asadzadeh - direct shear testing on a rockfill material

Mozhgan Asadzadeh and Abbas Soroush

October 2009 The Arabian Journal for Science and Engineering, Volume 34, Number 2B 379

DIRECT SHEAR TESTING ON A ROCKFILL MATERIAL Mozhgan Asadzadeh

Department of Civil and Environmental Engineering Amirkabir University of Technology

Tehran, Iran [email protected]

and

Abbas Soroush*

Department of Civil and Environmental Engineering Amirkabir University of Technology

Tehran, Iran [email protected]

:ةـالخالصادات اإل مستويات مختلفة من ىعل عدد من االختبارات جريتأ ردميات المواد الصخريةىعل والترطيب جهاداإل ىمستو ثيرأت التحقق من جلأمن جه

.الجافة المشبعة باستخدام اختبار القص المباشر متوسط الحجم و المشبعة و في الحاالت الجافة

ى ل فوعل ة المماث ن التوج رغم م وةيمال امالت الق ق بمع ساء ا يتعل ستوي ةوالج ة بم اداإل ذات العالق ي جه االت ف ثالث الح ذآورة أال الة الم ن و. ع م . للمادة ة الترطيب يقلل من معامالت القوة والجساءنأالواضح

ىإ إضافة ك ل إن ذل ةف ارات الجاف ائج االختب ين -نت شبعة تب شبع نأ الم اتالعّي ت ة ن ل الحجز ( الجاف شيدة ولاألمث سدود الم ي ال ى ف واد عل ات الم ردمي .وهذا ما يسمي بظاهرة االنهيار المشبع في المقاالت العلمية . واد المعلى ًامفاجئ ًاهبوط القص ويفرض جهاداتإيقلل ) الصخرية الجافة

ؤات نفس ال على) triaxial( مقارنه نتائج االختبارات بنتائج اختبار ثالثي المحاور نإف ذلك علىوعالوة ادة تظهر تنب ر أم ة آث ارات طةاس بو دق اختب .) triaxial(القص المباشر منها باختبار ثالثي المحاور

*Corresponding author: Dept. of Civil and Env. Eng., Amirkabir University of Technology No. 424, Hafez Ave., Tehran, Iran P.O.Box: 15875-4413, Tel. +98 (21) 64543009; Fax: +98 (21) 66414213 E-mail: [email protected] Classification: Civil Engineering, Geotechnical Engineering, Material Behavior

Paper Received January 29, 2009; Paper Revised April 18, 2009; Paper Accepted May 27, 2009


The Arabian Journal for Science and Engineering, Volume 34, Number 2B October 2009 380

ABSTRACT

In order to investigate the possible effects of two main factors (stress level and wetting) on a rockfill material, a number of medium-scale direct shear tests have been conducted with different stress levels in dry, saturated, and dry-saturated conditions. The results suggest that despite the similar trend of the strength and stiffness parameters related to the stress level in the three conditions, wetting decreases the strength and stiffness parameters of the material. In addition, the results of the dry-saturated tests showed that saturation of the dry specimens (modeling the first impounding of dry-constructed rockfill dams) decreases shear stresses and imposes sudden settlements on the material. This phenomenon is named saturation collapse in the literature. In addition, a comparison of the tests results with the results of the triaxial tests conducted on the same material indicates comparatively higher predictions by the direct shear tests.

Key words: direct shear test, rockfill material, shear strength, normal stress level, saturation collapse, particle breakage



DIRECT SHEAR TESTING ON A ROCKFILL MATERIAL

1. INTRODUCTION

The vast application of rockfill materials in geotechnical engineering, as in rockfill dams, makes the precise knowledge of the behavior of these materials indispensable. Testing large sizes of prototype rockfill materials is difficult and expensive. Therefore, to conduct a feasible test, the size of these materials must be scaled down. For this purpose, two major points should be considered: 1) determination of the maximum particle size, and 2) selection of a proper gradation modeling technique. In a laboratory specimen, the maximum particle size (d) is determined according to the minimum dimension of the specimen (D). For materials with broad gradations, D/d=4 [1] and for materials with narrow gradations, D/d=6 [2] are proposed. The specimen gradation can be selected by one of the four modeling techniques: scalping [3], parallel gradation [4], quadratic grain-size distribution [5], and replacement technique [6].

