Bayesian Sequential Experiments for Tensile Testing of Multi-material Fused Filament Fabricated Dogbones

This dataset contains raw tensile testing data of multimaterial dogbone samples made of polylactic acid (tough PLA) and Acrylonitrile Butadiene Styrene (ABS). The data were collected for 35 different parameter settings with the fused filament fabrication (FFF) process. The parameter settings were adjusted using a Bayesian exploratory experimental design technique, with a Gaussian Process Regression surrogate model to predict toughness from processing parameters and predictive variance as the acquisition function. The Python notebook used for experimental design is provided in the supporting files, along with the code that tests the design technique with an analytical function. Experiments for each parameter setting were repeated 3-times to account for the variability of the FFF process. The dataset includes optical microscopy images from both the fractured surfaces for two specimens that showed the highest toughness values. All dogbones were fabricated using UltiMaker S7. Tensile tests were conducted using a Shimadzu universal testing machine at the High Performance Materials Institute (HPMI). Optical microscope images were collected with Keyence VHX-X1. Raw data contains tensile tested data from all the experiments, the optical microscopy images, some standard images of the speciments, and the data collectors' notes. The supporting files include all the supporting code and metadata.