Prediction of unsaturated shear strength of an adobe soil from the soil–water characteristic curve

Construction and Building Materials 98 (2015) 892–899

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Prediction of unsaturated shear strength of an adobe soil from the soil–water characteristic curve Umaima Al Aqtash, Paola Bandini ⇑ Civil Engineering Department, New Mexico State University, PO Box 30001 MSC 3CE, Las Cruces, NM 88003-8001, USA

h i g h l i g h t s  Adobe masonry exists in unsaturated conditions and its strength is affected by water content.  Unsaturated shear strength of adobe was predicted from SWCC, effective strength and available models.  This prediction does not require shear strength testing with suction control.  The predicted failure envelopes in the shear strength–matric suction plane were nonlinear.  After the air-entry value, the slope of the failure envelopes decreased with soil suction.

a r t i c l e

i n f o

Article history: Received 1 November 2014 Received in revised form 16 July 2015 Accepted 29 July 2015 Available online 16 September 2015 Keywords: Adobe Unsaturated soil Shear strength Direct shear Suction Friction angle Earthen construction

a b s t r a c t This paper studied the unsaturated shear strength properties of a soil with the gradation characteristics of the material typically used in adobe construction. The main goal was to predict the unsaturated shear strength of the adobe soil using the soil–water characteristic curve (SWCC) and the effective cohesion and friction angle of the material. Specimens were trimmed from adobe bricks made according to the traditional technique of the southwestern region of the United States. The SWCC of the adobe soil was constructed using data obtained with the filter paper test. The effective cohesion and friction angle were found to be 11.7 kPa and 31.4°, respectively, from the results of consolidated drained direct shear tests. In addition, unconsolidated undrained direct shear tests were done on specimens at different water contents to determine the apparent shear strength parameters. The predicted shear strength was found to increase significantly with decreasing water content, and the strength increase was approximately 350 kPa between near saturation condition and water content of 6.01%. The predicted failure envelopes in the shear strength–matric suction plane were nonlinear; their initial slope was close to the (saturated) effective friction angle, and the slope decreased as the suction increased. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction In the southwestern parts of the United States of America (USA), many historic landmarks as well as contemporary buildings and residential construction are built with sun-dried mud bricks, regionally called adobe. The traditional adobe brick production is environmentally friendly and sustainable as it uses locally available soils and requires very little energy and water. The soils used to make adobe bricks typically contain fractions of clay, silt and sand. Cut straw is often added to the adobe mixture to help obtain even drying and in turn reduce shrinkage cracking of the bricks. Adobe walls are constructed by laying the brick units and mud mortar in an alternating fashion as in conventional masonry construction. Mud mortar is made of the same soil as the adobe bricks. ⇑ Corresponding author. E-mail address: [email protected] (P. Bandini). http://dx.doi.org/10.1016/j.conbuildmat.2015.07.188 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.

A serious concern related to adobe construction is its susceptibility to moisture penetration and weakening caused by seasonal wetting and drying cycles, capillary rise and rain [1]. Significant moisture damage in adobe walls has been reported in historic landmark buildings in the USA, such as the Franciscan church at the Mission San Jose de Tumacacori National Monument, New Mexico, built in 1828 [2], the Amador Hotel in Las Cruces, New Mexico, built in 1866 and expanded in 1885 (Fig. 1), and the Andres Pico Adobe residence, built circa 1816 to 1860 with several later additions and restorations [3]. The strength of sun-dried adobe bricks, and consequently adobe masonry, can be reduced considerably with increasing water content of the soil material within the walls [4]. The detrimental effects of moisture in adobe include material weakening due to substantial loss of soil suction and adobe weathering. Cracks in the plaster and capillary rise from the foundation are means for

893

U. Al Aqtash, P. Bandini / Construction and Building Materials 98 (2015) 892–899

sf ¼ c0 þ ðr  ua Þf tan /0 þ ðua  uw Þf tan /b

ð1Þ

which could be written as

sf ¼ c0 þ ðr  ua Þf tan /0 þ c00

ð2Þ 0

(a)

(b) Fig. 1. The Amador Hotel in Las Cruces, New Mexico: (a) street view of the building in 2013, and (b) lower part of an adobe wall showing extensive brick degradation and material loss due to moisture damage (the plaster was removed during restoration work).

