Вы находитесь на странице: 1из 7

Journal of Structural Engineering

Vol. 44, No. 4, October 2017  pp. 377-383 No. 44-36

Numerical and experimental simulation of liquefaction for scenario earthquake


P. Kamatchi*,, M. Gayathri** and M. Dilna**
 Email: kamat@serc.res.in

*CSIR-Structural Engineering Research Center, CSIR Campus, Taramani, Chennai - 600 113, India
**Rajalakshmi Engineering College, Chennai - 602 105, India

Received: 30 May 2017; Accepted: 09 August 2017

In this paper, numerical simulation of liquefaction has been done and comparisons are made with the results from
experiment on shake table. Properties of river sand sample which are required for numerical simulations are obtained
through laboratory tests. Scenario earthquake is artificially generated with seismological parameters of Mw 7.6 Bhuj
earthquake January 2001 and numerical simulations are carried out using OpenSeesPL computer program. The excess
pore pressure ratios, shear stress-strain histories and stress paths are studied. River sand sample passing through 1.18
mm sieve is filled in Perspex glass box for a dimension of 1.15 m × 1 m × 0.54 m and tested on shake table. A model
frame of 7 storey building is placed in the soil specimen and instrumented for the observation of settlement. The
time of occurrence of liquefaction observed from numerical simulation is 4.8s and found to be in agreement with the
observation from the experiment on shake table. Further, the acceleration response at 110 mm from base observed in the
experiment and numerical simulation are compared and are found to be in good agreement. This study, demonstrates
the applicability of numerical simulations in prediction of liquefaction phenomenon and dynamic response analysis of
typical soil sites for scenario earthquakes.

Keywords: Liquefaction; experiment on shake table; numerical simulation; scenario earthquake.

Number of studies are reported in literature on post the chosen sample. Three dimensional finite element
earthquake damage due to liquefaction1-4. Damages modelling of soil specimen is subjected to earthquake
occurred during Mw 7.6 Bhuj earthquake, January shaking in OpenSeesPL computer program15 and the
2001 in the form of ground deformation and settlement time history dynamic response of the soil specimen
of foundation are reported by Rastogi5 and Dash, et from numerical simulations are studied. To simulate
al.6,7. Various studies on laboratory experiments and the settlement of building during liquefaction, a typical
numerical simulation of liquefaction are also reported in model frame has been placed in the liquefiable soil set
literature8-14. In the present study, numerical simulation up on shake table and the horizontal displacement and
of liquefaction has been done and the comparisons are vertical settlement are observed during experiment for
made with the results from experiment on shake table. the scenario earthquake.
Experimental and numerical simulation of liquefaction
phenomenon are carried out for an artificial ground Properties of sample
motion simulated at Kandla port for a scenario
Properties of sample which are required for numerical
earthquake of Mw 7.6 with seismological parameters
simulation are estimated through laboratory
of Bhuj 2001 earthquake. Through different tests
experiments. Grain size distribution test of the sample
carried out at laboratory, properties of the sample viz.,
was carried out through mechanical sieve analysis for
specific gravity, relative density, permeability, grain
1 kg of sample passing through 1.18 mm sieve size
size distribution and unit weight, are determined for

Journal of Structural Engineering 377


Vol. 44, No. 4, October - Novemeber 2017
and percentage finer values are estimated. For 600, proposed by Motazedian and Atkinson22. Ground
300, 150, 75 micron sieves the % finer values are 51%, motions are simulated for latitude and longitude of
18%, 4%, 0%, respectively. From the gradation curve, Kandla port site for rock outcrop as given in Fig. 1. Peak
sample chosen has been classified as uniformly graded. ground acceleration of the ground motion simulated is
The relative density of the sample has been estimated 0.49g which is in close agreement with the maximum
as 50.1% from natural, maximum and minimum void acceleration value reported for the main shock of Mw
ratios. The specific gravity, average permeability and 7.6 for stress drop of 200 bars reported by Singh, et
angle of internal friction of the sample are given al.21, for Kandla port site. Present study do not include
in Table 1. The shear modulus, bulk modulus and wave propagation effects and the rock outcrop ground
combined bulk modulus of the sample are estimated motions are used typically for the experimental and
through analytical expressions and given in Table 1. numerical analysis.

