ARTICLE IN PRESS

Geotextiles and Geomembranes 24 (2006) 116128 www.elsevier.com/locate/geotexmem

A simple method to evaluate the pullout resistance of extruded geogrids embedded in a compacted granular soil
` Nicola Moraci, Domenico Gioffre
` Universita Mediterranea di Reggio Calabria, Dip. MECMAT, via Graziella Loc. Feo di Vito, I-89060 Reggio Calabria, Italy Received 28 February 2005; received in revised form 26 September 2005; accepted 15 November 2005 Available online 10 January 2006

Abstract Pullout tests are necessary in order to study the interaction behaviour between soil and geosynthetics in the anchorage zone; hence, the resulting properties have direct implications on the design of reinforced soil structures. Several experimental studies showed the inuence of different parameters (reinforcement stiffness, geometry and length, applied vertical effective stress, and geotechnical properties of soil) on the peak and on residual pullout resistance. On the basis of the results of the tests carried out by Moraci and Recalcati [2005. Factors affecting the pullout behaviour of extruded geogrids embedded in a compacted granular soil. Geotextiles and Geomembranes, submitted for publication], a new theoretical method was developed to determine the peak and the residual pullout resistance of extruded geogrids embedded in a compacted granular soil. The method is capable of evaluating both the bearing and the frictional components of pullout resistance, taking into account the reinforcement extensibility and geometry as well as the non-linearity of the failure envelope of backll soil. The comparison between theoretical and experimental results was favourable, thus conrming the suitability of the proposed approach. r 2005 Elsevier Ltd. All rights reserved.
Keywords: Soil dilatancy; Reinforcement extensibility; Pullout resistance; Skin friction; Bearing resistance

1. Introduction The main interaction mechanisms affecting the pullout resistance of extruded geogrids are the skin friction, between soil and reinforcement solid surface, and the bearing resistance, that develops against transversal elements (Fig. 1). The pullout resistance of a geogrid, assuming that the different interaction mechanisms act at the same time with maximum value and that they are independent of each other, may be evaluated using the following equation: PR PRS PRB , (1)

The frictional component of pullout resistance, for a geogrid of length LR and unit width WR (Fig. 2), may be evaluated from the following expression: PRS 2aS LR t 2aS LR s0n tan d, (2) where s0n is the normal effective stress, d the skin friction angle between soil and geogrid, t the shear stress acting at soilreinforcement interface and aS the fraction of geogrid surface area that is solid. To evaluate the bearing component of pullout resistance, Jewell (1990) proposed the following expression:   LR PRB (3) aB s0b B, S where S is the spacing between geogrid bearing members, LR =S the number of geogrid bearing members, aB the fraction of total frontal area of geogrid available for bearing, B the bearing member thickness and s0b the effective bearing stress on the geogrid bearing members. For granular soils, the bearing stresses s0b on geogrid bearing members are linked to the soil shear strength angle,

where PRS is the skin friction component of pullout resistance and PRB the bearing component of pullout resistance.
Corresponding author. Tel.: +39 09658 75263; fax: +39 09658 75201.

E-mail addresses: moraci@ing.unirc.it (N. Moraci), ` gioffre@ing.unirc.it (D. Gioffre). 0266-1144/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.geotexmem.2005.11.001

ARTICLE IN PRESS
` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128 117

Nomenclature d aB s0b s0n aS Ab Ar At B Br Bt Beq CaS d50 fb L LR skin friction angle between soil and geogrid (deg.) fraction of total frontal area of geogrid available for bearing (dimensionless) effective bearing stress on the geogrid bearing members (kN/m2) normal effective stress (kN/m2) fraction of geogrid surface area that is solid (dimensionless) area of each rib element (mm2) node embossment area (mm2) bar portion between two nodes area (mm2) bearing member thickness (mm) node thickness (mm) thickness of the bar portion between two nodes (mm) strip of uniform thickness reduction coefcient of geogrid area where skin friction develops (aS) average grain size (mm) soilgeosynthetic pullout interaction coefcient (dimensionless) reinforcement length in the anchorage zone (m) specimen length (m)

nt ntb PR PRB PRR PRS PRRS S U wopt Wr Wt f0 f0cv f0p gdmax mR S=GSY mS/GSY

number of geogrid bearing members number of nodes in a transversal element pullout resistance (kN/m) bearing component pullout resistance (kN/m) residual pullout resistance (kN/m) skin friction component pullout resistance (kN/m) skin friction component pullout resistance under residual conditions (kN/m) spacing between geogrid bearing members (mm) uniformity coefcient (dimensionless) optimum water content (%) node width (mm) width of the bar portion between two nodes (mm) soil shear strength angle (deg.) soil shear strength angle at constant volume (deg.) peak shear strength angle (deg.) maximum dry unit weight (kN/m3) soilgeosynthetic residual interface apparent coefcient of friction (dimensionless) soilgeosynthetic peak interface apparent coefcient of friction (dimensionless)

Fig. 2. Denition of terms for a geogrid (Jewell, 1990). Fig. 1. The two mechanisms for bond between reinforcement and soil (Jewell et al., 1985).