the water to enter the adobe walls. Greater moisture damage is often reported in the lower portions of walls [2,5] as penetration of capillary water at the base is a major source of moisture. The problem may be aggravated in adobe walls treated with impervious surface coating because the water is forced higher into the wall before it can reach the surface and evaporate [5]. Moisture affects negatively the static and seismic performance of adobe walls through different mechanisms. For example, a through-wall shear plane may develop in a wall with weakened wet base during ground shaking and result in sliding or collapse of the part of the wall above this plane [6,7]. Wetting of the wall base at water content near or over the plastic limit may induce bulging or sagging of the lower part of the wall under self-weight and service loads. Despite the importance of the changes in material properties resulting from wetting of adobe walls during service, these changes are not considered explicitly in the practice and the material strength properties are usually determined by testing air-dry adobe bricks. Adobe is mostly in unsaturated conditions in its natural environment. The authors aimed at emphasizing the importance of considering unsaturated soil principles when studying the material and structural performance of adobe masonry. However, laboratory testing to determine unsaturated soil strength with controlled suction is expensive and requires specialized equipment and training. In the practice, direct suction measurements may not be required as they could be evaluated more easily from the soil–water characteristic curve (SWCC) of the soil [8,9]. In this context, the main goal of this paper is to predict the shear strength of unsaturated adobe soil using the SWCC and the effective shear strength parameters of the adobe soil.

where sf is the shear stress on the failure plane at failure, c and /0 are the effective shear strength parameters (effective cohesion and effective friction angle, respectively), (r  ua)f is the net normal stress on the failure plane at failure, (ua  uw)f is the matric suction at failure, r is the total normal stress, ua is the pore-air pressure, uw is the pore-water pressure, /b is a variable that describes the rate of change in shear strength relative to changes in matric suction [11], and c00 is the capillary cohesion. The net normal stress and the matric suction are two independent stress-state variables. In this model, the shear strength failure envelope is a planar surface in the space of net normal stress, matric suction and shear stress [10]. Experimental results reported in the literature indicate that the failure envelope in the shear strength–suction plane is nonlinear for matric suction values greater than the air-entry value and within the desaturation zone [11–14]. In this suction range, the rate of increase in shear strength with suction decreases [11,12]. After reaching the residual state, the shear strength of unsaturated soils may increase, remain constant or decrease depending on the amount of water remaining in the pores to transmit suction among the soil particles [12]. Bai and Liu [14] showed that the shear strength–matric suction curves for compacted Nanyang soil were nonlinear, with /b decreasing from 11.1° (=/0 ) to about 2° at matric suction of approximately 2000 kPa. In compacted specimens of Ankara clay [13], the shear strength–total suction curves were nonlinear too. Under normal pressure of 150 kPa, the shear strength increased from 50 kPa for specimens with water content (w) greater than optimum (w = 22.8%, 24.8%, 26.8%) to about 250 kPa for w drier than optimum (w = 14.8%, 16.8%, 18.8%). Modified triaxial or direct shear testing equipment in which the matric suction is controlled by axis-translation or osmotic technique is used to determine the shear strength of unsaturated soils [15,16]. Alternatively, results from conventional direct shear test on saturated specimens and the SWCC can be used as input in predicting models of the shear strength of unsaturated soils, e.g. [11,12,17]. The SWCC describes the relation between suction and water content of the soil. This curve is dependent on the soil type. To establish the SWCC, soil suction must be determined directly or indirectly using one or several techniques, such as pressure plate extractors, filter paper test, tensiometers, and psychrometers. The filter paper test is a reliable, inexpensive and relatively simple method to measure soil suction [18] and thus was used in this study. The semi-empirical model proposed by Vanapalli et al. [12] predicts the unsaturated soil shear strength as a function of matric suction using the effective (saturated) soil shear strength parameters and the SWCC. Vanapalli et al. [12] proposed modifications to Eq. (1) as follows:

sf ¼ c0 þ ðr  ua Þf tan /0 þ ðua  uw Þf



 h  hr tan /0 hs  hr

ð3Þ

where (ua  uw)f is the matric suction (wm) from the SWCC at the water content of the specimen at failure, h is the volumetric water content at the corresponding suction, hs is the volumetric water content of the saturated soil, and hr is the soil residual volumetric   r is the normalized water content. water content. The term hhh s hr

2. Shear strength of unsaturated soil 3. Experimental program The shear strength of unsaturated soil can be described using the extended Mohr–Coulomb failure criterion proposed by Fredlund et al. [10], given by

The experimental program consisted of laboratory measurements of total and matric suction and shear strength parameters

