Development of a Cyclic Simple Shear Apparatus

Considering the increasing incidence of cyclic loadings on engineering structures and the enhancement of design analysis, soil testing under cyclic conditions has renewed its importance. Laboratory tests are conducted to simulate as near as possible field conditions. Assumed conditions aid on the choice of the tests to be conducted in order to determinate the relevant geotechnical parameters to each situation observed on the field. The simple shear test is highlighted among the typical tests in Geomechanics. This is the only laboratory test capable of submitting the specimen to plane strain conditions under constant volume while allowing the rotation of principal stresses. Such conditions are representative of situations such as the adjacent shear mechanism to the shaft of piles or under offshore platforms. In this sense a simple shear testing apparatus was developed. Contrasting with commercial equipment where confinement is made by means of a rigid membrane, specimens are confined by cell pressure in the developed apparatus. Consolidation can be conducted under isotropic or anisotropic paths and shearing, under monotonic and cyclic conditions (either stress or strain controlled). Validation tests were conducted on the equipment using an well-known material. The results obtained were satisfactory, validating the developed apparatus.


Introduction
The direct simple shear apparatus has been successfully employed to characterize static and dynamic soil properties for many years (Duku et al., 2007).This test is often preferred when the continuous rotation of the principal stress directions during shearing is a field condition.In the conventional apparatus, initial stresses can be applied to simulate at-rest field conditions when wire reinforced membranes are used to minimize lateral distortion of specimens (Bjerrum & Landva, 1966).A few of the most common applications of the simple shear testing are the vertical shear wave propagation through a soil column, the mode of shearing to a pile shaft (Randolph & Wroth, 1981) and under an offshore gravity base platform (Andersen et al., 1980).Advantages and limitations of simple shear tests relative to other types of laboratory tests have been described by many authors (e.g.Lucks et al., 1972;Saada et al., 1983;Vucetic & Lacasse, 1982;Airey & Wood, 1984Budhu & Britto, 1987;Boulanger et al., 1993).
Most simple shear apparatuses, in order to impose no lateral distortion, enclose the soil in a rubber-reinforced membrane (Kjellman, 1951, Bjerrum & Landva, 1966).A near K 0 condition is assumed to be obtained in this type of equipment.Differently, recent devices, such as the University of Western Australia (UWA) equipment, enclose specimen in an unreinforced latex membrane inside a pressurized cell (Mao & Fahey, 2003).The vertical and cell pressures are controlled independently.With the equipment software routine, total vertical stress and sample height are kept constant during shearing.In order to achieve such conditions, the vertical loading ram is locked and the cell pressure varies to keep total vertical stress constant.Once height and volume are constant (undrained tests), average cross-sectional area is likely to remain constant.The UWA equipment is frequently called as simple shear apparatus.
This paper describes the design of a simple shear apparatus, based on the UWA apparatus.The equipment was designed, manufactured and calibrated at the Federal University of Rio Grande do Sul.Tests were carried out using a well-known uniform sand to validate the equipment.

