CN-120620222-B - Digital twin-based ship plane section welding multi-robot task distribution system

CN120620222BCN 120620222 BCN120620222 BCN 120620222BCN-120620222-B

Abstract

The invention discloses a digital twin-based ship plane segment welding multi-robot task distribution system, which relates to the technical field of intelligent manufacturing of ships, and comprises the following steps: and the four-dimensional digital twin body module is used for collecting workpiece assembly errors, welding seam positions and robot state data in real time and constructing a twin model comprising a physical engine layer, a process knowledge base and a dynamic topological graph. According to the digital twin-based ship plane segment welding multi-robot task distribution system, a welding line task topological graph is dynamically constructed by fusing workpiece assembly errors, robot motion states and welding process rules in real time through a digital twin body, and the multi-robot task accurate distribution is realized by combining an anti-collision auction algorithm. The space-time meshing cooperative mechanism is utilized to effectively resolve motion conflicts between the large gantry and the small gantry mechanism, zero interference of welding paths is guaranteed, and the local path optimization module adjusts the posture and the dry elongation of the welding gun in a self-adaptive manner aiming at a narrow space, so that physical collision risks are eliminated.

Inventors

HU QING
YOU RUICHAO
ZHU JINTONG
YU DONG
LIU ZHE

Assignees

招商局金陵船舶(江苏)有限公司
唐山开元自动焊接装备有限公司

Dates

Publication Date: 20260508
Application Date: 20250804

Claims (8)

1. Digital twin-based ship plane segment welding multi-robot task distribution system is characterized by comprising: The four-dimensional digital twin body module is used for collecting workpiece assembly errors, welding seam positions and robot state data in real time and constructing a twin model comprising a physical engine layer, a process knowledge base and a dynamic topological graph; the dual-cycle optimization engine module dynamically distributes welding seam tasks through the global task distribution ring and generates a conflict-free path through the local path fine adjustment ring; the multiple verification module is used for carrying out digital twin previewing verification and physical-virtual data comparison on the optimization result to ensure that paths are free of conflict and the welding quality reaches the standard; the load balancing module is used for calculating the task load index of each robot in real time and dynamically adjusting task allocation; the global task allocation ring adopts an anti-collision auction algorithm: each robot bidding a welding seam task with an efficiency value, wherein the efficiency value is calculated based on the welding speed, the welding seam matching degree and the path conflict risk; Discretizing the travel of the large gantry travelling mechanism into space-time grids, and dynamically avoiding path conflicts through a virtual time window; the anti-collision auction algorithm is as follows: When the distance between the two robots is lower than a safety threshold value, forcibly reducing the speed and inserting a waiting time window; The weight coefficient of the efficiency value is dynamically adjusted according to the type of the welding seam, and the vertical welding task is preferentially distributed to the robot with large Z-axis travel.
2. The digital twin based marine plane segment welding multi-robot mission distribution system of claim 1, wherein the four-dimensional digital twin body module comprises: the real-time data layer is used for capturing workpiece assembly errors and weld position deviations through fusion of the laser scanning sensor and the visual sensor; The physical engine layer integrates a multi-body dynamics simulation model, and previews the motion path of the robot to the influence of welding thermal deformation; Embedding SMARTWELD welding rules generated by the process knowledge base, and converting the welding leg range and the dry elongation constraint into an optimization objective function; and the dynamic topological graph takes a welding seam as a node and a robot reachable path as an edge, and the weight comprises welding time and quality risk coefficients.
3. The digital twin based marine plane segment welding multi-robot mission distribution system of claim 1, wherein the local path trim ring comprises: based on a self-adaptive fuzzy RRT algorithm, generating a collision-free path in a narrow space by combining a welding line type and a rotational freedom degree theta axis of a robot with +/-185 degrees; And the welding gun attitude optimizer is used for adjusting the dry extension length and angle of the welding gun to ensure the weld plumpness.
4. The digital twin based marine plane segment welding multi-robot mission distribution system of claim 1, wherein the multiple verification module performs: digital twin preview verification, namely injecting real-time error data into KCONG simulation environment, and detecting path deviation and joint moment overrun; and (3) comparing physical-virtual data, namely inverting the state of a molten pool through arc data of a welding power supply, comparing the state with a predicted value of a twin model, and triggering process parameter re-optimization.
5. The digital twin based marine plane segment welding multi-robot mission distribution system of claim 1, wherein the load balancing module: Calculating a robot load balancing index LBI, and reassigning tasks of the high-load robot to the idle units when the index is lower than a set threshold; The index is dynamically updated based on the ratio of the variance of the time length of each robot task to the average time length.
6. The digital twin based marine plane segment welding multi-robot mission allocation system of claim 1, further comprising a quality defect prediction unit: constructing a weld seam air hole and crack probability prediction model based on real-time arc data of a welding power supply; and when the defect probability exceeds the limit, suspending the task and triggering the local path fine tuning ring to re-plan.
7. The digital twin-based ship plane segment welding multi-robot task distribution system according to claim 1, wherein the system transmits laser scanning data and control instructions through a 5G communication module, and the task distribution response period is lower than a set threshold.
8. A method of multi-robot task allocation for welding of marine plane segments according to the system of any of claims 1-7, comprising the steps of: s1, capturing workpiece errors and robot states in real time through a digital twin body; s2, generating an initial task allocation scheme through a global task allocation ring; s3, optimizing the path and the posture of the welding gun through the local path fine adjustment ring; s4, performing double verification through digital twin previewing and physical data comparison, and then performing welding; And S5, dynamically adjusting task allocation based on the load balancing index.

