The quality of welds is critical to the safety of structures in construction, so early detection of irregularities is crucial. Advances in machine vision inspection technologies, such as deep learning models, have improved the detection of weld defects. This paper presents a new CNN model based on ResNet50 to classify four types of weld defects in radiographic images: crack, pore, non-penetration, and no defect. Stratified cross-validation, data augmentation, and regularization were used to improve generalization and avoid over-fitting. The model was tested on three datasets, RIAWELC, GDXray, and a private dataset of low image quality, obtaining an accuracy of 98.75 %, 90.255 %, and 75.83 %, respectively. The model proposed in this paper achieves high accuracies on different datasets and constitutes a valuable tool to improve the efficiency and effectiveness of quality control processes in the welding industry. Moreover, experimental tests show that the proposed approach performs well on even low-resolution images.

The tribological behavior of a new composite material containing graphene nanosheets (GNS) is presented. The composite material was obtained by Friction Stir Processing, using as metallic matrix the AA1050-H24 alloy. Different tool rotation and advancing speeds were tested in friction stir processing (FSP). The worn surfaces of obtained materials were analyzed by Scanning Electron Microscopy (SEM). Raman spectroscopy demonstrated that graphene reinforcements are successfully mixed into the aluminum matrix. The results proved the feasibility of using GNSs to obtain nanocomposites by FSP. The coefficient of friction of the aluminum alloy was 0.57, decreasing to 0.38 for the nanocomposite GNSs/AA1050. These values decrease for samples obtained at lower tool rotation speeds. The weight losses of the composites are less than that of unreinforced AA1050-H24 alloy for conditions with lower advancing speeds (40 and 60 mm/min) and 1120 rpm.