Welding Stability Detection Using Weld Spot Center Variance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the stability of welding equipment are inaccurate due to interference from foreign matter and impurities, leading to defects such as misalignment of welding spots, and existing noise filtering methods are ineffective in monitoring welding states.
Innovation Solution
A method involving the acquisition and analysis of initial welding images using detection models to determine welding spot positions and center positions, which assesses the stability of the welding equipment by processing images of welded workpieces and calculating variance or standard deviation of welding spot centers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If mathematical models are used to detect welding stability by monitoring current and arc changes, then the welding state can be monitored, but the accuracy of determining stability is low due to inability to effectively filter various noises
Solution Approach 1:
The patent replaces traditional electrical monitoring methods (current and arc changes) with optical detection methods. Specifically, it uses a camera to capture welding images and employs deep learning algorithms to analyze these images for stability detection, substituting the mechanical/electrical sensing approach with an optical-computational approach that is more resistant to electrical noise and interference.
Solution Approach 2:
The patent introduces an intermediary processing layer between the raw welding signals and the stability determination. It uses image capture as an intermediary step, converting welding parameters into visual data, then applies deep learning models as an intermediary computational layer to extract meaningful stability information while filtering out noise and interference.
2Productivity
If welding equipment operates for a long time, then productivity increases, but foreign matter and impurities accumulate on the equipment interfering with geometric modules and vibrations, leading to welding errors and misalignment
Solution Approach 1:
The patent implements a feedback mechanism where welding images are captured in real-time or near-real-time during the welding process. The deep learning model analyzes these images to detect deviations in welding spot positions and equipment stability. This information feeds back to operators or control systems, enabling timely adjustments to maintain precision even during extended operation periods when contamination accumulates.
Solution Approach 2:
The patent performs preliminary detection of welding quality issues by analyzing images captured during the welding process. The deep learning model identifies potential misalignment or equipment instability before they result in defective welds, allowing for preventive corrective actions rather than waiting for errors to manifest in the final product.
Data Source
AI summary
The present application provides a method for determining a stability of a welding equipment. The method includes acquiring initial welding images of the welding equipment; obtaining at least one welding spot position of each of at least one welded workpiece in each initial welding image by processing the initial welding images; determining a welding center position of each welded workpiece based on the at least one welding spot position of each welded workpiece, and obtaining welding center positions of all welded workpieces comprised in the initial welding images; and determining a stability of welding equipment based on the welding center positions of all welded workpieces. The method determines whether the welding equipment is stable by analyzing the welding images, thereby improving an accuracy of a detection of a stability of the welding equipment.