For general shear failure mechanism, the ratio s0b =s0n may be dened as follows: the initial stress state, the interface roughness and the reinforcement depth in relation to the sizes of the bearing members (Rowe and Davis, 1982). For punching failure mechanism, the ratio s0b =s0n depends only on soil shear strength angle and may be dened as following (Jewell et al., 1985):   s0b p f0 p=2f0 tan f0 e tan . (4) 4 2 s0n   s0b p f0 p tan f0 e tan . 4 2 s0n (5)

According to Jewell (1996), the Eqs. (4) and (5) represent a lower bound and an upper bound for the bearing component of resistance in pullout conditions. In order to evaluate the bearing component of pullout resistance, Matsui et al. (1996) and Bergado and Chai (1994) proposed other relationships.

ARTICLE IN PRESS
118 ` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128

In particular, Matsui et al. (1996) proposed an equation based on a Prandtls mechanism, as shown in Fig. 3:     s0b p f0 p f0 ep tan f tan cos 4 2 4 2 s0n  0  p f 1 sin f0 sin . 6 4 2 The comparison between the values s0b =s0n obtained through expression (6) and the pullout test results performed by the authors on steel grid reinforcement (diameter 6 mm) embedded in granular soil (f0 37:31) showed a good agreement (Fig. 4). Assuming that there was an uniform distribution of shear stress applied along the whole surface of reinforcement, Jewell (1990) obtained a general theoretical relationship to evaluate the pullout resistance of a geogrid:   LR PR 2aS LR s0n tan d aB Bs0b S 2f b LR s0n tan f0 , 7 where fb is the interaction coefcient in pullout conditions. This coefcient can be evaluate based on geometrical parameters of the reinforcement and on the soil shear strength characteristics (Jewell, 1990). Others studies (Palmeira, 2004; Palmeira and Milligan, 1989) emphasized the inuence of scale, shape and interference effects. The scale effects are related to the ratio between the transverse element thickness (B) of the reinforcement and soil grain size D50 and to the ratio between the spacing between geogrid bearing member (S) and soil grain size D50. These effects are relevant if B/D50p10 (Fig. 5,

Fig. 4. Comparison between proposed equation and available pullout test data of anchor and grid reinforcements (Matsui et al., 1996).

100 Bearing resistance per unit width of reinforcement (kN/m) N 75 Grid geometry 225x150x6 (mm) 50 Normal pressure n = 98.1k Pa

N= 1 N= 2

25 N= 3 0 0

20

40 60 80 Pullout displacement (mm)

100

120

Fig. 5. Scale effect on bearing capacity (Palmeira and Milligan, 1989).

Fig. 3. Assumed failure surface for the evaluation of s0b (Matsui et al., 1996).

Palmeira and Milligan, 1989) and if S/D50 is less than 3 (Jewell et al., 1985). The shape effects are related to the geometry of the reinforcement transverse elements. Jewell (1996), in order

ARTICLE IN PRESS
` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128 119

to take into consideration the shape effects, suggested using a shape coefcient that can be assumed to be equal to 1.0 in the case of a circular shape and equal to 1.2 in the case of rectangular shape. Milligan et al. (1990), by means of photo-elastic studies, showed that the bearing action reduced the friction between the soil and the reinforcement (interference effect). This means that, under similar conditions, the fraction of the ultimate pullout load due to skin friction may be considerably smaller than that due to bearing, and smaller than the value produced between grid area available for friction, normal stress and the skin friction coefcient between the soil and grid. As the ratio between the distance between grid bearing members and bearing member thickness increases, the interference between these members will decrease to an extent that they will behave as a series of isolated bearing members being pulled out of the soil mass. Palmeira (2004) observed metal grids embedded in dense sand and saw that, for S/B ratios of above 40, the grid bearing members behaved in isolation, under the experimental conditions adopted. As mentioned above, the passive failure surfaces that developed against bearing members cause a reduction of the skin friction component of the pullout resistance. This effect can be taken into account in terms of reduction of geogrid area where skin friction develops (aS). For this reason a reduction coefcient, CaS, may be introduced. The limits of theoretical expression used to evaluate the soilgeosynthetic pullout interaction coefcient, fb, have been investigated by different researchers (Palmeira and Milligan, 1989; Wilson-Fahmy and Koerner, 1993; Moraci and Montanelli, 2000; Ghionna et al., 2001). In particular, previous experimental studies (Palmeira and Milligan, 1989; Moraci and Montanelli, 2000; Ghionna et al., 2001) have shown that the values of fb are largely inuenced by the reinforcement geometry, extensibility and soil dilatancy. Thus, it is important to develop a new theoretical expression that is able to include the evaluation of all the parameters that inuence the mobilization of the interaction mechanisms (frictional and passive) during pullout, as emphasized by previous works (Moraci et al., 2002, 2003, 2004; Moraci and Recalcati, 2005). In the present paper, on the basis of the test results obtained by Moraci and Recalcati (2005), a new theoretical method was developed to evaluate the pullout resistance of extruded geogrids embedded in a compacted granular soil. The method is able to evaluate both the passive and the frictional components of pullout resistance taking into account the reinforcement extensibility and geometry, as well as, the non-linearity of the failure envelope of the backll soil. 2. The method Test results obtained by Moraci and Recalcati (2005) showed the inuence of different parameters (reinforce-

