Metal fused filament fabrication (MF 3 ) combines fused filament fabrication and sintering processes to fabricate complex metal components. To design for MF 3 , an understanding of part geometry, printing parameters and material properties' effects on processability, part quality and ensuing properties is required. However, such investigation is a complex problem having several linked geometry, process and material variables to be considered that influence the process outcome. Moreover, such investigations through the experimental trial-and-error approach are costly and time-consuming, and sometimes not even feasible due to so many input variables involved. This study investigated the sensitivity of key output parameters toward each of the input parameters in MF 3 . FEA-based thermomechanical process simulations were used to estimate the process outcome in response to variable inputs, and a systematic procedure for sensitivity analysis has been successfully developed for the printing phase of the MF 3 process. Dimensionless sensitivity values for all output parameters were calculated in the response of each input parameter, which allows parameters with different units to be compared quantitatively with a single yardstick. Moreover, three different part geometries were studied to identify how the process sensitivity varies with part geometry. For each output parameter, the most influential input parameters were identified from the whole set of input parameters and their influence trends were evaluated for different part geometries. The present sensitivity analysis procedure is expected to be an invaluable tool not only for process parameters optimization but also for the development of material and part geometry for MF 3 , hence enabling design for MF 3 (DfMF 3 ).

No abstract

No abstract

Koopman spectral theory has provided a new perspective in the field of dynamical systems in recent years. Modern dynamical systems are becoming increasingly non-linear and complex, and there is a need for a framework to model these systems in a compact and comprehensive representation for prediction and control. The central problem in applying Koopman theory to a system of interest is that the choice of finite-dimensional basis functions is typically done apriori, using expert knowledge of the systems dynamics. Our approach learns these basis functions using a supervised learning approach where a combination of autoencoders and deep neural networks learn the basis functions for any given system. We demonstrate this approach on a simple pendulum example in which we obtain a linear representation of the non-linear system and then predict the future state trajectories given some initial conditions. We also explore how changing the input representation of the dynamic systems time series data can impact the quality of learned basis functions. This alternative representation is compared to the traditional raw time series data approach to determine which method results in lower reconstruction and prediction error of the true non-linear dynamics of the system.