Efficient Navigation of Additive Manufacturing Process Space via Data-Driven and Bayesian Inference Methods

P. Balachandran
University of Virginia,
United States

Keywords: additive manufacturing, 3D printing, cerebral palsy, upper extremity, orthosis, arm, hand


In our group, we have been exploring two independent computer modeling routes to interface with the additive manufacturing (AM) process of functional and structural materials. The first approach takes a data-driven motivation, where the underlying assumption is that there is no simple analytical or numerical model to guide the AM process towards the desired regions in the vast search space. We apply machine learning (ML) methods to establish the relationship between the AM parameters and the property of interest. The ML model predictions, along with the uncertainties, are used to efficiently inform the AM process. The second approach combines experimental data with numerical or analytical models (e.g., Eagar-Tsai Models) within the framework of Bayesian inference. One of the objectives is to rapidly construct printability maps that delineate various sources of porosities in laser AM processes as a function of melt pool geometries. In this talk, I will focus on specific examples that highlight the potential of both approaches.