Search for: All records

Abstract Additive manufacturing (AM) techniques, such as fused deposition modeling (FDM), are able to fabricate physical components from three-dimensional (3D) digital models through the sequential deposition of material onto a print bed in a layer-by-layer fashion. In FDM and many other AM techniques, it is critical that the part adheres to the bed during printing. After printing, however, excessive bed adhesion can lead to part damage or prevent automated part removal. In this work, we validate a novel testing method that quickly and cheaply evaluates bed adhesion without constraints on part geometry. Using this method, we study the effect of bed temperature on the peak removal force for polylactic acid (PLA) parts printed on bare borosilicate glass and polyimide (PI)-coated beds. In addition to validating conventional wisdom that bed adhesion is maximized between 60 and 70 °C (140 and 158 °F), we observe that cooling the bed below 40 °C (104 °F), as is commonly done to facilitate part removal, has minimal additional benefit. Counterintuitively, we find that heating the bed after printing is often a more efficient process for facile part removal. In addition to introducing a general method for measuring and optimizing bed adhesion via bed temperature modulation, these results can be used to accelerate the production and testing of AM components in printer farms and autonomous research systems. 

Vat polymerization is a type of additive manufacturing that is used extensively to produce micro‐architected structures for mechanical applications, which brings the mechanical properties of photopolymerized resins into sharp focus. However, it is known that photopolymerization is sensitive to a number of factors, perhaps the most notorious of which is oxygen inhibition. Herein, the degree to which oxygen inhibition influences the macroscopic and microscopic properties of structures made using vat polymerization is explored. This work is motivated by an observation of lattices being >4 times softer in the experiment than predicted by simulation, which is hypothesized to be due to the material at the surface being incompletely cured. This hypothesis is supported by four‐point bending tests in which flexural modulus is found to increase with beam thickness. Nanoindentation and bulk compression studies show that this surface softening is present for three distinct resins. Importantly, it is observed that structures post‐print cured in nitrogen are stiffer than those post‐print cured in air, however, regardless of the post‐print curing environment, printing samples in the presence of oxygen makes them softer than samples photocured in nitrogen. Collectively, these results show the outsized influence of oxygen inhibition on micro‐architected structures realized using vat polymerization.

 

Bayesian optimization (BO) has been leveraged for guiding autonomous and high-throughput experiments in materials science. However, few have evaluated the efficiency of BO across a broad range of experimental materials domains. In this work, we quantify the performance of BO with a collection of surrogate model and acquisition function pairs across five diverse experimental materials systems. By defining acceleration and enhancement metrics for materials optimization objectives, we find that surrogate models such as Gaussian Process (GP) with anisotropic kernels and Random Forest (RF) have comparable performance in BO, and both outperform the commonly used GP with isotropic kernels. GP with anisotropic kernels has demonstrated the most robustness, yet RF is a close alternative and warrants more consideration because it is free from distribution assumptions, has smaller time complexity, and requires less effort in initial hyperparameter selection. We also raise awareness about the benefits of using GP with anisotropic kernels in future materials optimization campaigns.

 

While additive manufacturing (AM) has facilitated the production of complex structures, it has also highlighted the immense challenge inherent in identifying the optimum AM structure for a given application. Numerical methods are important tools for optimization, but experiment remains the gold standard for studying nonlinear, but critical, mechanical properties such as toughness. To address the vastness of AM design space and the need for experiment, we develop a Bayesian experimental autonomous researcher (BEAR) that combines Bayesian optimization and high-throughput automated experimentation. In addition to rapidly performing experiments, the BEAR leverages iterative experimentation by selecting experiments based on all available results. Using the BEAR, we explore the toughness of a parametric family of structures and observe an almost 60-fold reduction in the number of experiments needed to identify high-performing structures relative to a grid-based search. These results show the value of machine learning in experimental fields where data are sparse.