894

U. Al Aqtash, P. Bandini / Construction and Building Materials 98 (2015) 892–899

of the adobe soil. The suction values were used to establish the SWCC for the adobe material. The index properties of the adobe soil were also determined, including particle gradation, liquid and plastic limits, and dry unit weight. 3.1. Material description and adobe brick making The adobe soil was composed of natural soil, poorly graded sand, and cut straw. The ratio of natural soil to poorly graded sand was 25–9 by weight, which was found to be an appropriate proportion after several mixing trials. This ratio produced a soil mixture with the gradation and consistency of the material traditionally used for making adobe bricks in New Mexico. The approximate ratio of straw to poorly graded sand was 0.12–9 by weight. The material was prepared by hand in a single batch following the traditional procedure of adobe brick making of New Mexico. The natural soil was mixed with 20% water by weight and left soaking for 24 h in covered buckets to preserve the water content. The next day, the material was transferred to a wheelbarrow and mixed with the poorly graded sand thoroughly using a shovel (no water was added). Cut straw was incorporated towards the end of mixing (Fig. 2a). The adobe bricks were formed in wooden molds using bare hands; the mud was compacted with the fists in a kneading motion. Before making the bricks, the molds were soaked in water. Excess material was removed with a wet hand to produce a flat top surface on each brick (Fig. 2b). Finally, the wooden molds were slowly lifted and separated from the adobe bricks. The mold dimensions were 25.4  40.6  10.2 cm (10  16  4 in.). The bricks were kept in the laboratory to prevent fast drying until the specimens were prepared. 3.2. Specimen preparation Within the first two days after removing the molds, the specimens were cut from the adobe bricks by gently but firmly pushing a thin-wall metal cylinder into the wet bricks. A sharp knife was used to cut straw fibers along the specimen sides as the metal cylinder was pushed. The specimens’ dimensions and initial weight

were recorded. Some specimens were not allowed to dry for testing corresponding to saturated conditions. The remaining adobe specimens were allowed to air-dry very slowly in the laboratory until reaching the moist weight corresponding to a target degree of saturation (S). Then, the specimens were stored in small plastic bags and sealed until their testing to allow moisture equilibration. Specimen dimensions and target S values are given in Table 1. 3.3. Suction measurements The total and matric suctions were determined on 24 adobe specimens using the filter paper method according to ASTM D5298 standard procedure [19]. Triplicate specimens were tested per each target S, with gravimetric water content (w) ranging from 2.5% to 15.5%. Each specimen was cut in half along a plane perpendicular to the longitudinal axis. The filter paper technique provides an indirect measurement of suction. The method is relatively simple and inexpensive and provides reliable results when performed carefully. The basic assumption in this method is that the suction values of the soil specimen and the filter paper will be the same if enough time is allowed for them to reach moisture equilibrium in a constant temperature environment. The filter paper and the soil will come to moisture equilibrium either by vapor flow (noncontact method) or by liquid flow (contact method). The water content of the filter paper in direct contact with the specimen was used to determine matric suction, and the water content of the filter paper suspended above the specimen (noncontact) was used to determine total suction (Fig. 3). A fine metal mesh was used to support the noncontact filter paper. Whatman No. 42 filter paper was used because it has shown to produce more consistent results compared to other common types of filter paper [18,20]. The Whatman No. 42 calibration curve (for wetting) provided in the ASTM D5298 standard procedure [19] was adopted. The calibration curve is given by two equations that reflect the different sensitivities of the filter paper at the low and high suction values:

log w ¼ 5:327  0:0779 wfp ; for wfp 6 45:3%

ð4aÞ

Table 1 Specimen dimensions and target degrees of saturation. Test

Height (mm)

Diameter (mm)

Number of specimens

Target S (%)

Filter paper test

62

50

24

Unconsolidated undrained direct shear test Consolidated drained direct shear test

25

62

24

10, 50, 80 10, 60,

25

62

3

20, 30, 40, 60, 70 and 30, 40, 50, 80, and 90

100

(a)

(a)

(b)

(c)

(d)

(b) Fig. 2. Adobe mixing and brick making: (a) addition of cut straw to the mud and (b) molds and bricks.

Fig. 3. Specimen setup for the filter paper method for indirect suction measurements: (a) placing of contact filter paper, (b) introducing the specimen into the glass jar, (c) placing of noncontact filter paper and (d) sealed jar.

