SUMMARY

Wire Arc Directed Energy Deposition (DED) efficiently fabricates large metal parts but faces challenges in geometric accuracy, surface finish, and mechanical properties. This research aims to address these issues through a comprehensive investigation of machining-based hybrid manufacturing, predictive modeling, and advanced scientific machine learning (SciML) techniques. The primary objectives are to understand the effects of machining-induced deformation and thermal cycling on the microstructure and mechanical properties of metallic structures fabricated using Wire Arc DED, develop interlayer machining-based strategy for improving geometric accuracy, surface quality, and mechanical properties, and to develop robust and computationally efficient models for three-dimensional bead geometry prediction using SciML approaches. The methodology includes systematic experimentation of interlayer machining interventions, analyzing their impacts on grain refinement and mechanical properties, and leveraging physics-informed neural networks (PINNs) and transfer learning for accurate and computationally efficient three-dimensional part geometry prediction. This work aims to enhance the quality, consistency, and performance of additively manufactured metal parts, thereby advancing the field and enabling model-based process planning strategies.