No abstract

No abstract

Friction stir welding is an advanced joining technology that is particularly suitable for aluminum alloys. Various studies have shown that welding quality depends significantly on the welding speed and the rotational speed of the tool. It is frequently possible to detect an unsuitable setting of these parameters by examining the resulting surface defects, such as increased flash formation or surface galling. In this work, Artificial Neural Networks were used to analyze process data in friction stir welding and predict the resulting quality of the weld surface. For this purpose, nine different variables were recorded during friction stir welding of EN AW-6082 T6 sheets: the forces and accelerations in three spatial directions, the spindle torque, and temperatures at the tool shoulder and tool probe. In Case 1, the welds were assigned to the classes good and defective on the basis of a human visual inspection of the weld surface. In Case 2, the welds were categorized into the two classes on the basis of a surface topography analysis. Subsequently, three different major Artificial Neural Network architectures were tested for their ability to predict the surface quality: Feed Forward Fully Connected Neural Networks, Recurrent Neural Networks and Convolutional Neural Networks. The highest classification accuracy was achieved when Convolutional Neural Networks were used. Thus, the evaluation of the force signal transverse to the welding direction yielded the highest accuracy of 99.1% for the prediction of the result of the human visual inspection. The result achieved for the prediction of the topography analysis was an accuracy of 87.4% when the spindle torque was evaluated. Using all nine different process variables to predict the topography analysis, the accuracy could be improved slightly to 88.0%. The sampling rate of the spindle torque was varied between 40 Hz and 9600 Hz and no significant influence was determined. The findings show that Convolutional Neural Networks are well suited for the interpretation of friction stir welding process data and can be used to predict the resulting surface quality. In future work, the results are to be used to develop a parameter optimization method for friction stir welding.

Preliminary studies have shown the superiority of convolutional neural networks (CNNs) compared to other network architectures for determining the surface quality of friction stir welds. In this paper, CNNs were employed to detect cavities inside friction stir welds by evaluating inline measured process data. The aim was to determine whether CNNs are suitable for identifying surface defects exclusively, or if the approach is transferable to internal weld defects. For this purpose, 120 welds were produced and examined by ultrasonic testing, which was the basis for labeling the data as “good” or “defective.” Different types of artificial neural network were tested for predicting the placement of the welds into the defined classes. It was found that the way of labeling the data is significant for the accuracy achievable. When the complete welds were uniformly labeled as “good” or “defective,” an accuracy of 98.5% was achieved by a CNN, which was a significant improvement compared to the state of the art. When the welds were labeled segment-wise, an accuracy of 79.2% was obtained by using a CNN, showing that a segment-wise prediction of the cavities is also possible. The results confirm that CNNs are well suited for process monitoring in friction stir welding and their application enables the identification of various defect types.

Friction stir welding is an advanced joining technology that is particularly suitable for aluminum alloys. Various studies have shown a significant dependence of the welding quality on the welding speed and the rotational speed of the tool. Frequently, an inappropriate setting of these parameters can be detected through an examination of the resulting surface defects, such as increased flash formation or surface galling. In this work, two different learning-based algorithms were applied to improve the surface topography of friction stir welds. For this purpose, the surface topographies of 262 welds, which were performed as part of ten studies, were evaluated offline. The aim was to use reinforcement learning and Bayesian optimization approaches to determine the most appropriate settings for the welding speed and the rotational speed of the tool. The optimization problem was solved using reinforcement learning, specifically value iteration. However, the value iteration algorithm was not efficient, since all actions and states had to be iterated over, i.e., each possible parameter combination had to be evaluated, to find the best policy. Instead, it was better to solve the optimization problem directly using the Bayesian optimization. Two approaches were applied: both an approach in which the information from the other studies was not used and an approach in which the information from the other studies was used. On average, both the Bayesian optimization approaches found suitable welding parameters significantly faster than a random search algorithm, and the latter approach improved the result even further compared with the former approach. Future research will aim to show that optimization of the surface topography also leads to an increase in the ultimate tensile strength.