Influence of workpiece geometry and natural frequencies on ultrasonic metal welding

Abstract

Ultrasonic metal welding is a well-established solid state joining process for electrical applications. The process relies on the friction between workpieces and welding tools for joint formation. This friction is generated by the process force and the ultrasonic oscillation of the welding tools imposed on the workpieces. At such high frequencies, the occurrence of resonances in actual workpiece geometries is not surprising. It is known that critical dimensions in length and width lead to nearly no bond, depending on the welding frequency and the mechanical properties of the material. In real applications, this limits the possible designs of terminals and leads to extensive testing of clamping devices. It is also known that machine learning (ML) models for quality prediction based on power signals or tool oscillation can account for changes in welding position. In this study, we investigated the impact of part resonance and antiresonance on horn and anvil oscillation, power consumption and bond strength to identify typical behaviors induced by the workpieces. The influence of material thickness and roughness was considered, and numerical analysis of the natural frequencies of the workpieces was conducted. It can be shown that the results allow a distinction between the welding positions and workpiece geometries without directly measuring the oscillation patterns of the workpieces, allowing a simple validation of geometry weldability and clamping device in applications. Furthermore, the investigation allows the knowledge based specific deduction of signal parameters for future ML models, allowing a consideration of welding position and workpieces geometry with reduced test data.