Laboratory tests have shown that the behavior of rockfill materials depends on a variety of factors such as stress level, void ratio, gradation, amount of fine particles , maximum grain size, moisture content, mineral composition, and particle shape. As the stress level increases, dilation and compressibility decrease [7,8]. whereas initial settlement and creep rate increase [9]. Failure envelopes of rockfill materials are usually non-linear, for which particle breakage is responsible [2]. The breakage, depending on the individual particles’ specifications, can occur even at relatively low stresses. Several factors responsible for the amount of particle breakage are proposed: stress level, stress path, particle size, relative density, particle angularity, mineral hardness, and water presence [10]. The particle breakage increases as the stress level increases [5].

Saturation degrades the strength parameters and deformation modulii of rockfill materials [11,12]. This degradation is due to (1) weakening of particles leading to their breakage [8], and lubrication effects of water which acts on the grain-to-grain contacts [13,14].

It is evident in the literature that large-scale laboratory tests such as triaxial, plane strain, direct shear, and odometer tests have been employed for studying the behavior of rockfill materials. In this paper, the main focus is on the saturation collapse phenomenon. For this purpose, a medium-scale direct shear apparatus is employed and several tests are conducted on dry, saturated, and dry-saturated specimens of a rockfill material. In the dry-saturated tests, the materials are compacted dry and subjected to normal stress and sheared; then, they are saturated in a specific shear stress level and thereafter shearing is continued. Having considered the effect of normal stress level, most of the tests were conducted with five different normal stress levels. The results of the dry and saturated tests were compared together to probe the effects of stress level and wetting. Furthermore, the results of the saturated tests were compared with the results of Consolidated Drained (CD) triaxial tests conducted on the material.

2. APPARATUS, MATERIAL, AND TESTING PROCEDURE

2.1. Apparatus

A medium-scale direct shear apparatus (30x30x15 cm) was used for testing. The apparatus had an upper and a lower shear box, and the sample was sheared strain-controlled by pushing the lower shear box horizontally. Two gauges were used to measure vertical displacements and shearing forces.

2.2. Rockfill Material

For the present research, rockfill materials were obtained from an under-construction rockfill dam. The material has a maximum particle size of 800 mm. Table 1 shows the characteristics of the material along with the standards employed for their determination. A number of point load tests were carried out on dry and saturated particles of the material (Table 2). According to Table 2, saturation does not have considerable effects on the material properties. The average uniaxial compression strength of the material is estimated to be about 84 MPa from correlations with the average results of the point load tests (Table 2). The geomechanical classification of rockfill materials by point load testing results [15] shows that this material is classified as pretty hard.

Table 1. Rockfill Material Characteristics Los Angeles

Abrasion Point Load

Index Gs

Water Absorption Water Content Shape Mineralogy

40% ASTM

(C535-96)

4 ASTM

(D5731-95)

2.75 ASTM

(C127-128)

3% ASTM

(C127-128)

0.1% ASTM

(D2216-92) Angular Limestone



Table 2. Results of Point Load Tests

σc

MPa

Is(50)

MPa F

Is

MPa

2eD

(mm2)

P

(kN)

W

(mm)

D

(mm)

W2

(mm)

W1

(mm) Test

72.9 3.7 0.84 4.4 1146 5 30 30 32 28 1 88.9 4.3 0.88 4.9 1433 7 37.5 30 43 32 2 78.1 3.7 0.90 4.1 1569 6.5 38.5 32 39 38 3 56.1 2.6 0.92 2.8 1758 5 34.5 40 46 23 4 91.0 4.3 0.92 4.7 1718 8 43.5 31 51 36 5 75.8 3.5 0.94 3.8 1862 7 39.5 37 41 38 6 89.8 4.4 0.88 5.0 1408 7 33.5 33 35 32 7 36.5 1.8 0.90 2.0 1534 3 43 28 54 32 8

Dry

71.3 3.3 0.96 3.4 2062 7 41.5 39 37 46 9 106.7 5.0 0.92 5.4 1754 9.5 40.5 34 36 45 10 105.4 5.1 0.89 5.7 1490 8.5 39 30 42 36 11 90.3 4.2 0.92 4.6 1738 8 44 31 41 47 12 58.2 2.8 0.91 3.0 1651 5 36 36 41 31 13 69.7 3.1 0.99 3.1 2410 7.5 43 44 40 46 14

Satu

rate

d

45.0

50 ⎟⎟⎠

⎞⎜⎜⎝

⎛= eD

F

)50()175.014( sec ID+=σ

2e

sDPI =

ss IFI *)50( =

221 WW

W+

=

πDWDe

42 =

W1 و W2: Specimen width perpendicular to the loading direction D: Distance between platen contact points P: Failure load Is : Point load index F: Size correction factor Is(50): The size corrected point load index for D= 50 mm σc: Uniaxial compression strength

The gradation curve of the material is shown in Figure 1, for which the values of the coefficient of uniformity (Cu) and the coefficient of curvature (Cc) are 17 and 1.7, respectively.