ment stiffness and structure, embedded length and vertical effective stress) on the pullout behaviour of mono-oriented extruded geogrids embedded in a compacted uniform medium sand. In particular, it was found that the dilatancy of the soil at the interface is the phenomenon that most inuences the pullout resistance and the interface apparent coefcient of friction (mS/GSY). Due to the dilatancy effects, the apparent coefcient of skin friction mobilized at low vertical effective conning pressures is higher than that at high conning pressures. Experimental results (Moraci and Recalcati, 2005) also showed that the reinforcement extensibility inuences the peak pullout resistance. In particular, extensibility effects were more evident in long reinforcements and in high vertical conning stresses. In residual conditions, the extensibility effects were negligible. Test results (Moraci and Recalcati, 2005) also showed an increase in peak and residual pullout resistance, and therefore in the mobilized interface apparent coefcient of friction, while increasing the competent bearing area of each node, upon which the bearing mechanisms are mobilized. The decrease of the pullout resistance after the peak is related to both reinforcement length and conning stress. Finally, the apparent coefcient of friction mobilized in residual conditions depends only on the applied vertical stress and geogrid geometry; in these conditions mR S=GSY does not depend on reinforcement length. In the case of long reinforcements and high effective vertical stresses, reinforcement extensibility induces a progressive mobilization of the elementary interaction mechanisms (skin friction and bearing resistance). Vice versa, in short reinforcements (independent of vertical effective stresses) and in long reinforcements (subjected to low vertical effective stresses) the longitudinal strains are small. In such cases, the reinforcement behaves in a rigid way rigid and the interaction mechanisms are effectively activated at the same time along the whole length of the reinforcement. Different experimental studies (Matsui et al., 1996; Palmeira, 2004) showed that, for the characteristic displacement eld of pullout tests, the bearing stress, after the displacement corresponding to the maximum value, remains almost constant with any increase in displacement (Figs. 6 and 7). In order to separate the two components of the pullout resistance, it is possible to perform pullout tests on geogrid specimens where a portion of transverse reinforcing elements are removed (Alagiyawanna et al., 2001; Alfaro et al., 1995; Matsui et al., 1996; Palmeira and Milligan, 1989). Due to their structure, the bearing resistance of the geogrids used in this study, develops both at the node embossments and at the transverse bars. With this reinforcement form it is not possible to determine experimentally the two components of pullout resistance.

ARTICLE IN PRESS
120
2.5

` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128

2.0

squa re sec t ion ro un d s ecti on

(b /n ) s/(b /n ) ' ' ' '

1.5

Fig. 8. Schematic cross-section AA of the geogrid bar.

1.0

0. 5

Table 1 Structural characteristics of the different geogrids (for symbol see Fig. 8) Geogrid Wr (mm) 11.26 11.86 12.36 Wt (mm) 6.6 6.0 5.5 Br (mm) 3.80 4.65 5.16 Bt (mm) 3.57 4.48 4.85 Ab (mm2) 66.35 82.03 90.45

0. 0

10

15

20

25

B/ D50
Fig. 6. Relationship between bearing resistance and pullout displacement (Matsui et al., 1996).

GG1 GG2 GG3

element (including the single node and the bar portion between two nodes) where the bearing resistance can be mobilized (Fig. 8) and s0b the bearing stress evaluated by Eq. (6) according to Matsui et al. (1996). 2. To take into account the particular structure of the elements on which the bearing resistance mobilizes, the soil dilatancy effects (non-linearity of the failure envelope of back ll soil) and the geogrid extensibility. 3. The comparison between theoretical (Moraci and Recalcati, 2005) and experimental values of the pullout resistance under different conditions. Pullout tests have been performed on three different HDPE extruded mono-oriented geogrids (described as GG1, GG2 and GG3, respectively) (Moraci and Recalcati, 2005). The three geogrids show similar geometrical characteristics when viewed in plan. They have the same number of tensile elements per unit width and longitudinal rib pitch, and similar elliptical aperture shape. On the contrary, the three geogrids have a different cross-sectional shape with major differences in rib and bar thickness. A more detailed analysis of the transversal bar geometry has shown a non-uniform shape with greater thickness at the rib intersection. The bearing interaction mechanisms develop both at the node embossments and at the transverse bars. Therefore, the node embossment and the transverse bar geometry have been carefully determined to evaluate the bearing resistance surfaces. The results of this analysis are reported in Table 1, where Wr and Br are the node width and thickness, respectively, Wt and Bt are the width and thickness of the bar portion between two nodes, respectively (Fig. 8), and Ab is the area of each rib element (including the node embossment and the bar portion between two nodes At Ar ) where the bearing resistance can be mobilized. The complex geometry of the transverse bars, including the areas Ab in the same transverse element, was assumed to be equivalent to that of a strip of uniform thickness (Beq) (Fig. 9).

Fig. 7. Load displacement curve for an isolated bearing member (Palmeira, 2004).

In order to validate the previous ndings, which may explain the different behaviour of the three geogrids used in the research (in terms of pullout resistance and in terms of apparent coefcient of friction), the following approach was used: 1. The use of a simple Eq. (8) for the determination of the pullout resistance in geogrids and soils for which the scale effects are negligible (i.e. S=B larger than 40 and S/D50 larger than 1000): PR 2C aS aS LR s0n tan d nt ntb Ab s0b , (8) where CaS is the reduction coefcient of geogrid area where skin friction develops, nt LR =S the number of geogrid bearing members, ntb the number of nodes in a transversal element, Ab At Ar the area of each rib

ARTICLE IN PRESS
` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128 121

Fig. 9. Assumed equivalent geometry.

