New model for crack growth using random walkers

TL;DR

This paper introduces a novel random walker algorithm for modeling crack growth in elastic media, efficiently identifying stress hot spots without full stress field calculations, enabling large-scale crack-cluster simulations.

Contribution

The paper presents a new random walker approach for crack growth modeling that improves efficiency and scalability over traditional lattice methods.

Findings

Successfully generated crack-clusters of up to 20,000 particles.

The method reduces computational complexity by focusing on crack surface hot spots.

Demonstrated feasibility on standard workstations.

Abstract

In close analogy to diffusion limited aggregation (DLA) and inspired by a work of Roux, a random walker algorithm is constructed to solve the problem of crack growth in an elastic medium. In contrast to conventional lattice approaches, the stress field is not calculated throughout the whole medium, but random walkers are used to detect only the hot sites on the surface of the crack. Using this new method we generate crack-clusters up to sizes of 20,000 particles on simple workstations within reasonable time.