895

U. Al Aqtash, P. Bandini / Construction and Building Materials 98 (2015) 892–899

log w ¼ 2:412  0:0135 wfp ; for wfp > 45:3%

100

where w is the soil suction in kPa, and wfp is the gravimetric water content (%) of the filter paper. Several models that relate soil suction and water content are found in the literature (e.g., Brooks–Corey model [21], van Genuchten model [22], Fredlund and Xing model [23]). In this research, the SWCC of the adobe soil was prepared by fitting the empirical model proposed by Fredlund and Xing [23] to the experimental data from the filter paper test. The Fredlund and Xing (FX) model relates the soil volumetric water content (h) and suction (w) with this expression:

90

Natural soil

80

Poorly graded sand

70

Adobe soil mixture

"

1   h ¼ CðwÞhs ln expð1Þ þ ðw=aÞn

Percent passing (%)

ð4bÞ

60 50 40 30 20 10 0

#m

10

1

0.1

0.01

0.001

0.0001

Particle size (mm)

ð5Þ

Fig. 4. Grain size distribution of the natural soil, poorly graded sand, and adobe soil.

where exp is the base of the natural logarithm (also called Euler’s number), a, m and n are curve-fitting parameters (m and n are dimensionless; a has the same units as w), and C(w) is a correction function to account for an upper limit in soil suction (approximately 106 kPa after the residual water content) that has been observed experimentally:

2 CðwÞ ¼ 41 

3 ln½1 þ ðw=wr Þ 5 h  i ln 1 þ 106 =wr

ð6Þ

where wr is the soil suction corresponding to hr. The fitting parameters a, m, and n define the shape and slope of the SWCC and their values were determined by minimizing the residual sum of squares. 3.4. Direct shear tests The unconsolidated undrained direct shear test without suction control was used to determine the shear strength parameters of 24 adobe specimens with varying water content. Triplicate specimens were tested for each target S. The specimens were sheared at a constant rate of 0.02 mm/s. The applied normal stresses were between 66.8 kPa and 510.7 kPa. The consolidated drained direct shear test was conducted on three saturated soil specimens according to ASTM D 3080-98 standard procedure [24]. The specimens were sheared at a constant rate of 0.001 mm/s. The applied normal stresses ranged from 74.2 kPa to 198.2 kPa. For target S = 100%, the shear box and specimen were submerged in water, and the soil was allowed to consolidate for at least 12 h. The final water contents were determined using the whole specimens immediately after completing each test. The Mohr–Coulomb failure criterion was applied to calculate the apparent shear strength parameters of the unsaturated adobe soil and the effective shear strength parameters of the saturated adobe soil. Specimen failure was defined as the attainment of the maximum (peak) shear stress for target S values of 10–40%, or 20% horizontal strain for target S values of 50–90%. For a target S value of 100% (consolidated drained direct shear test), the shear strength was taken as the shear stress at 12% horizontal strain.

Table 2 Total and matric suction results from the filter paper test. Specimen ID

FP-1 FP-2 FP-3 FP-4 FP-5 FP-6 FP-7 FP-8 FP-9 FP-10 FP-11 FP-12 FP-13 FP-14 FP-15 FP-16 FP-17 FP-18 FP-19 FP-20 FP-21 FP-22 FP-23 FP-24

Soil water content

Degree of saturation

w (%)

h

Target S (%)

Actual S (%)

2.89 2.67 2.83 4.53 4.74 4.30 6.69 6.60 5.52 7.46 7.94 8.01 9.58 9.64 10.06 12.18 11.15 13.56 13.65 13.70 13.93 15.38 16.02 16.63

0.05 0.05 0.05 0.08 0.08 0.07 0.12 0.11 0.10 0.13 0.14 0.14 0.17 0.17 0.17 0.21 0.19 0.23 0.24 0.24 0.24 0.27 0.28 0.29

15 15 15 20 20 20 30 30 30 40 40 40 50 50 50 60 60 60 70 70 70 80 80 80

14.1 13.0 13.8 22.1 23.1 21.0 32.6 32.2 26.9 36.4 38.7 39.0 46.7 47.0 49.0 59.3 54.3 66.1 66.5 66.8 67.9 74.9 78.1 81.0

Total suction, wt (kPa)

Matric suction, wm (kPa)

54,126 46,898 57,892 16,034 11,565 13,406 2957 2830 6624 912 895 433 577 568 401 194 197 171 139 210 109 154 101 82

53,642 58,816 63,997 11,470 7574 5548 307 812 5443 253 469 523 319 63 66 102 138 62 90 42 52 38 34 13

sandy, silty clay (CL-ML) according to the Unified Soil Classification System (USCS) [25]. The adobe soil mixture was composed of 54% sand, 33% silt, and 13% clay. The adobe soil was classified as a sandy loam based on the USDA classification and as silty, clayey sand (SC-SM) according to the USCS [25]. The adobe specimens had on average a dry unit weight of 16.7 kN/m3 (106 lb/ft3), corresponding to a void ratio of 0.55. The fully saturated adobe soil had w = 20.5% and hs = 0.355. 4.2. Soil suction and SWCC

