Systems and methods for welding vehicle component assemblies

Patent metadata

Field	Value
Publication number	US-2024286232-A1
Application number	US-202318114312-A
Country	US
Kind code	A1
Filing date	Feb 27, 2023
Priority date	Feb 27, 2023
Publication date	Aug 29, 2024
Grant date	—

Abstract

A system for welding vehicle component assemblies includes a measurement sensor configured to scan first and second vehicle components, a welding apparatus configured to weld the first vehicle component and the second vehicle component together, and at least one processor configured to execute computer-executable instructions to access scan data of the first and second vehicle components, generate a CAD model of an as-scanned assembly of the first and second vehicle components, obtain input parameters associated with one or more features of the first and second vehicle components, generate predicted optimal welding parameters, using the machine learning model, based on CAD model and the input parameters, and control the welding apparatus to perform at least one welding operation on the first and second vehicle components according to the predicted optimal welding parameters.

First claim

What is claimed is:

1 . A system for welding vehicle component assemblies, the system comprising:
a measurement sensor configured to scan a first vehicle component and a second vehicle component on a mechanical handling device; a welding apparatus configured to weld the first vehicle component and the second vehicle component together; memory configured to store a machine learning model configured to output predicted optimal welding parameters, virtual computer-aided design (CAD) design data for multiple vehicle components, and computer-executable instructions; and at least one processor configured to execute the computer-executable instructions to:
access scan data of the first vehicle component and the second vehicle component from the measurement sensor;
generate a CAD model of an as-scanned assembly of the first vehicle component and the second vehicle component according to the scan data;
obtain input parameters for the machine learning model, the input parameters associated with one or more features of the first vehicle component and the second vehicle component;
generate predicted optimal welding parameters, using the machine learning model, based on CAD model and the input parameters; and
control the welding apparatus to perform at least one welding operation on the first vehicle component and the second vehicle component according to the predicted optimal welding parameters.

2 . The system of claim 1 , wherein the at least one processor is configured to:
access a virtual design CAD assembly of the first vehicle component and the second vehicle component according to the CAD design data; and determine an adjusted CAD assembly based on a combination of the CAD model and the virtual design CAD assembly, the adjusted CAD assembly indicative of one or more manufacturing differences between the first vehicle component and a specified virtual design of the first vehicle component, and between the second vehicle component and a specified virtual design of the second vehicle component; wherein generating the predicted optimal welding parameters includes generating the predicted optimal welding parameters based on the adjusted CAD assembly.

3 . The system of claim 2 , wherein the at least one processor is configured to model a predicted weld distortion of a welded assembly including the first vehicle component and the second vehicle component, after completion of the at least one welding operation, using finite element analysis.

4 . The system of claim 3 , wherein:
the at least one processor is configured to compare the predicted weld distortion to one or more specified weld distortion thresholds; and controlling the welding apparatus includes controlling the welding apparatus to perform the at least one welding operation in response to the predicted weld distortion being less than the one or more specified weld distortion thresholds.

5 . The system of claim 3 , wherein the at least one processor is configured to:
compare the predicted weld distortion to one or more specified weld distortion thresholds; obtain a set of alternative predicted optimal welding parameters from the machine learning model in response to the predicted weld distortion being greater than the one or more specified weld distortion thresholds; and in response to a number of multiple sets of alternative predicated optimal welding parameters being equal to a specified iteration value:
determine which one of the multiple sets of alternative predicted optimal welding parameters has a lowest predicted weld distortion; and
control the welding apparatus to perform the at least one welding operation according to the determined set of alternative predicted optimal welding parameters having the lowest predicted weld distortion.

6 . The system of claim 3 , wherein the at least one processor is configured to:
access scan data of the welded assembly including the first vehicle component and the second vehicle component as captured by the measurement sensor or a post-weld scan sensor; generate a welded assembly CAD assembly according to the scan data of the welded assembly; and calculate a distortion difference according to a comparison between the welded assembly CAD assembly and the adjusted CAD assembly, the distortion difference indicative of a difference between actual distortion in the welded assembly and the predicted weld distortion.

7 . The system of claim 6 , wherein the at least one processor is configured to supply the distortion difference to the machine learning model to perform reinforcement learning on the machine learning model.

8 . The system of claim 1 , wherein the predicted optimal weld parameters include at least one of parameters configured to minimize distortion of a welded assembly including the first vehicle component and the second vehicle component, or parameters configured to minimize a predicted deviation between the welded assembly including the first vehicle component and the second vehicle component and a nominal engineering CAD model.

9 . The system of claim 1 , wherein the predicted optimal welding parameters include one or more welding parameters and one or more welding operation parameters.

10 . The system of claim 9 , wherein the one or more welding operation parameters include at least one of a number of tack welds, a location of tack welds, a number of each weld section, a location of each weld segment, a length of each weld segment, or a direction of each weld segment.

11 . The system of claim 9 , wherein:
the welding apparatus includes a gas metal arc welding (GMAW) welding apparatus, and the one or more welding parameters include at least one of a weld amperage, a weld volume, a weld frequency a travel speed, a weave schedule, a work distance, a work angle, a travel angle, a joint root offset, a shielding gas type, a shielding gas flow rate, a filler wire type or a filler wire diameter.

12 . The system of claim 1 , wherein the input parameters for the machine learning model include at least one of metal combinations of the first vehicle component and the second vehicle component, a total weld length per area or block, a ratio of yield strength in materials of the first vehicle component and the second vehicle component, or a ratio of thickness in in materials of the first vehicle component and the second vehicle component.

13 . The system of claim 2 , wherein the at least one processor is configured to calculate a residual stress of the first vehicle component and the second vehicle component using finite element analysis, based on the adjusted CAD assembly.

14 . The system of claim 1 , wherein the input parameters include at least one of dimensional measurements of the first vehicle component and the second vehicle component as scanned by the measurement sensor, or a dimensional deviation between a nominal engineering CAD model and the first vehicle component and the second vehicle component as scanned by the measurement sensor.

15 . The system of claim 1 , wherein the measurement sensor includes at least one of a camera, a point-scanning sensor, or a line-scanning laser.

16 . The system of claim 1 , wherein the mechanical handling device includes at least one of a conveyor, a pick and place module, a robot, a part holding fixture, a part nest, a table, and a robot gripper.

17 . The system of claim 1 , wherein the welding apparatus includes a robotic welding torch and a part holding fixture.

18 . The system of claim 1 , wherein the welding apparatus includes a robotic fixtureless assembly having

Assignees

• Gm Global Tech Operations Llc

Inventors

• Ghassemi-Armaki Hassan
• Spicer John Patrick
• Mellas Spyros P
• Engel Mark
• Saez Miguel Arturo
• Carlson Blair E

Classifications

• B23K31/02

  relating to soldering or welding · CPC title

• G06F2119/02

  Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA] · CPC title

• B23K2101/006

  Vehicles · CPC title

• G05B19/4097

  characterised by using design data to control NC machines, e.g. CAD/CAM (G05B19/4093 takes precedence) · CPC title

• G06F30/27

  using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model · CPC title