A granular soil was used in the tests. The soil was classied as uniform medium sand with a uniformity coefcient U d 60 =d 10 1:5 and an average grain size d 50 0:22 mm. Standard Proctor compaction tests gave a maximum dry unit weight gdmax 16:24 kN=m3 at an optimum water content wopt 13:5%. Direct shear tests, performed at an initial unit weight equal to 95% of gdmax (obtained at a water content of 9%), yielded high single values of the peak shear strength angle f0p , in the range 481 (for s0v 10 kPa) to 421 (for s0v 100 kPa). The shear strength angle at constant volume, f0cv , was 341. The soil shear strength angle used to determine of the skin friction component of the pullout resistance based on previous experimental researches on smooth HDPE geomembranes was assumed to have a value of d equal to 1/3 f0 (Fannin and Raju, 1993; Raju, 1995). In order to take into account the reinforcement extensibility, the following assumptions were made: 1. In long reinforcements (LR 0:921:15 m) and high effective vertical stresses, the reinforcement extensibility induces a progressive mobilization of the two elementary interaction mechanisms. Under these conditions, the skin friction was evaluated using an average value of the shear strength angle between the peak and the constant volume values, assuming a non-linear failure envelope for the backll soil. 2. In short reinforcements (LR 0:4 m), independent of the applied vertical effective stresses, and in long reinforcements (subjected to low vertical effective stresses), the longitudinal strain is small. In such cases, the reinforcement behaves as a rigid material and the interaction mechanisms are activated simultaneously along the whole reinforcement. Under these conditions, the peak shear strength angle can be used to evaluate both components of the pullout resistance, assuming a non-linear failure envelope for the backll soil and a suitable stress level. On the basis of the experimental results obtained by Matsui et al. (1996) and Palmeira (2004), the bearing resistance component of pullout resistance was evaluated using the peak shear strength angles corresponding to the

different vertical effective stresses, in order to take into account the non-linearity of the failure envelope (due to dilatancy effects) of the backll soil. Eq. (8) permits the evaluation of the residual pullout resistance PRR. In this case in order to evaluate the skin friction component of pullout strength, the soil shear strength angle at constant volume f0cv was used. In order to evaluate the reduction of the skin friction component induced by the passive failure surfaces developed on bearing members, a reduction coefcient, CaS, of the geogrid area, where skin friction develops (aS), was used. This value, derived from the assumption that the maximum extensions of passive failure surfaces are equal to 40 times the thickness of the equivalent bearing members (Fig. 9), is given by C aS Seff S 40nBeq . (9) S S This reduction is only applied under residual conditions.

3. Experimental validation of proposed method In order to validate the proposed method, the theoretical values of the peak and residual pullout resistances obtained using Eq. (8) were compared with the experimental results obtained by Moraci and Recalcati (2005) reported in Table 2. Tables 35 show the peak (Pexp ) and residual (Pexp ) R RR experimental pullout resistances, calculated by the present method (Ptheor and Ptheor ), the theoretical values of the R RR peak and residual skin friction resistance (Ptheor and Ptheor ) RS RRS and the ratio between the skin friction resistance and pullout resistance both at the peak and in residual conditions (Ptheor =Pexp and Ptheor =PRexp ). RS RRS R R Figs. 1012 show (in the different reinforcement lengths) the comparison between experimental and theoretical values of the peak pullout resistances, evaluated for the different applied vertical effective conning stresses. The skin friction component of the peak pullout resistance is small in comparison to the bearing component. On the basis of the theoretical analysis (Eq. (8)), the skin friction component PRS 2aS LR s0n tan d represents less than 20% of the peak pullout resistance. In particular, the values vary between 9% and 19%, for GG1, and between

ARTICLE IN PRESS
122 ` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128 Table 2 Peak (PR) and residual (PRR) pullout resistance (kN/m) measured in the tests (Moraci and Recalcati, 2005) Geogrid Specimen length (m) Normal stress s0v 10 (kPa) PR GG1 GG1 GG1 GG2 GG2 GG2 GG3 GG3 GG3
a

25 (kPa) PRR 5.63 12.14 14.76 8.44 15.43 19.53 7.36 15.48 19.61 PR 20.26 34.55 37.13 24.76 39.99 51.43 22.72 41.80 47.75 PRR 13.29 29.79 34.32 15.43 32.14 44.00 13.64 34.69 43.79

50 (kPa) PR 30.95 52.53 62.79 41.18 70.07 75.62 37.68 72.95 81.77 PRR 18.93 50.34 62.79 24.04 62.46 75.62 25.18 61.27 81.77

100 (kPa) PR 39.79 78.44a 72.48a 56.59 103.91 106.91a 58.68 97.59 115.19 PRR 26.43 37.51 103.91 49.04 97.59 115.19

0.40 0.90 1.15 0.40 0.90 1.15 0.40 0.90 1.15

9.62 16.62 20.00 13.42 21.32 26.96 12.84 19.85 24.35

Specimen failure.