4. Results and discussion 4.1. Gradation and soil classification The grain size distributions of the natural soil, poorly graded sand, and adobe material are shown in Fig. 4. The natural soil was composed of 37% sand, 45% silt, and 18% clay. The liquid limit and plasticity index of the fraction passing sieve No. 40 of the natural soil were 27 and 6, respectively. The natural soil was classified as a loam based on the USDA soil classification system and as

The experimental results of total and matric suction (wt and wm respectively) determined with the filter paper (FP) test for the adobe soil are given in Table 2. The suction data were best-fitted with power functions (w in percentage), which are shown in Fig. 5 with the corresponding coefficients of determination (R2). As the water content decreased, the total and matric suction curves seem to converge. Houston et al. [18] explained that this is likely caused by the use of the logarithmic scale for the suction and does not mean that the osmotic potential, which is the difference

896

U. Al Aqtash, P. Bandini / Construction and Building Materials 98 (2015) 892–899

1000000

100000

Fredlund and Xing model

Matric suction 10000

ψt = 3x106w-3.77

1000

R² = 0.97 100

ψm = 5x106w-4.44

10000

Adobe soil

1000 100

Silt

10 1

Sand

0.20

0.40

0.60

0.80

1.00

Normalized water content, Θ

10 5

Clay

0 0.00

R² = 0.95 0

Data from Filter Paper Method

100000

Matric suction, ψm (kPa)

Total or matric suction (kPa)

Total suction

10

15

20

Gravimetric water content, w (%)

Fig. 7. The SWCC for the adobe soil compared with typical curves for sand, silt and clay soils provided by Fredlund and Xing [23].

Gravimetric water content, w (%)

Fig. 5. Total and matric suction data versus water content.

25 FP test data Fredlund and Xing model

20 15 10 5 0 0.1

1

10

100 1000 10000 Matric suction, ψm (kPa)

(/) and apparent cohesion (c) for the adobe soil with water content (w) and the best-fit equations are shown in Fig. 9a and b respectively, and the results are summarized in Table 3. The values of / and c increased with decreasing w. The / value of the adobe soil increased from 11.0° when w was 17.22% (nearly saturated condition) to 52.1° when w was 2.24% (air-dry condition). For the same w range, c increased from 36.8 kPa to 461.2 kPa. Fig. 10 shows the effective failure envelope for the saturated adobe soil obtained from the consolidated drained direct shear test. The effective strength parameters were /0 = 31.4° and c0 = 11.7 kPa, which are in the range of expected values for sands containing fines. These results were within the expected ranges of values for this adobe soil.

100000 1000000

Fig. 6. Matric suction data and Fredlund and Xing model.

between the total and matric suctions, is zero. For soil suction up to about 1000 kPa, the FP test can be used to measure total and matric suction reliably. For higher suction values, the FP test provides total suction only regardless of whether the filter paper is in contact or not with the soil because the moisture movement occurs through vapor transfer rather than capillary transfer [26]. For the adobe soil, this was observed for w in the range of 2–3%. The FX model [Eqs. (5) and (6)] was fitted to the matric suction data from the FP test to construct the SWCC of the adobe soil (Fig. 6). This corresponds to the drying curve. For hs = 0.355 and wr = 600 kPa, the curve-fitting parameters in Eq. (4) were a = 12.8, n = 1.473, and m = 0.508. The bubbling pressure for the adobe soil was approximately 5 kPa; at this suction value, the air starts entering the pores and the soil becomes unsaturated. The SWCC for the adobe soil was between typical curves for clay and silt soils [23] (Fig. 7), which is a reflection of the particle gradation of the adobe mixture. 4.3. Shear strength parameters of the adobe soil from testing The results of the unconsolidated undrained direct shear tests on the adobe soil are given in Table 3. The resulting failure envelopes and the mean (final) water contents are shown in Fig. 8. The mean water content varied from 2.24% for very dry specimens to 17.22% for nearly saturated specimens. The slope of the failure envelopes progressively increased as the water content decreased; this means that as the specimens dried, their apparent shear strength increased. The variation of the apparent friction angle

