Assessment of liquefaction potential of soils using ...

0 downloads 0 Views 2MB Size Report
(10b) in which, Mw is moment magnitude and amax is the peak ground acceleration in units of gravity. Magnitude correlated duration weighting factor (DWFM) is ...
Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Assessment of liquefaction potential of soils using field tests-based methods: case study Issa Shooshpasha, Ali Hasanzadeh, Sadegh Rezaei & Yasser Ebrahimian Ghajary Department of Civil Engineering, Babol Noshirvani University of Technology, Babol, Mazandaran, Iran Corresponding author: Ali Hasanzadeh, [email protected]

1/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Assessment of liquefaction potential of soils using field tests-based methods: case study Abstract Evaluation of soil liquefaction potential is an important and challenging issue in many geotechnical investigations in earthquake-prone regions. Liquefaction is a common cause of ground failure during earthquakes. The prediction of liquefaction potential of soil is an important step for decreasing earthquake hazard. Babol city is located in a high seismic area. Hence, estimation of liquefaction is very important in this city. In this study, using field-based methods and geotechnical data of 60 available boreholes in Babol, three liquefaction microzonation maps were provided. Finally, one comprehensive liquefaction microzonation map was presented for soil of Babol city. The obtained results in this paper are in good agreement with the previous studies. The results indicate that application of different field tests in evaluation of liquefaction is of great importance. Key words: liquefaction, Andrus and Stokoe method, Idriss and Boulanger method, Moss et al. method, microzonation map

1 Introduction Earthquake-induced liquefaction has caused extensive damages to residential lands and houses, as well as to infrastructures, such as roads, ports and water supply/sewage systems. This phenomenon is associated with the generation of large pore-water pressures in soils due to cyclic loading effects of earthquakes, resulting in a reduction of effective stress, a sudden loss in stiffness and a consequent loss of strength. The investigation of failure of soil masses during earthquakes requires sciences of geology and engineering [1]. To confront liquefaction destructive effects, evaluation of liquefaction potential of soils and recognition of liquefiable regions are essential. There are several methods for determination of liquefaction potential. The liquefaction of soils can be estimated using laboratory tests such as cyclic triaxial and cyclic torsional shear tests. Since the cost of collecting high quality undisturbed samples is considerably high and the laboratory conditions can not simulate actual conditions of field, methods based on in-situ tests such as the Standard Penetration Test (SPT), the Cone Penetration Test (CPT) and the shear wave velocity test (Vs) are widely accepted by geotechnical engineers for estimating the liquefaction of soils. The Standard Penetration Test (SPT), due to its simplicity of execution, is the most commonly used insitu test to gain idea about the stratigraphic profile at a site [2]. SPT-based methods have been adopted for liquefaction evaluation of soils for decades and Standard Penetration resistance has been used as an index of soil liquefaction resistance during earthquakes in engineering practice. Development of SPT-based methods began in Japan by studies performed by some investigators such as Kishida [3] and Ohsaki [4]. Then, many researchers studied and recommended procedures for estimation of liquefaction using SPT [5-11]. The Cone Penetration Test (CPT) is an advantageous test in characterizing subsurface conditions and estimating various soil properties. CPT provides a continuous record of the penetration resistance. In comparison with SPT, CPT is less vulnerable to operator error and can find thin liquefiable strata that are missed by SPT. However, by CPT, no sample can be obtained. CPT is a reliable test that has found widespread use as a tool for evaluating the liquefaction resistance of soils. Development of

