A new dynamic method for the implementation of faults in finite-element models

Faults strongly affect the seismotectonic but in geodynamic modelling the Earth is often represented as a continuous material. The challenge in representing these structures in FE models is still open. We propose a new method that enables the dynamic identification of the fault during the system evolution, without predefining either its geometry or the slip. Our method is an advancement of Marotta et al. (2020) method, which, in turn, modified the classical split node method (Jungles and Frazier, 1973; Melosh and Raefsky, 1981) by replacing the prescribed slip with a coupling factor along the fault plane. We developed an algorithm that, in the frame of a finite element approach, identifies the faults as the envelope of nodes on which a rupture criterion is satisfied. The breaking point is identified as the node on which the rupture condition is mostly exceeded; then the propagation proceeds along a line of neighboring nodes. The elements adjacent to the fault are classified as left or right, and the coupling factor is assigned. An AMR has been developed in such a way that where the fault cuts an element, the grid is recalculated. We show the results of some benchmarks performed to test the correctness of the propagation algorithm, and the localization of a shear zone in a complex tectonic context. References Jungles and Frazier, 1973, DOI: 0.1029/JB078i023p05062 Marotta et al., 2020, DOI: 0.1093/gji/ggaa029 Melosh and Raefsky, 1981, DOI: 10.1785/BSSA0710051391