2.3. Modeled Rockfill Material

Testing the prototype rockfill material was almost impossible because of its coarseness and the limitations of the shear box dimensions. Therefore, the laboratory test specimens were scaled down (in terms of particle sizes) by some degrees. The prototype gradation was broad (Figure 1); so, based on the literature, the maximum particle size of about 1.5 inches (3.8 cm) was selected by using D/d=4 and the scalping method [3] was employed for modeling the gradation curve.

In order to exclude the existence of an excess amount of fine particles in the specimens that may impede drainage during shearing, it was necessary to eliminate the lower end of the gradation curve (particles under 0.075 mm which were about 3%). The gradation of the modeled rockfill material is presented in Figure 1; this gradation has Cu=19 and Cc=1.2.



0

10

20

30

40

50

60

70

80

90

100

0.01 0.1 1 10 100 1000Particle Size (mm)

% P

assi

ngmodeled materialsOriginal materials

Figure 1. Prototype and modeled rockfill gradation curves

Figure 2 shows the individual coarse particles of different sizes of the material (Figure 2a) along with a guide for the particle shape identification (Figure 2b). A comparison of these two figures suggests the angularity of the particles.

4# ″

41

″

83

″

21

″

43

1 ′′

a

b

Figure 2. a) Coarse particles used based on individual sieve size b) Identification and classification of soils and rocks



2.4. Experimental Procedure

The quantity of various sizes of the materials required to achieve the modeled gradation for preparing the specimen at the specified density, γd=2 gr/cm³, was determined by weight. The individual fractions were divided into three parts and each part was mixed thoroughly (in order to achieve a more homogenous sample) and compacted in the shear box to achieve the required density. After the specimen was prepared, it was first subjected to the selected normal stress and then sheared by applying horizontal load at a rate of 0.5 mm/min. This rate was in the range of the rate proposed by ASTM D3080, i.e., 0.0025 to 1.0 mm/min, and was selected based on the relatively high permeability of the material. Some extra tests on this material proved that the rate of shearing has no meaningful effect on the results.

For the saturated shear tests, the specimens were compacted dry in the shear box and then saturated for 48 hours. In the dry-saturated tests, the specimen was first sheared (in dry conditions) up to 90 percent of the minimum shear strength of the dry tests in the same stress level; then, the test was stopped, the specimen was saturated for three hours, and shearing was continued. Five normal stress levels (111, 222, 444, 666, 777 kPa) for the dry and saturated conditions, and three normal stress levels (444, 666, 777 kPa) for the dry-saturated condition were applied. In order to verify the reproducibility and repeatability of the results, the dry and saturated tests in each normal stress were repeated. Also in the dry-saturated condition, the test with 777 kPa normal stress was carried out twice.

3. RESULTS

3.1. Dry Tests

Figure 3 presents the shear stress-horizontal displacement and vertical displacement-horizontal displacement behavior of the dry specimens. In this figure, TD stands for the dry condition, the first number is the test number, and the last three digits represent the normal stress value in kPa. In TD1-111 and TD2-111, the final horizontal displacements were less than the others because the shear stress stabilized sooner and the tests were stopped. In this figure and hereafter, dilation is considered positive. It can be seen that all of the tests show post-peak softening behavior and in all of the normal stress levels, the materials dilate after initial contraction.

0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50a) Horizontal Displ. (mm)

Shea

r Stre

ngth

(kPa

)

TD1-111TD2-111TD1-222TD2-222TD1-444TD2-444TD1-666TD2-666TD1-777TD2-777



-2

-1

0

1

2

3

4

5

6

0 10 20 30 40 50

b) Horizontal Displ. (mm)

Verti

cal D

ispl

. (m

m)

TD1-111TD2-111TD1-222TD2-222TD1-444TD2-444TD1-666TD2-666TD1-777TD2-777

Figure 3. a) Shear stress-horizontal displacement b) vertical displacement-horizontal displacement behavior of dry specimens

3.2. Saturated Tests

Figure 4 shows the shear stress-horizontal displacement and vertical displacement-horizontal displacement behavior of the saturated specimens. TS stands for the saturated condition, the first number is the test number, and the last three digits represent the normal stress value in kPa. Again, TS1-111 and TS2-111 were stopped at less horizontal displacements like TD1-111 and TD2-111. Similar to the dry tests, the materials show post-peak softening behavior after reaching a peak and dilate after the initial contraction.

