Page 31 - IJAMD-1-2

International Journal of AI for
Materials and Design
AI-driven quality assurance in AM

Table 1. Overview of literature reviews on AM and AI
Review Focus Contributions Gaps
Plathottam et al. 5 Feature recognition for Feature recognition, process optimization Limited to manufacturability
manufacturability assessment in AM assessment
Kim et al. 9 Traditional QC methods in AM Detailed inspection methods, statistical process Limited to AI integration
control
Qin et al. 10 ML applications in AM Comprehensive ML applications, process General ML focus, not specific to QC
optimization
Bonatti et al. 11 ML applications in bioprinting QC Specialized in bioprinting, AI-driven QC Niche focus, limited to general AM
methods application
Jin et al. 12 Automated defect detection and Real-time defect detection, predictive modeling Limited to specific defect detection
prediction in AM
Bikas et al. 13 Design optimization for directed Design optimization, manufacturability Limited to specific AM process
energy deposition processes assessment
Stavropoulos et al. 14 Automated feature recognition for Feature recognition, process optimization Limited to manufacturability
manufacturability assessment in AM assessment
Wang et al. 15 Overview of ML applications in AM Overview of ML applications, future directions General overview, not specific to QC
Stavropoulos et al. 16 Chatter detection in machining Chatter detection in machining, AI-based Limited to machining applications
operations methods
Stavropoulos et al. 17 Impact of AM process complexity on Modeling in AM, complexity analysis Limited to the modeling aspect
modeling
Rojek et al. 19 Design considerations in AM Design considerations, AM processes Limited to the design aspect
Current review AI-driven QC methods in AM AI applications across AM processes, practical Integrative overview, connecting AI
challenges, and future directions techniques with QC
Abbreviations: AI: Artificial intelligence; AM: Additive manufacturing; ML: Machine learning; QC: Quality assurance.

2.3. Build orientation effects mitigating process-induced defects require a combination
The orientation of a part during the AM process can of process monitoring, in-situ sensing, real-time feedback
have a significant impact on its mechanical properties, control, and post-processing techniques. Implementing
surface finish, and dimensional accuracy. Parts printed in robust process control strategies and defect detection
different orientations may exhibit variations in material algorithms can help identify and address defects early in
density, residual stress, and anisotropic behavior, making the manufacturing process, minimizing scrap rates and
it challenging to establish uniform quality standards. 18,21,23 ensuring consistent part quality.
Optimizing build orientation to minimize distortion, 2.5. Regulatory and certification challenges
improve mechanical performance, and enhance surface
quality requires a thorough understanding of process As AM technologies continue to evolve and gain
physics, computational modeling tools, and experimental acceptance in safety-critical industries such as aerospace
validation techniques. Moreover, incorporating build and healthcare, ensuring compliance with regulatory
1,2
orientation considerations into design guidelines and standards and certification requirements becomes
automated part orientation algorithms can help mitigate increasingly important. Establishing confidence in the
orientation-related defects and inconsistencies. quality, reliability, and traceability of AM parts requires
adherence to industry-specific regulations, quality
2.4. Process-induced defects management systems, and certification processes. 11,12
AM processes are susceptible to various defects and However, existing regulatory frameworks may not fully
anomalies that can compromise part quality and address the unique characteristics and complexities of
structural integrity. Common defects include porosity, AM, leading to ambiguity and uncertainty in regulatory
lack of fusion, warping, delamination, surface roughness, compliance.
and dimensional inaccuracies, which can arise from AI, particularly ML techniques, can play a significant
factors such as improper process parameters, material role in addressing these regulatory and certification
degradation, and thermal gradients. Detecting and challenges in several ways.
26

Volume 1 Issue 2 (2024) 25 doi: 10.36922/ijamd.3455