This study explores the application of deep learning (DL) to evaluate welder qualification specimens. The aim is to establish an objective assessment method and reduce inspectors' workload. The study employed a convolutional neural network (CNN) based on the ResNet architecture for image classification, focusing on medium-thickness bend test specimens from the Japan Welding Engineering Society (JWES) examination. Initial binary classification experiments distinguishing between “pass” and “fail” demonstrated that accuracy improved with training sample size, achieving over 90% on average. A confidence score, defined from CNN output values, enabled quantitative evaluation of classification reliability and the identification of specimens that could potentially be misclassified. Subsequently, multi-class classification was conducted to approximate real testing standards. Due to the presence of data imbalance, the analysis was conducted on six representative defect categories instead of the original 13. The accuracy exceeded 70% with 40 images per class and further improved to 76.7% after additional training iterations. Furthermore, the augmentation of the dataset, even under imbalanced conditions, led to a substantial enhancement in performance, culminating in 97.1% accuracy when all available images were utilized. The proposed confidence score demonstrated efficacy in multi-class settings, enhancing reliability-informed decision-making. These findings demonstrate that CNN-based approaches can accurately and objectively evaluate welder qualification test specimens. The proposed methodology could improve inspection efficiency, reduce examiner workload, and contribute to the development of standardized, automated evaluation systems for welding certification.

View full abstract

This paper deals with the ductile crack growth in the heat affected zone of butt welded joints which are made by multi-layer CO2 gas shielded arc welding and made with steel for building structure. The purpose of this study is to evaluate the ductile crack growth in heat affected zone under large cyclic deformation. The test welded joint whose heat affected zone has the inhomogeneity of strength and ductility was fabricated, and in order to confirm the ductile crack growth behavior in the heat affected zone, the large deformation cyclic loading test for notched element specimens taken from the heat affected zone was carried out. Then, FEM analysis simulating the loading test was carried out. It was found that the ductile crack growth behavior in the heat affected zone can be evaluated by using the ductile damage model proposed by the authors. As a result of FEM analysis, it was indicted that the ductile crack growth behavior in the heat affected zone was greatly affected by the strength inhomogeneity. FEM analysis was also carried out using the assignment of material properties to the heat affected zone as a parameter, and it was found that the ductile crack growth behavior can be evaluated even if the assignment of material properties to the heat affected zone is simplified, if the effect of the strength inhomogeneity on the strain distribution can be reproduced.

View full abstract

Dissimilar joining of carbon-fiber-reinforced thermoplastics (CFRTP) and metallic materials has received increasing attention for reducing the weight of transportation equipment. Although practical joint and temperature-dependent evaluations have been conducted on joints obtained via spot joining and thermocompression bonding, respectively, the dominant factors governing joint strength in relation to the resin flow during spot joining are insufficiently understood. This study investigated the influence of plastic flow of molten resin derived during dissimilar friction stir spot welding (FSSW) of CFRTP to a 6000-series Al alloy on the strength and fracture behavior of the resulting joint. The relationship between the flow of the molten resin during the FSSW process and the joint strength was evaluated by measuring the flow distance and cross-tension strength. This strength tended to increase with increasing flow distance of the molten resin during the joining process and subsequently decreased despite a gradual increase in flow distance. The joint cross-section contained large voids within the CFRTP in the tool-plunged region, and the flown molten resin was sufficiently bonded to Al owing to the penetration of the resin into the surface microstructure in the outside the tool-plunged region. Fracture and thermal analyses revealed that the fracture progressed from the boundary between the molten CFRTP and the CFRTP matrix, which was formed as a result of insufficient fusion caused by inadequate heat input to the flowing resin, leading to reduced wettability. Thus, this study demonstrated that the weldability of Al alloy/CFRTP joints during FSSW can be improved by maintaining the interface temperature during joining from the melting point of the CFRTP resin matrix to the pyrolysis-initiation temperature and enabling the CFRTP to efficiently flow outward.

View full abstract

Toward the realization of a carbon-neutral society, the use of high-tensile steel plates, which combine lightweight and high strength, is expanding in the body frames of electric vehicles (EVs) and hybrid electric vehicles (HEVs). Accordingly, triple-layer joint structures consisting of thin/thick/thick plates are becoming increasingly common. However, uneven heat balance makes it difficult to ensure sufficient weldability on the thin plate side while suppressing expulsion on the thick plate side. To address this issue, this study proposed a new resistance spot welding method that combines pulse current waveform control and automatic off-time control. This method improves weldability on the thin plate side by applying a high peak current, while suppressing excessive nugget growth on the thick plate side by providing the off-time. Regarding the nugget formation process and expulsion suppression mechanism of this welding method, macrostructural observations and welding CAE analysis revealed effective expansion and growth of the corona bond and the existence of high heat generation areas on the thin plate side. Furthermore, the effect of expanding the optimum welding current range was verified using test pieces, and the effects of expanding the weld diameter and suppressing expulsion were demonstrated through application to actual A-pillar components on a mass production line.

View full abstract