This work explores an additive-manufacturing-enabled combination-of-function approach for design of modular products. AM technologies allow the design and manufacturing of nearly free-form geometry, which can be used to create more complex, multi-function or multi-feature parts. The approach presented here replaces sub-assemblies within a modular product or system with more complex single parts that are designed and manufactured using AM technologies. This approach can increase the reliability of systems and products by reducing the number of interfaces, as well as allowing the optimization of the more complex parts during the design. The smaller part count and the ability of users to replace or upgrade the system or product parts on-demand should reduce user risk, life-cycle costs, and prevent obsolescence for the user of many systems. This study presents a detailed review on the current state-of-the-art in modular product design in order to demonstrate the place, need and usefulness of this AM-enabled method for systems and products that could benefit from it. A detailed case study is developed and presented to demonstrate the concepts.

The deep energy method (DEM), a type of physics-informed neural network, is evolving as an alternative to finite element analysis. This method employs the principle of minimum potential energy to predict deformations under static loading conditions. However, the model’s accuracy is contingent upon choosing the appropriate architecture for the model, which can be challenging due to the high interactions between hyperparameters, large search space, difficulty in identifying objective functions, and non-convex relationships with the objective functions. To improve DEM’s accuracy, we first introduce random Fourier feature (RFF) mapping. RFF mapping helps with the training of the model by reducing bias towards high frequencies. The effects of six hyperparameters are then studied under compression, tension, and bending loads in planar linear elasticity. Based on this study, a systematic automated hyperparameter optimization approach is proposed. Due to the high interaction between hyperparameters and the non-convex nature of the optimization problem, Bayesian optimization algorithms are used. The models trained using optimized hyperparameters and having Fourier feature mapping can accurately predict deflections compared to finite element analysis. Additionally, the deflections obtained for tension and compression load cases are more sensitive to variations in hyperparameters than bending.

This paper investigates the influence of sample size and layer height on the elastic modulus of samples 3D printed by fused filament fabrication (FFF). Cubic samples of different sizes from either a stiff thermoplastic polymer (polycarbonate (PC)) or a soft polymer (thermoplastic polyurethane (TPU)) were printed at three different layer heights: 0.1 mm, 0.2 mm, and 0.4 mm. The samples were imaged using micro-computed tomography to visualize porosity and tested under compression. The results indicate that the elastic modulus of TPU varied with sample size, demonstrating size effects due to high porosity. In contrast, PC samples had a low void volume fraction and did not demonstrate statistically significant size effects on the elastic modulus. The variations in elastic modulus due to layer height were statistically significant for both materials. The highest elastic modulus was observed at 0.2 mm layer height for both materials across different sample sizes, which can be attributed to low void volume fractions. The study develops process-structure-property relations for PC and TPU samples manufactured by FFF. It provides new insights into the design guidelines for FFF based on the size effects. The results indicate that variations in mechanical properties due to size effects must be considered during the design stage for TPU to accurately predict the material’s behavior. Additionally, the locations of the printed parts that are subjected to higher loads should be printed closer to the heated print bed due to the lower void volume fraction observed in the section of the samples printed adjacent to the heated print bed.