CN-122007725-A - Programming-free multilayer multi-channel welding robot device and welding method thereof

CN122007725ACN 122007725 ACN122007725 ACN 122007725ACN-122007725-A

Abstract

The application relates to the technical field of welding robot control and discloses a programming-free multilayer multichannel welding robot device and a welding method thereof, wherein the method comprises the steps of constructing a region to be filled of a current layer to be welded based on groove boundary parameters, target layer upper boundary parameters and current completed layer surface profile, and determining effective width and area to be filled; the method comprises the steps of forming a candidate single-pass section template set through a prediction model according to candidate welding process parameters, generating candidate pass combinations based on area distribution, boundary constraint relation and side wall wetting risk, determining the transverse center position and welding sequence of each welding bead, combining a current layer of predicted surface profile and a terminal layer of target boundary to obtain target pass combinations, and generating welding tracks to control a robot and a welding machine to execute welding.

Inventors

SUN KE
ZHANG HUAJIAN
WEI ZENGKUN
WANG QUAN
ZHANG XIQI

Assignees

中钰匠鑫机械制造有限公司

Dates

Publication Date: 20260512
Application Date: 20260325

Claims (10)

1. A programming-free multi-layer multi-channel welding robot welding method is characterized by comprising the following steps: constructing a region to be filled of a current layer to be welded based on boundary parameters of a groove of the workpiece to be welded and boundary parameters on a target layer, combining the surface profile of the layer to be welded at present, and determining the effective width and the area to be filled corresponding to the region to be filled; predicting the single-pass effective cross-section area, the surface width and the residual height of a candidate welding bead through a prediction model according to the candidate welding process parameters to form a candidate single-pass cross-section template set; Judging the type of the current region to be welded layer based on the area distribution and boundary constraint relation of the region to be filled, and calculating the side wall wetting risk according to the effective residual boundary width, the welding gun posture allowance and the welding position category of the corresponding side of each candidate welding bead; generating candidate pass combinations of the current to-be-welded layer according to the to-be-filled area, the candidate single-pass section template set and the side wall wetting risk, and respectively determining the transverse center position and the welding sequence of each welding bead aiming at each candidate pass combination, wherein the transverse offset of the welding beads at the left side and the right side is independently determined; Calculating the residual area to be filled of the subsequent layer according to the predicted surface profile of the current layer corresponding to each candidate pass combination, and screening each candidate pass combination by combining the target boundary of the terminal layer to determine the target pass combination of the current layer to be welded; and generating a welding track based on the target pass combination and controlling the robot and the welding machine to execute welding.
2. The method according to claim 1, wherein the constructing the region to be filled of the current layer to be welded according to the boundary parameters of the groove of the workpiece to be welded and the boundary parameters of the target layer in combination with the surface profile of the current completed layer comprises: Taking the upper boundary of a target layer of the current to-be-welded layer as an upper boundary, taking the surface profile of the current completed layer as a lower boundary, and taking the groove boundaries on the left side and the right side as side boundaries to form a to-be-filled region of the current to-be-welded layer; And carrying out longitudinal segmentation on the area to be filled along the length direction of the welding seam, and respectively calculating the area to be filled and the effective width for each segmentation, wherein the effective width is the transverse distance after the unsafe area is deducted.
3. The method of claim 2, wherein determining the type of the region of the current layer based on the area distribution and boundary constraint relationship of the region to be filled comprises: When the effective residual boundary width of any side is smaller than the minimum arrangement width of the corresponding side welding bead, the current subsection is judged to be a side wall constraint leading area, when the proportion of the area to be filled in the middle part to the total area to be filled in the current subsection exceeds a preset threshold value, the current subsection is judged to be a middle volume leading area, and when one side is limited by the groove boundary and the other side is limited by the surface bulge of the leading welding bead, the current subsection is judged to be a transition area.
4. The method for welding a multilayer multi-channel welding robot without programming according to claim 3, wherein the prediction model is a structured industrial data regression model, and the structured industrial data regression model comprises a feature preprocessing module, a plurality of decision tree modules which are cascaded in sequence and an output analysis module; the feature preprocessing module performs normalization or box division processing on continuous input features, performs coding processing on discrete input features, sequentially fits the residual error of the previous round by the decision tree modules, and the output analysis module gathers the output of the decision tree modules to obtain a single-channel effective cross-sectional area, a surface width and a residual height.
5. A method of programming-free multi-layered multi-pass welding robot welding according to claim 3, wherein said calculating sidewall wetting risk from the effective remaining border width, gun pose margin, and welding position category of the corresponding side of each candidate bead comprises: And respectively establishing a side wall wetting risk value aiming at the left welding bead and the right welding bead, wherein the side wall wetting risk value is determined by the relation between the surface width of the corresponding side candidate welding bead and the effective residual boundary width of the corresponding side, the attitude allowance of the welding gun on the corresponding side and the welding position category, and the side wall wetting risk value of the left welding bead and the side wall wetting risk value of the right welding bead are used as input conditions of transverse offset and welding sequence.
