Experimental investigation of reverse fault rupture propagation through cohesive granular soils

Experimental investigation of reverse fault rupture propagation through cohesive granular soils

Geomechanics for Energy and the Environment ( ) – Contents lists available at ScienceDirect Geomechanics for Energy and the Environment journal ho...

NAN Sizes 1 Downloads 75 Views

Geomechanics for Energy and the Environment (

)



Contents lists available at ScienceDirect

Geomechanics for Energy and the Environment journal homepage: www.elsevier.com/locate/gete

Experimental investigation of reverse fault rupture propagation through cohesive granular soils Mohammad Ahmadi *, Mojtaba Moosavi, Mohammad Kazem Jafari Civil Engineering, International Institute of Earthquake Engineering and Seismology (IIEES), No. 21, Arghavan St., North Dibaji St., Farmanyeh, Tehran, Iran

highlights • • • •

Fault rupture through cohesive soil was investigated using 1g physical models. Cohesive soil were modeled by adding different amount of clay to sand. Changes in cohesion affect all aspects of the faulting compared with dry soil. Fault displacement required for outcropping is increased by increase of clay content.

article

info

Article history: Received 25 July 2017 Received in revised form 27 April 2018 Accepted 29 April 2018 Available online xxxx Keywords: Surface fault rupture Physical modeling Cohesive soil Sand Shear band

a b s t r a c t Naturally occurring dry cohesionless soil is rarely found in urban areas; however, previous studies on surface fault rupture propagation using physical modeling has usually concentrated on dry cohesionless soil. In this investigation, the effects of cohesion on fault rupture propagation through granular soil were studied. Physical models were developed in which inherent cohesion was produced by adding different percentages of clay to the sand. A dry test also was conducted for comparison. The results show that changes in cohesion affect all aspects of the behavior of faulting, including fault rupture propagation, fault scarp at surface and required displacement at bedrock for outcropping. It was found that the vertical fault displacement required for outcropping (h0 /H) increased as the percentage of clay increased. © 2018 Elsevier Ltd. All rights reserved.

1. Introduction Cohesive soil is a commonly observed constituent of sedimentary layers. Major cities are usually constructed on cohesive layers. The presence of clay minerals in the topsoil layers can affect surface fault rupture propagation by changing the geomechanical behavior of the soil. Research on fault rupture propagation that is limited to the investigation of cohesionless granular soil using physical modeling or numerical approaches cannot cover all situations in which fault rupture propagates through soil. Some studies have focused on dry cohesionless soil, especially by laboratory testing.1–4 In others, water has been added to the soil for model construction, but the investigation has not focus on the effects of induced cohesion.5–8 Still others have investigated wet and cohesive soil.9–13 Less effort has been devoted to understanding the effect of cohesion on fault rupture propagation. This is could

be related to difficulties in the preparation of laboratory models or problems relating to consolidation in fine materials. An appropriate method for geotechnical studies is 1-g physical modeling. In this type of modeling, as the confining stress of soil decreases (with a decrease in the dimensions), the cohesion of the soil also should be reduced based on scaling laws. This reduced cohesion sometimes cannot be satisfied by modeling or presents difficulties. To overcome these difficulties, the current study added a small percentage of clay to granular soil (sand) to produce cohesive soil. 2. Physical modeling The main purpose of modeling was to understand and recognize the response of cohesive soil to fault rupture propagation. The 1g physical modeling approach was adopted in the present study, because it is economical and accessible. 2.1. Geometry of model and experimental apparatus

* Corresponding author.

E-mail addresses: [email protected] (M. Ahmadi), [email protected] (M. Moosavi), [email protected] (M.K. Jafari).

A split box was used to reproduce reverse faulting in this investigation. Faulting was applied to the model using an electromotor

https://doi.org/10.1016/j.gete.2018.04.004 2352-3808/© 2018 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Ahmadi M., et al., Experimental investigation of reverse fault rupture propagation through cohesive granular soils, Geomechanics for Energy and the Environment (2018), https://doi.org/10.1016/j.gete.2018.04.004.