4.4. Predictions of shear strength of the unsaturated adobe soil Using the SWCC, /0 and c0 of the adobe soil, Eq. (3) was applied to predict the shear strength as a function of matric suction. The shear strength was calculated for various normal stresses. Changes in shear strength were assumed to occur immediately with suction changes. The resulting failure envelopes in the shear strength–matric suction plane are shown in Fig. 11 for matric suction values up to 2000 kPa. As expected, the shear strength increased significantly with soil suction. The predicted shear strength increased by about 350 kPa from the near saturation condition to wm = 2000 kPa (w = 6.01%). The predicted failure envelopes in the sf wm plane were nonlinear. Similar observation was reported in the literature, e.g. [11,13,14]. The slope angle (/b) of the failure envelope in the sf  wm plane was calculated as a function of matric suction (Fig. 12). The slope of the initial part of the failure envelopes was equal or very close to /0 (=31.4°) of the adobe soil at very small suction values (when the soil was saturated or nearly saturated) until the air-entry suction level (5 kPa). The slope of the shear strength–suction curve decreased as the soil suction increased, taking a value of /b = 3.8° at wm = 58.2 MPa (w = 2.24%) (Fig. 12). The contribution of matric suction to the shear strength of the unsaturated soil is represented by the capillary cohesion (c00 ). Fig. 13 shows the considerable increase in c00 with increasing wm in the adobe soil. The calculated capillary cohesion took values of 14 kPa at wm = 32 kPa (near saturation), 81 kPa at wm = 303 kPa, and 3897 kPa at wm = 58200 kPa (air-dry soil). 5. Summary and conclusions The unsaturated shear strength of an adobe soil was predicted using the model of Vanapalli et al. [12]. The SWCC and the effective

897

U. Al Aqtash, P. Bandini / Construction and Building Materials 98 (2015) 892–899 Table 3 Results of the unconsolidated undrained direct shear (DS) tests. Test no.

Specimen ID

1

DS-1 DS-2 DS-3 DS-4 DS-5

2.16 2.79 1.99 1.95 2.30

0.04 0.05 0.03 0.03 0.04

2

DS-6 DS-7 DS-8

4.30 3.85 4.43

3

DS-9 DS-10 DS-11

4

Target S (%)

Applied normal stress, r (kPa)

Measured shear strength, sf (kPa)

Apparent cohesion, c (kPa)

Apparent friction angle, / (°)

2.24

10 10 10 10 10

153.6 159.6 306.2 428.0 448.1

665.6 624.0 934.8 955.2 1049.3

461.2

52.1

0.07 0.07 0.08

4.19

30 30 30

151.1 292.9 451.9

484.2 695.7 859.5

308.4

51.2

5.51 6.08 6.44

0.10 0.11 0.11

6.01

40 40 40

147.9 300.5 442.7

408.9 593.6 659.8

300.0

40.5

DS-12 DS-13 DS-14

8.80 8.43 9.59

0.15 0.15 0.17

8.94

50 50 50

178.6 345.9 510.7

318.3 472.5 555.4

202.3

35.5

5

DS-15 DS-16 DS-17 DS-18 DS-19

11.80 10.38 11.10 10.49 11.24

0.20 0.18 0.19 0.18 0.19

11.00

60 60 60 60 60

164.9 185.0 335.1 346.5 506.5

224.1 213.1 288.7 300.5 449.3

92.4

33.4

6

DS-20 DS-21 DS-22

14.50 14.61 15.60

0.25 0.25 0.27

14.90

80 80 80

171.3 338.6 467.1

136.9 198.6 247.2

72.9

22.6

7

DS-23 DS-24

17.33 17.11

0.30 0.30

17.22

90 90

66.8 163.9

49.8 68.8

36.8

11.0

Soil water content w (%)

Mean for test w (%)

h

1100

60

1000 900

700

φ (degrees) = -2.63 w + 59.5

50

R² = 0.97

40 30 20 10 0

600

0

5

10

15

20

Gravimetric water content, w (%)

500

(a)

400

500

300

400

200 100 0 0

100

200

300

400

500

Normal stress, σ (kPa)

600

700

Fig. 8. Mohr–Coulomb failure envelopes for the unsaturated adobe soil from the unconsolidated undrained direct shear tests.

(saturated) shear strength parameters (/0 and c0 ) were input into the model. The SWCC was constructed by fitting the FX model [23] to the matric suction results (drying curve) obtained with the filter paper test. The adobe soil was found to have effective friction angle of 31.4° and effective cohesion of 11.7 kPa. In addition,

Apparent cohesion, c (kPa)

Apparent shear strength, τf (kPa)

800

Apparent friction angle, φ (degrees)

w = 2.2% w = 4.2% w = 6.0% w = 8.9% w = 11.0% w = 14.9% w = 17.2%

c = 1.57 w2 - 57.36 w + 564.7 R² = 0.97

300 200 100 0 0

5

10

15

20

Gravimetric water content, w (%)

(b) Fig. 9. Variation of apparent shear strength parameters with water content from the direct shear test.

apparent shear strength parameters (/ and c) for adobe soil specimens at different water contents were determined experimentally. The w values ranged from very low water contents of

898

U. Al Aqtash, P. Bandini / Construction and Building Materials 98 (2015) 892–899

5000

Capillary cohesion, c'' (kPa)

Shear strength, τf (kPa)

150 125

τ = 0.614 σ + 11.7

100

R² = 1.0

75 50 25

4000 3000 2000 1000 0

0 0

25

50

75

100

125

150

175

200

225

250

Normal stress, σ (kPa)

Predicted shear strength, τf (kPa)

700 Normal stress (kPa) σ=0 σ = 100 σ = 200 σ = 300 σ = 400

500 400 300 200

φ ' = 31.4o

100

c' = 11.7 kPa

0 0

500

1000

Matric suction, ψm (kPa)

1500

2000

Fig. 11. Failure envelopes in the sf  wm plane predicted with Eq. (3).