Table 1 0.6

Acceleration (g)
0.4
Properties of the sample through laboratory 0.2
0
experiments -0.2
Properties Test Results -0.4
-0.6
Relative density (Dr) 50.1 % 0 5 10 15 20 25
Time (sec)
Maximum void ratio (emax) 0.795
Fig. 1 Artificial ground motion time history for the scenario
Minimum void ratio (emin) 0.472 earthquake of Mw 7.6 simulated at Kandla port with focal
Void ratio (e) 0.633 parameters of Bhuj 2001 earthquake
Maximum unit weight (γmax) 16.87 kN/m³
Minimum unit weight (γmin) 13.93kN/m³ Table 2
Unit weight (γ) 14 kN/m3 Seismological parameters for the simulation
Angle of internal friction (f) 34.4º of Bhuj 2001 earthquake

Confining pressure 100 kPa Parameter Value Parameter Value


Shear modulus (Gmax) 6.186e4 kPa Sub fault width
Fault strike 660 8.25
(km)
Bulk modulus (K) 1.34e5 kPa
hypocenter at sub
Combined bulk modulus 3.46e5 kPa Fault dip 640 34
fault
Permeability 5.42e-5 m/s Fault depth to
Specific gravity 2.59 10 Magnitude 7.6
upper edge (km)
Fault length (km) 44 FFT points 4096
Scenario earthquake Fault width (km) 33 dt (sec) 0.02
Shear wave
Due to the scarcity of recorded ground motions,
velocity of rock
generation of site-specific artificial strong motions Latitude of source 23.410
near the source (β)
3.6
using stochastic models by identifying major fault (km/sec)
zones and propagating seismic waves generated at Longitude of
70.180 Density (ρ), g/cm3 2.85
these potential sources to the sites of interest is well source
accepted in literature16,17. In this process, path effects Latitude of site 23.030 Pulsing percentage 25
and anelastic attenuation effects predicted by the Longitude of site 70.220 Stress drop (bars) 200
empirical and theoretical models18 are used. For source No. of sub faults
5 fmax (Hz) 35
representation, point source models19 or finite source along strike
models20 are widely used. In this paper, artificial ground No. of sub faults
4 Q0 508.00f0.48
motions are generated for Mw 7.6 earthquake with along dip
parameters21 of Bhuj earthquake as given in Table 2, Sub fault length Saragoni Hart
8.8
using extended finite source stochastic model (EXSIM) (km) window

378 Journal of Structural Engineering


Vol. 44, No. 4, October - Novemeber 2017
Numerical analysis 2.00

presssure (kPa)
Excess pore
1.50
The numerical simulation reported in this paper is 1.00
carried out using open source computational platform 0.50
0.00
OpenSees23 with the graphical user-interface24 0 5 10 15 20 25
“OpenSeesPL”. In OpenseesPL, both 8-4 node (8- Time (s)
solid translational degrees of freedom with 4 corner (a)

nodes describing fluid pressure), 20-8 node (20- solid 3.00

presssure (kPa)
Excess pore
translational degrees of freedom with 8 corner nodes 2.00

describing fluid pressure) effective stress fully coupled 1.00


(solid-fluid) 3D brick elements are implemented24. In 0.00
the present study, 8-4 node brick elements are adopted 0 5 10 15 20 25
Time (s)
for all the simulations. The soil model of size 1.15m (b)
× 1m × 0.54 m is created and is subjected to one 2.50

presssure (kPa)
simulation of artificial ground motion generated for 2.00

Excess pore
1.50
the scenario earthquake of Mw 7.6 simulated at Kandla 1.00
port with focal parameters of Bhuj 2001 earthquake. 0.50
0.00
Rigid box boundary condition is adopted for the 0 5 10 15 20 25
numerical study simulating the Perspex box used for Time (s)
(c)
experiment. The deformation contour mesh of sample
6.00
from numerical model of the soil sample during the
presssure (kPa)
Excess pore 4.00
initiation of occurrence of liquefaction when the excess
pore pressure ratio approaches 1 at 4.8s is shown in 2.00

Fig. 2. Excess pore pressure time histories at different 0.00


0 5 10 15 20 25
depths from surface are given in Fig. 3. It can be seen Time (s)
that the maximum excess pore pressure sets at around (d)
5s which lasts upto 10 s then drops. The shear stress 8.00
presssure (kPa)
Excess pore

6.00
and strain variation along different depths are shown 4.00
in Fig. 4. From Fig. 4 maximum shear stress and shear 2.00
strain are observed near the base. Shear stress path 0.00
and the variation of shear stress with mean effective 0 5 10
Time (s)
15 20 25

stress at different depths are shown in Fig. 5. From (e)


