Weld Surface Image Analysis for Spatter-Based Quality Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing welding process quality assessment techniques are complex, costly, and require significant investment, especially for industrial-scale production, due to the need for advanced equipment and software for real-time dynamic analysis of molten metal spatter during welding.
Innovation Solution
A method using a smartphone or similar image capture device for post-weld analysis, employing statistical processing and a convolutional neural network to classify and quantify spatter on the weld surface, reducing equipment and software complexity by analyzing images after the welding process is completed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time dynamic analysis of molten metal spatter during welding is performed using advanced equipment and software, then welding quality assessment accuracy is improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent uses a digital image copy of the weld surface instead of real-time physical analysis of molten metal. A photograph captures the spatter pattern, which is then analyzed through image processing to assess welding quality. This replaces complex real-time monitoring equipment with simple imaging devices while maintaining assessment accuracy.
Solution Approach 2:
The patent replaces mechanical/optical real-time analysis systems with a digital image processing system. Instead of using complex sensors and real-time optical devices to analyze molten metal dynamics, the system uses a photograph and computational algorithms to assess spatter characteristics, thereby reducing device complexity.
2Measurement precision
If real-time monitoring equipment and complex software are deployed for welding analysis, then welding performance determination accuracy is improved, but investment cost increases prohibitively for small structures
Solution Approach 1:
The patent employs inexpensive, readily available devices such as smartphone cameras or simple digital cameras instead of expensive industrial-grade real-time monitoring equipment. The temporary nature of the photograph (a single captured image) replaces the need for持续 expensive real-time monitoring systems, making the solution economically viable for small structures.
Solution Approach 2:
By using a digital image copy of the weld surface, the patent eliminates the need for expensive proprietary software and hardware. Standard image processing techniques can be applied to the photograph, allowing implementation on budget-friendly devices while maintaining accurate welding performance assessment.
3Device complexity
If post-weld image analysis is performed using simple devices, then device complexity and cost are reduced, but the ability to assess spatter during the welding process is lost
Solution Approach 1:
The patent performs the welding operation first, then captures an image of the completed weld for analysis. The spatter pattern is already formed on the solidified weld surface, providing sufficient information for quality assessment. This preliminary completion of the welding process followed by image capture eliminates the need for complex real-time monitoring during the actual welding operation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for determining the performance of a welding method carried out on a metal workpiece (1), in particular an electric arc welding or laser welding method, with the following steps: introducing one or more extracts of the initial image (2) each comprising at least one presumed projection (21), as input to at least one neural network, in particular a convolutional neural network, so as to classify the presumed projections as confirmed or unconfirmed projections, carrying out a second digital processing operation on the initial image (2) comprising the projections classified in step c) so as to determine at least one parameter representative of the quantity of confirmed projections chosen from the surface of one or more projections.