Experimental evaluation of the shear behavior of large landslide slip zone using Multistage - Multiphase shearing method.


 Abstract
The study of shear behaviors comprises of the analysis of shear deformation characteristics and strength parameter determinations. Understanding the constitutive behaviour of shear zone soil samples is essential for evaluation of the critical shear behaviors of reactivated landslides, through considerate experimental tests. The experiment was carried out based on the multistage - multiphase shear technique via ring shear apparatus. The main advantage of this new method is to examine the same soil specimen for progressive loading and the continuous phases of shearing. The result shows that two major shear characteristics under the multiphase shear mode; the first is a post-peak shear weakening behavior of the soils on the last phase shear process, it indicated a decrease in friction resistance on fast rate effects, which enables to narrate with fastly moved coseismic landslide behaviors. The other one is a practical increase in friction coefficient values (hardening behavior) on the entire shear process as the rate increases. This shear resistance scenario can also be related with the stabilization mechanism of landslide slip surface soils against sliding.


Introduction
The shear strength of soils depicted great attention on its signi cance to various engineering structures and natural slopes (Bhattacharya et al. 2017;Eid et al. 2015;Kalteziotis 1993;Mesri and Shahien 2003).
Numerous studies were carried out to investigate the shear behavior of largely displaced landslides through different experimental mechanisms (Gratchev and Sassa 2015;Habibbeygi and Nikraz 2018). It is expected that the soil skeleton, particle size distribution, and displacement rate have considerable effect on the frictional strength of soils (Gratchev and  . Therefore, analysis of shear behavior of natural slip zones should reveal a more realistic and complex history of the landslide processes. Moreover, the shear behavior of slide zone soils usually results from the particle level characteristics and the shear processes, it is intuitively recognized that particle size and shape affects soil shear behavior.
The conventional shear approaches used by several researchers circle on a single stage or multi stage loading, and apply a certain shear displacement to an individual samples, it uses more sample specimens for planned tests. There is no research so far considered the incremental phase of shearing in a different normal stresses. We designed a multistage -multiphase shearing technique that can test a single soil specimen to be sheared for continuous incremental large displacement with simultaneous fast rate, under different normal loading. It assesses the shearing behavior of soils under such progressive deformation characteristics and failure conditions.
The study mainly aims to evaluate and learn the multistage -multiphase shearing technique on the determination of critical shear strength behavior of shear zone soil samples using an automated suction controlled ring shear test apparatus. With the application of multiphase shear mode, it assesses the effects of shear rate on the residual shear behaviors, internal friction angle, and shear deformation characteristics of soil specimens.

Experimental Methods
Long runout coseismic landslides do exhibit progressive increment of slide mobility from detachment and initiation phase to a higher speed of slide mass with down throw and ejection seismic forces, as the displacement increases (Getahun et al. 2019). Moreover, after large displacement, braking and stabilization of the runout landslides occurs. To emulate this real phenomenon and conceptual model of the landslide runout, a multiphase shear mode was designed. A multistage multiphase shear technique is implemented on a single soil specimen, with incremental normal loads to consolidate and applies a continuous progressive large shear displacement and rate. This new approach was applied on samples collected from the reactivated landslide shear zone soils.
The study consists of detail eld inspection, measurements, and soil samplings. The soil samples were collected from the shear zone of the Chenjiaba Landslide, which was induced by the 2008 Wenchuan earthquake. Soil physical property tests comprising of the moisture content, density, grain size distribution of the samples were carried out rst in the index laboratory. Disturbed samples were used to prepare specimens for ring shear tests according to the standard procedures. The Uni ed Soil Classi cation System (USCS) standards for grading soils were used to characterize the soil sample type (ASTM-D2487-11 2011). In this research, the samples used are natural soils trimmed from shear zone of Chenjiaba landslides, and graded as mixture of some gravel and nes with sandy soils.
To observe and understand the particle arrangement after shearing test, scanning electron microscope (SEM) imaging tests were used to know the micro fabric nature. In SEM, the surface of the soil is scanned using a high energy beam of electrons which provides an `image' of the soil, then allowing a view of the soil's microstructure. It is a method used for high-resolution imaging of samples surfaces. The SEM imaging tests of samples from the shear zone of the landslide are analyzed at different scales (1 mm, 500 µm, 200 µm, 100 µm, and 50 µm) magni cations to understand the particles microfabric setup and arrangement. The Scanning electron microscope (SEM) equipment used in this research was Nova Nano SEM 650.

