Seismic performance of loess-mudstone slope in Tianshui – Centrifuge model tests and numerical analysis

Highlights

• We model a loess-mudstone slope under earthquake by centrifuge test and numerical simulation.

• Dynamic behaviors and failure mechanisms of the slopes during earthquakes are studied.

• Acceleration responds of the loess-mudstone slope showed several amplifying characteristic: the height effect, lithologic effect, surface effect, and the intensity of input motions effect.

Abstract

Dynamic behavior and failure mechanism of slope during earthquake are hot topics in geotechnical engineering community. In Tianshui area, Northwest of China, the deformation features of loess-mudstone slopes induced by earthquakes are still not well understood. The current study aims to discuss the dynamic behavior of the slope under earthquake, by performing a series of centrifuge shaking table tests and numerical simulations. The results are processed for PGA amplificatory effect study and to evaluate the earthquake induced deformation mechanism. Better agreement is obtained between the results of numerical simulations and centrifuge tests. The results indicated that the acceleration response of the loess-mudstone slope showed several amplifying characteristic: the height effect, lithologic effect and surface effect. It also proves that the displacements in the loess layer are much larger than mudstone. As the input amplitude increased, the shear plastic zone developed along a further slide surface. Tension plastic zone mainly occurred at the crest. Roughly, the regions of earthquake induced deformation state increased with the increase in amplitude of earthquake wave. Deformations occurred at the upper part of the slope. Tensile cracks evolved on the crest and the upper part of slope in loess layer, moreover, failures occurred at slope surface, and only slight deformation at slope toe. Deformations with continuous shear cracks and slope surface upheavals developed at the middle part of slope.

Introduction

Failure mechanisms of slopes during earthquakes have become the focus of studies dealing with the complexity of geotechnical earthquake engineering at the present (Wasowski et al., 2011, Loáiciga, 2015). The lessons learned from the earthquakes, which have occurred over the last 50 years (Goodman and Seed, 1966, Sarma, 1975, Chang et al., 1984, Ashford et al., 1997, Martino and Mugnozza, 2005, Lenti and Martino, 2012, Zugic et al., 2014, Gischig et al., 2015), have highlighted the importance of slope dynamic response behaviors and failure mechanisms. At present, investigations of the slope dynamic response behaviors are mainly performed through several methods, including field investigation, seismic station monitoring, numerical simulation, and physical model test (Wang et al., 2003, Yin et al., 2009, Wang et al., 2010, Bozzano et al., 2011, Gischig et al., 2015, Dong and Kutter, 2015, Molnar et al., 2015, Zhang et al., 2015, Facciorusso et al., 2016, Martino et al., 2016). The theory and application in this area are still in the developing stages.

Physical modeling can be performed by shaking table tests. Some earthquake response behaviors studies of large-scale shaking table tests on slope, dam and embankment have been documented in literatures (Lin and Wang, 2006, Wang and Lin, 2011, Srilatha et al., 2013, Huang et al., 2013). However, the large-scale shaking table tests could not satisfy the actual stress levels. As a result, centrifuge model tests play an important role in analyzing the behavior and failure mechanisms of slopes because they can reproduce the gravity stress field and the gravity-related deformation process (Take et al., 2004). Such testing method has been widely used for studying the behavior of reinforced slopes and rainfall induced slope deformation (Take et al., 2004, Thusyanthan et al., 2007), and has been proved to be an effective tool for exploring the evolvement of slopes. Recently, reduced shaking table equipment has been made and installed inside in-flight centrifuge apparatus (Massimino and Michele, 2013). The seismic performances of slopes have been studied by shaking table equipment on centrifuge, and it also has been proved to be an efficient way (Yu et al., 2008, Wang and Zhang, 2014). Nevertheless, in the previous centrifuge tests, the input earthquake wave is only in horizontal direction due to the restriction of the equipment, which may cause insufficient dynamic input.

Recently numerical modeling has become a powerful tool in the characterization of rock slope deformation, failure and post-failure (Stead et al., 2006, Marcato et al., 2012, Alfaro et al., 2012). The stress-strain numerical analysis performed under dynamic conditions can take into account the interaction between slopes and seismic inputs in terms of local ground motion amplification due to the topography of the slope and the geological setting (Lenti and Martino, 2012, Lenti and Martino, 2013). The Itasca FLAC2D (ITASCA, 2011) program could perform non-linear numerical analyses for slope strata under earthquake conditions. FLAC2D could consider different real acceleration time histories as input motion. This method has been widely used for slope, dam and embankment dynamic response study under earthquakes (Chugh and Stark, 2006, Andrianopoulos et al., 2014, Stamati et al., 2016).

China is one of the most seismically active countries in the world with a complex geology. In history, several earthquakes with a moment magnitude more than 6.0 struck Northwest of China and induced extensive slope failures, which led to severe damage to the highway system and private houses (Wang et al., 2015). However, the failure mechanisms and dynamic behaviors of the loess-mudstone slope in Tianshui, Northwest of China are still not well understood. Meanwhile, forecasting the location and shaking conditions needed to trigger geologic hazard is an important element in regional seismic hazard assessment (Zhang et al., 2015), and it is also of particular interest to predict the deformation of natural slopes under seismic shaking for engineering practice (Wasowski et al., 2011, Lenti and Martino, 2013).

The present study aims to describe a series of 0.1–0.5g large-scale shaking table model tests on the centrifuge performed at China Water Resources and Hydropower Research Institute (CWRH) on the loess-mudstone slopes under seismic conditions. The experimental results are recorded by a great quantity of accelerometers and displacement transducers. The loess and mudstone are pluviated into a flexible container maintaining a constant controlled relative density by means of an accurate procedure. The shaker on the centrifuge at CWRH is a most advanced apparatus in the world which could achieve vertical and horizontal input wave. The experimental results are, then, compared with the results of numerical modeling by FLAC2D code. The comparison makes it possible to validate the proposed numerical modeling, to evaluate their ability to analyse the dynamic response behaviors of loess-mudstone slope, as well as to capture interesting aspects of the tested slope deformation and failure mechanisms.