2/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

CPT-based methods for evaluation of liquefaction began with work by Zhou [12]. Then, various investigators assessed CPT-based liquefaction methods [13-20]. Moreover, applying shear wave velocity (Vs) measurements for evaluating the liquefaction resistance of soils is an effective method because shear wave velocity and liquefaction resistance are both influenced by similar factors (such as void ratio, geologic age and state of stress). Shear wave velocities can be measured in situ by several tests such as cross-hole, downhole and Spectral Analysis of Surface Waves (SASW). Vs measurements are possible in soils that are difficult to penetrate with CPT and SPT (such as gravelly soils). Generally, the precision of different kinds of shear wave velocity tests is higher than that of penetration tests. However, shear wave velocity testing does not produce samples for classification or may not be performed with sufficient details to detect thin liquefiable layers if the measurement interval is too large. Numerous studies have been performed to investigate the relationship between shear wave velocity and liquefaction resistance [21-30]. Although some researchers conducted studies about soil liquefaction potential of Babol city [31, 32], the obtained results were different because they applied only one method in their investigations. Therefore, in this paper, a focus is made on the most widely accepted methods utilized for estimating liquefaction. For this purpose, three different analysis methods were selected in this study to evaluate the liquefaction potential: (1) Idriss and Boulanger [33] method (which is a SPT-based method); (2) Andrus and Stokoe [34] method (which is a Vs-based method); and (3) Moss et al. [35] method (which is a CPT-based method). In this study, first, seismology and geology of Babol is presented. Then, the utilized methods for assessment of soil liquefaction potential of this city are briefly reviewed. Finally, soil liquefaction potential of Babol city is studied using the mentioned methods and the obtained results are compared.

2 Seismology and geology of Babol The study area in this paper is Babol city, which is located in Mazandaran province in the north of Iran. This city is located in front of Alborz mountain which is tectonically an active region. The tectonic of Alborz mountain is controlled by boundary conditions due to convergent motion between Arabia and Eurasia, which probably started in the Cretaceous [36]. The area around Babol has repeatedly experienced earthquakes. The first historically reported major earthquake in this region was Amol earthquake that took place in 1809. This earthquake was felt in a very large area and damaged Babol city [1]. Chahar dange earthquake destroyed many villages in 1935. Band pey earthquake killed 1600 people and razed many structures to the ground with over 25 million dollars economical toll in 1957 [37]. Recently, Babol was affected by the occurrence of the moderate shaking at Kojoor and Marzi kola earthquakes. Table 1 shows the locations, years of occurrence, magnitudes and intensities of earthquakes occurred in and around Babol.

3/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Table 1. List of earthquakes in and around Babol city [1] Location

Year

Source

Intensity

Magnitude

Liquefaction occurrence

Amol

1809

20 km west of Babol

IX

6.5

Yes

Talar rood

1935

35 km southeast of Babol

VII

5.7

No

Chahar dange

1935

60 km southeast of Babol

VIII

6.3

Yes

Band pey

1957

10 km west of Babol

IX

6.8

Yes

Babol

1971

Babol

VI

5.2

No

Kojoor

2004

60 km northwest of Babol

VIII

6.3

Yes

Marzi kola

2012

Babol

VI

5

No

Babol region consists of soft deposits and is situated in a high seismic area. In addition, this city is located on the east bank of Babolrood river and receives abundant rainfall annually. Therefore, the assessment of liquefaction potential in this area is very important. Figure 1 depicts the ground water level distribution in Babol city based on the underground water level data collection in the geotechnical boreholes.

4/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Figure 1. The ground water level distribution in Babol city According to the geolithological variations, the subsurface soil column at Babol city can be classified into 5 groups: 1) Extremely loose to medium sand deposits from the surface to 15 m depth with a groundwater table less than 1 m. 5/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

2) Thin top layer of silt (3–5 m with SPT-N values from 15 to 20) underlain by thick layer of loose fine sand with some gravel (10–15 m with SPT-N values between 10-20) below the ground level. 3) Thick top layer of clay (8–15 m with SPT-N values from 10 to 15) underlain by thin layer of loose fine sand with some gravel (3–6 m with SPT-N values between 15-20) below the ground level. 4) Thick layer of clay (20–30 m with SPT-N values from 20 to 25). 5) Thick layer of sand (15–20 m with SPT-N values from 15 to 25). Figure 2 shows classification of the subsurface soil at Babol city in which G.W.T. is ground water table and SPT-N is Standard Penetration Test number (Figure 2 is typical and not to scale).

