AI-Enabled Post-Process Surface Inspection in Laser-Welded Al–Cu Battery Interconnects: A Critical Review

The growth of electric vehicles has increased the need for reliable, low-resistance Al–Cu battery interconnects. Laser welding provides localized heat input, high productivity, and automation potential, but Al–Cu joining remains limited by thermophysical mismatch, unstable energy coupling, and brittle intermetallic compound formation, which can lead to visible defects, hidden discontinuities, and performance variability. This review focuses on artificial intelligence (AI)-enabled post-process surface inspection of laser-welded Al–Cu battery interconnects. Process–structure–property relationships, defect mechanisms, process enablers, and in-process monitoring are discussed as supporting context to distinguish surface-visible evidence from attributes requiring complementary validation. The review critically examines controlled imaging, illumination and reflectivity effects, defect taxonomy, annotation, dataset design, AI task selection, and risk-based quality decisions. Classification, detection, segmentation, and anomaly detection are compared in terms of annotation requirements, traceability, metrics, and limitations. The main contribution is a post-process inspection framework that links controlled image acquisition, defect taxonomy, weld-level datasets, AI inference, benchmarking, and accept–review–reject decision logic with functional validation. The framework supports reproducible, risk-sensitive, and functionally validated quality assurance for Al–Cu battery welds while preserving expert review for uncertain cases and for the acceptance of welding technologies, process windows, and critical quality decisions.