Previous Work
Simple shear tests have been used for many engineering and geology purposes, such as the study of mechanical behavior of sands (De Alba et al., 1976, Mao & Fahey, 2003), clays (Chu & Vucetic, 1992, Boulanger et al., 1993) and mine tailings (Wijewickreme et al., 2005, Wijewickreme et al., 2010, Festugato et al., 2013, Festugato et al., 2015), and the modeling of folding and fracture patterns (Price & Torok, 1989).The equipment development started with an apparatus by Kjellman (1951) to overcome some of the shortcomings of the traditional direct shear test, which suffers from non-uniform stress distribution throughout the specimen.Typically, the test consists of a circular specimen, consolidated to a stress level under K 0 conditions.
There are a number of different configurations in which direct simple shear devices were developed (Doherty & Fahey, 2011).Each tries to solve the most difficult problem of the test: to apply normal and shear stresses at the lateral boundaries of the specimen while preventing vertical and horizontal deformation.During shearing, the different designs try to keep volume and dimensions of the crosssectional area of the specimen constant (Franke et al., 1979).Roscoe (1953) resolved the problem by enclosing a square-shaped specimen within rigid metallic walls (Cambridge apparatus).The sidewalls parallel to the directions of the shear deformation are fixed relative to the base, while the walls perpendicular to the deformation are attached to the base by hinges.The movable lid remains parallel to the base during the shear phase as well as during consolidation.Authors as Roscoe & Burland (1967), Ansell & Brown (2008), Peacock & Seed (1968) and Finn et al. (1971) have improved aspects of the equipment develop by Roscoe (1953).Once the specimen is enclosured by a sealed rubber membrane, a sand specimen can be fully saturated with water through backpressure, so the pore water pressure can be measured during undrained simple shear tests.
Another configuration of simple shear was developed by the Norwegian Geotechnical Institute (NGI) to study quick clays (Bjerrum & Landva, 1966).In this device, a cylindrical specimen is enclosured by a rubber membrane reinforced by a spiral winding of steel wire.The steel spiral is supposed to prevent any change in the diameter of the specimen during the test, while allowing for vertical strains during consolidation.Undrained shear tests conducted using backpressure to saturate specimens are not possible with this type of equipment.That is because pore water pressure and a sufficiently high backpressure would lead to bulging of the reinforced rubber membrane (Franke et al., 1979).The constant volume of the specimen is obtained by adjusting the vertical load to maintain a constant height.The change in vertical stress is then assumed to correspond to the change in pore water pressure of an equivalent test with pore water pressure measurements.Casagrande (1976), DeAlba et al. (1976) and Ansell & Brown (1978) made some adjustments to the equipment.
The aim of all these apparatuses was to apply a simple shear mode of deformation to a soil specimen, but the need for the ends of the specimen to extend during shearing means that complementary shear stresses are not generated on the ends (Boylan & Long, 2009).Because of this, the shear stress is non-uniform across the top and bottom of the specimen, falling to zero at the corners.The resulting unbalancing couple has to be counteracted by an opposite couple generated by a non-uniform distribution of normal stress on the top and bottom of the specimen surface (Airey et al., 1985).
These apparatus devices have been criticized as they only measure the total vertical normal stress and the total horizontal shear stress on the specimen during shearing and give no idea of the uniformity of these stresses and true stress state in the specimen (Perazzolo, 2008).To overcome these shortcomings, researchers (e.g.Budhu, 1984, Airey & Wood, 1984) have developed apparatuses, that surround the test specimen with an array of load cells to measure the complete state of stress around the specimen.Radiographic techniques were used to monitor lead shot embedded in the test specimen to give a measure of the internal strains, the uniformity of strains and allow the development of ruptures to be monitored (Budhu, 1984).Research conducted by Airey & Wood (1984) on kaolin showed that direct simple shear tests in a routine apparatus with only the total horizontal shear stress and total vertical stress measured underestimated the simple shear values measured in an elaborately instrumented apparatus by only 10%.It has been suggested on the basis of experimental results that simple shear tests on clay can be presented with more confidence than those conducted on sand (Airey & Wood, 1984).Mao & Fahey (2003) presented the simple shear apparatus manufactured in the University of Western Australia (UWA).In this equipment, the sample is enclosed in an unreinforced latex membrane and contained in a pressurized cell, very similar to a triaxial apparatus.The vertical and cell pressures are controlled independently.A feedbacked system allows total vertical stress to be kept constant during shearing phase while maintaining a constant sample height.This is achieved by locking the vertical loading ram, and using the feedback system to vary the cell pressure to keep the total vertical stress constant.As the height and volume are both constant (for undrained tests), the average cross-sectional area is likely to remain constant also.
As a number of different simple shear apparatuses have been developed over the past few decades, Doherty & Fahey (2011) investigated two different aspects of the devices.The different total stress paths followed by devices that impose constant cross-sectional area using a stiff external boundary, and those that use a constant total stress lateral boundary condition were explored.This was done by conducting finite element analysis of a single cubic element.The authors observed that this element might be subjected to perfect simple shear using four different boundary condition types.Each boundary condition type results in the same effective stress path, but different total stress paths and excess pore pressures.However, significant differences in total stress path and excess pore pressures occur among the four boundary condition types.

