Abstract—Resistance Spot Welding (RSW) is a key technology for joining car body parts. During the process, the upper and lower electrodes of the welding gun are subject to wear, which affects the electric current flow and thus, the quality of the spot weld. To counteract this effect, the electrode tip is dressed after a predefined number of spot welds. The frequency for tip dressing is based on experience and does currently not directly reflect the actual state of the electrode. This work provides an approach to inline electrode state monitoring to determine the ideal and demand-based point in time for electrode tip dressing based on process data. We avoid costly and time-intensive experimental labeling of electrode states and utilize changes in the dynamic electrical resistance between dressing cycles to represent the electrode wear state. To describe the changes in the dynamic electrical resistance curve, new features such as the peak time delay are calculated and visualized. We evaluate our approach with a real-world data set that stems from a dynamic and complex environment out of a series production line, which, in contrast to laboratory data, ensures a successful application of the proposed methods in an industrial setting.
 

Index Terms—industrial data analytics, resistance spot welding, electrode wear state monitoring, feature construction, data mining

Cite: Samiha Durnagöz, Marco F. Huber, Mathias Mayer, and Peter Reimann, "An Approach to Inline Monitoring of the Electrode State in Resistance Spot Welding," International Journal of Electrical and Electronic Engineering & Telecommunications