Fig. 5 consolidation is observed at top layers and Fig. 3 Excess pore pressure time histories at different depths
consolidation and dilation both are observed at 0.548m from surface (a)0.108m (b) 0.216m (c)0.323m (d)0.431m
bottom most layer. (e)0.55m

Model for experiment


The experimental study on liquefaction is done on
shake table of size 2m × 2m at advanced seismic testing
laboratory of CSIR-Structural Engineering Research
center located at Chennai, India. Sand sample for
liquefaction experiment is filled in a perspex box of
size 1.15m × 1m × 0.9 m (Fig. 6) for a depth of 0.54 m
and saturated with water. Model frame of the building
with the typical seven stories is fabricated by using
wooden boards of size 600 mm × 600mm × 16mm as
Fig. 2 Deformation contour mesh of liquefaction at 4.8s
slab and reinforcement bars of 10 mm diameter to be

Journal of Structural Engineering 379


Vol. 44, No. 4, October - Novemeber 2017
acting as columns as shown in Fig. 6. The test set up the scenario earthquake of Mw 7.6 simulated at Kandla
is instrumented with accelerometers, dynamic pressure port with focal parameters of Bhuj 2001 earthquake
transducers and laser based displacement measuring and the acceleration responses at different levels are
Optho NCDT instruments and connected to data obtained. During the experiment liquefaction got
acquisition system (Fig. 7). initiated in the first 5 seconds as shown in Fig. 8. The
settlements, horizontal displacement and acceleration at
20.00
top floor of model frame are shown in Fig. 9. Variation
Shear stress (kPa)

15.00
10.00 in the settlement of the two edges of the frame indicates
5.00
0.00
the order of tilting of building which is a typical
-5.00 failure mode due to liquefaction. The peak values of
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
response of the structural model are recorded through
Shear strain (%)
(a) instruments and are given in Table 3. The comparison
20.00 of acceleration response from experiment and from
Shear stress (kPa)

15.00 numerical analysis at 110 mm from base is given in


10.00
5.00
Fig. 10. The peak ground acceleration(PGA) observed
0.00 from experiment is in close agreement with the PGA
-5.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
from numerical simulation.
Shear strain (%)
10.00

Shear stress (kPa)


(b) 8.00
20.00 6.00
Shear stress (kPa)

4.00
15.00 2.00
10.00 0.00
5.00 -2.00
0.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
-5.00 Mean effective stress (kPa)
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 (a)
Shear strain (%)
10.00
Shear stress (kPa)

(c) 8.00
25.00 6.00
Shear stress (kPa)

20.00 4.00
15.00 2.00
10.00 0.00
5.00 -2.00
0.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
-5.00 Mean effective stress (kPa)
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
(b)
Shear strain (%)
10.00
Shear stress (kPa)

(d) 8.00
Fig. 4 Shear stress-strain at different depths from surface 6.00
4.00
(a) 0.085m (b) 0.193m (c) 0.305 (d) 0.548m 2.00
0.00
-2.00
Two numbers of accelerometers are placed at 11 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
cm and 28 cm from the bottom of the soil to measure Mean effective stress (kPa)
(c)
acceleration response. Three Optho NCDT instruments
viz., two for measuring vertical displacement and one 10.00
Shear stress (kPa)

8.00
for measuring horizontal settlement are used as shown 6.00
4.00
in Fig. 7 for measurement of settlement and horizontal 2.00
displacement of model frame during shaking. 0.00
-2.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
Results from experiment Mean effective stress (kPa)
(d)
The perspex glass box with sand and model frame is Fig. 5 Stress path at different depths from surface (a) 0.085m
exited with the artificial ground motion generated for (b) 0.193m (c) 0.305 (d) 0.548m

380 Journal of Structural Engineering


Vol. 44, No. 4, October - Novemeber 2017
15
10

Displacement (mm)
5

-5

-10
-15
0 5 10 15 20 25
Time (sec)
(a) Horizontal displacement
8
6

Displacement (mm)
4
2
0
-2
Fig. 6 Sand in perspex box with model frame on shake table
-4
-6
0 5 10 15 20 25
Time (sec)
(b) Vertical displacement at 3rd floor level - left side
4
3
Displacement (mm)

2
1
0
-1
-2
-3
-4
-5
0 10 5 15 20 25
Time (sec)
(c) Vertical displacement at 3rd floor level – right side

Fig. 7 Instrumentation set up (left side) 6


Acceleration (m/sec2)

