SCEC Award Number 18157 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Towards Dynamic Rupture Simulations with High Resolution Fault Zone Inelasticity
Investigator(s)
Name Organization
Ahmed Elbanna University of Illinois at Urbana-Champaign
Other Participants Xiao Ma
Setare Hajarolasvadi
SCEC Priorities 2d, 2e, 5a SCEC Groups FARM, CS, Seismology
Report Due Date 03/15/2019 Date Report Submitted 03/14/2019
Project Abstract
Active fault zones are homes for a plethora of complex structural and geometric features
that are expected to affect earthquake rupture nucleation, propagation, and arrest, as well as
interseismic deformation. Simulation of these complexities have been largely done using continuum plasticity or scalar damage theories. In this paper, we use a highly efficient novel hybrid
finite element-spectral boundary integral equation scheme to investigate the dynamics of fault
zones with small scale pre-existing branches as a first step towards explicit representation of
anisotropic damage features in fault zones. The hybrid computational scheme enables exact
near-field truncation of the elastodynamic field allowing us to use high resolution finite element discretization in a narrow region surrounding the fault zone that encompasses the small
scale branches while remaining computationally efficient. Our results suggest that the small
scale branches may influence the rupture in ways that may not be realizable in homogenized
continuum models. Specifically, we show that these short secondary branches significantly
affect the post event stress state on the main fault leading to strong heterogeneities in both
normal and shear stresses and also contribute to the enhanced generation of high frequency
radiation. The secondary branches also affect off-fault plastic strain distribution and suggest
that co-seismic inelasticity is sensitive to pre-existing damage features. We discuss our results
in the larger context of the need for modeling earthquake ruptures with high resolution fault
zone physics.
Intellectual Merit This intellectual merits of this project lied in blending advances in computational mechanics and earthquake physics to address the problem of influence of small scale heterogeneities and geometric complexities on the dynamics of earthquake ruptures. A newly developed computational scheme by our group ( the hybrid FEM-SBI approach) has enabled us to model fault zones with high resolution physics and incorporate explicitly some small scale features that have been previously homogenized or smeared out due to computational challenges. The results of this project shed lights on sources of stress heterogeneities on major faults as well as origin of enhanced high frequency radiation and contributes to better understanding of earthquake source complexity.
Broader Impacts Funding of this project has contributed support to two PhD students: Xiao Ma (Male) and Setare Hajarolasvadi (Female). The results have been disseminated in two journal publications with the two graduate students taking a leading role in them. The results were also presented in several talks (at AGU, Stanford, and USNCTAM conference) and poster presentations (at SCEC). The results of this project will contribute to development of higher resolution physics based dynamic rupture model which is critical for next generation physics based seismic hazard platforms.
Exemplary Figure Figure 3: Contours of the bulk velocity field and an example acceleration spectrum at a near field station. (a) Bulk velocity field in a Homogeneous medium. (b) Bulk velocity field in a Domain with fish bone structure. Coherent high frequency generation emerge in the case of the fault with secondary branches (fish bone structure) and propagate away from the fault plane as concentric fringes. These high frequency waves are generated as a result of the constructive interference between the waves emitted by the main fault and the secondary branches. In the homogeneous case the high frequency wave field is localized near the rupture fronts. (c) Fault normal acceleration spectral amplitude at station x∗ = 15R and y ∗ = −2R showing increased high frequency content and a flat frequency spectrum in the range of 1-20 Hz for the fault zone with short branches.