On-Line Inner Porosity Defect Detection of Aluminum Alloy Robotic Welding for Aerospace
ZHANG Zhifen1, ZHANG Linjie2, YANG Zhe1, REN Wenjing1, WEN Guangrui
( 1. Institute of Aero Engine, School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an 710049, China; 2. State Key Laboratory of Metal Material Strength, School of Materials Science and Engineering, Xi’an Jiaotong University, Xi’an 710049, China)
Abstract：Robotic argon tungsten arc welding (GTAW) of aluminum alloy is the main forming manufacturing method of key parts in aerospace, in which inner porosity is a typical defect. Due to its weak correlation to sensing data, accurate on-line porosity detection is still challenging. The porosity was qualitatively and quantitatively characterized in terms of distribution, composition and size using SEM and EDS. Two types of porosity were discovered as well as the importance of Mg during the process of porosity generation and growth. The strong correlation between porosity and hydrogen spectrum line (656.28nm) was investigated and established using principal component analysis. A new real-time porosity detection for Al alloy was proposed by means of PCoefHS (PCA Coefficient of HI Spectrum). Finally, different levels of porosity were visualized by combining PCA–Tsne. The verification shows that three types of welding seams with no porosity, moderate porosity and chain-type severe porosity can be effectively detected and visualized.
2019, 