4
2

0
Left side

-2

-4

-6
0 510 15 20 25
Time (sec)
(d) Acceleration response at top floor of test model
Fig. 9 Time history response measured from experiment on model
Fig. 8 Occurrence of liquefaction
frame

Journal of Structural Engineering 381


Vol. 44, No. 4, October - Novemeber 2017
Table 3 soil are studied. The shear stress, strain and stress path
Peak values of response parameters from are observed for different depths of soil stratum.
experiment A model frame simulating a seven storey building
Parameter Peak values has been placed in the liquefiable soil set up on
Settlement of model frame measured at shake table and the horizontal displacement, vertical
4.41 mm
3rd floor - left side settlement and acceleration responses are observed
Settlement of model frame measured at during experiment for the scenario earthquake. This
6.46 mm
3rd floor – right side methodology followed in the present study can be
Horizontal displacement of model adopted to estimate the performance of the buildings
13 mm
frame measured at 5th floor and possible settlement under liquefaction during future
Acceleration response measured at 1st
0.507 g earthquakes. Further, the possibility of occurrence
floor level of model frame of liquefaction phenomenon, maximum strains and
Acceleration response measured at 3rd displacements can be estimated through numerical
0.594 g
floor level of model frame
simulations as reported in the present study.
Acceleration response measured at 7th
0.468 g
floor level of model frame
Acknowledgement
Acceleration(m/s2)

6.00
4.00 110mm from base This paper is being published with the kind permission
2.00
0.00 of the Director, Prof. Santosh Kapuria CSIR-SERC.
-2.00 The authors are thankful to Dr. K. Balaji Rao,
Advisor[M], Dr. N. Gopalakrishnan, Chief scientist,
-4.00
-6.00
0 5 10 15 20 25 Dr. K. Ramanjaneyulu, Chief scientist, Smt. R. Sreekala
Time (sec) Principal scientist and technical staff of Advanced
(a) Acceleration response from Open Sees PL
seismic testing and research laboratory, CSIR-SERC
for the encouragement and support extended during the
Acceleration(m/s2)

6.00
4.00 110mm from base
2.00 work reported in this paper. First author acknowledges
0.00
-2.00 Prof. Ahmed Elgamal, USA for providing ‘OpenSeesPL’
-4.00 program for carrying out numerical simulations.
-6.00
0 5 10 15 20 25
Time (sec) References
(b) Acceleration response from shake table test
1. Khan, M.A., Abbast, I.A., Hadi, S., Laghari, A.
Fig. 10 Comparison of acceleration response from numerical
analysis and from experiment and Bilham, R., “Bhuj earthquake of January 26,
2001: Effects in the Thar-Nagar Parkar region of
Conclusions Sindh SE Pakistan”, Geological Bulletin Univ.
Peshawar, Vol. 35, 2002 pp 9–26.
In the present study, experimental and numerical 2. Rathje, E.M., Kelson, K., Ashford, A., Kawamata,
simulation of liquefaction phenomenon are carried out Y., Towhata, I., Kokusho, T. and Barder, J.P.,
for an artificial ground motion simulated at Kandla port “Geotechnical Aspects of the 2004 Niigata Ken
of Gujarat for a scenario earthquake of Mw 7.6 with Chuetsu, Japan Earthquake” Earthquake spectra,
seismological parameters of Bhuj 2001 earthquake. As Vol. 22, (S1) 2006, pp S23–S46.
it has been simulated in the three dimensional finite 3. Ishihara, K., “Liquefaction of Subsurface Soils
element model, liquefaction phenomenon occured at During Earthquakes”, Jl. of Disaster Res., Vol. 1,
around 4.8sec during experiment. The acceleration (2), 2006, pp 245–261.
response of the soil model during shaking from 4. Huang, Y. and Yu, M., “Review of soil liquefaction
experiment and numerical simulation are compared characteristics during major earthquakes of the
and found to be in good agreement. The variation of twenty-first century” Nat Hazards. Vol. 65, 2013,
excess pore pressure with time at different depths of pp 2375–2384.

