The aim of this paper is to introduce and characterize, both experimentally and numerically, three classes of non-traditional 3D infill patterns at three scales as an alternative to classical 2D infill patterns in the context of additive manufacturing and structural applications. The investigated 3D infill patterns are biologically inspired and include Gyroid, Schwarz D and Schwarz P. Their selection was based on their beneficial mechanical properties, such as double curvature. They are not only known from nature but also emerge from numerical topology optimization. A classical 2D hexagonal pattern has been used as a reference. The mechanical performance of 14 cylindrical specimens in compression is quantitatively related to stiffness, peak load and weight. Digital image correlation provides accurate full-field deformation measurements and insights into periodic features of the surface strain field. The associated variability, which is inherent to the production and testing process, has been evaluated for 3 identical Gyroid specimens. The nonlinear material model for the preliminary FEM analysis is based on tensile test specimens with 3 different slicing strategies. The 3D infill patterns are generally useful when the extrusion orientation cannot be aligned with the build orientation and the principal stress field, i.e., in case of generative design, such as the presented branching structure, or any complex shape and boundary condition.

Increasingly available open medical and health datasets encourage data-driven research with a promise of improving patient care through knowledge discovery and algorithm development. Among efficient approaches to such high-dimensional problems are a number of machine learning methods, which are applied in this paper to pressure ulcer prediction in modular critical care data. An inherent property of many health-related datasets is a high number of irregularly sampled time-variant and scarcely populated features, often exceeding the number of observations. Although machine learning methods are known to work well under such circumstances, many choices regarding model and data processing exist. In particular, this paper address both theoretical and practical aspects related to the application of six classification models to pressure ulcers, while utilizing one of the largest available Medical Information Mart for Intensive Care (MIMIC-IV) databases. Random forest, with an accuracy of 96%, is the best-performing approach among the considered machine learning algorithms.

3D concrete printing technology (3DCP) is a relatively new technology that was first established in the 1990s. The main weakness of the technology is the interface strength between the extruded layers, which are deposited at different time intervals. Consequently, the interface strength is assumed to vary in relation to the time of concrete casting. The proposed experimental study investigated the behavior of a hardened concrete mixture containing coarse aggregates that were up to 8 mm in size, which is rather unusual for 3DCP technology. The resulting direct tensile strength at the layer interface was investigated for various time intervals of deposition from the initial mixing of concrete components. To better understand the material behavior at the layer interface area, computed tomography (CT) scanning was conducted, where the volumetric and area analysis enabled validation of the pore size and count distribution in accordance with the layer deposition process. The analyzed CT data related the macroscopic anisotropy and the resulting crack pattern to the temporal and spatial variability that is inherent to the additive manufacturing process at construction scales while providing additional insights into the porosity formation during the extrusion of the cementitious composite. The observed results contribute to previous investigations in this field by demonstrating the causal relationships, namely, how the interface strength development is determined by time, deposition process, and pore size distribution. Moreover, in regard to the printability of the proposed coarse aggregate mixture, the specific time interval is presented and its interplay with interface roughness and porosity is discussed.

As energy distribution systems evolve from a traditional hierarchical load structure towards distributed smart grids, flexibility is increasingly investigated as both a key measure and core challenge of grid balancing. This paper contributes to the theoretical framework for quantifying network flexibility potential by introducing a machine learning based node characterization. In particular, artificial neural networks are considered for classification of historic demand data from several network substations. Performance of the resulting classifiers is evaluated with respect to clustering analysis and parameter space of the models considered, while the bootstrapping based statistical evaluation is reported in terms of mean confusion matrices. The resulting meta-models of individual nodes can be further utilized on a network level to mitigate the difficulties associated with identifying, implementing and actuating many small sources of energy flexibility, compared to the few large ones traditionally acknowledged.