In-situ Monitoring Techniques and Closed-loop Control Strategies in Additive Manufacturing

Abstract

Additive manufacturing (AM) enables the fabrication of complex components with high design freedom, but its layer-by-layer processing nature often leads to defects such as porosity, cracking, warping, residual stress, and dimensional inaccuracy. To improve process stability and part quality, this paper reviews in-situ monitoring techniques and closed-loop control strategies in AM. The study first summarises major AM processes and explains why real-time observation is necessary during material deposition, melting, and solidification. It then compares three main monitoring modalities: optical monitoring for surface morphology and geometric deviation, thermal monitoring for melt-pool temperature and heat distribution, and acoustic monitoring for process instability and defect-related events. The paper further discusses how machine learning methods can support defect detection, process-state classification, and quality assessment. More importantly, it emphasises that effective closed-loop control requires transforming monitoring signals into interpretable control variables, such as melt-pool temperature, cooling rate, layer height, or acoustic features. Finally, the paper identifies future directions, including multi-modal sensing, control-oriented feature extraction, physics-data hybrid modelling, and real-time edge computing.