Multistage -Multiphase shearing approach
The stabilization of long runout landslide mostly occurs along shear zone, after large displacements. So, the experimental design to achieve the residual shear behavior of soils is considered on the progressive large displacement rate test to the reactivated soils, and evaluates the shear response of the soils under the experiment. The shear resistance developed for long runout landslide was, reasonably, from the critical shear strength of the shear zone soils (Hu et al. 2018). Unlike the conventional shear strength tests that use several soil specimens, a multistage test use a single soil specimen and shear the sample in stages with increasing con ning stresses (Mairaing 2008;Nam et al. 2011). In order to learn the stabilization mechanisms of long runout landslides along the shear zone, we modi ed and designed a multi stage multiphase shear process to evaluate the continuous shear behavior of reactivated slip zone soils.
The multistage ring shear test type was one of the procedures used to measure the residual shear behavior of soils (Bhat et  The general approach of this research method consists of three stages of normal loading, and four phases of shearing on each stages of normal loading. After the initial consolidation attained by intended normal stress, shearing proceeds with designed shear rates, times and targeted angular degree or shear displacement. The new approach of multistage -multiphase shearing tests were carried out through progressive increasing normal loads, with an initial loading of 60 kPa, and double increased for the second and third stage of consolidation. The consolidation time given to each stage of loading was equal with varied rates. These loading plans were made based on the slide soil average thicknesses (10 m) and average density (2.4 g/cm 3 ) of the soil samples from the eld sites investigation. The tests were conducted in drained conditions. , and experiences of some studies that practiced shear tests on large landslides are used as a baseline to plan for multistage multiphase shear approach. The multiphase shear approach considers an initial shear displacement of 40 mm and rate of 13.2 mm/min on the rst phase, and these proceeds with a double increment to next phases (Table_1).

Ring shear apparatus and test procedures
The frequent challenges of studying the deformation characteristics in large shear displacements and The ring shear apparatus used herein was the Modi ed form where originally based from Ring shear apparatus developed by Bromhead (1979). It consists of two automated servo-controlled actuators: one pneumatic actuator for normal loads and one electrical actuator for shear loads.
The apparatus consists of three main components: (a) Main cell with rotational shear system; (b) data acquisition/ process control (DA/PC) system with performance/ data reduction software for real-time calculations of shear stress and average linear displacements; (c) Suction control panel. Main cell includes upper and lower platen sample boxes, Pneumatic servo-controlled actuator for application of vertical loads, electrical servo-motor actuator for application of torsional loads, linear variable differential transducer (LVDT) to measure vertical deformations, electrical sensors for real-time measurements of shear torque, shear angular deformation and vertical load, adjustable top and bottom stainless-steel loading platens, and, reinforced-acrylic con ning cell with 1000-kPa air pressure capacity (Hoyos et al. 2010, 2014) (Fig. 2). .
The experiment begins with the preparation of samples, depending on the intended test, samples need to be oven dried for basic physical property and drained ring shear tests. During ring shear tests, soil samples required to conform to the annular space between the two concentric rings of the lower platen.
The setup of the lower sample annular stainless steel platen has 152.4 mm outer diameter and 96.5 mm inner diameter and an average height of specimen 20 mm. The inside to outside diameter ratio is 0.63 as per ring shear testing standard (ASTM-D6467-13 2013).
In the apparatus, an air actuator applies the axial load (normal force), torque, axial shortening, and rotation speed were recorded depending on experiment duration (Kitajima et al. 2010). The designed input parameters for consolidation and shearing are set. According to the plan to this experiment, consolidation of specimen for speci ed normal stress (kPa) and rate initially done, after consolidation nished at the intended time then shearing test takes place according to the designed shear rate (System can do 0.001°/Min (minimum rate) and 360°/ min (maximum rate)) and targeted angular degree (shear displacement). Lastly, using the automated DA/PC system all required data are properly checked and saved on the computer data storage, and then the data's were extracted for further analyses, plots and interpretation.