0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50a) Horizontal Displ. (mm)

Shea

r Stre

ngth

(kPa

)

TS1-111TS2-111TS1-222TS2-222TS1-444TS2-444TS1-666TS2-666TS1-777TS2-777



-2

-1

0

1

2

3

4

5

6

0 10 20 30 40 50

b) Horizontal Displ. (mm)

Verti

cal D

ispl

. (m

m)

TS1-111TS2-111TS1-222TS2-222TS1-444TS2-444TS1-666TS2-666TS1-777TS2-777

Figure 4. a) Shear stress-horizontal displacement b) vertical displacement-horizontal displacement behavior of saturated specimens

3.3. Dry-Saturated Tests

The dry-saturated shear tests in this research were carried out to simulate the saturation collapse phenomenon. Saturation collapse is a complex phenomenon involving settlement and shear strength reduction in constant stress levels due to wetting (saturating) of dry granular materials. Three reasons can be found responsible for the saturation collapse: 1) sudden reduction in confining pressure, 2) easier sliding of particles because of the lubrication effect of water, and 3) particle breakage due to interaction of water and surficial microcracks (if any) of the particles. Saturation collapse during first impounding of rockfill dams imposes relatively large displacements on the upstream shell of the dams which are constructed dry [16].

Figures 5, 6, and 7 show the shear stress-horizontal displacement and vertical displacement-horizontal displacement behaviors of the specimens tested under 444, 666, and 777 kPa normal stresses in the dry-saturated condition. The results are compared in these figures with the average results of their corresponding dry and saturated tests. Here, TDS refers to the dry-saturated condition and the last three digits again indicate the normal stress level in kPa.



Figure 5. Shear stress-horizontal displacement and vertical displacement-horizontal displacement behavior of the specimen

with 444 kPa normal stress in dry-saturated condition compared with the dry and saturated average results

Figure 6. Shear stress-horizontal displacement and vertical displacement-horizontal displacement behavior of the specimen with

666 kPa normal stress in dry-saturated condition compared with the dry and saturated average results



Figure 7. Shear stress-horizontal displacement and vertical displacement-horizontal displacement behavior of the specimens

with 777 kPa normal stress in dry-saturated condition compared with the dry and saturated average results

The above figures show that in the first part of dry-saturated shearing (which is in dry condition), the shear stress-horizontal displacement behavior is pretty similar to the shear stress-horizontal displacement behavior of the material tested in dry conditions. This similarity was expected and further supported the repeatability of the tests results. After saturation, shear stresses reduce about 30% in a constant horizontal displacement. This sudden reduction recovers about 50% within about 1 mm horizontal displacement. After that, shear stresses reach their peak and then show a post-peak softening behavior. Likewise, the vertical displacement-horizontal displacement behavior of the material in the first part of the dry-saturated tests is similar to its corresponding behavior in the dry tests. The saturation causes sudden settlements. After that, the vertical displacement-horizontal displacement behavior is more or less similar to their corresponding behavior in the saturated tests.

4. ANALYSIS OF THE RESULTS

Figure 8 shows the variations of peak shear strength (τp) and peak friction angle (φp) of the material versus normal stresses (σn) for the dry, saturated, and dry-saturated tests. The peak friction angle, assuming c=0, is calculated as

⎟⎟⎠

⎞⎜⎜⎝

⎛= −

n

pp σ

τϕ 1tan (1)

Turning to Figure 8, it is seen that the failure envelopes (τp:σn) are non-linear. The non-linearity is more evident in low stress levels. The shear strength of the saturated material is less than that of the dry material, and their difference increases as the normal stress increases. The peak friction angle (φp) decreases as the normal stress (σn) increases. φp ranges between 57° and 48° for the dry specimens and between 56° and 46° for the saturated specimens. The peak friction angle in the dry-saturated tests is less than that of the saturated tests; however, the difference is not considerable.



