Authors

Gianluca Cavallo, Annamaria di Lernia, Gaetano Elia

Affiliation

Department of Civil, Environmental, Land, Building Engineering and Chemistry (DICATECh), Technical University of Bari, via Orabona 4, 70125 Bari, Italy

g.cavallo@phd.poliba.it

annamaria.dilernia@poliba.it

gaetano.elia@poliba.it

The seismic risk assessment is typically conducted through the numerical modelling of the seismic site response to predict the shaking at the ground surface of a specific site. The estimation of the seismic site response may be conducted by adopting one-dimensional, two- or three-dimensional schemes, as a function of the site conditions, and implementing linear visco-elastic or nonlinear elasto-plastic constitutive assumptions for the description of the cyclic soil behaviour. The predictive capacity of these numerical schemes depends on several factors, such as the extension of the domain, number of elements, boundary conditions, soil constitutive model and features of the input motions. Accurate site response predictions imply the adoption of detailed and sophisticated numerical models, which may be extremely time-consuming and require huge computational resources. Very often, the complexity of such numerical models is reduced in favour of simplified, faster but also less accurate schemes.

Within this framework, the Italian program of the National Research Centre in High-Performance Computing (HPC), Big Data and Quantum Computing (ICSC) “Environment and Natural Disasters (Spoke 5)” aims at developing advanced numerical tools for the real-time simulation of natural disaster-inducing phenomena, such as seismic wave propagation processes, with the primary objective of reducing the associated risks. In this context, the work illustrates the results of a preliminary numerical investigation aimed at assessing the performance of nonlinear 2D finite element (FE) analyses for the assessment of the seismic response of a natural slope. The FE slope model has been inspired by the western slope of Chieuti (Foggia), a small village located in the south of Italy, very well characterised from a geotechnical point of view [1]. The simulations have been conducted with the FE code OpenSees [2], taking advantage of the National Research Centre HPC resources to improve their efficiency and using an in-house  python pre-processor[8] [9] to compile the executable files. The cyclic response of the slope soils has been described by the Pressure Independent Multi Yield (PIMY) model [3], a nonlinear elasto-plastic constitutive model accounting for both isotropic and kinematic hardening. The slope model developed in OpenSees adopts 9_8_QuadUP finite elements, implementing the u-p formulation for the solid-fluid interaction during fully-coupled dynamic simulations. As benchmark, the same slope has been modelled in the FE code PLAXIS 2D [4]. In this latter case, the nonlinear elasto-plastic constitutive model HSsmall [5] has been adopted to describe the cyclic soil response. Both constitutive models have been calibrated based on the laboratory and in-situ tests available for the site of reference.

The performance of the OpenSees and PLAXIS dynamic analyses has been assessed in terms of displacement field and acceleration time histories predicted at the ground surface. The comparison shows a good agreement between the numerical predictions obtained with the two codes, with a clear advantage of reducing the computational cost when using OpenSees. The results prove how the HPC resources may enable cost-effective analyses of complex geotechnical problems, demonstrating the potential of advanced computational techniques to improve the seismic resilience of critical infrastructure.

Keywords: seismic response analysis, FE modelling, nonlinear constitutive models, CN_HPC project

References:

1.     A. di Lernia, C. Buono, G. Elia (2023). Evaluation of seismic site effects in a real slope through 2D FE numerical analysis. 9th ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering – COMPDYN2023

2.     McKenna F.T. (1997). Object-oriented finite element programming: Frameworks for analysis, algorithms and parallel computing. Ph.D. thesis, Univ. of California, Berkeley, CA.

3.     Yang Z. (2003). Computational model for cyclic mobility and associated shear deformation. Journal of Geotechnical and Geoenvironmental Engineering, 129(12) pages 1119-1127

4.     Brinkgreve R., Kumarswamy S., Swolfs W. (2022), PLAXIS 2D Connect Edition V22.1, Reference manual, Bentley.

5.     Benz. T, Schwab R., Vermeer P. (2009) Small-strain stiffness in geotechnical analysis, 86(51), pages 16-27.

6.     Petracca M., Candeloro F., Camata G. (2017), STKO user manual. ASDEA Software Technology, Pescara, Italy

7.     Yang Z., Lu J., Elgamal A. (2008), OpenSees soil models and solid-fluid fully coupled elements. User's Manual.

8.     Cavallo G., OpenSToolsV2 - A python code to pre-process the finite element models in the OpenSees and OpenSeesSP software (https://zenodo.org/records/14056127)

9.     Cavallo G., OpenSToolsV2: example fem models (https://zenodo.org/records/14174620)