Soil gradation characterizations
The soil shear behavior analysis has been linked with grading, shape and other physical properties. Grain size distribution (GSD) is the most fundamental type of physical property in soil studies (Yong 2017). The gradation characteristics parameters of the soils are described based on the uni ed soil classi cation system; it includes the effective grain diameters passing through 10% (D 10 ), 30% (D 30 ), 60% (D 60 ) and calculated coe cient of uniformity (Cu) and coe cient of curvature (Cc) ( Table 2). The soil samples were collected at different horizontal basal slip surface, where exposed at deep river cut in contact with massive sandstone unit during eld inspections.  The soil samples of the study area were characterized by poorly graded sandy soil with some gravel and ne mixtures, which have dark yellow to light gray weathered colors. The grain size distribution plot shows that soils are uniformly sorted with certain difference in its grading parameters. Moreover, sample Ch-07 has shown ne sand dominated soils compared to other samples from the soils grading curve. During the multiphase ring shear experiment, the grain size, shape, and their arrangement has showed substantial effects in the shear stress values corresponding with the continuous shear displacement rate on the soil samples. The grain size distribution curve of the soil samples are shown in Fig. 3. The internal friction angle values (ϕ r ) for each phase of the increasing shear displacement rate are shown in Fig. 4. Sample Ch-04 has the ϕ r value of 32.96 0 (at initial phase), 35.65° (at second phase), 33.36 0 (at third phase), and 31.62° (at last phase). The result indicates that the multiphase shear displacement rate caused the friction angle value to decrease progressively and reach residual state. The other sample (Ch-06) has the ϕ r value of 28.28 0 (at initial phase), 28.51° (at second phase), 29.78 0 (at third phase), and 29.82° (at last phase). In this case, the sample showed a steady state resistance developed to the entire shear process. However, sample Ch-07 has the ϕ r value of 29.27 0 (at initial phase), 28.44° (at second phase), 28.42 0 (at third phase), and 33.38° (at last phase). Sample Ch-07 has shown turbulent frictional resistance that has developed hardening effect at the last phase of shear mode compared to others (Fig. 5).

Shear displacements and rate effects
The shear deformation characteristics of samples are evaluated from the relationship between shear stress and shear displacement plots on different normal stresses, and continuous phases of shear mode. The results from the plot (Figs. 5 and 6) presents that samples do have varied post peak shear behaviors; soil softening (Ch-04), slight hardening (Ch-07) and remain steady state (Ch-06) property in the last phase of shear process. The shear stress against displacement shows some jagged features from peak to residual stresses during multiphase shear processes, for instance samples Ch-04 and Ch-07 have some uctuations in shear stresses on the entire shear process, and sample Ch-06 looks steady comparatively. Li et al, (2017) reported that to determine the value of residual strength under high shear rate was di cult because of the obvious uctuation of shear resistance in the shearing process from peak strength down to residual strength.
Many researchers indicated increment of the shear strength with the increment of an applied normal stresses (Alshameri et al. 2017). In addition, the decrement of the shear strength with the increment of the shear rate (Gratchev and Sassa 2015). Results of shear rate effects on the rst stage of loading indicate that the shear weakening behavior for samples Ch-04 and Ch-06, and an increasing strength phenomenon (hardening effect) for sample Ch-07 under last phase shear processes. Meanwhile, on second stage of loading, samples Ch-06 and Ch-07 has shown increasing shear strength to the entire shear process. Moreover, on the last stage of loading, the residual friction resistance showed decreasing trends to all samples compared to initial loadings.

Soil microstructure analysis and their implication on shear test
Soil structure is characterized by observing the shape, size and arrangement of soil aggregates either microscopically or macroscopically. In microscopic technique, thin soil sections are examined under various types of microscopes for shape and size of the aggregates and voids. Further SEM imaging tests was performed to observe the post shear soil particle structures that may exhibit the particles arrangements and possible shear resistance behavior of shear zone soils. We plan to observe the micro structure grain arrangements of some samples that indicate the existence of signi cant shear hardening behavior, for instance sample Ch-07 on large shear displacement and rate effects.
The study site samples have showed some difference in terms their post peak shear behaviors under the multiphase shear processes. We plan to observe the micro structure grain arrangements of some samples that indicate the presence of signi cant shear hardening i.e. sample Ch-07 on large shear displacement and rate effects. In this study, the SEM images of sample Ch-07 was taken at the scales of 30 µm in detail magni cation (Fig. 8).
The SEM post shear image of Sample Ch-07 shows the existence of coarse to ne fractions, fragmentary grain alignments with spaces between particles, and some small sized particles aligned in a course of drive patterns. The particles have an elongated contact on varied ways of their faces and edges in plane view. The image view shows the different size and shape of particles that may contribute to the shear resistance developed. The resistance to crushing and rearrangement of particles along the shearing direction may cause to increase the shear torque during shear tests; it resulted in increased shear strength at high rate effects.