0

200

400

600

800

1000

1200

0 200 400 600 800 1000

σn (kPa)

τ p (k

Pa)

0

10

20

30

40

50

60

TDTSTDSTDTSTDS

φp

(deg

)

Figure 8. Peak shear strength and peak friction angle versus normal stress in dry, saturated and dry-saturated conditions

Figures 3 and 4 show that the residual shear strength (τres) versus normal stress (σn) in the dry and saturated tests has the same trend as peak shear strength versus normal stress (it increases as normal stress increases). Moreover, the residual shear envelope was again non-linear.

A comparison of φp and φres values versus normal stress for the dry and saturated tests is presented in Figure 9. The average difference between φp and φres is about 6 degrees for both of the dry and saturated specimens.

35

40

45

50

55

60

0 200 400 600 800 1000σn (kPa)

φ (d

eg)

φp-TD

φp-TS

φres-TD

φres-TS

φp-TD

φp-TS

φres-TD

φres-TS

Figure 9. Comparison of φp and φres versus normal stress in dry and saturated conditions



With reference to Figures 3 to 7, a maximum dilation angle (ψmax) during shearing can be defined according to the following equation:

max

1max tan ⎟⎟

⎠

⎞⎜⎜⎝

⎛= −

h

v

ddδδ

ψ (2)

where dδv, and dδh represent the vertical and horizontal displacements increments, respectively. Variations of ψmax versus σn (normal stress) are shown in Figure 10. It can be seen that ψmax decreases as σn increases; ψmax reduces from 17° to 7° in the dry tests and from 16° to 6° in the saturated tests. For the dry-saturated tests, ψmax is approximately 2 to 3 degrees less than those of the saturated tests.

0

5

10

15

20

0 200 400 600 800 1000

σn (kPa)

ψm

ax (d

eg)

TDTSTDSTDTSTDS

Figure 10. Maximum dilation angle versus normal stress in dry, saturated and dry-saturated conditions

Plots of the Marsal’s particle breakage index (Bg) [2] versus normal stress (σn) for the tests are shown in Figure 11. The breakage index used here involves the changes in individual particle sizes between the pre-test and post-test grain-size distributions. The difference in the percentage retained is computed for each sieve size. This difference will be either positive or negative. Marsal’s breakage index (Bg) is the sum of the differences having the same sign.



0

2

4

6

8

10

0 200 400 600 800 1000

σn (kPa)

Bg

(%)

TDTSTDSTDTSTDS

Figure 11. Particle breakage index versus normal stress in dry, saturated and dry-saturated conditions

Figure 11 shows that Bg increases almost linearly with increasing of σn. The figure also shows that the breakage in dry conditions is more than the breakage in saturated conditions. In the saturated tests, the easier sliding of particles due to the lubrication effect of water may have been responsible for the comparatively lower particle breakages. The particle breakage in the dry-saturated tests is generally less than those of the dry tests and more than those of the saturated tests.

In order to evaluate the stiffness of materials in the direct shear tests, the following equation may be applied.

ph

ppeakG

τδτ

)(* = (3)

For the dry and saturated tests, horizontal displacement values corresponding to the peak shear strength [(δh)τp] and their corresponding G*peak values are presented in Table 3. It is seen that (δh)τp increases almost linearly as σn increases. For both conditions, G*peak, however, increases with σn; in other words, the increase of normal stress generally makes the materials stiffer. In addition, the dry materials have higher G*peak values than the saturated materials.

Table 3. Horizontal Displacements Corresponding to the Peak Shear Strength [(δh)τp] and G*Peak for Different Normal Stresses

σn

(kPa)

τp

(kPa)

(δh)τp

(mm)

G*peak

(kPa/mm)

111 181 14 13 222 321 16 20 444 563 21 27 666 771 21 37

Dry

777 838 25 34 111 163 13 12 222 293 17 18 444 531 20 26 666 717 22 33 Sa

tura

ted

777 810 25 32 values are in average



As mentioned before, in the dry-saturated condition, saturation imposes a sudden settlement collapse (∆Dv) on the rockfill materials. The values of this settlement are cited in Table 4, which reveals that the settlement increases almost linearly with the increase of normal stress.