Table 3 Theoretical and experimental peak (PR) and residual (PRR) pullout resistance (kN/m) for geogrid GG1 LR (m) 0.40 0.40 0.40 0.40 0.90 0.90 0.90 0.90 1.15 1.15 1.15 1.15
a

s0v (kPa)

Pexp (kN/m) R

Ptheor (kN/m) R 7.33 14.29 22.78 37.03 14.88 29.10 45.51 74.28 18.66 36.51 57.10 93.28

Ptheor (kN/m) RS 0.87 2.08 3.98 7.58 1.96 4.69 7.89 15.36 2.50 5.99 10.08 19.63

Ptheor =Pexp RS R (%) 9.06 10.28 12.85 19.04 11.79 13.56 15.02
a

Pexp (kN/m) RR

Ptheor (kN/m) RR 6.61 12.59 19.57 30.98 13.27 25.28 39.33 62.34 16.59 31.62 49.21 78.02

Ptheor (kN/m) RRS 0.15 0.38 0.76 1.52 0.34 0.86 1.71 3.42 0.44 1.09 2.19 4.38

Ptheor =Pexp RRS RR (%) 2.70 2.87 4.02 5.76 2.82 2.88 3.40
a

10 25 50 100 10 25 50 100 10 25 50 100

9.62 20.26 30.95 39.79 16.62 34.55 52.53

a

5.63 13.29 18.93 26.43 12.14 29.79 50.34

a

20 37.13 62.79
a

12.52 16.14 16.06

a

14.76 34.32 62.79

a

2.97 3.19 3.49

a

Specimen failure.

Table 4 Theoretical and experimental peak (PR) and residual (PRR) pullout resistance (kN/m) for geogrid GG2 LR (m) 0.40 0.40 0.40 0.40 0.90 0.90 0.90 0.90 1.15 1.15 1.15 1.15
a

s0v (kPa)

Pexp (kN/m) R

Ptheor (kN/m) R 9.11 17.78 28.38 46.19 18.51 36.23 56.68 92.65 23.20 45.46 71.13 116.37

Ptheor (kN/m) RS 1.12 2.69 5.13 9.77 2.53 6.04 10.18 19.81 3.23 7.72 13.00 25.32

Ptheor =Pexp RS R (%) 8.37 10.86 12.45 17.26 11.86 15.11 14.52 19.07 11.98 15.01 17.19
a

Pexp (kN/m) RR

Ptheor (kN/m) RR 8.19 15.59 24.23 38.38 16.42 31.29 48.71 77.26 20.54 39.15 60.95 96.69

Ptheor (kN/m) RRS 0.20 0.49 0.98 1.96 0.44 1.10 2.21 4.42 0.56 1.41 2.82 5.64

Ptheor =Pexp RRS RR (%) 2.33 3.18 4.08 5.23 2.86 3.44 3.54 4.25 2.89 3.20 3.73
a

10 25 50 100 10 25 50 100 10 25 50 100

13.42 24.76 41.17 56.59 21.32 39.99 70.07 103.91 26.96 51.43 75.62
a

8.44 15.43 24.04 37.51 15.43 32.14 62.46 103.91 19.53 44.00 75.62
a

Specimen failure.

8% and 19% for GG2 and GG3 (Tables 35). The lower values are related to the lower vertical effective stress and to short reinforcements; the higher values are related to

the higher vertical effective stresses and to long reinforcements. Figs. 1315 show the same comparison in terms of residual pullout resistance.

ARTICLE IN PRESS
` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128 Table 5 Theoretical and experimental peak (PR) and residual (PRR) pullout resistance (kN/m) for geogrid GG3 LR (m) 0.40 0.40 0.40 0.40 0.90 0.90 0.90 0.90 1.15 1.15 1.15 1.15 s0v (kPa) Pexp (kN/m) R Ptheor (kN/m) R 9.79 19.00 30.14 48.73 19.84 38.59 60.21 97.70 24.86 48.39 75.51 122.61 Ptheor (kN/m) RS 0.99 2.36 4.50 8.57 2.22 5.30 8.93 17.39 2.83 6.78 11.41 22.22 Ptheor =Pexp RS R (%) 7.68 10.37 11.94 14.61 11.17 12.69 12.24 17.82 11.65 14.19 13.95 19.29 Pexp (kN/m) RR Ptheor (kN/m) RR 8.98 17.08 26.50 41.88 18.01 34.26 53.22 84.20 22.52 42.85 66.58 105.35 Ptheor (kN/m) RRS 0.17 0.43 0.86 1.72 0.39 0.97 1.94 3.88 0.50 1.24 2.48 4.95 Ptheor =Pexp RRS RR (%) 2.34 3.16 3.42 3.51 2.51 2.79 3.16 3.97 2.52 2.83 3.03 4.30 123

10 25 50 100 10 25 50 100 10 25 50 100

12.84 22.72 37.68 58.68 19.85 41.8 72.95 97.59 24.35 47.74 81.77 115.15

7.36 13.64 25.18 49.04 15.48 34.69 61.27 97.59 19.61 43.79 81.77 115.19

100
experimental Pullout failure theoretical

experimental

theoretical

120
80

Pullout failure

P [kN/m]

60 P [kN/m]
LR = 1.15m

LR = 0.90m

80
LR = 1.15m LR = 0.90m LR = 0.40m

40
LR = 0.40m

40
20

GG1 - peak

GG2 - peak

0 0 20 40 60 80 100 120 140 v [kPa]

Fig. 10. Comparison between experimental and theoretical values of peak pullout resistance for GG1.

