A Surrogate Model for Numerical Evaluation of Elastic Properties of Particulate Composites with Rotationally Symmetric Particles

The development of new materials is an expensive process, especially in the case of composite materials, because their microstructure can be very complex. This process is usually supported by numerical analysis of the considered material microstructures to predict their overall mechanical properties and optimize performance. However, this numerical modeling can become challenging when the reinforcement particles have complex shapes. The computational approach proposed in this work employs particle swarm optimization in combination with the finite element method to substitute the original complex particles with equivalent “surrogate” particles of simpler shape (circular in 2D) that produce the same contribution to the overall elastic properties. Usage of such surrogate particles significantly reduces computational time and, thus, can minimize the development cost of new particle reinforced composite materials, while increasing the spectrum of evaluated material features.