6. The method according to claim 5, wherein generating a candidate pass combination of the current weld layer according to the area to be filled, the candidate single-pass cross-section template set and the sidewall wetting risk, and determining a lateral center position and a welding sequence of each weld bead for each candidate pass combination, wherein lateral offset amounts of the weld beads at the left and right sides are determined independently, comprises: Determining candidate channel number intervals according to the areas to be filled of the current welding layer, generating corresponding candidate channel combinations for each candidate channel number interval, respectively determining the center position of a middle channel and the transverse distance from the middle channel to a left welding channel and the transverse distance from the middle channel to a right welding channel for each candidate channel combination, and respectively correcting the transverse offset of the left welding channel and the right welding channel based on the distribution difference of the areas to be filled on the left side and the wetting risks of the side walls on the right side, wherein the transverse offset of the left welding channel and the transverse offset of the right welding channel are respectively calculated and determined according to the distribution of the areas to be filled on the corresponding sides and the wetting risks of the side walls on the corresponding sides, and the center position of the middle channel is used as a reference standard for determining the transverse positions of the welding channels on the left side and the right side.
7. The method of claim 6, wherein the welding sequence is determined based on zone type and sidewall wetting risk, comprising: When the current subsection is a side wall constraint leading area, a welding bead on one side with higher side wall wetting risk is firstly arranged, then a middle channel is arranged, and finally a welding bead on the other side is arranged.
8. The method of claim 7, wherein calculating the remaining area to be filled of the subsequent layer according to the predicted surface profile of the current layer corresponding to each candidate pass combination, and screening each candidate pass combination with the target boundary of the end layer, determining the target pass combination of the current layer to be welded comprises: And generating a current layer predicted surface profile according to the single-pass effective cross-sectional area, the surface width, the residual height and the transverse center position of each weld bead aiming at each candidate pass combination, calculating a residual region to be filled of the next weld layer based on the current layer predicted surface profile and the groove boundary parameter, and determining a target pass combination from each candidate pass combination by taking a terminal layer target boundary, a terminal layer residual width and a terminal layer pass arrangement condition as screening constraint.
9. The method of claim 8, wherein generating a welding trajectory based on the target pass combination and controlling the robot and welder to perform welding comprises: And generating a welding path center track according to the transverse center position of each welding bead, generating a swing superposition track according to the swing width, the swing frequency and the left and right end residence time proportion corresponding to each welding bead, synthesizing the welding path center track and the swing superposition track into a robot execution track, and controlling the robot and the welding machine to sequentially execute welding based on the welding sequence of each welding bead in the target pass combination and corresponding technological parameters.
10. A programming-free multi-layer multi-pass welding robot device and a welding device thereof, comprising: The to-be-filled region construction module is configured to construct a to-be-filled region of a current to-be-welded layer based on groove boundary parameters of the to-be-welded workpiece and boundary parameters on a target layer in combination with the surface profile of the current completed layer, and determine the effective width and the to-be-filled area corresponding to the to-be-filled region; the single-channel section prediction module is configured to predict the single-channel effective section area, the surface width and the residual height of a candidate welding bead through a prediction model according to the candidate welding process parameters to form a candidate single-channel section template set; The region judging and risk calculating module is configured to judge the type of the region of the current welding layer to be filled based on the area distribution and boundary constraint relation of the region to be filled, and calculate the side wall wetting risk according to the effective residual boundary width, the welding gun posture allowance and the welding position category of the corresponding side of each candidate welding bead; The pass combination generating module is configured to generate candidate pass combinations of the current to-be-welded layer according to the to-be-filled area, the candidate single-pass section template set and the side wall wetting risk, and respectively determine the transverse center position and the welding sequence of each welding bead for each candidate pass combination, wherein the transverse offset of the welding beads at the left side and the right side is independently determined; the pass combination screening module is configured to calculate the remaining area to be filled of the subsequent layer according to the predicted surface profile of the current layer corresponding to each candidate pass combination, screen each candidate pass combination by combining the target boundary of the terminal layer, and determine the target pass combination of the current layer to be welded; and the track generation and execution module is configured to generate a welding track based on the target pass combination and control the robot and the welding machine to execute welding.

