LS-DYNA Machine Learning–Based Multiscale Method for Nonlinear Modeling of Short Fiber–Reinforced Composites

Abstract

Short-fiber-reinforced composites (SFRC) are high-performance engineering materials for lightweight structural applications in the automotive and electronics industries. Typically, SFRC structures are manufactured by injection molding, which induces heterogeneous microstructures, and the resulting nonlinear anisotropic behaviors are challenging to predict by conventional micromechanical analyses. In this work, we present a machine learning-based multiscale method by integrating injection molding-induced microstructures, material homogenization, and Deep Material Network (DMN) in the finite element simulation software LS-DYNA for structural analysis of SFRC. DMN is a physics-embedded machine learning model that learns the microscale material morphologies hidden in representative volume elements of composites through offline training. By coupling DMN with finite elements, we have developed a highly accurate and efficient data-driven approach, which predicts nonlinear behaviors of composite materials and structures at a computational speed orders-of-magnitude faster than the high-fidelity direct numerical simulation. To model industrial-scale SFRC products, transfer learning is utilized to generate a unified DMN database, which effectively captures the effects of injection molding-induced fiber orientations and volume fractions on the overall composite properties. Numerical examples are presented to demonstrate the promising performance of this LS-DYNA machine learning-based multiscale method for SFRC modeling.

Graph-enhanced deep material network: multiscale materials modeling with microstructural informatics

“…Fully-connected H (4) = ReLU W (2) • H (3) + b (2) (64, 64) Fully-connected y (1) = tanh W (3) • y (0) + b (3) (64, 64)…”

“…• F denotes the Frobenius norm. The Frobenius norm A F of a matrix A is defined as Fully-connected y (2) = ReLU W (4) • concat(X f eats , p) + b (4) (127, 127)…”

“…Traditionally, multiscale modeling has been performed using the finite element method (FEM) or the fast Fourier transform (FFT). Well-established software applications in the industry that allow to run multiscale simulations, such as LS-DYNA ® [2], are based on these methods. However, FEM and FFT-based multiscale modeling can be daunting due to their high computational cost.…”

“…This is because the homogenization behavior considered in DMN is independent of material nonlinearity. Due to its unique characteristics and capacity for fast and reliable predictions, DMN even found its way into legacy software in the industry [23,24].…”

Effects of process parameters on mechanical properties and interface of 3D printed bamboo‐inspired CCFR‐PLA/TPU composites

“…These kinds of composites optimized by different materials are widely used in aerospace, automotive, and marine fields due to their excellent properties. [8][9][10][11] According to the type of reinforcing phase, composites are divided into particle reinforced composites, 12,13 short fiber reinforced composites, 14,15 and continuous fiber reinforced (CFR) composites. [16][17][18][19][20] Among them, CFR composites, in particular, have been extensively studied for their superior mechanical properties.…”