Today, the quality of welded seams is often examined off-line with either destructive or non-destructive testing. These test procedures are time-consuming and therefore costly. This is especially true if the welds are not welded accurately due to process anomalies. In manual welding, experienced welders are able to detect process anomalies by listening to the sound of the welding process. In this paper, an approach to transfer the “hearing” of an experienced welder into an automated testing process is presented. An acoustic measuring device for recording audible sound is installed for this purpose on a fully automated welding fixture. The processing of the sound information by means of machine learning methods enables in-line process control. Existing research results until now show that the arc is the main sound source. However, both the outflow of the shielding gas and the wire feed emit sound information. Other investigations describe welding irregularities by evaluating and assessing existing sound recordings. Descriptive analysis was performed to find a connection between certain sound patterns and welding irregularities. Recent contributions have used machine learning to identify the degree of welding penetration. The basic assumption of the presented investigations is that process anomalies are the cause of welding irregularities. The focus was on detecting deviating shielding gas flow rates based on audio recordings, processed by a convolutional neural network (CNN). After adjusting the hyperparameters of the CNN it was capable of distinguishing between different flow rates of shielding gas.

This research presents a hybrid approach to generate sample data for future machine learning applications for the prediction of mechanical properties in directed energy deposition-arc (DED-Arc) using the GMAW process. DED-Arc is an additive manufacturing process which offers a cost-effective way to generate 3D metal parts, due to its high deposition rate of up to 8 kg/h. The mechanical properties additively manufactured wall structures made of the filler material G4Si1 (ER70 S-6) are shown in dependency of the t_8/5 cooling time. The numerical simulation is used to link the process parameters and geometrical features to a specific t_8/5 cooling time. With an input of average welding power, welding speed and geometrical features such as wall thickness, layer height and heat source size a specific temperature field can be calculated for each iteration in the simulated welding process. This novel approach allows to generate large, artificial data sets as training data for machine learning methods by combining experimental results to generate a regression equation based on the experimentally measured t_8/5 cooling time. Therefore, using the regression equations in combination with numerically calculated t_8/5 cooling times an accurate prediction of the mechanical properties was possible in this research with an error of only 2.6%. Thus, a small set of experimentally generated data set allows to achieve regression equations which enable a precise prediction of mechanical properties. Moreover, the validated numerical welding simulation model was suitable to achieve an accurate calculation of the t_8/5 cooling time, with an error of only 0.3%.

In addition to the unique weld seam properties, friction stir welding (FSW) includes specific challenges such as comparatively high static and dynamic forces and torque during welding. In the literature review to date, the design of the tools is mostly based on empirical values, which can result in over- and undermatching. The aim of this study is to systematically analyze the dynamic tool damage and the respective failure mechanisms in relation to the tool dimensions and process temperature, accounting for weld seam length and quality. The determination of partial tool damage enables classification of the maximum tolerable tool life, considering the impact of process temperatures on permissible stresses of FSW tools made of H13 steel. The results indicate that dynamic stresses can be significantly affected by rotational speed and welding speed. Linear damage accumulation was used to predict the maximum tolerable tool life by summing up partial damage over time, which has never been used in the context of friction stir welding on the basis of experimental measurement data. For this purpose, forces and torques were converted into stresses on the probe and initially compared with analytically determined S–N curves for the main stresses during welding. Although the resulting weld seam lengths currently indicate a clear overestimation of the tool life, they allow for an estimation of the tool damage depending on the used probe diameter and welding speed. Particularly when a rotational speed of 2500 min ⁻¹ is considered, an eleven times overestimation of tool life results for the 5 mm probe and eight times for the 6 mm probe. The most favorable outcomes were observed when the 5 mm probe was operated at 4000 min ⁻¹ , with an estimated tool life of 238 m and an experimental result of 213 ± 35 m, as an exact determination of the tool life was possible in this instance. A sensitivity analysis with regard to the recording frequency demonstrates a strong dependence on the pre-processing of the measurement data. For instance, adjusting the recording frequency and using a Butterworth low-pass filter enhances the prediction by approximately 47% (6 mm probe at 2500 min ⁻¹ ). The incorporation of experimentally determined S–N curves could further enhance the precision of the prediction in the course of further investigations. The experiments were carried out with a force-controlled robotized welding setup from Grenzebach Maschinenbau GmbH in which AA 6060 T66 sheets with a thickness of 5 mm were joined.

Zusammenfassung: Die additive Fertigung mittels DED-Arc-Verfahren ist durch die hohen Fertigungszeiten ein kostenintensiver Prozess. Die klassischen Qualitätsprüfverfahren, die der Fertigung nachgelagert sind, geben erst nach Fertigstellung des Werkstückes Aufschluss über die erreichte Qualität. Nacharbeiten sind häufig ausgeschlossen, sodass bei unzureichender Qualität das Bauteil dem Ausschuss zugeführt wird. Die Anwendung von in situ Qualitätskontrollen ermöglicht es, Unregelmäßigkeiten frühzeitig bereits während des eigentlichen Aufbauprozess zu erkennen und eine direkte Nacharbeit kann stattfinden. Die akustischen Emissionen werden bereits erfolgreich von erfahrenen Schweißern als Qualitätsmerkmal für den Prozess eingesetzt. In diesem Beitrag soll mit Hilfe der akustischen Signale, die während des Schweißens entstehen, die Prozessqualität vorhergesagt werden. Dabei liegt der Fokus auf der Erkennung des Schutzgasflusses und eventuell vorhandener Oberflächenverunreinigung durch Öl. Dies kann maßgeblich zur Bildung von Poren innerhalb der entstehenden Schweißraupen beitragen. Zur Erkennung dieser Parameter wird die Verwendung eines Convolutional Neural Networks (CNN) zur Auswertung des emittierten Luftschalls diskutiert. Als wesentliches Merkmal zur Klassifizierung werden die, in der Sprecher- und Spracherkennung verbreiteten, Mel-Cepstralkoeffizienten (MFCC) herangezogen. Des Weiteren wird der Einfluss der Netzwerkparameter des CNN auf die Klassifizierungsgüte des resultierenden Netzwerkes dargestellt. Es zeigt sich, dass die ausschließliche Verwendung der MFCC dem CNN ermöglicht Prozessabweichungen zu detektieren. Aufgrund der hohen Informationsdichte der MFCC gegenüber dem STFT-Spektrum bieten erstere die Möglichkeit die Größe des verwendeten CNN erheblich zu reduzieren.