DEF-RSFA: an integrated nondestructive testing and reliability-aware SRAM design with fault analysis for additive manufacturing

Additive manufacturing (AM) enables the rapid fabrication of complex electronic packages and structural components; however, defects such as pores, cracks, voids, and interlayer discontinuities can compromise structural integrity and electronic reliability. This study proposes a novel Defect Evaluation and Reliability-Conscious SRAM Fault Analysis (DEF-RSFA) framework that integrates non-destructive defect inspection with circuit-level reliability assessment. The framework initially employs X-ray Computed Tomography (CT) to acquire three-dimensional volumetric representations of additively manufactured components. AI-assisted image analysis is then used to identify voxel-level defects and quantify their spatial extent and severity. The extracted defect characteristics are mapped to circuit stress parameters and subsequently evaluated using Static Random-Access Memory (SRAM) as a representative circuit-level reliability model. Based on defect severity, adaptive fault-mitigation strategies, including guard-banding, transistor resizing, and redundancy, are applied to reduce defect-induced reliability degradation. Fault-injection analysis is performed to evaluate bit-error rates, timing errors, and failure probabilities under different defect conditions. Experimental results demonstrate a defect detection coverage of 98.42%, while the reliability analysis indicates a strong relationship between defect severity and circuit failure probability. The proposed DEF-RSFA framework provides an interpretable and scalable approach for linking manufacturing defects with electronic reliability, supporting defect-aware design optimization and manufacturing-process improvement for aerospace, automotive, and high-performance computing applications.