Description of the Apparatus
The simple shear device of the Federal University of Rio Grande do Sul (UFRGS) was designed to confine samples with a membrane through confining pressure.In the developed apparatus, sample height is kept constant during shearing and total vertical stress is constant.The maximum horizontal displacement a sample can achieve during shearing is 25 mm, while in UWA equipment the maximum displacement is 10 mm.
Figure 1 presents the basic definitions of the simple shear apparatus.In Fig. 1, s y is the vertical stress, s x is the horizontal stress, D is the diameter of the specimen, h represents the height of the specimen, t xy is the shear stress, e x is the horizontal strain, e y is the vertical strain and g xy is the shear strain.
Shear stress t refers to the shearing loads in the horizontal direction, while the strain caused by shearing, g, is the ratio between the horizontal displacement and height of the specimen.The principal stresses of can be determined through Eq. 1 and 2, being dependent on the vertical and horizontal effective stresses, s' V and s' where CP is the confining pressure, PP is the pore pressure and q is the deviator stress.
The principal effective stresses (s' 1 , s' 2 , s' 3 ) can be obtained from the equation: where s' V and s' H are the effective vertical and horizontal stresses and t is the shear stress on the horizontal plane (Mao & Fahey, 2003).
q and p' can be obtained from Eq. 3 and 4 respectively. (3) The principal requirement of the apparatus design was that all test phases could be controlled and monitored by software.Also, the instrumentation should be as close as possible to the soil sample.
A schematic view of the apparatus and instrumentation is shown in Fig. 2. The specimen is cylindrical and three different diameter samples can be tested in the developed equipment: 50, 75 and 100 mm (the validation tests performed in this research used 100 mm diameter samples).The sample height is usually not higher than half of the diameter to ensure uniform stress distribution.To confine the sample, a latex membrane is used.The specimen is enclosed both top and bottom by two steel plates.These plates present 8 mm salient edges that contain the porous stone and the sample during shearing.Hoses are connected to each plate to allow drainage.This group of pieces is called shear cell, and its schematic view is shown in Fig. 3.The two plates and the specimen are placed in the equipment and connected to pins linked to the load cell.
The horizontal displacement system, vertical force application system, confining pressure and backpressure system, instrumentations and control system are described in the next sections.

Horizontal displacement system
The horizontal displacement is applied with a servo motor system.The rotation of the motor is converted into horizontal displacement through a recirculating ball screw.A piston is linked to the ball screw and connected to the load cell, responsible for measuring the shear stress supported by the sample.The load cell is involved by metal pieces connected at the bottom to a slider and a rail system that allows horizontal displacement, and at the top to the shear box, linked to the load cell through a rigid pin.In addition, a linear displacement transducer is connected to this piece assembly to measure horizontal displacements.

Vertical force application system
The same configuration of servo motor system is used to apply vertical force on the top of specimen.The rotation of the motor is converted into vertical translation through a recirculating ball screw.A piston is linked to the ball screw and connected to a load cell, responsible to measure the vertical force applied to the sample.A linear transducer is fixed to the reaction system used to impose constant height during shearing.Linear guides guide the system up and down movement.

Confining and back pressures
The apparatus was design to receive air as fluid to apply confining pressure.The air was used as confining fluid to allow all the instrumentations be placed inside the chamber without the need of waterproof instruments, also the large volume of fluid required to fill the chamber contributed to this option.
Compressed air is received by a proportional valve, which sends the pressure requested by the control system to the chamber.The pressure imposed on the system is measured and checked by a pressure transducer located at the top of the chamber.
The application of backpressure was based on the consolidation systems used in triaxial tests.Water is pressurized by a piston moving inside a cylinder.The piston is actuated by a ball screw and a servomotor guided by sliders and rails.
This system is responsible for applying backpressure and measuring volume change.A pressure transducer linked to the system controls the backpressure imposed to the specimen.As the volume of the cylinder and the position of the servomotor are known, it is possible to measure any volume change in the specimen.Digital limitations of movement are used to prevent the piston to exceed the limits of the cylinder.The servomotor stops when the digital limitation is triggered.

