Advances in additive manufacturing enable virtually unlimited design complexity and thus require far more design decisions. Moreover, increased complexity and flexibility in a system add significant amounts of uncertainty associated with the larger number of design variables. This results in significant deviations between an original design and the printed product due to current limitations in the additive manufacturing process. Accordingly, there is an increasing pressure to enable the certification of additively manufactured parts by analysis without conducting customized testing for the printed parts. This award supports scientific investigations on the behavior of additively manufactured parts at different length scales under uncertainty. The knowledge generated from this research will play a large role in the design and certification of additively manufactured parts in the context of practical engineering systems. Ultimately, the results from this research will facilitate the delivery of high-quality additively manufactured parts with predictable performance, which is critical in aerospace, medical device, and automotive industries, among others.
The goal of this research is to enable engineers to certify additively manufactured parts according to desired properties across a broad range of processes and materials. This goal will be accomplished by enhancing our knowledge on the behavior of additively manufactured parts based on advanced statistical analysis with physical experiments. Thus, the objective of this research is to understand the relationship between mechanical behavior of fabricated parts and their inherent uncertainties in geometrical deviations (e.g., stacked position and length, and surface quality), and process parameters (e.g., build orientation, layer thickness, raster angle, air gap, contour width, etc.). Various sizes and types of cellular structures will be fabricated to analyze the range of microstructures capable of being produced by fused deposition modeling. The experimental results will be correlated to geometrical deviations, process variables and their uncertainties as well as mechanical properties. Appropriate representation schemes will also be identified for uncertainty parameters of manufacturing processes and parts. These uncertainties will be used to obtain mathematical formalisms by way of stochastic expansions, which will be synthesized into a multi-level, multi-scale model that describes the behavior of fabricated parts. The obtained model can be used to accurately describe the behavior of fabricated parts at different length scales under uncertainty. Ultimately, this additional effort will identify new knowledge regarding key uncertainties in process parameters that affect the performance of additively manufactured parts.
