axis and a torch axis; the swing axis is used to control the displacement of the welding torch parallel to the pipe wall; the crawl axis is used to control the longitudinal displacement and the transverse displacement of the welding robot parallel to the pipe wall; the wire feeding axis is used to deliver the welding wire; and the torch axis is used to control the displacement of the welding torch perpendicular to the pipe wall; The welding method comprises: establishing a groove model according to the shape and size parameters of the groove cross section; establishing each sub-bead model according to the shape and size parameters of each sub-bead cross section, in combination with the wire feeding speed, the welding speed, the deposition coefficient and the welding wire diameter; calculating the number of welding layers for filling the groove, the layer height of each welding layer and the number of sub-beads, and correcting the layer height of each welding layer and the number of sub-beads according to the calculation results, so as to establish a bead model; wherein the cross section of the cover welding layer is fitted as a rectangle; each welding layer of the lower welding layer comprises a plurality of sub-beads with parallelogram cross sections and one sub-bead with trapezoidal cross section, and the parallelogram and the trapezoid are of the same height, and the formula is as follows: ； ； wherein is the area of the parallelogram, is the diameter of the welding wire, is the wire feed speed, is the deposition coefficient, is the welding speed, is the height of the weld layer, is the length of the base of the parallelogram; for each sub-bead, a correction factor m1 and m2 are set at a distance from the edge of the fitted sub-bead; m1 and m2 are 2-3 mm, and the single-side swing width is less than half of the sub-bead width; establishing a coordinate system based on the bead model and determining the starting point coordinates of each sub-bead; wherein in the parallelogram sub-bead, the welding starting point is the intersection of two opposite sides; and in the trapezoidal sub-bead, the welding starting point is the intersection of the vertical bottom side of the intersection of two opposite sides and the straight line parallel to the starting point and the bottom side of the parallelogram; the fitting mode of the upper welding layer is the same as that of the lower welding layer; installing the welding robot at the position to be welded, inputting a pulse, controlling the displacement of the swing axis, the crawl axis, the wire feeding axis and the torch axis by the controller, measuring the corresponding stroke of the displacement in the coordinate system, and recording the pulse value of the pulse and the stroke data of the stroke; preprocessing the stroke data: removing the intercept term and transposing the two columns of data composed of the stroke data and the pulse value; performing polynomial fitting on the stroke data to establish the relationship between the stroke and the pulse value; associating the relationship with the shape and size parameters of the groove cross section, the shape and size parameters of each sub-bead cross section, the number of welding layers for filling the groove, the layer height of each welding layer and the number of sub-beads after correction, based on the coordinate system, to obtain an automatic welding motion model; the welding robot performs welding according to the automatic welding motion model.

2. A method of welding a complex groove of a thick walled pipe according to claim 1, characterized in that, The establishment of the coordinate system based on the bead model comprises: establishing a three-dimensional rectangular coordinate system based on the bead model.

Images