Figure 2. Classification of the subsurface soil at Babol city

3 Liquefaction analysis approaches used in the study Idriss and Boulanger [33] re-examined semi-empirical procedures for evaluating the liquefaction potential of soils during earthquakes and suggested relations for use in practice. Andrus and Stokoe

6/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

[34] presented a method for the evaluation of liquefaction potential through measurement of shear wave velocity. Their method was based on field performance data from 26 earthquakes and in situ shear wave velocity measurements at over 70 sites. Moss et al. [35] evaluated the probability of liquefaction using CPT and proposed a correlation for CPT-based assessments of seismically induced soil liquefaction hazard. 3.1 Idriss and Boulanger method Idriss and Boulanger [33] used the Seed and Idriss [38] simplified procedure to estimate the cyclic shear stress ratios (CSR) induced by earthquake ground motions using the following expression:

σ  CSR = 0.65  vo  rd a max  σvo 

(1)

in which amax is the peak horizontal ground surface acceleration (as a fraction of gravity), is the total vertical stress at depth z, is the effective vertical stress at the same depth and r d is the shear stress reduction coefficient. rd is adopted to consider the soil profile as a deformable body. The parameter rd is expressed as:

Ln  rd  = α  z  +β  z  M

(2)

 z  α  z  = -1.012 -1.126 Sin  + 5.133   11.73 

(2a)

 z  β  z  = 0.106 + 0.118 Sin  + 5.142   11.28 

(2b)

in which z is depth below the ground surface in meters and M is moment earthquake magnitude. These equations are considered suitable to a depth z ≤ 34 m. For z > 34, the following expression is applicable:

rd = 0.12exp  0.22M 

(2c)

Magnitude scaling factor (MSF) is used to account for shaking duration or equivalent number of stress cycles and is defined as:  -M  MSF = 6.9 exp   - 0.058  1.8  4 

(3)

Overburden correction factor for cyclic stress ratio (Kσ) is determined by the following relation:

 σ K σ = 1- Cσ ln  vo  Pa

   1.0 

(4)

in which σvo is the effective vertical stress at depth z and P a is the reference stress of 101 kPa or about atmospheric pressure. The coefficient Cσ is obtained by:

Cσ =

1 18.9 - 2.55

 N1 60

 0.3

(4a)

where (N1)60 is the modified Standard Penetration Test (SPT) number. (N1)60 is limited to maximum value of 37. (N1)60 is adjusted to an equivalent clean sand value ((N1)60cs) as:

 N1 60cs =  N1 60 + Δ  N1 60

(5)

7/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Δ  N1 60

2  9.7  15.7   = exp 1.63 + -    FC + 0.01  FC + 0.01   

(5a)

where FC is fines content. The cyclic resistance ratio (CRR) can be expressed as: 2 3 4   N     N1 60cs    N1 60cs    N1 60cs  1 60cs CRR = exp  +   -   +   - 2.8  14.1   126   23.6   25.4 

(6)

The factor of safety (FS) in this method is determined by:

FS =

CRR × MSF× K σ CSR

(7)

Liquefaction is predicted to occur when FS  1 (i.e., the loading exceeds the resistance). 3.2 Andrus and Stokoe method Andrus and Stokoe [34] used Eq.(1) and Eq.(2) for determination of cyclic stress ratio (CSR) and parameter rd, respectively. They suggested the following relation for determination of cyclic resistance ratio (CRR):

  Vs1 2  1 1   CRR = a  + b * - *   MSF    100   Vs1 - Vs1 Vs1 

(8)

where Vs1 is overburden stress-corrected shear wave velocity and V*s1 is the limiting upper value of Vs1 for cyclic liquefaction occurrence. a and b are curve fitting parameters taken to be 0.022 and 2.8, respectively and MSF is magnitude scaling factor which is calculated by Eq. (3). Vs1 is obtained by:

P  Vs1 = Vs CV = Vs  a   σv 

0.25

(8a)

where Vs is shear wave velocity, CV is overburden stress correction factor, Pa is atmospheric pressure and σ'v is initial effective overburden stress (kPa). The maximum value of C V is 1.4 which is generally applied to shear wave velocity data at shallow depths. They expressed the relationship between V*s1 and fines content (FC) as: *

V s1=215 m/s, *

V s1=215-0.5(FC-5) m/s, * V s1=200

m/s,

for sands with FC ≤ 5%

(8b)

for sands with 5% < FC < 35%

(8c)

for sands with FC ≥ 35%

(8d)

It should be mentioned that if Vs1> V*s1, no liquefaction is predicted to occur in this method. The factor of safety (FS) against liquefaction can be defined by: FS =

CRR CSR

(9)

when FS  1, liquefaction happens. 3.3 Moss et al. method In Moss et al. [35] method, cyclic stress ratio (CSR) is obtained by Eq. (1). The nonlinear shear mass participation factor (rd) for d < 20 m (d = depth in meters) is defined as:

8/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

9.147  4.173.a max  0.652.M w   0.089  3.28.d  7.760a max  78.576   1  10.567  0.089.e  rd  d, M w , a max    9.147  4.173.a max  0.652.M w   0.089  7.760a max  78.576   1   10.567  0.089.e 

(10a)

and for d ≥ 20m, rd is defined as:

9.147  4.173.a max  0.652.M w   0.089 3.28.d  7.760a max  78.567   1  10.567  0.089.e    0.0014. 3.28.d  6.5 rd  d, M w ,a max       9.147  4.173.a  0.652.M  max w  1  0.089 7.760a max  78.567     10.567  0.089.e 

(10b)

in which, Mw is moment magnitude and amax is the peak ground acceleration in units of gravity. Magnitude correlated duration weighting factor (DWFM) is calculated using the following equation:

DWFM  17.84.Mw1.43

(11)

In this method, qc1 is the normalized tip resistance (MPa):

qc1  Cq qc

(12)

c

P  Cq   a   1.7  v 

(12a)

in which Cq is tip normalization factor, qc is raw tip resistance (MPa) obtained by Cone Penetration Test (CPT), Pa is atmospheric pressure (101 kPa), σ'v is effective overburden stress (kPa) and c is tip normalization exponent:

R  c  f1  f   f3 

f1  x1.qcx2 ,

f2

x1  0.78,

f 2    y1.qcy2  y3  ,

y1  0.32,

f3  abs log 10  qc  , z1

(12b)

x 2  0.33

y2  0.35,

(12c)

y3  0.49

z1  1.21

(12d) (12e)

where Rf is friction ratio in CPT (the ratio of sleeve to tip resistance, in percent). The cyclic resistance ratio (CRR) can be expressed as:

 q1.045  0.11qc1R f  0.001R f  c 1  0.85R f   0.848ln  M w   0.002ln  v   1 (PL )  CRR  exp  c1  7.177  

(13)

where 1 is the inverse cumulative normal distribution function and PL is the probability of liquefaction. The factor of safety (FS) against liquefaction can be defined by:

FS =

CRR × DWFM CSR

(14)

Similar to Idriss and Boulanger [33] and Andrus and Stokoe [34] methods, FS  1 shows soil liquefaction.

4 Evaluation of the liquefaction potential in the study area The reliability of any liquefaction estimation depends on the quality of the site characterization. In order to evaluate the liquefaction potential of Babol soil using three mentioned methods, a total 9/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

number of 60 borehole logs were collected for this study. Figure 3 shows the location of available geotechnical boreholes in Babol region. In order to measure the shear wave velocity, downhole tests were performed in boreholes. Moreover, CPT tests were conducted at the nearest possible locations to boreholes.

Figure 3. Location of geotechnical boreholes Based on site investigations, the most liquefiable layers were found in some boreholes such as B1 (Figure 4) in which qc and Vs show cone tip resistance and shear wave velocity, respectively:

10/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Figure 4. Exploratory boring log (borehole B1)

Since presenting complete results for all boreholes is not possible, one borehole (B23) is selected and the results obtained for this borehole are described completely. The obtained results for other boreholes are presented by liquefaction microzonation maps. Figure 5 and Table 2 show stratigraphy and properties of soil recognized by borehole B23, respectively. In Table 2, w, γ, LL, PL, PI, FC, D 50 , Dr, N, Rf, qc and Vs depict water content, unit weight of soil, liquid limit, plastic limit, plasticity index, fine content, mass-median diameter, relative density, SPT number, friction ratio, cone tip resistance and shear wave velocity, respectively. As observed in Figure 5, the depth of borehole B23 is 20 m and the groundwater table is 4 m below the ground surface. In addition, PGA (peak ground acceleration) values were selected in each borehole position according to Standard 2800 [39]. At the location of borehole B23, PGA has been 0.35 g (g is the acceleration of gravity). In this study, the earthquake magnitude (Mw) is assumed 8. Therefore, the calculated MSF and DWF M values required for liquefaction analysis are 0.87581 and 0.9119, respectively.