2

M. Ahmadi et al. / Geomechanics for Energy and the Environment (

)



Table 1 Geomechanical properties of Firoozkooh no. 161 sand in dry condition. Gs

emax

emin

d50 (mm)

Fine percent (%)

ϕ ′ (degree)

c ′ (kPa)

Uniformity Coefficient, Uc

2.65

0.943

0.548

0.27

1

37

0

1.87

Fig. 1. Faulting box and geometry of model.

in quasi-static mode to reproduce what occurs during a real fault rupture.14 Free field conditions were studied in this research. The model simulated a soil layer with a height of 25 cm that was subjected to reverse faulting at a dip angle of 45◦ . The maximum vertical fault displacement was limited to 5 cm in all models. Fig. 1 shows the faulting box used and the geometry of the model. The total height of the soil layer, vertical fault displacement during faulting and the vertical fault displacement at which the fault rupture reaches the surface are denoted by H, h and h0 , respectively. Particle image velocimetry (PIV) was used to recognize shear band formation and its propagation upward to the surface during faulting. The software used was Vision Strain Gauge (VSG) developed by IIEES.15 To capture the propagation of faulting, an 18 MP camera (Canon EOS 650D) was mounted in front of the Plexiglas side to record high-resolution photographs for image processing. 2.2. Soil Firoozkooh #161 sand was used in all tests as the main component of the soil and cohesion was implemented by adding different percentages of clay. Table 1 lists the geomechanical properties of the Firoozkooh sand. Fig. 2 shows the gradation curves for the different specimens. CL clay was added to pure sand for preparation of cohesive soil such that the cohesion of the mixed soil can be controlled according to the percentage of fine material. Liquid limit (LL), plastic limit (PL) and plastic index (PI) of CL clay (commercially named as Kaolinite ZMK2) are 29%, 17% and 12%, respectively. Soil having a value of Dr = 50% was used in all tests. The wet tamping method was employed to construct the model. First, clay and sand were mixed precisely and then 5% water was added to produce homogeneous soil. The soil was poured in layers 2.5 cm in height and tamped to achieve the desired level of density. The density of the layers was controlled by its height after compaction. The number of tamps was determined in pilot tests. Under-compaction was employed to ensure a homogeneous soil layer. Tamping was increased on the top layer to assess the effect of compaction energy. 2.3. Physical modeling programs In this study, four models were conducted. In three, the percentage of clay in the soil was varied to investigate the cohesiveness

Fig. 2. Gradation curve of Firoozkooh no. 161 sand and CL clay.

Table 2 Test programs and characteristics. Test No. Water Content (%) Sand (%) Clay (%)

CSF-10 0 (dry) 100 0

CSF-14 5 95 5

CSF-15 5 90 10

CSF-16 5 85 15

of the granular soil. To compare the response of cohesive and noncohesive soil against faulting, one dry test using pure sand was also carried out. The characteristics of each test are listed in Table 2. 3. Results and discussion 3.1. Fault rupture propagation Fig. 3 shows the results of cohesive testing with 5% clay at h/H = 4%, at the initiation of faulting, microshear bands were produced from the fault tip at bedrock to the surface. Some of these microshear bands reached the surface and deform the surface slightly. Before vertical fault displacement at h/H = 4%, the main fault rupture shear band reached the surface and a distinctive scarp was created. In this situation, development of other microshear bands that did not run through the main fault trace appears to have ceased. At h/H = 8%, localization of shear strain occurred and the overall width of the shear band tended to decrease. Slight bifurcation or branching of the fault occurred in the upper half of the soil layer. At the end of faulting, at h/H = 20%, a new shear band formed from the tip of fault and began to propagate upward. This bifurcation indicates that the amount of relative displacement is likely the most important parameter when determining how a fault rupture can become hazardous. This shows the importance of the prediction of fault relative displacement before the design of a structure when a known fault exists at the construction site. The main difference between the results of cohesive testing with the 10% and 15% clay model with the 5% model was diffusion of the shear bands in the upper layer of soil. An increase of more than 10% in the percentage of clay significantly increased the required fault displacement at bedrock for outcropping. Fig. 3 shows the 10% and 15% clay tests at h/H = 4% in which shear band localization formed above the fault tip at bedrock, but only microshear bands were created and reached the surface over a

Please cite this article in press as: Ahmadi M., et al., Experimental investigation of reverse fault rupture propagation through cohesive granular soils, Geomechanics for Energy and the Environment (2018), https://doi.org/10.1016/j.gete.2018.04.004.

