For the calculation of mechanical properties of large magnet systems in 3-dimensional space, a very fine subdivision of the magnetstructure is necessary. In the finite element programms, this will lead to unacceptably lang computing times and to the limits of computer-storage capacity. This limitation requires a simplification of the structure model. The objective of this work is to find such s...

In this paper, a numerical multiscale method is proposed for computing the response of structures made of linearly non-aging viscoelastic and highly heterogeneous materials. In contrast with most of the approaches reported in the literature, the present one operates directly in the time domain and avoids both defining macroscopic internal variables and concurrent computations at micro and macro...

Abstract Material modeling using modern numerical methods accelerates the design process and reduces costs of developing new products. However, for multiscale heterogeneous materials, well-established homogenization techniques remain computationally expensive high accuracy levels. In this contribution, a machine learning approach, convolutional neural networks (CNNs), is proposed as efficient s...

Scale separation is often assumed in most homogenization-based topology optimization (TO) frameworks for design of material microstructures. This work goes beyond the mainstream TO contributions by abandoning scale hypothesis. First, it puts to evidence limits approach when size Representative Volume Element (RVE) not negligible with respect structure. Then, a re-localized scheme bridging RVE a...

A model to predict the high temperature ultimate strength of a continuous fiber metal matrix composite (CFMMC) has been developed. The model extends the work of Rosen by including high temperature processes such as matrix creep, fiber-matrix debond~ and the effects of randomly spaced fiber breaks which typically exist in the MMC prior to loading. A finite element model (FEM), developed in the f...

Fiber failures which may exist in a continuous fiber· composite before the .composite is loaded or generated while the composite is loaded, introduce an additional strain component not included in existing continuous fiber models. Equations are developed which can be used to calculate the magnitude of this additional strain. A Finite Element Model (FEM), in the form of a Representative Volume E...

Smart composite materials have been an active field of research in the last few decades. Magnetorheological elastomers (MREs) are examples such smart composites and they show a coupling between magnetism elasticity. MREs heterogeneous consist magnetic particles silicone-based elastomer. To describe predict behaviour macroscopic continuum accurately, microstructural information needs to be taken...

This work focuses on the effects of inclusion proximity elastic behavior dilute matrix-inclusion composites. Rigid or soft monodisperse spherical inclusions are considered with moderate volume fractions. To conduct this study, Representative Volume Elements (RVE) an effective local minimum distance between varying sphere’s radius and one-tenth built. Numerical finite element calculations RVE pe...

This paper presents the development of a Representative Volume Element (RVE) for Ultra-high-Molecular-Weight-Polyethylene (UHMWPE) composites. The numerical models were based on fibrillar nature UHMWPE fibres, which consist smaller scale, continuous through-the-length macro-fibrils. A three-dimensional constitutive model macro-fibrils was developed and implemented in LS-DYNA explicit finite ele...