0 0 20 40 60 80 100 120 140 v [kPa]

Fig. 11. Comparison between experimental and theoretical values of peak pullout resistance for GG2.

Thus, the following conclusions can be drawn. The skin friction component of residual pullout resistance is also small in comparison to the bearing component. On the basis of the theoretical analysis (Eq. (8)), the skin friction PRS 2C aS aS LR s0n tan d component represents less than 6% of the residual pullout resistance (Tables 35). Such small values are due to the reduction of the skin friction component caused by the bearing failure surfaces (interference effects). Figs. 1012 show that the proposed method is in close agreement with the experimental data. In particular, an underestimation of the peak pullout resistance was observed which was more evident for short reinforcements. This situation could be attributed to the local increment of

the vertical effective stress due to the constrained dilatancy, which is not considered in the simple proposed model. For short reinforcements (LR 0:40 m), the percentage differences between experimental results and theoretical values in terms of peak pullout resistance, ranges between 7% and 32%; for long reinforcements (LR 0:902 1:15 m), such differences are quite small (0% and 19%) (Table 6). Similar results were obtained in terms of residual pullout resistance. In this case, the method agrees well with the experimental data (Figs. 1315). In short reinforcements, the differences between the experimental results and the theoretical values vary between 1% and 26% (Table 6).

ARTICLE IN PRESS
124 ` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128

160

experimental Pullout failure

120

theoretical

experimental 120 Pullout failure

theoretical

P [kN/m]

P [kN/m]

80

LR = 1.15m

80

LR = 1.15m

LR = 0.90m LR = 0.40m

LR = 0.90m

40

40

LR = 0.40m

GG3 - peak
0 0 20 40 60 80 100 120 140

GG2 - residual 0 0 20 40 60 80 100 120 140

v [kPa]
Fig. 12. Comparison between experimental and theoretical values of peak pullout resistance for GG3.

v [kPa]
Fig. 14. Comparison between experimental and theoretical values of residual pullout resistance for GG2.

100 experimental Pullout failure

Pullout failure 120

theoretical

160 experimental theoretical

80

P [kN/m]

60 LR = 1.15m 40

P [kN/m]

LR = 0.90m

LR = 1.15m 80 LR = 0.90m

LR = 0.40m
LR = 0.40m

20

40

GG1 - residual 0 0 20 40 60 80 v [kPa] 100 120 140

0 0 20 40 60 80 v [kPa] 100 120 140 GG3 - residual

Fig. 13. Comparison between experimental and theoretical values of residual pullout resistance for GG1.

Fig. 15. Comparison between experimental and theoretical values of residual pullout resistance for GG3.

For the tests in which conned tensile failure occurs the method gives higher values of pullout resistance than in-air tensile resistance evaluated at the same pullout test rate (Moraci and Recalcati, 2005). Thus, the method is also useful in the evaluation of the combination of s0v and LR that produces conned reinforced pullout failures. To evaluate the soilgeosynthetic interface apparent coefcient of friction, mS/GSY, the following equation

can be used: mS=GSY 2aS LR s0n tan d nt ntb Ab s0b PR . 0 2LR sv 2LR s0v (10)

The analysis was performed both in terms of the peak and the residual soilgeosynthetic interface apparent coefcient of friction.

ARTICLE IN PRESS
` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128 Table 6 Percentage differences between experimental results and theoretical values LR (m) s0v (kPa) GG1   Ptheor Pexp 
R R

125

GG2 %   Ptheor Pexp 

RR RR

GG3
R

PR

exp

PRR

exp

  Ptheor Pexp 
R

PR

exp

  Ptheor Pexp 
RR RR

PRR

exp

  Ptheor Pexp 
R R

PR

exp

  Ptheor Pexp 
RR RR

PRR

exp

0.40 0.40 0.40 0.40 0.90 0.90 0.90 0.90 1.15 1.15 1.15 1.15

10 25 50 100 10 25 50 100 10 25 50 100

24 29 26 7 10 16 13

17 5 3 17 9 15 22

7 2 9

12 8 22

32 28 31 18 13 9 19 11 14 12 6

3 1 1 2 6 3 22 26 5 11 19

24 16 20 17 0 8 17 0 2 1 8 6

22 25 5 15 16 1 13 14 15 2 19 9

Specimen failure.

1.4

experimental Lr = 0.40 m experimental Lr = 0.90 m experimental Lr = 1.15 m

theoretical Lr = 0.40 m theoretical Lr = 0.90 m theoretical Lr = 1.15 m

experimental Lr = 0.40 m experimental Lr = 0.90 m experimental Lr = 1.15 m

theoretical Lr = 0.40 m theoretical Lr = 0.90 m theoretical Lr = 1.15 m

1.2

GG2 - peak

GG1 - peak

1.6

S/GSY

S/GSY

1.2

0.8
0.8

0.6

0.4

0.4 0 20 40 60 80 100 120 140

20

40

60

80

100

120

140

v [kPa]
Fig. 16. Comparison between experimental and theoretical values of peak soilgeosynthetic interface apparent coefcient of friction for GG1.

v [kPa]
Fig. 17. Comparison between experimental and theoretical values of peak soilgeosynthetic interface apparent coefcient of friction for GG2.