M. Ahmadi et al. / Geomechanics for Energy and the Environment (

)



3

Fig. 3. Cumulative shear strain produced due to faulting at different stage for cohesive soil with various amount of clay.

relatively large area. This means that the ability of the soil layer to absorb the fault rupture increased. At this stage of faulting (h/H = 4%), only a smooth surface profile was created at the surface and no sharp scarp was observed. Another significant difference between the 5% clay test and 10% and 15% clay tests was the size of the faulting zone created by the shear band. In the 5% clay test, shear bands formed relatively uniformly along all parts of the soil layer. However, it is known that localization occurs more often at depth due to the confining pressure of the soil. In the 10% and 15% clay tests, shear strain localization was concentrated at one-third the depth of the soil layer above the bedrock. This is because the confining pressure of the soil decreased at the top soil layer, increasing the effect of cohesion. An increase in confining pressure increased strain localization. As the soil depth increased, the effect of cohesion decreased in comparison with the confining pressure. Thus, localization occurred easily at depth. Near the surface, the confining pressure decreased and cohesion effected microshear band formation. This means that shear strain occurred across a wider zone and strain localization required increased fault displacement to occur. For cohesive testing with 5% clay, the maximum cumulative shear strain calculated by PIV was about 60%. The maximum shear strain measured was 840% and 340% for the 10% and 15% clay tests, respectively. It should be noted that the calculated value for maximum shear strain can be used qualitatively; however, the range of values must be sufficiently reliable. In dry soil tests, as shown in Fig. 4, the maximum shear strain of soil was limited to about 70%, meaning that rupture diffused across a wider zone. Secondary main fault rupture also occurred at the end of faulting and this secondary fault widened the distorted area at the surface.

3.2. Fault scarp at surface As shown in Fig. 5, the surface deformation of dry (with no clay) and cohesive soil at various stages of faulting differed widely,

which demonstrates the effect of cohesion on fault scarp creation. For the dry test, a smooth profile was created as the gradient at the surface and in the vicinity of the scarp increased as faulting progressed. A sharp scarp was created in the test with 5% clay. The sharp change in the gradient of the soil in the narrow zone around the scarp could cause serious damage to aboveground structures. Scarp formation in this test appeared as a crack. The deformation profiles for 10% and 15% clay were relatively similar. The formation of a distinctive scarp at the surface was the result of increased fault displacement. Changes occurred in the surface elevation on both sides of the fault. The relatively large amount of cohesion prevented formation of a crack at the scarp and a transition area was created that caused the surface gradient to change slightly, as in the dry test. The biggest difference between them was the width of the gradient. 3.3. Vertical fault displacement at bedrock for outcropping The fault displacement required at bedrock (h0 ) for the fault rupture to reach the surface is an important parameter for designers. An obvious scarp formed at the surface and further displacement occurred in this narrow area. It is important in the design of structures to protect against fault rupture. If the probable displacement of a specific fault during the design life of a project is less than h0 , then the fault poses little threat to the performance of the project. Of course, the normalized form of this parameter is generally used. This parameter is usually normalized using H (soil layer height) and is expressed as percentage. In brittle material, h0 /H is less than ductile. Fig. 6 shows h0 /H for dry and cohesive tests versus percentage of clay. The value of h0 /H for dry soil was about 3.5%. With an increase in clay, h0 /H initially slightly decreased and then increased significantly. This suggests that a small amount of clay and the existence of intergranular water cause the soil to be slightly more brittle than in the dry condition. The clay content was insufficient

Please cite this article in press as: Ahmadi M., et al., Experimental investigation of reverse fault rupture propagation through cohesive granular soils, Geomechanics for Energy and the Environment (2018), https://doi.org/10.1016/j.gete.2018.04.004.