Instrumentation and control system
The digital control system for the simple shear device serves two purposes.The first is to provide control signals to the drives that control the three servomotors of the equipment (horizontal displacement, vertical force and backpressure system) and the proportional valve (confining pressure).The second purpose is to acquire data from the load cells, linear transducers and pressure transducers.
A PCI-2517 board was used to control the test and measure the instrumentation.This enables guaranteed sampling frequencies with an internal feedback loop of 1 kHz.Differential analog input channels are used for the two load cells, two LVDTs and two pressure transducers.
The board has four analog output channels with resolution of 16 bits.Accuracy of analog input is 0.031%.The board channels control the three motors and the proportional valve.The proportional valve is from Norgren.Is has a total error smaller than 1% and the response time is smaller than 100 ms, maximum pressure is 10 bar.The servomotors used in the apparatus are from Delta Instruments.They have a power of 0.74 kW, torque of 2.39 Nm and maximum rotation of 3000 rpm.
A LabVIEW algorithm was implemented to control and read the instrumentation during tests.The simple shear device operates under strain control or stress control which requires two different shear routines.
In strain controlled tests, the reading of the horizontal LVDT is required to set the direction and way of motion of the servomotor.In stress controlled tests, the servomotor operates according to the measurement of the horizontal load cell.In addition, the confining pressure transducer governs the operation of the pressure valve.Differing from the confining pressure, the backpressure is applied with a servo-controlled motor, which allows the measurement of volume change of sample during tests.Instrumentation is read and recorded every 20 ms.
Before conducting any test with the apparatus, all measuring instruments, such as load cells, pressure transducers and LVDTs were calibrated.The servo-controlled motors were also calibrated, so routine test could be conducted through a linear velocity, instead of an angular one.Figure 4 presents the developed simple shear apparatus.

Test Procedure
Once the size of the specimen is chosen, the corresponded plates are positioned on the equipment.The sample is positioned between the bottom and top caps and the latex membrane is placed.The set is positioned in the apparatus.The vertical piston is positioned in order to touch the set.With the entire fixation finished, the chamber is closed and all the zeros of the test are read.
The phases of percolation (upwards water flow is established through the sample), saturation (incremental steps of backpressure are imposed to achieve saturation) and consolidation (sample is conducted to the effective stress state desired before shearing) are conducted with the according software specific routine.Once the first steps of the test are concluded, the sample is ready to be sheared.The routines of monotonic, strain controlled or stress controlled loading are then opened and conducted.
The sand is classified as uniform fine sand.Quartz corresponds to 99% of the mineralogical composition.Osorio sand has a specific gravity of solids of 2.62; uniformity coefficient, C u , of 2.1; curvature coefficient, C c , 1.0; its effective diameter, D 10 , is 0.09 mm; mean diameter, D 50, 0.16 mm, minimum void ratio, e min , 0.6 and maximum void ratio, e max , 0.9.
Specimens were prepared with a split mold.The sand was manually mixed with 10% of water and compacted in two layers by tamping inside the mold at a relative density of 50%.The size of the specimens was fixed with diameter of 100 mm and 50 mm height.The latex membrane was positioned inside the mold, which has a hollow tube to apply vacuum.A vacuum pump was used to approximate the latex membrane to the mold during sample compaction.Orings were used for sealing.Once the specimen was prepared, the top cap was positioned and the set was put in the apparatus.Figure 5 presents the method of preparation of specimens and the test procedure.

Simple shear testing
The main focus of the tests was to validate the equipment.Thus, tests were carried out with monotonic and cyclic loading.Specimens were tested with relative density of 50%.The initial effective stresses were 50, 100 and 150 kPa for monotonic loading and 100 kPa for cyclic loading.For all tests the backpressure was 300 kPa and the confining pressure was changed according to the initial effective stress.This range of effective stresses was chosen so that the results observed in this research could be compared against previous works that studied similar levels of effective stresses (Casagrande, 2005, Festugato, 2008, Marcon, 2005).Shearing was performed under undrained conditions.For monotonic tests, the displacement rate of 0.1 mm/s was adopted based on previous work of Festugato et al. (2013).Cyclic testing was conducted under strain controlled conditions with initial effective stress of 100 kPa, a shear strain amplitude of ± 2.5% and frequency of 0.1 Hz (Festugato et al., 2013).Monotonic simple shear tests were analyzed through shear stress-shear strain, pore pressure variation-shear strain, variation of vertical effective stress-shear strain curves and stress paths.Cyclic simple shear tests were studied through shear stress-shear strain curves; shear stress, shear strain, variation of pore pressure and variation effective vertical against the number of cycles curves, and stress paths.From the curves, strength and stiffness parameters are defined: effective internal friction angle, g', and shear modulus, G, respectively.

Simple Shear Test Results
In order to demonstrate the performance of the simple shear apparatus, tests were conducted on Osório sand specimens.The obtained results were compared to results of other authors.