Figs. 1618 show the trends of the experimental and theoretical values of the peak pullout interface apparent coefcient of friction mS=GSY , as a function of the vertical effective applied stress, for the three different reinforcement specimen lengths used. In all cases, it is possible to observe a reduction in the mobilized peak pullout interface apparent friction coefcient with an increase in the applied vertical effective stress. Moreover, it is possible to note that the lower values of mS/GSY are given with the longer reinforcement specimens. These results are due to two different phenomena:

The rst, of greater importance, is related to soil dilatancy that develops in conjunction with the three-dimensional

passive failure surfaces that arise at the node embossments and at the geogrid transverse reinforcing elements. Due to soil dilatancy, which decreases with an increase in the conning vertical effective stress, two main effects develop: the rst is due to the different work made to expand the dilatancy surface at different vertical effective conning stresses; the second effect is due to the restriction of the dilatancy connected to the nearby soil stiffness (constrained dilatancy), which produces a local increment of the effective conning stress. The second effect, of less intensity, is due to the extensibility of the reinforcement which modies the interface tangential stress distribution and the corresponding pullout strength.

ARTICLE IN PRESS
126 ` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128

Moraci and Recalcati (2005) compared the experimental results of the tests carried out on the three different geogrids, with the same anchorage lengths and normal stress, so that they were not inuenced by the reinforcement extensibility and the dilatancy effects. They observed
2
experimental Lr = 0.40 m experimental Lr = 0.90 m experimental Lr = 1.15 m theoretical Lr = 0.40 m theoretical Lr = 0.90 m theoretical Lr = 1.15 m

GG3 - peak
1.6

1.2

0.8

0.4 0 20 40 60 80 100 120 140

v [kPa]
Fig. 18. Comparison between experimental and theoretical values of peak soilgeosynthetic interface apparent coefcient of friction for GG3.

that the experimental results, interpreted as a function of the different longitudinal tensile stiffnesses, do not show a specic correlation. Vice versa, the test results conrmed that the values of the soilgeosynthetic peak interface apparent coefcient of friction, mS/GSY, are mainly inuenced by the structural characteristics (geometry and shape) of the geogrids. In particular, the maximum percentage differences of the values of mS/GSY are close to the percentage differences of the competent bearing areas (Ab) between geogrid types against which the passive resistance is mobilized (Fig. 19). Figs. 2022 show the same curves obtained in terms of the residual soilgeosynthetic interface apparent coefcients of friction mR S=GSY . The experimental results obtained by Moraci and Recalcati (2005) showed that the residual pullout interface apparent coefcient of friction does not depend on the reinforcement length but only on the applied conning stress. Comparison of the results obtained for the three different geogrids shows that mR S=GSY depends on geogrid geometry. The differences between the predicted and the experimental values range from 0% to 32% under peak conditions and from 1% to 26% under residual ones. The results indicate that the proposed model is suitable to predict the interface apparent coefcient of friction, particularly in the case of extensible reinforcements. In the case of rigid reinforcements, the proposed method underestimates the interface apparent coefcient of friction.
1.6
LR = 0.90 m GG3 GG2

S/GSY

2
LR = 0.40 m GG3 GG2 GG1

1.6 1.2 0.8 0.4 0 0 20 60 80 Normal stress v [kPa] 1.6 40 100

1.2 S/GSY

GG1

S/GSY

0.8

0.4

0 120 0 20 40 60 80 100 120

Normal stress v [kPa] LR = 1.15 m GG3 GG2

1.2 S/GSY

GG1

0.8

0.4

0 0 20 40 60 80 100 120
Normal stress v [kPa]

Fig. 19. Peak interface apparent coefcient of friction vs. s0 for different geogrids (Moraci and Recalcati, 2005).

ARTICLE IN PRESS
` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128
experimental Lr = 0.40 m experimental Lr = 0.90 m experimental Lr = 1.15 m theoretical Lr = 0.40 m theoretical Lr = 0.90 m theoretical Lr = 1.15 m experimental Lr = 0.40 m experimental Lr = 0.90 m experimental Lr = 1.15 m

127
theoretical Lr = 0.40 m theoretical Lr = 0.90 m theoretical Lr = 1.15 m

0.9

1.2

0.8

GG1 - residual 1

GG3 - residual

0.7

RS/GSY

RS/GSY

0.6

0.8

0.5

0.6
0.4

0.3 0 20 40 60 80 v [kPa] 100 120 140

0.4 0 20 40 60 80 v [kPa] 100 120 140

Fig. 20. Comparison between experimental and theoretical values of residual soilgeosynthetic interface apparent coefcient of friction for GG1.

Fig. 22. Comparison between experimental and theoretical values of residual soilgeosynthetic interface apparent coefcient of friction for GG3.

1.2

experimental Lr = 0.40 m experimental Lr = 0.90 m experimental Lr = 1.15 m

theoretical Lr = 0.40 m theoretical Lr = 0.90 m theoretical Lr = 1.15 m

GG2 - residual 1

0.8

0.6

stresses and reinforcement length) predicts the experimental data well, both in terms of the pullout resistance and in terms of the interface apparent coefcient of friction, especially for extensible reinforcements. In the case of extruded geogrids embedded in compacted uniform medium sand, the skin friction components of the peak pullout resistance are small in comparison to the bearing component. The skin friction component represents less than 20% of the peak pullout resistance. The skin friction components of the pullout resistance are small in comparison to the bearing component, in residual conditions. The proposed method can be used also to evaluate the combination of s0v and LR relating to the conned reinforcement pullout failure.