35

φ b = φ ' when φ = φs

30 25

φ b (degrees)

100

1000

10000

100000

Matric suction, ψm (kPa)

Fig. 10. Mohr–Coulomb failure envelope for the saturated adobe soil from the consolidated drained direct shear test.

600

10

Fig. 13. Capillary cohesion predicted with Eq. (2).

slope /b equal to /0 (= 31.4°) of the adobe soil and then decreasing progressively to 3.8° as wm increased to 58.2 MPa. The contribution of soil suction to the unsaturated shear strength was also presented through the variation in capillary cohesion (c00 ). The latter showed a significant rate of increase when wm was greater than 2000 kPa. The c00 value increased by 350 kPa from near saturation to matric suction of about 2000 kPa, and from that to the air-dry zone, c00 increased by about 3540 kPa. The approach to predict unsaturated shear strength of adobe soil applied in this paper could be used to obtain estimates useful when evaluating the strength of partially saturated material in existing adobe walls, and thus reducing the need for destructive sampling and laboratory testing. This work was part of a larger study aiming at understanding and quantifying the influence of moisture on the structural strength and performance of adobe walls. The study included a parametric analysis to determine the effects of moisture on the adobe wall’s ability to resist in-plane and out-of-plane lateral loading using finite element analysis (FEA). It was shown in this paper that moisture can reduce the shear strength of the adobe material, and that this drop in shear strength can be considerable. The results from this paper were used to identify input parameters for the material constitutive model of the FEA as functions of water content. Acknowledgements

20

The authors would like to thank Dr. Sonya Cooper, Regents Professor of New Mexico State University, for her expert advice on the material selection and the traditional technique of adobe brick making. Her technical assistance and collaboration in this matter were crucial to the study.

15 10 5

References 0 1

10

100

1000

10000

100000

Matric suction, ψm (kPa) Fig. 12. Angle /b calculated with Eq. (1).

air-dry adobe to the very high water contents near saturation condition. The predicted unsaturated shear strength of the adobe soil increased considerably with increasing soil suction (and decreasing water content). Even though the SWCC corresponded to a drying curve, it is anticipated that the drop in shear strength following a wetting curve would be also significant. The predicted failure envelopes in the sf wm plane were nonlinear, having an initial

[1] L. Rainer, Deterioration and pathology of earthen architecture, in: E. Avrami, H. Guillaud, M. Hardy (Eds.), Terra Literature Review, an Overview of Research in Earthen Architecture Conservation, The Getty Conservation Institute, Los Angeles, 2008, pp. 45–61. [2] A. Crosby, Monitoring moisture at Tumacacori, Assoc. Preserv. Technol. Int. Bull. 19 (4) (1987) 32–44. [3] E.L. Tolles, F.A. Webster, A. Crosby, E.E. Kimbro. Survey of damage to historic adobe buildings after the January 1994 Northridge Earthquake. GCI Scientific Program Reports, Los Angeles, The Getty Conservation Institute, 1996. [4] J.R. Clifton, F.L. Davis, Mechanical properties of adobe. U.S. Nat. Bur. Stand. Tech. Note 996. U.S. Department of Commerce, Washington DC, U.S. Government Printing Office, May 1979. [5] A. Crosby, A preservation monitoring system at Tumacacori National Monument, Assoc. Preserv. Technol. Int. Bull. 10 (2) (1978) 47–76. [6] F.A. Webster, E.L. Tolles, Earthquake damage to historic and older adobe buildings during the 1994 Northridge, California Earthquake, in: Proc. 12th World Conf. on Earthquake Engrg, Auckland, New Zealand, vol. 14, Balkema, New Zealand, 2000. paper 628 (CD-ROM).

