Towards robust dissimilar joining and overall process monitoring in resistance welding

Leveraged by political, economic, and ecologic incentives, manufacturing industry shifts towards multi-material structures, combining material properties, optimising costs, enabling lightweight design, and reducing emissions during the production and the lifetime of the product. In automotive production, lightweight hybrid design reduces fuel consumption, hence the emission of greenhouse gasses. Amongst other measures, this is achieved by combining steel and aluminium alloys in body-in-white manufacturing, as well as using both copper and aluminium alloys for optimised weight, cost and performance of electrical assemblies, such as battery packs for electric vehicles. This inevitably introduces the need for joining highly dissimilar materials. To safeguard the production output, high-value production lines call for traceability and process monitoring, becoming increasingly important when implementing more complex dissimilar material combinations.

Joining sheet-like metal components in mass production environments is often performed by resistance (spot) welding processes, known for their cost- efficiency, automatability and high throughput. Yet, robust joining of highly dissimilar material combinations by such processes has proven challenging due to discrepancies in their physical and metallurgical properties. In contrast to joining similar alloys, resistance spot welding of such material combinations is not characterised by the formation of a common molten zone (nugget) joining the workpieces after solidification, but relies on a metallurgical connection based on the formation of brittle intermetallic compounds at the dissimilar interface, compromising the joints mechanical properties, hence hindering their practical application.

Despite the introduction of alternative joining methods in manufacturing industry, such as self piercing riveting and friction welding processes, economical and operational benefits of resistance welding encourage developments of process adaptations for highly dissimilar joining. On the one hand, these comprise alterations to the heat balance of dissimilar welding stacks, compensating for discrepancies in physical material properties and optimising the temperature field at the dissimilar interface, thereby influencing physical and metallurgical interactions between the joining members. On the other hand, adaptations in chemical composition at the dissimilar interface effectively alter the formation rate, morphology and composition of intermetallic compounds at this interface, hence joint properties. Yet, none of these developments has led to (wide) adoption by industry.

The present work contributes to this field by the development of a novel resistance welding process variant for highly dissimilar materials, that is, materials significantly differing in physical (i.e. melting temperature and electrical and thermal resistance) and metallurgical properties. The process, locked projection welding, extends the metallurgical bond by a mechanical connection, hereby achieving robust joints characterised by ductile failure and significant mechanical strength in otherwise unfavourable loading directions. The novel process is experimentally validated on copper-aluminium connections, similar to those found in electrical applications.

Additionally, in the current industrial landscape where resistance welding processes are employed in the manufacturing of high-end products, these processes are pushed to the limits of their operation range to gain efficiency as well as by joining new (combinations of) materials, increasing their susceptibility to defects. This, combined with the increased importance of traceability systems in modern production lines has put research on in-process monitoring systems high on the agenda. The latter has led to the development and commercialisation of in-line resistance spot weld monitoring systems based on dynamic resistance, electrode reactions and ultrasonic signals. However, the downside of these systems is that their working principle limits the scope of imperfections they can observe, as well as the resistance welding process variants they are applicable on.

Acoustic emission has been brought forward in multiple academic publications and patent applications as a potential tool for resistance spot welding process monitoring. As the captured signals are influenced by a wide range of phenomena occurring around the studied welding setup, the technique has been proposed for the detection of multiple imperfections. However, quantitative predictions of nugget diameter, arguably the most significant quality indicator of a resistance spot weld, have never been presented. As such, despite the implementation of acoustic emission based quality monitoring techniques in multiple industries, no industrial implementations are to be found in resistance welding industry.

This work presents an experimental programme revealing correlations between joint diameter and acoustic emission signals in resistance welding independent from variations in machine parameters, and links those to physical events related to joint formation during the welding process. This, together with the validation of expulsion detection in industrial production environments, prove the potential of acoustic emission as a monitoring tool for resistance welding processes.