11/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Figure 5. Exploratory boring log (borehole B23)

Table 2. Soil properties (borehole B23) Depth (m)

w

γ

LL

PL

PI

(kN/m )

(%)

(%)

(%)

FC (%)

D50 (mm)

Dr (%)

N

(%)

2

12

17.1

33

27

6

90.5

0.015

25

4

14

17.5

32

25

7

92.8

0.012

6

15

17.3

32

25

7

91.5

8

17

17.2

23

20

3

10

-

17.3

-

-

12

20

17.3

27

14

22

16.5

16

-

18 20

(%)

qc (MPa)

(m/s)

17

1.60

5.1

210

25

24

3.50

7.9

230

0.0065

95

19

3.40

7.1

207

22.6

1.6

92

25

2.80

9.8

198

-

14.3

2.0

99

22

2.30

8.9

200

21

6

10.1

2.2

95

27

2.60

10.2

190

30

23

7

55.5

0.05

55

15

2.80

7.8

165

16.8

-

-

-

60.4

0.11

52

21

2.90

9.2

180

-

17.6

-

-

-

51.8

0.42

99

29

3.00

11.4

195

23

18.1

-

-

-

54.6

0.42

92

34

2.90

14.3

200

3

Rf

Vs

12/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Now, liquefaction potential for borehole B23 is evaluated using Idriss and Boulanger [33] method (Table 3). Tables 4 and 5 show liquefaction potential for borehole B23 using Andrus and Stokoe [34] and Moss et al. [35] methods, respectively. In the liquefaction analysis, factors of safety against liquefaction at different depths of boreholes are calculated.

Table 3. Liquefaction analysis using Idriss and Boulanger [33] method for borehole B23 Depth (m)

rd

CSR





(N1)60cs

CRR

FS

2

0.99

0.22

0.13

1.00

26.0

0.31

1.22

4

0.98

0.22

0.15

1.00

29.3

0.45

1.75

6

0.96

0.27

0.13

1.00

25.1

0.29

0.94

8

0.95

0.30

0.15

1.00

28.8

0.42

1.22

10

0.93

0.32

0.13

0.98

23.9

0.27

0.72

12

0.91

0.33

0.15

0.96

25.7

0.31

0.78

14

0.88

0.34

0.10

0.97

18.2

0.19

0.46

16

0.86

0.34

0.12

0.95

22.7

0.24

0.59

18

0.84

0.34

0.15

0.92

28.9

0.42

0.99

20

0.81

0.34

0.17

0.89

32.3

0.68

1.55

Table 4. Liquefaction analysis using Andrus and Stokoe [34] method for borehole B23 Depth (m)

rd

CSR

2

0.99

4

Vs1

*

(m/s)

V s1 (m/s)

CRR

FS

0.22

275.29

200

-

No Liquefaction

0.98

0.22

252.80

200

-

No Liquefaction

6

0.96

0.27

216.63

171.75

-

No Liquefaction

8

0.95

0.30

198.99

206.2

0.40

1.34

10

0.93

0.32

194.03

210.35

0.21

0.66

12

0.91

0.33

178.72

200

0.16

0.49

14

0.88

0.34

151.42

189.75

0.09

0.28

16

0.86

0.34

161.35

200

0.10

0.29

18

0.84

0.34

170.69

200

0.12

0.37

20

0.81

0.34

171.09

215

0.10

0.30

13/25

Časopis GRAĐEVINAR

Journal CIVIL ENGINEER

Table 5. Liquefaction analysis using Moss et al. [35] method for borehole B23 Depth (m)

rd

CSR

c

CRR

FS

2

0.95

0.21

0.42

0.30

1.27

4

0.89

0.20

0.28

0.43

1.90

6

0.82

0.23

0.30

0.31

1.23

8

0.74

0.23

0.29

0.45

1.74

10

0.67

0.23

0.32

0.32

1.24

12

0.61

0.22

0.29

0.39

1.58

14

0.58

0.22

0.31

0.24

0.98

16

0.55

0.22

0.29

0.30

1.25

18

0.54

0.22

0.27

0.44

1.80

20

0.52

0.21

0.25

0.68

2.85

Figure 6 shows the comparison between CRR values determined for borehole B23 using these methods. As seen, CRR values obtained using Moss et al. [35] and Idriss and Boulanger [33] methods are in good agreement with each other. However, the obtained CRR values using Andrus and Stokoe [34] method are different. It should be noted that case history data and suggested CRR-VS1 curves by Andrus and Stokoe [34] are limited to relatively level ground sites with average depths