Detecting Process Anomalies in the GMAW Process by Acoustic Sensing with a Convolutional Neural Network (CNN) for Classification

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