Monotonic test results
The monotonic tests were performed under undrained conditions, with constant displacement speed of 0.1 mm/min, equivalent to a shear strain rate constant of approximately 0.2%/min.
Figure 6 presents results of the monotonic test at the initial effective vertical stress of 50, 100 and 150 kPa.The sandy matrix under undrained simple shear conditions presents slightly pronounced strength peak, followed by a shear stress reduction associated with the increase of pore pressure (Fig. 7).
The shear stress rises up to a level around 50 kPa falling to about 40 kPa until 10% strain for the specimen tested with an initial vertical effective stress of 50 kPa.After this fall, the sand specimen regains strength.
The specimen tested with 100 kPa initial vertical effective stress exhibited similar behavior to the 50 kPa test.Under undrained monotonic simple shear conditions, the sandy soil presented a slightly pronounced strength peak.The shear stress increased up to a level of 85 kPa falling to about 80 kPa at 22% strain.Analogous behavior was observed for 150 kPa initial effective vertical stress.
The variation of pore pressure and effective vertical stress is presented in Fig. 7.For the test conducted with initial effective vertical stress of 50 kPa, the variation of pore pressure increases to around 7 kPa at 3% strain.After that, the pore pressure increment decreases until the end of the test when it comes to -90 kPa.In response to pore pressure variation, to guarantee plane strain conditions with constant volume, effective vertical stress initially at 50 kPa is reduced to 43 kPa at 3% strain.It then undergoes a gradual increase, reaching 140 kPa at a strain of 33%.
In the test of 100 kPa initial effective vertical stress, the pore pressure increment increased up to 15 kPa, reducing afterwards to -90 kPa upon reaching the deformation 20% and continued to reduce until the end of the test.In response to increasing pore pressure, to guarantee conditions of plane strain with constant volume, the effective vertical stress initially at 100 kPa reduced to 85 kPa, reaching afterwards 190 kPa at 20% strain.
With an initial effective stress of 150 kPa, pore pressure increases to a value of 36 kPa and remains constant until the end of the test.Effective vertical stress reduces to 114 kPa in response.
In Fig. 8 the results are presented as p' vs. q.Results show the same behaviors for all tests until the maximum value of q is reached in each test.
The ratio between shear stress and effective vertical stress vs. shear strain is presented in Fig. 9.It was observed the expected behavior.For all monotonic tests, as strength of cohesionless materials is essentially derived from friction, shear stress normalized by the vertical effective stress is shown to be similar.

Cyclic test results
Cyclic simple shear tests were conduced with the initial effective vertical stress of 100 kPa.Due to a problem in data acquisition, the first readings were not recorded by the software.The data obtained starts for Test 1 at vertical effective stress of 65 kPa and for Test 2 of 85 kPa.Cyclic loading was performed under strain controlled conditions.The frequency was 0.1 Hz and the shear strain amplitude was ± 2.5%.
Figure 10     cycles present higher values of strength.With increasing number of cycles, strength decreases with the reduction of effective vertical stress.The decrease of the effective vertical stress during the test can be seen in Fig. 11.This reduction is quite pronounced during the test.This decrease is a response to the pore pressure variation.It is observed the increase in pore pressure increment with increasing cycles, resulting in reduction of the effective vertical stress.Although the pore pressure during cyclic test has presented distinct behavior between Test 1 and Test 2, both tests resulted in the same strength envelope (Fig. 12).

Comparison of results
The studied sand behavior was analyzed through testing at different initial effective vertical stresses.For monotonic tests, initial effective vertical stress were carried with 50, 100 and 150 kPa.Cyclic tests were conducted with 100 kPa initial effective vertical stress.Figure 12 gathers the shear stress-shear strain curves of the three monotonic tests and the cyclic tests.The strength peaks for monotonic tests are observed at shear strains varying from 5 to 7%.After these peaks, the strength falls.Such behavior is expected for medium dense sands (Atkinson, 1993, Wood, 1990, Lambe & Whitman, 1979).In the initial cycles of testing, it is verified that the stress path of cyclic tests did not reach the failure envelope.After the third cycle, the stress path reaches the failure envelope.The test was conducted until the effective vertical stress was reduced to zero for both tests.
To analyze the strength parameter, a straight line was fitted to the points of maximum shear stress.The internal friction angle of 35°was obtained and the cohesive intercept was zero.
The shear modulus (G) variation with shear strain is presented in Fig. 13.It can be observed that the three analyzed stresses had very similar shear modulus variation trend.As shear strain evolves the shear modulus degrades.Atkinson & Sallfors (1991) and Mair (1993) observed also a stiffness degradation for a wide range of shear strain.