4

M. Ahmadi et al. / Geomechanics for Energy and the Environment (

)



Fig. 6. Required bedrock fault displacement normalized to height of the soil layer (h0 /H) for outcropping for dry and cohesive soil with different clay content.

to force the soil to behave in a plastic manner. An increase in the clay content increased the soil plasticity and ductility; thus, h0 /H increased. The values of h0 /H for the 10% and 15% clay tests were approximately equal. 4. Limitations

Fig. 4. Cumulative shear strain at different stage of faulting for dry sandy soil.

Due to reduction of stress level in 1g physical modeling, the results of such investigations usually are used for a qualitative description of the phenomenon. This means that the results should be confirmed by physical models at true scale or in centrifuge (similar to Refs. 16, 17). If naturally cohesive soil is used in 1g smallscale modeling with no decrease in its cohesion, then it cannot represent real cohesive soil. This is because the cohesion part of shear strength will remain constant in a small-scale model in comparison with the prototype, however frictional part is decreased as dimensions of the small-scale model decreased. So, small amount of cohesion in 1g physical modeling tests leads to a significant cohesion in prototype model. Satisfying this reduced cohesion creates limitations for preparation of a 1g model, especially when a large amount of water must be added to the clayey soil to reduce its cohesion.18 It shows the importance of using centrifuge modeling for validation of 1g physical modeling results. It is also known that, the mechanical behavior of granular soils among others also depends on the breakage of grains. The latest is a function of vary factors among which the principals are the shape and dimension of grains, grading, mineralogical constitution of particles, stress level, stress path, presence or absence of water, packing density, grain contact, type of contacts, etc.19–21 Considering the type of sand used in the tests and the low stress level reached in the 1g models, probably the breakage of particles is not very important, but in the real scale, in which the stress level in the fault rupture propagation can be extremely high, it cannot be overlooked. 5. Conclusion

Fig. 5. Progressive scarp formation in various stage of faulting for dry (with no cohesion) and cohesive tests.

In this investigation, the effects of cohesion on fault rupture propagation through granular soil were studied. Physical models were developed in which inherent cohesion was produced by adding different percentages of clay to the sand. Fault rupture behaved differently in all aspect of faulting, including fault rupture propagation, fault scarp at surface and required displacement at bedrock to outcropping. The normalized vertical displacement required for the fault to outcrop (h0 /H) increased with an increase in the percentage of clay. For the 5% clay test, h0 /H decreased slightly in comparison with the dry test. Beyond this limit it is increased. Results of 1g tests must be verified by physical models at true scale or in centrifuge or by numerical models, before to draw general deductions, in special mode for problem that involve fault rupture, in which the difference between the stress level in 1g models and that in true scale is very high.

Please cite this article in press as: Ahmadi M., et al., Experimental investigation of reverse fault rupture propagation through cohesive granular soils, Geomechanics for Energy and the Environment (2018), https://doi.org/10.1016/j.gete.2018.04.004.

