Multi-signal quality monitoring of aluminium resistance spot welding using principal component analysis

The current migration to lighter materials in car bodies, such as aluminium, has resulted in significant challenges for joining in production. Resistance Spot Welding (RSW) is the primary sheet metal joining technique in the automotive industry due to its quick cycle time, low cost and high strength. However, aluminium RSW suffers from problems with quality consistency compared to steel, requiring more frequent interventions. This results in a higher cost in production through increased cycle times and the use of consumable electrodes. To address this issue, a new multi-signal quality monitoring technique is proposed to allow for complete real-time quality monitoring of aluminium spot welds in production. The proposed solution utilises multiple signals during welding and an efficient algorithm using Principal Component Analysis to determine the signal shapes of interest. It was found that an RMSE of 119N could be achieved when predicting the strength of aluminium spot welds using multiple signals, which is approximately ±5% of the mean strength of the welds and an improvement on previous attempts.