Application of multimodal process monitoring for enhanced quality control and defect detection in laser-based directed energy deposition

Abstract

Laser-based directed energy deposition (also known as laser material deposition, LMD) is an additive manufacturing (AM) process that uses laser radiation to create a melt pool on a component's surface while feeding filler material through a nozzle. LMD is established in various industrial applications for coatings, repairs, and AM. However, it involves numerous parameters, necessitating complex process management and expertise. Variable conditions, such as differing workpiece geometries, materials and varying powder quality, complicate the process, potentially leading to time-consuming and costly development efforts along with deviations in component quality. This work applies a multimodal process monitoring workflow to analyze and actively improve process behavior toward defect mitigation. Reference cubes built under nominal process parameters are compared to cubes deposited with varying laser power. During deposition, spatially synchronized data streams from a pyrometer, melt-pool camera, and optical coherence tomography are acquired. A laser scanner combined with post-build computed tomography provides 3D-defect masks that are co-registered with the process data. Extensive data analysis revealed a pronounced directional dependence. Based on these findings, additional test cubes are deposited using targeted, local parameter adjustments to mitigate the directional bias with an open-loop strategy.