Description

Digital twin-based ship plane section welding multi-robot task distribution system Technical Field The invention relates to the technical field of intelligent manufacturing of ships, in particular to a digital twin-based ship plane section welding multi-robot task distribution system. Background The ship plane sectional welding adopts robot assembly line operation, such as sectional construction and deck sectional welding assembly line, a plurality of Kawasaki RA005L robots are carried by a large gantry, a small gantry and other mechanisms, welding programs are generated by means of SMARTWELD and KCONG systems, and various welding operations such as flat fillet welding, vertical fillet welding and the like are realized. The system needs multiple robots to cooperatively finish the welding tasks of a large-size 15000mm long 15000mm wide and multiple types of welding seams, but the current task allocation depends on a preset program or simple scheduling, and is difficult to respond to workpiece assembly errors, robot running state changes and welding seam distribution differences in real time. This results in that multiple robots are prone to task overlapping and path collision, or part of robots are idle due to uneven load, so that not only welding efficiency is reduced, but also weld quality such as welding leg deviation and appearance defects may be affected due to interference, and it is difficult to meet the production requirements of high precision and high beat of ship welding, and a system capable of dynamically optimizing task allocation is needed to solve the above problems. Disclosure of Invention In order to achieve the purpose, the invention is realized by the following technical scheme that the digital twin-based ship plane segment welding multi-robot task distribution system comprises: The four-dimensional digital twin body module is used for collecting workpiece assembly errors, welding seam positions and robot state data in real time and constructing a twin model comprising a physical engine layer, a process knowledge base and a dynamic topological graph; the dual-cycle optimization engine module dynamically distributes welding seam tasks through the global task distribution ring and generates a conflict-free path through the local path fine adjustment ring; the multiple verification module is used for carrying out digital twin previewing verification and physical-virtual data comparison on the optimization result to ensure that paths are free of conflict and the welding quality reaches the standard; And the load balancing module is used for calculating the load index of each robot task in real time and dynamically adjusting the task allocation. Preferably, the four-dimensional digital twin body module includes: the real-time data layer is used for capturing workpiece assembly errors and weld position deviations through fusion of the laser scanning sensor and the visual sensor; The physical engine layer integrates a multi-body dynamics simulation model, and previews the motion path of the robot to the influence of welding thermal deformation; Embedding SMARTWELD welding rules generated by the process knowledge base, and converting the welding leg range and the dry elongation constraint into an optimization objective function; and the dynamic topological graph takes a welding seam as a node and a robot reachable path as an edge, and the weight comprises welding time and quality risk coefficients. Preferably, the global task allocation ring adopts an anti-collision auction algorithm: each robot bidding a welding seam task with an efficiency value, wherein the efficiency value is calculated based on the welding speed, the welding seam matching degree and the path conflict risk; and discretizing the travel of the large gantry travelling mechanism into a space-time grid, and dynamically avoiding path conflict through a virtual time window. Preferably, the local path fine tuning loop includes: based on a self-adaptive fuzzy RRT algorithm, generating a collision-free path in a narrow space by combining a welding line type and a rotational freedom degree theta axis of a robot with +/-185 degrees; And the welding gun attitude optimizer is used for adjusting the dry extension length and angle of the welding gun to ensure the weld plumpness. Preferably, the multiple verification module performs: digital twin preview verification, namely injecting real-time error data into KCONG simulation environment, and detecting path deviation and joint moment overrun; and (3) comparing physical-virtual data, namely inverting the state of a molten pool through arc data of a welding power supply, comparing the state with a predicted value of a twin model, and triggering process parameter re-optimization. Preferably, the load balancing module: Calculating a robot load balancing index LBI, and reassigning tasks of the high-load robot to the idle units when the index is lower than a set threshold; The index is dynamically