Data Analysis of Aluminum Alloy Defect Online Detection to Provide Data Support for AI Model

The application of AI technology is becoming increasingly widespread across various domains. However, its application in the field of internal crack detection in aluminum alloys remains uncommon. This scarcity can be attributed to the heavy reliance of AI models and algorithms on vast amounts of data, which are challenging to obtain through conventional experimental approaches. To address this issue and provide substantial data for AI models, this study proposes a novel constraint force testing method based on bending deformation, aimed at enabling online detection of internal cracks. By establishing a comprehensive model and conducting meticulous data analysis, we investigate the influence of cracks on constraint forces. The findings reveal that even small internal cracks induce fluctuations in the Y-axis constraint force of the pressing roller, ranging between 553.1-631.0 N. Furthermore, the Mises stress plots and the maximum Mises stress across the entire model exhibit dynamic alterations as the specimen is in motion, with varying crack positions. Notably, the peak Mises stress throughout the analysis occurs when the crack lies directly beneath the compression roller. Through a combined approach of modeling and numerical calculations, this study provides empirical evidence supporting the establishment of AI models for online crack detection. It elucidates the patterns of evolving constraint forces and Mises stress in relation to crack positions. These findings contribute to the development of efficient and accurate AI-based crack detection systems, addressing the limitations of traditional methods.