Comprehensive strategies for defect mitigation and process optimisation in laser beam welding of aluminium alloys: a systematic review

In recent years, the demand for electric vehicles (EVs) has increased due to their crucial role in mitigating climate change. Moreover, laser beam welding (LBW) of aluminium alloys has been increasingly adopted in electric vehicle manufacturing because of its high precision and productivity; however, defects such as porosity and hot cracking remain major challenges. While previous studies have examined strategies for suppressing defects, existing reviews remain narrow in scope and typically address only a subset of these approaches. To bridge this gap, this study analyses the full spectrum of emerging mitigation strategies through a systematic evaluation of peer-reviewed journal articles published over the past decade (2015–2025). The review reveals that porosity is primarily governed by keyhole instability and hydrogen entrapment and is most effectively mitigated through controlled heat input, filler addition, and beam shaping. In contrast, hot cracking is strongly influenced by alloy volatility, eutectic segregation, solidification behaviour and tensile strain, with filler selection, heat input, beam oscillation, and weld-pool agitation showing greater effectiveness in mitigating crack formation. Quantitatively, porosity reductions to as low as 0.08% have been achieved using TiCnp/AA7075 powder fillers, while optimised shielding gas (Ar + 5%He) and heat input (~ 75 kJ m⁻¹) reduced porosity to 0.03% and 0.20%, respectively. Similarly, beam shaping decreased porosity from 2.49% to 0.222%, highlighting its strong defect-suppression capability. Numerical simulations and convolutional neural network (CNN)-based monitoring have advanced defect prediction and process understanding, however, real-time adaptive control remains limited by data availability and generalisation challenges. Future research should prioritise the controlled integration of non-autogenous (powder-assisted) LBW with welding-head delivery systems, multi-strategy optimisation, development of transferable machine-learning models, and the validation of fatigue and durability performance for powder-assisted LBW to enable reliable industrial deployment in automated EV manufacturing lines.