Table 4. Values of Saturation-Induced Settlement (∆dv) in Dry-saturated Tests

σn

(kPa)

∆Dv

(mm)

444 0.33 666 0.55 777 0.61

4.1. Failure Envelope

According to Figure 8, the material has non-linear failure envelopes. This non-linearity can be explained by the reduction of dilation rate, which is a function of particle breakage, related to the stress level increase [17,18]. Therefore, the Mohr-Coloumb linear criterion is inappropriate to define the behavior of this material. In this regard, different equations have been suggested in the literature (Charles and Watts, 1980; Sarac and Popovic, 1985; Indraratna, 1994) [3,19]. Charles and Watts (1980) [3] presented the following equation:

bnp Aστ = (4)

where A and b depend on the type of the material. This equation is simple enough and can predict the behavior of rockfill materials fairly well [17]. Based on the results (Figure 8), Equation 4 for the material of this research will be as follows:

821.072.3 np στ = dry condition (5)

824.04.3 np στ = saturated condition (6)

4.2. Internal Friction Angle

Among the reported equations in the literature for predicting the friction angle of rockfill materials, equations suggested by Gharavy (1996) [17] seem to be simpler than the others and involve fewer parameters. These equations are:

Dn

p Pa

−

⎟⎟⎠

⎞⎜⎜⎝

⎛=

σϕϕ 0 (7)

⎟⎟⎠

⎞⎜⎜⎝

⎛−=

PaBLog n

pσ

ϕϕ 0 (8)

where D and B are constants related to the material type and φ0 is the materials friction angle corresponding to σn equal to one atmospheric pressure. Table 5 shows the calculated values of φ0 and D (Equation 7) and φ0 and B (Equation 8) for the tested materials.

Table 5. Values of the Parameters Related to Friction Angle (Equations 7 and 8)

Equation 8 Equation 7

B φ0 D φ0 Test Condition

11.27 58.6 0.093 58.8 Dry

11.30 56.6 0.096 56.8 Saturated

4.3. Dilation Angle

The following correlation between the peak friction angle (φp) and the maximum dilation angle (ψmax) can be proposed for all of the dry, saturated, and dry-saturated tests.

ψmax≈φp-40° (9)



4.4. Residual Shear Strength

The results of all of the tests in dry and saturated conditions are considered together to correlate the residual and peak shear strength. Based on the correlation, a linear relationship is defined as follows:

8.24867.0 −= pres ττ (10)

5. COMPARISON OF DIRECT SHEAR AND TRIAXIAL TESTS RESULTS

The results of the dry and saturated tests carried out in this research are compared with the results of the triaxial tests on the same material (in saturated conditions) carried out by the Building and Housing Research Centre (BHRC). Moreover, the results are compared with the direct shear and triaxial tests results conducted on two other rockfill materials: 1) triaxial and direct shear tests on angular basalt materials with two different gradations G1 and G2 [17], and 2) triaxial and direct shear tests on an angular limestone material [20]. The characteristics of these tests are introduced in Table 6, where DS and TR stand for direct shear and triaxial tests, respectively. The gradations of the above tests are shown in Figure 12.

0

10

20

30

40

50

60

70

80

90

100

0.01 0.1 1 10 100Particle Size (mm)

% P

assi

ng

DS-TD, DS-TSTR-BDS-G1, TR-G1DS-G2, TR-G2DS-RTR-R

Figure 12. Gradation curves of different material used for comparison

Table 6. Characteristics of Rockfill Materials and Tests Compared in This Research

Reference Symbol Test

Condition

Apparatus

Dimensions

(cm)

Mineralogy

Maximum

Particle

Size (mm)

Relative

Density

(gr/cm³)

Rate of

Shearing

(mm/min)

This

Research

DS-TD

DS-TS

Dry

CD 30x30

Angular

limestone 38.1 2 0.5

BHRC TR-B CD Dia.=30 Angular

limestone 50.8 2.1 0.5

Rayhani

(2000)

DS-R

TR-R CD

30x30

Dia.=30

Angular

limestone

25.4

50.8 Dr=0.95 0.5

Gharavy

(1996)

DS-G1

TR-G1 CD

30x30

Dia.=20

Angular

basalt 37.5 G1=2

0.275

0.254

Gharavy

(1996)

DS-G2

TR-G2 CD

30x30

Dia.=20

Angular

basalt 37.5 G2=1.84

0.275

0.254



Figure 13 shows the shear strength envelopes resulting from the above tests. According to this figure, the shear strength envelope resulting from the direct shear tests of this research (DS-TS) stands generally at a higher level than the shear strength envelope resulting from the triaxial tests (TR-B). The same conclusion is evident for the other two materials [17,20]. Different stress paths and different boundary conditions (direct shear test versus triaxial test) are responsible for this difference in the behavior [21].