Description

Programming-free multilayer multi-channel welding robot device and welding method thereof Technical Field The application relates to the technical field of welding robot control, in particular to a programming-free multilayer multichannel welding robot device and a welding method thereof. Background The multilayer multi-pass welding is widely applied to the filling and cover surface welding of thick plate butt joints, narrow-gap groove joints and large-thickness components. Existing robotic welding systems typically require operators to pre-complete weld pass number settings, track teaching for each layer, weld sequence arrangement, and swing parameter adjustment when performing such tasks. For the working conditions of regular groove shape, higher pairing precision and smaller change of the section of the welding seam, the method can meet the basic production requirement, but in actual production, the conditions of slight change of the groove along the length direction of the welding seam, fluctuation of pairing clearance, local convergence or relaxation caused by spot welding, local filling space change caused by heat input of preamble welding and the like often exist in the workpiece to be welded, and the filling requirements of different positions are difficult to consider by adopting a fixed-die-type welding bead planning method. In the prior art, one type of scheme mainly generates a multi-layer and multi-channel welding track according to preset groove parameters and an empirical process, and the other type of scheme relies on a visual measurement or contour detection device to acquire interlayer geometric information and then adjusts the welding track. The former scheme has higher requirements on the consistency of workpieces, when the local section is changed, the problems of unreasonable distribution of welding beads in certain areas, limited distribution of side wall positions, uneven residual space of a subsequent layer and the like easily occur, and the latter scheme can acquire more field information, but has more complex system composition and higher hardware cost and maintenance requirements. Especially in the multi-layer and multi-channel welding process, the effective widths, the areas to be filled and the side wall constraint relations of different positions of the same layer are not consistent, and if welding beads are still arranged in a symmetrical and fixed sequence mode, the conditions of transverse bias mismatch, welding sequence mismatch, unstable reservation conditions of the tail end layer and the like are easy to occur in a local area. Therefore, it is necessary to propose a robot welding control scheme suitable for programming-free multi-layer multi-pass welding, which combines groove boundaries, interlayer surface profiles and welding process parameters to coordinate the current region to be filled of the weld layer, candidate bead arrangement and subsequent layer reservation state. Disclosure of Invention The application aims to provide a programming-free multi-layer multi-channel welding robot device and a welding method thereof, which are used for solving the problems in the background technology. According to an aspect of the present application, there is provided a programming-free multi-layered multi-pass welding robot apparatus and a welding method thereof, including the steps of: constructing a region to be filled of a current layer to be welded based on boundary parameters of a groove of the workpiece to be welded and boundary parameters on a target layer, combining the surface profile of the layer to be welded at present, and determining the effective width and the area to be filled corresponding to the region to be filled; predicting the single-pass effective cross-section area, the surface width and the residual height of a candidate welding bead through a prediction model according to the candidate welding process parameters to form a candidate single-pass cross-section template set; Judging the type of the current region to be welded layer based on the area distribution and boundary constraint relation of the region to be filled, and calculating the side wall wetting risk according to the effective residual boundary width, the welding gun posture allowance and the welding position category of the corresponding side of each candidate welding bead; generating candidate pass combinations of the current to-be-welded layer according to the to-be-filled area, the candidate single-pass section template set and the side wall wetting risk, and respectively determining the transverse center position and the welding sequence of each welding bead aiming at each candidate pass combination, wherein the transverse offset of the welding beads at the left side and the right side is independently determined; Calculating the residual area to be filled of the subsequent layer according to the predicted surface profile of the current layer corresponding to each candidate pass combina