Investigation of Transport Phenomena in Microcasting Shape Deposition Manufacturing via Experiments Designed using Optimal Sampling

Abstract

A manufacturing technique called macrocasting is being developed as part of the Shape Deposition Manufacturing process. Using microcasting, multi-material metal artifacts can be crated by successively depositing layers of material into the desired configuration. By combining microcasting with machining operations and the use of sacrificial support materials, complex, multi-material geometries and composite structures can be manufactured. The advantage of the SDM process lies in directly generating free-form, near net-shape structures that traditional machining, casting or spraying methods cannot readily achieve. Ongoing research efforts seek to improve the microcasting process and to minimize manufacturing defects in the deposition layer. Numerical prediction and experimental testing are required to relate microcasting parameters with measurable deposition quality. Previous experiments focused on layers of material; the current work focuses on the individual drops that make up the layers. Adaptive Bayesian surrogate models, in conjunction with optimal sampling, provide and efficient method for selecting and evaluating experiments and exploring the influence of different application parameters. Sequential updating of surrogates along with optimal sampling for data collection improves surrogate accuracy in our multistage approach and reduces the amount of data collection required. The use of statistical techniques helps the development of an understanding of the relationship between process characteristics and the final quality of the deposition layer.