0

200

400

600

800

1000

1200

0 200 400 600 800 1000 1200σn (kPa)

τ p (k

Pa)

DS-TDDS-TSTR-BDS-G1DS-G2TR-G1TR-G2DS-TTR-T

Figure 13. Maximum shear strength versus normal stress in different tests

For all of the above tests, Figure 14 compares the variations of the peak friction angle values versus normal stress. This figure evidences that the friction angle (φp) values decrease at decreasing rates. Also, the figure indicates that φp values resulting from the direct shear tests are generally larger than their corresponding φp values resulting from the triaxial tests.

35

40

45

50

55

60

65

70

0 200 400 600 800 1000 1200σn (kPa)

φp

(deg

)

DS-TDDS-TSTR-BDS-G1DS-G2TR-G1TR-G2DS-TTR-T

Figure 14. Peak friction angle versus the normal stress in different tests



The literature suggests that the peak friction angle values of rockfill materials from direct shear tests are generally about 3 to 4 degrees higher than those from triaxial tests [20]. On the other hand, the maximum difference between φp values resulting from plane strain tests and those resulting from triaxial tests is about 8 degrees (φps ≈ φtr+8) [2]. Therefore, owing to the fact that the stress conditions in the field is generally more similar to plane strain conditions, direct shear tests may suggest more realistic values for peak friction angle of rockfill materials.

Figure 15 illustrates the Marsal’s breakage index (Bg) versus normal stress for some of the above tests. The results of particle breakage for the other references were not available. The figure shows that the particle breakage in TR-B is more than the particle breakage in DS-TS, which are, respectively, triaxial and direct shear tests on the same material. Most probably, the difference in the gradations (Figure 12), stress paths, and boundary conditions are together responsible for the above different Bg values.

0

2

4

6

8

10

0 200 400 600 800 1000 1200σn (kPa)

Bg

(%)

DS-TDDS-TSTR-BTR-G1TR-G2Linear (DS-TD)Linear (DS-TS)

Figure 15. Marsal’s breakage index versus normal stress in different tests

6. SUMMARY AND CONCLUSIONS

A medium-scale direct shear apparatus has been employed to investigate the mechanical behavior of an angular rockfill material in dry, saturated, and dry-saturated conditions at five normal stress levels. The effects of stress level, wetting, and saturation collapse were explored. Finally, some of the results were compared with the results of triaxial tests having been carried out on the same material.

The tests results on the dry specimens showed that stress level affects the behavior of the rockfill material. The shear strength of the material increased non-linearly as the normal stress increased. Therefore, a non-linear relationship should be applied for defining the failure envelope of this material. The increase of the normal stress reduced the peak friction angle (in decreasing rates) and the dilation angle. Also, the comparison of the materials’ gradations before and after the tests showed that particle breakage occurred during the tests and the breakage amounts increased with increasing the normal stress. A number of correlations have been suggested to estimate the behavior of this material as well.

A similar trend was observed in the saturated specimens; however, wetting reduced the strength parameters’ values. The shear strength, peak friction angle, dilation angle, and particle breakage of the saturated specimens were less than those of the dry specimens.

Although the trend was the same in the dry-saturated tests, the strength parameters were less, but the particle breakage was more than those of the saturated tests. In these tests, also, saturation induced sudden settlement and shear stress reduction.

Moreover, the comparison of the results of these tests with the results of the triaxial tests on the same material showed that the shear strength and peak friction angle from the direct shear tests were higher than those of the triaxial tests.

Overall, in order to properly design a rockfill structure (e.g., a rockfill dam), it is of crucial importance to think over the type of the tests and pay close attention to the two significant factors (stress level and wetting), and their effects on the mechanical behavior of rockfill materials. Also, saturation collapse phenomenon should be considered as far as the behavior of dry-constructed rockfill dams is concerned.



ACKNOWLEDGMENTS

The tests were performed at the soil mechanics laboratory of Amirkabir University of Technology. The authors are grateful to Mr. T. Bahrami and Mr. R. Javadi (the laboratory staff) for their useful comments and help in conducting the tests.

REFERENCES

[1] A. D. M. Penman, Rockfill, Building Research Station. Paper 15–1, BRE, Garston, Watford, 1971.

[2] R. J. Marsal, Mechanical Properties of Rockfill, Embankment Dam Engineering. New York, Casagrande vol. Wiley, 1973, pp. 109–200.