Discussions
The sliding zone soils are an essential part of landslides, and its studies thoroughly allied to the evolutionary development and landslide stability. Reactivated landslides could result from very slow to very rapid movements, such as those caused by earthquakes (Suzuki et al. 2017). The shear strength characteristics of soil samples from slip surface of reactivated landslide is evaluated by ring shear experimental method using multistage multiphase shear mode, it performs large displacement movement and increasing shear rate tests on shear zone samples.
The experimental results and plots show that the shear stress reach a peak value at initial or second phase shear displacement for the samples, and shown different post peak shear behaviors that consists of soil softening, slight hardening and remain steady state property in the last phase of shear process.  Wang et al. 2017). In this study, the increased shear rate mode of experiment has disclosed the shear weakening behavior, which is similar with the above mentioned scholars' outcome and can narrate the long runout landslides.
Numerous frictional sliding studies of soils from slip surfaces yields a wide range of reported strengths. Sample Ch-07 has showed a shear hardening behavior during the entire shear processes, the friction resistance increased abruptly after third phase of shearing with turbulent ow of particles and this resistance persisted under the different normal stresses; however, it showed decrease in friction coe cient value on last stage loading, even though slight hardening phenomenon continued. According to Hu et al., (2018), the rate effect can play a fundamental role in the interpretation and forecasting of landslide behaviour (Leroueil 2001). In reactivated landslides, a rate-strengthening behaviour can prevent dramatic acceleration and long runout (Hu et al. 2018;Leroueil 2001;Wang et al. 2010). This hardening behavior can relate with the stabilization of long runout landslides. Li et al., (2013) informed that the varying test results obtained for the relationship between the shearing rate and the residual shear strength of soil could possibly be due to material texture and modes of shear deformation. Based on ring shear tests on arti cial mixtures of medium sand and powdered mica Lupini et al., (1981) reported observations on shear surfaces, and proposed three modes of residual shear behaviour: sliding, transitional and turbulent. They attributed the differential effects of shearing rates on residual shear strength to the shear mode, which is controlled by particle shape and the coe cient of inter particle friction (Li et al. 2013;Lupini 1981). Residual shear stresses vary depending on the soil particles characteristics, applied normal loads and shearing rates.
The SEM analysis has helped to observe and realize the post shear sample specimen microstructure and particle arrangement conditions in relation to their shear resistance on the image surface features. It can be observed that the particles arrangement is fragmented and has mixed grain size and shapes. Since the soil particles are sub-angular, the tendency of rolling or sliding between particles contacts may lead to increase shear resistance (Abd Rahman et al. 2016). The shear zone soils might expand with shearing processes, when the chains of grain particles rotate and produce an axial displacement comparable to the length of the grain chains and shearing displacement (Ferri et al. 2011). The fragmented and some parallel oriented particle arrangement after shear shows some general view on the shear resistance developed on particles resisted for sliding (Ch-07).
The shear strength of dense soils can show high shear resistance due to dilative behavior at large shear deformations and can fall back to the residual value when subjected to large displacement shear. Angularity and roughness enhance the dilative tendency of dense packing's. The shear strength and plastic ow (dilatancy) depend on the granular structure, particle sphericity and shear direction (Radjai and Az´ema 2009; Santamarina and Cho 2004). The particles interlocking creates resistance to grain to grain shear, rotate, roll and reorient along the shear directions, this kind of shear in uences are recognized on microstructure features of the landslide shear zone soils.

Conclusions
Experiments demonstrate the residual shear behaviors which can re ect the kinematics and evolution of landslides. The following concluding points are made from this study.
The use of multistage -multiphase shearing approach was suitable to evaluate residual shear behavior of slip zone samples, and enables to narrate the residual shear behavior of soil samples with the reactivation process of coseismic long runout landslides characteristics.
The critical state shear strength behavior of the slip zone samples have unveiled normally the shear weakening behaviors to last phase of shear mode. However, there is also a practical increase in residual friction coe cient values (hardening behavior) on the entire shear process as the rate