M. Ahmadi et al. / Geomechanics for Energy and the Environment (

References 1. Moosavi SM, Jafari MK, Kamalian M, Shafiee A. A experimental investigation of reverse fault rupture –rigid shallow foundation interaction. Int J Civil Eng. 2010;8(2):85–98. 2. Fadaee M, Anastasopoulos I, Gazetas G, Jafari M, Kamalian M. Soil bentonite wall protects foundation from thrust faulting: analyses and experiment. Earthq Eng Eng Vib. 2013;12(3):473–486. http://dx.doi.org/10.1007/s11803-013-0187-8. 3. Bransby MF, Davies MC, Nahas AEl, Nagaoka S. Centrifuge modelling of reverse fault–foundation interaction. Bull Earthq Eng. 2008;6:607–628. http://dx.doi. org/10.1007/s10518-008-9080-7. 4. Lin M, Chung C, Jeng F. Deformation of overburden soil induced by thrust fault slip. Eng Geol. 2006;88(1–2):70–89. 5. Ashtiani M, Ghalandarzadeh A, Towhata I. Centrifuge modeling of shallow embedded foundations subjected to reverse fault rupture. Canadian Geotech J. 2016;53(3):505–519. http://dx.doi.org/10.1139/cgj-2014-0444. 6. Rojhani M, Moradi M, Galandarzadeh A, Takada S. Centrifuge modeling of buried continuous pipelines subjected to reverse faulting. Canadian Geotech J. 2012;49:659–670. http://dx.doi.org/10.1139/T2012-022. 7. Kiani M, Akhlaghi T, Ghalandarzadeh A. Experimental modeling of segmental shallow tunnels in alluvial affected by normal faults. Tunnelling Underground Space Technol. 2016;51:108–119. http://dx.doi.org/10.1016/j.tust.2015.10.005. 8. Kiani M, Ghalandarzadeh A, Akhlaghi T, Ahmadi M. Experimental evaluation of vulnerability for urban segmental tunnels subjected to normal surface faulting. Soil Dyn Earthq Eng. 2016;89:28–37. http://dx.doi.org/10.1016/j.soildyn.2016. 07.012. 9. Bray J, Seed R, Ciuff L, Seed H. Analysis of earthquake fault rupture propagation through cohesive soil. J Geotech Eng. 1994;120(3):562–580. 10. Ng CW, Cai QP, Hu P. Centrifuge and numerical modeling of normal faultrupture propagation in clay with and without a preexisting fracture. J Geotech Geoenv Eng. 2012;138(12):1492–1502. http://dx.doi.org/10.1061/(ASCE)GT. 1943-5606.0000719.

)



5

11. Ahmadi M, Moosavi M, Jafari MK. Water content effect on the fault rupture propagation through wet soil-using direct shear tests. Adv Lab Test Model Soils Shales (ATMSS) Springer Series Geomech Geoeng. 2017;131–138. http://dx.doi. org/10.1007/978-3-319-52773-4_14. 12. Johansson J, Konagai K. Fault induced permanent ground deformations-an experimental comparison of wet and dry soil and implications for buried structures. Soil Dyn Earthq Eng. 2006;26:45–53. http://dx.doi.org/10.1016/j.soildyn. 2005.08.003. 13. Ahmadi M, Moosavi M, Jafari MK. Experimental investigation of reverse fault rupture propagation through wet granular soil. Eng Geol. 2018;239:229–240. https://doi.org/10.1016/j.enggeo.2018.03.032. 14. Ahmadi, M, Fadaee, M, Jafari, MK. A geotechnical investigation on slip rate of surface fault rupture. In: Proc. 7th Int Conf on Seismology and Earthq Eng; 2015 [in Persian]. 15. Fadaee M, Jafari MK, Kamalian M, Mustafa SA. Fault rupture propagation in alluvium and its interaction with foundation: New insights from 1g modelling via high resolution optical image processing techniques. J Seismology Earthq Eng. 2012;14(4):271–283. 16. Valore C, Ziccarelli M, Muscolino SR. The bearing capacity of footings on sand with a weak layer. Geotech Res. 2017;4(1):12–29. 17. Ziccarelli M, Valore C, Muscolino SR, Fioravante V. centrifuge tests on strip footings on sand with a weak layer. Geotech Res. 2017;4(1):47–64. 18. Bray JD. The effect of Tectonic Movements on Stresses and Deformations in Earth Embankment [Ph.D. thesis], University of California Berkeley; 1990. 19. Lade PV, Yamamuro JA, Bopp PA. Significance of particle crushing in granular materials. J Geotech Eng. 1996;122(4):309–316. 20. Valore, C, Ziccarelli, M, The evolution of grain-size distribution of sands under 1D compression. In: Hamza M, et al. eds. 17th Int. Conf. on Soil Mech. and Geotech. Eng. - XVII ICSMGE –Alexandria, Vol. 1; 2009:84-88. http://dx.doi.org/10.3233/ 987-1-60750-031-5-84. 21. Celauro C, Ziccarelli M, Parla G, Valore C. An automated procedure for computing the packing properties of dense and locked sands by image analysis of thin sections. Granular Matter. 2014;16(6):867–880. http://dx.doi.org/10.1007/ s10035-014-0532-2.

Please cite this article in press as: Ahmadi M., et al., Experimental investigation of reverse fault rupture propagation through cohesive granular soils, Geomechanics for Energy and the Environment (2018), https://doi.org/10.1016/j.gete.2018.04.004.