[3] J. Zeller and R. Wullimann, “The Shear Strength of the Shell Materials for the Go-Schenenalp Dam, Switzerland”, Proc 4th J. on SMFE, London, 2(1957), pp. 399–404.

[4] J. Lowe, “Shear Strength of Coarse Embankment Dam Materials”, Proc., 8th Int. Congress on Large Dams, 3(1964), pp. 745–761.

[5] E. Fumagalli, “Tests on Cohesionless Materials for Rockfill Dams”, Journal of Soil Mechanics and Foundation Engineering, ASCE, 95(SMI)(1969), pp. 313–330.

[6] R. J. Frost, “Some Testing Experiences and Characteristics of Boulder-Gravel Fills in Earth Dams”, ASTM, STP (1973), pp. 523, 207.

[7] J. A. Charles and K. S. Watts, “The Influence of Confining Pressure on the Shear Strength of Compacted Rockfill”, Geotechnique, 30(4) (1980), pp. 353–367.

[8] R. J. Marsal, “Large Scale Testing of Rockfill Materials”, Journal of the Soil Mechanics and Foundations Div., ASCE, 93(SM2)(1967), pp. 27–43.

[9] M. Aghamajidi, “Laboratory Investigation of Creep in Rockfill Materials”, MSc thesis, Tarbiat Modarres University, Tehran, Iran, 2004, (in Persian).

[10] P. V. Lade, J. A. Yamamuro, and P. A. Bopp, “Significance of Particle Crushing in Granular Materials”, Journal of Geotechnical and Geoenvironmental Engineering, 122(4)(1996), pp. 309–316.

[11] E. Alonso, “Exploring the Limits of Unsaturated Soil Mechanics: The Behavior of Coarse Granular Soil and Rockfill”, The Eleventh Spenser J.Buchanan Lecture, Collage Station, TX77840 USA, 2003.

[12] A. Soroush and A. Aghaei Araei, “Analysis of Behavior of a High Rockfill Dam”, Proceeding of the Institution of Civil Engineering, Geotechnical Engineering, 159(GEI)(2006), pp. 49–59.

[13] I. W. Farmer and P. B. Attewell, “The Effect of Particle Strength on the Compression of Crushed Aggregates”, Rock Mechanics, 5(1973), Springer-Verlag, pp. 237–248.

[14] B. N. Touileb, S. Bonelli, P. Anthiniac, A. Carrere, O. Debordes, G. LA Berbera, A. Bani, and G. Mazza, “Settlement by Wetting of the Upstream Rockfills of Large Dams”, Proc. of the 53rd Canadian Geotechnical Conf., Montreal, 1(2000), pp. 263–270.

[15] Z. T. Beiniawski, “Engineering Classification of Jointed Rock Masses”, Trans. S. Afr. Instn. Civ. Engrs. 15(1973), pp. 335–343.

[16] A. Soroush and A. Aghaei Araei, “Uncertainties in Mechanical Behavior of Rockfills During First Impounding of Rockfill Dams”, 73rd Annual Meeting of Icold, Tehran, Iran, 2005.

[17] M. Gharavy, “Experimental and Numerical Investigations into the Mechanical Characteristics of Rockfill Materials”, Ph.D. thesis, Univ. of Newcastle Upon Tyne, Dep. of Civil Engineering, U.K., 1996.

[18] B. Indraratna, D. Ionescu, and H. D. Christie, “Shear Behavior of Railway Ballast Based on Large-Scale Triaxial Tests”, Journal of Geotechnical and Geoenvirnomental Engineering., 124(5)(1998), pp. 439–449.

[19] K. J. Douglas, ”The Shear Strength of Rock Masses”, Ph.D. thesis, University of New South Wales, Sydney, Australia, 2002.

[20] M. H. T. Rayhani, “Investigation of the Behavior of Earth and Rockfill Dam Shell Materials Based on Triaxial and Direct Shear Tests”, MSc thesis, Tarbiat Moallem University, Tehran, Iran, 2000 (in Persian).

[21] S. H. Liu, D. Sun, and H. Matsuoka, “On the Interface Friction in Direct Shear Test”, Computers and Geotechnics, 32(5)(2005), pp. 317–325.

asadzadeh - direct shear testing on a rockfill material

Documents

normal stress levels

peak friction angle

marsals breakage index

point load tests

direct shear testing

versus normal stress

maximum particle size

maximum dilation angle