I-HEAL AM: AI-Driven In-Situ Healing for Zero-Defect Additive Manufacturing

Objective

"Laser powder bed fusion (PBF) additive manufacturing (AM) is hampered by process-induced defects that degrade mechanical properties, preventing its use in safety-critical applications. Current in-situ monitoring systems are reactive, offering passive data logging but no autonomous, real-time correction. The research targets Inconel 718, an aerospace superalloy notoriously prone to such defects. The I-HEAL AM project will overcome this bottleneck by developing a truly autonomous “see-think-act-verify” workflow for proactive, in-situ healing with a <10-second overhead per layer. Its novelty is a fully closed loop of intelligent, surgical intervention and verification. A two-stage AI processes multi-modal sensor data: an unsupervised autoencoder flags anomalies, and a supervised CNN performs precise characterization. A custom software architecture then executes adaptive healing, including dynamic beam modulation for minor porosity and a novel 'Partial Recoat and Healing' strategy. This targeted approach minimizes heat input and detrimental residual stresses. The unique in-situ verification step confirms each repair, creating a data-rich “digital twin” for unprecedented part qualification. The system's success will be validated on an industrial demonstrator, improving functional performance. This project will establish a new benchmark for ""zero-defect"" manufacturing, enabling the qualification of reliable AM parts for European aerospace and energy sectors. A strategic secondment at the IAM@PoliTO, Politecnico di Torino, will advance the project’s science by integrating the methodology with the host's deep AM expertise to generate fundamental new insights into in-situ healing. For the researcher, this fellowship provides unique skills, paving the way for a leadership role in advanced manufacturing."