Validation of the apparatus
In order to validate the developed apparatus, the strength parameters derived from the testing of this study were compared against literature results (Table 1).It was found compatibility of the results obtained for the Osório sand with previous studies on this same material.tinct researches were evaluated.Three of these performed triaxial tests (Casagrande, 2005, Festugato, 2008, Marcon, 2005) and one of them carried out direct shear tests (Marcon, 2005).Drained triaxial tests conducted by Marcon (2005) found the friction angle was 35.9°.The mean void ratio in the study was 0.71 with relative density of 67% and the initial confining stresses were 50, 100 and 200 kPa.The author also conducted direct shear tests under normal stresses of 50, 100 and 200 kPa, obtaining a friction angle of 34.9°.Casagrande (2005) examined the behavior of Osório sand with relative density of 50% carrying out consolidated drained triaxial tests.The author used confining pressures of 20, 100, 200 and 400 kPa and obtained a friction angle of 33.5°.Festugato (2008) carried out drained triaxial tests with Osorio sand (at relative density of 50%) under the confining pressures of 50, 100 and 200 kPa.The effective friction angle was of 37.0°.
Such testing results are presented in Table 1 along with the current study data.
Through the results comparison, it can be seen that the effective internal friction angle obtained through the simple shear equipment is consistent with the reference values found in the literature.There was a slight variation of the values, which can be explained by the different stress paths performed.

Conclusions
A simple shear apparatus has been developed to test soil samples.The apparatus uses internal monitoring instruments.Shear stress can be applied considering monotonic or cyclic loading.Cyclic loading can be conducted under strain or stress controlled conditions.
The results of the simple shear tests presented in this paper indicate that the equipment has consistent and adequate quality results.When results of the simple shear tests are compared against results on this same Osorio sand from literature, values are comparable.Unlike usual triaxial tests, the simple shear test presents the advantage of allowing the simulation of complete rotation of the stress state by imposing a plane strain condition.
Despite the limited number of performed tests, the observed results were consistent and presented good agreement.A sound indication of that is the same strength enve-lope limiting all stress paths .The results presented in Table 1, show very low variety from the internal friction angle obtained on the simple shear.From literature, minimum value found was 33.5°and maximum was 37°, the average was 35.3°.Simple shear tests performed obtained a strength parameter of 35°.The shear strength parameter f' obtained in triaxial tests showed minimum variations compared with the value found with the simple shear apparatus.When compared to the direct shear test, the internal friction angle obtained through the simple shear tests varied very little.
It was observed that the behavior of the proportional valve, with an associated pressure transducer, showed optimum performance, maintaining the confining pressure with little variation.The behavior of the servomotor assembly and the associated pressure transducer was suitable for the requirements.Pressures were kept stable throughout the tests without leaks.
Load cells and displacement transducers showed expected performance.Small oscillations could be found during the analysis of the results, due to electromagnetic induction created by the drivers of the engines.These oscillations did not affect the results.The calibration of the engines was adequate and their behavior was satisfactory to perform simple shear tests.
The preparation of specimens, assembly and completion of simple shear tests showed no major difficulties.The developed equipment allowed the evaluation of soil behavior when subjected to cyclic loading..All analyzes obtained in this type of device were made in effective shear stresses.
Due to the reasons cited above, the equipment developed was considered satisfactory for the execution of simple shear tests.New researches will be developed using this same equipment.
Figure 1 -Basic definitions of the simple shear apparatus.

Figure 2 -
Figure 2 -Schematic view of the apparatus.

Figure 3 -
Figure 3 -Schematic view of the shear cell.

Figure 5 -
Figure 5 -Specimen preparation and test procedure: (a) insertion of saturated porous stone on top and bottom cap; (b) placing of saturated filter paper; (c) preparation of latex membrane inside split mold; (d) vacuum application; (e) compaction of soil specimen; (f) finalization of specimen preparation and cleaning of the membrane surface; (g) insertion of the top cap; (h) insertion of specimen on the simple shear apparatus; (i) specimen after shearing.

Figure 12 -
Figure 12 -Strength envelope for cyclic and monotonic tests.

Table 1 -
Tests analyzed to compare the strength parameters.