382 Journal of Structural Engineering


Vol. 44, No. 4, October - Novemeber 2017
5. Rastogi, B.K., “Ground deformation study of Mw 15. Lu, J., Yang, Z. and Elgamal, A., “OpenSeesPL 3D
7.7 Bhuj earthquake of 2001”, Episodes, Vol. 24, lateral pile-ground interaction user’s manual”,
(3) 2001, pp 160–165. University of California San Diego USA, 2010.
6. Dash, S.R., Govindaraju, L. and Bhattacharya, S., 16. Boore, D.M., “Stochastic simulation of high
“On the probable cause of the failure of Kandla frequency ground motions based on seismological
port and customs office tower during the 2001 Bhuj models of the radiated spectra”, Bull. Seism. Soc.
earthquake”, The 14th World Conf. on Earthquake Am., Vol. 73, (6), 1983, pp 1865–1894.
Engg., October 12-17, 2008, Beijing China. 17. Boore, D.M., “Simulation of ground motion using
7. Dash, S.R., Govindaraju, L. and Bhattacharya, the stochastic method”, Pure App. Geophy., Vol.
S., “A case study of damages of the Kandla port 160, (3-4), 2003, pp 635–676.
tower supported on a Mat-Pile foundation in 18. Beresnev, I.A. and Atkinson, G.M., “Source
liquefied soils under the 2001 Bhuj earthquake”, parameters of earthquakes in eastern and western
Soil Dyn. and Earthquake Engg., Vol. 29, 2009, north America based on finite fault modeling”,
pp 333–346. Bull. Seism. Soc. Am., Vol. 92, (2), 2002, pp 695–
8. Elgamal, A-W., Yang, Z., “Numerical modelling 710.
of liquefaction-induced lateral spreading”, 12th 19. Boore, D.M. and Atkinson, G.M., “Stochastic
World Conf. on Earthquake Engg., Paper No. prediction of ground motion and spectral response
2817, 2000. at hard-rock sites in Eastern North America”, Bull.
9. Mitsuji, K., “Numerical simulations for Seism. Soc. Am., Vol. 77, (2), 1987, pp 440–467.
development of liquefaction countermeasures by 20. Hartzell, S.H., “Earthquake aftershocks as Green’s
use of partially saturated sand”, The 14th World functions” Geophy. Res. Lett., Vol. 5, (1), 1978, pp
Conf. on Earthquake Engg., October 12-17 2008 1-4.
Beijing China, 2008. 21. Singh, S.K., Bansal, B.K., Bhattacharya, S.N.,
10. Elgamal, A-W., Lu, J. and Forcellini, D., “Mitigation Pacheco, J.F., Dattatrayam, R.S., Ordaz, M,
of liquefaction-induced lateral deformation in Suresh, G.K. and Hough, S.E., “Estimation of
a sloping stratum: three-dimensional numerical ground motion for Bhuj, (26 January 2001, Mw
simulation”, Jl. of Geotech. and Geoenvir. Engg., 7.6) and for future earthquakes in India”, Bull.
ASCE., Vol. 135, (11), 2009, pp 1672–1682. Seism. Soc. Am., Vol. 93, (1), 2003, pp 353–370.
11. Shahir, H. and Pak, A., “Estimating liquefaction 22. Motazedian, D. and Atkinson, G.M., “Stochastic
induced settlement of shallow foundation by Finite-Fault Modeling Based on a Dynamic
numerical approach”, Comp. and Geotech., Vol. Corner Frequency”, Bull. Seism. Soc. Am., Vol.
37, 2010, pp 267–279. 95, (3), 2005, pp 995–1010.
12. Mhaske, S.Y. and Deepankar, C., “GIS-based soil 23. Mazzoni, S., McKenna, F. and Fenves, G.L.,
liquefaction susceptibility map of Mumbai city “Open system for earthquake engineering
for earthquake events”, Jl. of Appl. Geophy., Vol. simulation user manual”, Pacific Earthquake
70, (3), 2010, pp 216–225. Engineering Research Center, University of
13. Wang, B., Zen, K., Chen, G.Q. and Kasama, K., California, Berkeley, 2006, (http://opensees.
“Effects of excess pore pressure dissipation on berkeley.edu/).
liquefaction-induced ground deformation in 1-g 24. Lu, J., “Parallel finite element modeling of
shaking table test”, Geomech. and Engg., Vol. 4, earthquake site response and liquefaction”, PhD
(2), 2012, pp 91–103. Thesis, University of California San Diego La
14. Maharjan, M. and Takahashi, A., “Centrifuge Jolla CA, 2006.
model tests on liquefaction-induced settlement
(Discussion on this article must reach the editor before
and pore water migration in non-homogeneous January 31, 2018)
soil deposits”, Soil dyn. and Earthquake Engg.,
Vol. 55, 2013, pp 161–169.

Journal of Structural Engineering 383


Vol. 44, No. 4, October - Novemeber 2017

Вам также может понравиться