U. Al Aqtash, P. Bandini / Construction and Building Materials 98 (2015) 892–899 [7] F.A. Webster, Application of stability-based retrofit measures on some historic and older adobe buildings in California, in: M. Hardy, C. Cancino, G. Ostergren (Eds.), Proc. Getty Seismic Adobe Project 2006 Colloquium, The Getty Conservation Institute, Los Angeles, 2006, pp. 147–158. [8] S.L. Houston, Applied unsaturated soil mechanics, state of the art report, in: J.F.T. Juca, T.M.P. de Campos, F.A.M. Marinho (Eds.), Unsaturated Soils: Proc. 3rd Int. Conf. on Unsaturated Soils, UNSAT 2002, Recife, Brazil, vol. 3, Swets & Zeitlingere, Lisse, 2002, pp. 1127–1134. [9] S.L. Houston, Characterization of unsaturated soils: the importance of response to wetting, in: A.J. Puppala, P. Bandini, T.C. Sheahan (Eds.), Geo-Congress 2014 Keynote Lectures: Geo-Characterization and Modeling for Sustainability, Geotechnical Special Publication 235, American Society of Civil Engineers, Atlanta, Georgia, 2014, pp. 77–96, http://dx.doi.org/10.1061/9780784413289.004. [10] D.G. Fredlund, N.R. Morgenstern, R.A. Widger, The shear strength of unsaturated soils, Can. Geotech. J. 15 (3) (1978) 313–321, http://dx.doi.org/ 10.1139/t78-029. [11] D.G. Fredlund, A. Xing, M.D. Fredlund, S.L. Barbour, The relationship of the unsaturated soil shear to the soil-water characteristic curve, Can. Geotech. J. 33 (3) (1996) 440–448, http://dx.doi.org/10.1139/t96-065. [12] S.K. Vanapalli, D.G. Fredlund, D.E. Pufahl, A.W. Clifton, Model for the prediction of shear strength with respect to soil suction, Can. Geotech. J. 33 (3) (1996) 379–392, http://dx.doi.org/10.1139/t96-060. [13] E. Çokça, H. Tilgen, Shear strength-suction relationship of compacted Ankara clay, Appl. Clay Sci. 49 (4) (2010) 400–404, http://dx.doi.org/10.1016/ j.clay.2009.08.028. [14] F.Q. Bai, S.H. Liu, Measurements of the shear strength of an expansive soil by combining a filter paper method and direct shear tests, Geotech. Test. J. 35 (3) (2013) 451–459, http://dx.doi.org/10.1520/GTJ103342. [15] N. Lu, W.J. Likos, Unsaturated Soil Mechanics, Wiley, New Jersey, 2004.

899

[16] C.W.W. Ng, Y.J. Cui, R. Chen, P. Delage, The axistranslation and osmotic techniques in shear testing of unsaturated soils: a comparison, Soils Found. 47 (4) (2007) 674–684. [17] N. Khalili, F. Geiser, G.E. Blight, Effective stress in unsaturated soils: review with new evidence, Int. J. Geomech. 4 (2) (2004) 115–126, http://dx.doi.org/ 10.1061/(ASCE)1532-3641(2004) 4:2(115). [18] S.L. Houston, W.N. Houston, A. Wagner, Laboratory filter paper suction measurements, Geotech. Test. J. 17 (2) (1994) 185–194, http://dx.doi.org/ 10.1520/GTJ10090J. [19] ASTM D5298, Standard test method for the measurement of soil potential (suction) using filter paper, in: ASTM Annual Book of Standards, ASTM International, West Conshocken, 2010, http://dx.doi.org/10.1520/D5298-10. [20] E.J. Murray, V. Sivakumar, Unsaturated Soils: A Fundamental Interpretation of Soil Behavior, Wiley, Ames, 2010. [21] R.J. Brooks, A.T. Corey, Hydraulic properties of porous media. Hydrology Papers No. 3, Colorado State Univ., Fort Collins, March 1964. [22] M.T. van Genuchten, A closed-form equation for predicting the hydraulic conductivity of unsaturated soils, Soil Sci. Soc. Am. 44 (5) (1980). 898-898.. [23] D.G. Fredlund, A. Xing, Equations for the soil-water characteristic curve, Can. Geotech. J. 31 (4) (1994) 521–532, http://dx.doi.org/10.1139/t94-061. [24] ASTM D 3080-98, Standard test method for direct shear test of soils under consolidated drained conditions, in: ASTM Annual Book of Standards, vol. 04.08, ASTM International, West Conshohocken, 1999. [25] ASTM D 2487, Standard classification of soils for engineering purposes, in: ASTM Annual Book of Standards, vol. 04.08, ASTM International, West Conshohocken, 2011, http://dx.doi.org/10.1520/D2487-11. [26] E.C. Leong, L. He, H. Rahardjo, Factors affecting the filter paper method for total and matric suction measurements, Geotech. Test. J. 25 (3) (2002) 322–333, http://dx.doi.org/10.1520/GTJ11094J.