Laser-based directed energy deposition for metallic materials (DED-LB/M) is a versatile additive manufacturing (AM) technique that facilitates the deposition of advanced protective coatings, the refurbishment of degraded components, and the fabrication of intricate metallic structures. Despite the technological advancements and potential, the presence of process-induced defects poses significant challenges to the repeatability and stability of the DED-LB/M process, limiting its widespread application, particularly in industries requiring high-quality products. In-situ process monitoring stands out as a key technological intervention, offering the possibility of real-time defect detection to mitigate these challenges. Focusing on the DED-LB/M process, this review provides a comparative analysis of various in-situ monitoring techniques and their effectiveness in identifying process-induced defects. The review categorises different sensing methods based on their sensor data format, utilised data processing techniques, and their ability to detect both surface and internal defects within the fabricated structures. Furthermore, it compares the capabilities of these techniques and offers a critical analysis of their limitations in defect detection. This review concludes by discussing the major challenges that remain in implementing in-situ defect detection in industrial practice and outlines key future directions necessary to overcome them.

This study evaluates the feasibility of airborne acoustic source localisation (ASL) for in situ crack localisation in industrial laser-based directed energy deposition (DED-LB/M) fabricated structures. A four-microphone array combined with a Generalised Cross-Correlation with Phase Transform (GCC-PHAT) algorithm was used to estimate crack positions from time differences of arrival (TDOAs) extracted from raw acoustic emissions during multi-layer single-track fabrication. Prior to experimentation, the microphone array geometry was numerically optimised under industrial placement constraints by introducing controlled TDOA perturbations and minimising three-dimensional localisation uncertainty using alpha-shape volume analysis. Experimental validation was performed on six-layer single-track structures, with estimated crack positions compared against post-process microscopic measurements. Localisation errors ranged from 12 to 68 mm in the X-direction, 0.7–32 mm in the Y-direction, and 5–100 mm in the Z-direction. While horizontal localisation demonstrated centimetre-scale accuracy for most cracks, depth estimation exhibited greater variability. The results confirm that airborne ASL can provide meaningful spatial information regarding crack formation during DED-LB/M. However, localisation performance remains sensitive to TDOA estimation accuracy, microphone array constraints, and the complex acoustic environment inherent to the process. This work demonstrates the industrial feasibility of ASL for in situ crack investigation while highlighting the need for further advancements in array design and signal processing to achieve robust three-dimensional defect localisation in additive manufacturing systems.