RS/GSY

0.4 0 20 40 60 80 v [kPa] 100 120 140

References
Alagiyawanna, A.M.N., Sugimoto, M., Sato, S., Toyota, H., 2001. Inuence of longitudinal and transverse members on geogrid pullout behaviour during deformation. Geotextiles and Geomembranes 19, 483507. Alfaro, M.C., Miura, N., Bergado, D.T., 1995. Soilgeogrid reinforcement interaction by pullout and direct shear tests. Geotechnical Testing Journal 18, 157167. Bergado, D.T., Chai, J.C., 1994. Pullout force/displacement relationship of extensible grid reinforcements. Geotextiles and Geomembranes 13, 295316. Fannin, R.J., Raju, D.M., 1993. Large-scale pull-out test results on geosynthetics. Proceedings of Geosynthetics 93 Conference, vol. 2. Vancouver, Canada, pp. 633643. Ghionna, V.N., Moraci, N., Rimoldi, P., 2001. Experimental evaluation of the factors affecting pullout test results on geogrids. In: Proceedings of the International Symposium: Earth Reinforcement, Fukuoka, Japan,

Fig. 21. Comparison between experimental and theoretical values of residual soilgeosynthetic interface apparent coefcient of friction for GG2.

4. Conclusions Comparison between the theoretical and experimental permits the following conclusions to be drawn:

The proposed method (which takes into account the effects of soil dilatancy, reinforcement extensibility, geogrid structure and geometry, vertical effective

ARTICLE IN PRESS
128 ` N. Moraci, D. Gioffre / Geotextiles and Geomembranes 24 (2006) 116128 Proceedings of the Seventh International Conference on Geosynthetics, Nice, France, vol. 4. Balkema Publisher, pp. 13451348. Moraci, N., Montanelli, F., Romano, G., 2003. Interface pullout behaviour of geogrids embedded in compacted granular soils. In: Proceedings of the Third European Conference on Soil Mechanics and Geotechnical Engineering, Praga, Rep. Ceka, vol. 1. CICE Publishers, pp. 837841. Moraci, N., Romano, G., Montanelli, F., 2004. Factors affecting the interface apparent coefcient of friction mobilised in pullout condition. In: DGGT, TUM-ZG (Eds.), Third European Geosynthetics Conference, March, Munich, pp. 313318. Palmeira, E.M., Milligan, G.W.E., 1989. Scale and other factors affecting the results of pull-out tests of grid buried in sand. Geotechinique 11 (3), 511524. Palmeira, E.M., 2004. Bearing force mobilization in pull-out tests on geogrids. Geotextiles and Geomembranes 22, 481509. Raju, D.M., 1995. Monotonic and cyclic pullout resistance of geosynthetic. Ph.D. Thesis. The University of British Columbia, Vancouver, Canada. Rowe, R.K., Davis, E.H., 1982. The behaviour of anchor plates in sand. Geotechinique 32 (1), 2541. Wilson-Fahmy, R.F., Koerner, R.M., 1993. Finite element modelling of soilgeogrid interaction with application to the behavior of geogrids in a pullout loading condition. Geotextiles and Geomembranes 12, 479501. 1416 November 2001IS Kyushu 2001 Landmarks in Earth Reinforcement, vol. 1. Balkema Publisher, pp. 3136. Jewell, R.A., 1990. Reinforcement bond capacity. Geotechnique 40 (3), 513518. Jewell, R.A., 1996. Soil Reinforcement with Geotextiles. CIRIA Special Publication 123, Thomas Telford. Jewell, R.A., Milligan, G.W.E., Sarsby, R.W., Dubois, D.D., 1985. Interactions between soil and geogrids. In: Proceedings from the Symposium on Polymer Grid Reinforcement in Civil Engineering. Thomas Telford, London, pp. 1830. Matsui, T., San, K.C., Nabesahirna, Y., Arnii, U.N., 1996. Bearing mechanism of steel reinforcement in pull-out test. In: Proceedings of the International Symposium: Earth Reinforcement, Fukuoka, Kyushu, Japan. Balkema Publisher, pp. 101105. Milligan, G.W.E, Earl, R.F., Bush, D.I., 1990. Observations of photoelastic pullout tests on geotextiles and geogrids. In: IV International Conference on Geotextiles, Geomembranes and Related Products, The Hague, The Netherlands, vol. 2. Balkema Publisher, pp. 747751. Moraci, N., Montanelli, F., 2000. Analisi di prove di sf` lamento di geogriglie estruse installate in terreno granulare compattato. Rivista Italiana di Geotecnica 4/2000, 521. Moraci, N., Recalcati, P.G., 2005. Factors affecting the pullout behaviour of extruded geogrids embedded in a compacted granular soil. Geotextiles and Geomembranes, in review. ` Moraci, N., Gioffre, D., Romano, G., Montanelli, F., Rimoldi, P., 2002. Pullout behaviour of geogrid embedded in granular soils. In: