TEXTILES.ORG
European experience in pullout tests: Part 2
By Daniele Cazzuffi, Lidia Sarah Calvarano, Giuseppe Cardile, Nicola Moraci, and Piergiorgio Recalcati
Introduction
Geogrids are one of the most common types of geosynthetic used for soil reinforcement. In particular, the use of geosynthetics has unique advantages over other soil strengthening techniques, because of technical, economic, and sustainability reasons (e.g., simplicity of construction, lower transportation costs, respectably low emission, and a range of physical and mechanical properties).
The redistribution of internal stresses within reinforced soil mass and its deformations depends on soil shear strength, reinforcement tensile strength and stiffness, and on interface stress mechanisms between soil and reinforcement. As changing the type of reinforcing material, their geometrical shape and mechanical properties, the soil-geosynthetic interaction mechanisms changes. Therefore, the use of geosynthetics as reinforcement requires an understanding of soil-geosynthetic interaction behaviour.
Pullout tests are helpful and widely used to study the complex 3-D interaction mechanisms between soil and geogrid in the anchorage zone. The apparent interface coefficient of friction and the pullout resistance obtained by means of pullout tests have important implications on the design of geosynthetic reinforced structures.
The main soil parameters that influence the pullout behavior are the same that typically affect the soil dilatancy, such as soil relative density and grain size distribution (in according to geogrid mesh size), and vertical effective stress (Moraci and Montanelli, 2000; Moraci and Recalcati, 2006). About the reinforcement, the geometry and thickness of the bearing ribs, the mesh size (in terms of spacing between elements), and the longitudinal and transversal members’ stiffness play an important role on the soil–geogrid interaction (Moraci and Recalcati, 2006; Calvarano et al., 2012 and 2013; Cazzuffi et al., 2011).
It is important to define the role of all design and test parameters on the mobilization of the interaction mechanisms (frictional and passive) in pullout condition, including geosynthetic length, tensile stiffness, geometry and shape, vertical effective stress, and the soil type.
Cazzuffi et al. (2011), analyzed the effect of the geogrid geometry on the pullout response and showed the existence of an optimum spacing between the transversal elements that maximizes the peak pullout resistance (Figure 1).
In this article the effects of vertical effective stress and of geosynthetic length on the soil-reinforcement interaction under static pullout loading conditions are analyzed. The results of pullout tests carried out on two extensible geogrids embedded in a compacted granular soil, in terms of peak pullout resistance at soil-geosynthetic interface, are analyzed. Some considerations about deformation behavior analysis are also reported.
Test apparatus
The pullout test apparatus (Figure 2a) is an automated device capable of applying a displacement controlled static pullout force.
It is basically composed of a rigid steel large pullout box (1700×680×600 mm3), a vertical load application system, a horizontal force application device, an internal special clamp, and all the required control and data acquisition instruments. A more detailed description of the test apparatus can be found in previous researches performed by the authors (Moraci and Recalcati, 2006; Moraci and Cardile, 2009, 2012).
The equipment incorporates two sleeves near the slot at the front wall of the pullout box to avoid front wall effects, as recommended by researchers (Farrag et al., 1993; Moraci and Montanelli, 2000).
The pullout apparatus is capable producing the confined failure of a geosynthetic specimen by using an internal clamp placed inside the soil beyond the sleeves. This clamp system keeps the geosynthetic specimen confined in the soil for the entire test, ensuring a constant anchorage length during the pullout test and measuring displacements and strains in confined conditions. The internal clamping system (Figure 2b) requires a series of preliminary evaluation clamping friction tests with the same boundary conditions, on the clamping systems without any reinforcement, to evaluate the pullout resistance developed by the clamping system alone. Moreover, it is important to point out that the use of the internal clamping system allows the study of the geogrids’ tensile resistance under soil confinement.
Friction between the soil and the side walls of the box, by the use of smooth Teflon® films, is minimized.
The specimen displacements were measured and recorded using inextensible steel wires connected to a minimum of six points along the geogrid specimen. The wires are connected to displacements transducers (RVDT) fixed to the external back side of the box. All the measurements were digitally recorded on a personal computer in real time.
Test materials
The test materials used in this research were characterized by standard laboratory tests that have allowed evaluating the main parameters necessary to develop the analysis and discussion of test results.
Soil
A granular soil was used. The soil (Figure 3) is a uniform medium sand (SP according to USCS classification system, A-3 according to CNR-UNI 10006 classification system), with grain shape from sub-rounded to rounded, uniformity coefficient, U, equal to 1.96, and average grain size, D50, equal to 0.32 mm.
The Standard Proctor compaction tests performed on the soil indicate a maximum dry unit weight, γmax =16.24 kN/m3, at an “optimum” water content wopt =13.5%.
The direct shear tests, as all the pullout tests, were carried out at an initial unit weight corresponding to the 95% of γmax for the sand studied. The peak shear strength angle, φ’p, at confining pressure equal to 100 kPa, was equal to 42°.
Geogrids
The pullout tests were performed on two HDPE extruded mono-oriented geogrids (called GGEM1 and GGEM2).
Figure 4 shows a schematic cross section of a generic geogrid-bearing member placed transversely to the direction of pullout force. 
Table 1 shows the geometrical characterization of the geogrids, where Wr and Br are the node width and thickness, respectively; S is the spacing between transversal bars in the pullout direction; Wt and Bt are the width and thickness of the bar portion between two nodes, respectively; 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 (Figure 4).
The mechanical properties of the different geogrids were evaluated by means of wide-width tensile tests according to EN ISO 10319 (see also Cazzuffi, 1996), but performed at the same rate of displacement used in the pullout tests (1 mm/min). The tensile test results are reported in Table 2.
Analysis of pullout test results
Pullout tests were carried out for each geogrid embedded in the compacted sand, the specimen length (LR = 0.40, 0.90, 1.15m), and the applied vertical effective pressure (σ’v = 10, 25, 50 and 100 kPa). The pullout test program is summarized in Table 3.
All pullout tests were performed at constant rate of displacement equal to 1.0mm/min. until geogrid tensile failure or until a total horizontal displacement of 100mm was achieved.
Factor affecting the pullout behavior
To study the influence of anchorage length and applied vertical stress on pullout behavior, the test results obtained on the mono-oriented extruded geogrids were analyzed.
Figure 5 shows, for the GGEM1 and GGEM2 geogrids, the classical pullout curves, that reporting the peak pullout force, PR, vs. the displacement measured at the edge attached to the clamp, for long (LR=1.15 m) reinforcement specimens. 
Figure 6, instead, shows the same pullout curves for short (LR=0.40m) reinforcement specimens. The different curves on the graphs refer to the different applied confining pressures. 
It is possible to observe that the pullout behavior is strongly influenced by applied confining stress and by geogrid length. In fact, for all the used geogrids, the tests performed with ‘‘long’’ specimens (LR =0.90-1.15m) and confining pressure higher than 25 kPa show a strain-hardening behavior, with a progressive increase of the pullout resistance with the increase of the displacement.
Moreover, in the tests carried out on the longer GGEM1 geogrid specimens (LR =1.15m) at the confining pressures equal to 50 kPa and 100 kPa, pullout tensile failure occurred. The tensile strength in pullout conditions was close to the tensile strength obtained by in air tensile tests performed at the same displacement rate of the pullout tests. This means that, under these test conditions, the influence of soil confinement on tensile strength was negligible.
The results of tests performed on “short” specimens (LR= 0.40 m) and on “long” specimens under low confining stresses (10 kPa and 25 kPa) show a strain-softening behavior, with a progressive decrease of pullout resistance after the peak value.
To analyze the influence of the anchorage length, and therefore of reinforcement extensibility, the pullout curves have been normalised with respect to double reinforcement length, 2LR. This ratio represents the apparent tangential tensions at the reinforcement interfaces.
Figure 7 shows the normalized pullout curves referring to the tests performed on mono-oriented geogrids for an applied confining stresses equal to 50 kPa. 
The shorter reinforcement specimens (LR=0.40 m) develop a greater normalised peak pullout resistance with respect to the longer ones. At large displacement the P/(2LR) values seem to be independent on the reinforcement length. Similar results were obtained for the other applied vertical stresses.
These results confirm the influence of the specimen length, and therefore the extensibility of the reinforcing element, on pullout behavior, particularly at peak load.
To analyze the influence of vertical confining stress on pullout behavior, pullout curves have been normalised with respect to σ’v (Figure 8). 
From these curves it is possible to notice an important reduction in the normalized resistance passing from low to high confinement stress.
The experimental results can be explained by means of the soil dilatancy phenomena.
The reinforcement extensibility effect can also be observed by means of nodal displacements analysis. Figure 9 shows the distribution of nodal displacements, along the reinforcement specimens, for the GGEM1 and GGEM2 geogrids. 
These curves refer to the peak pullout resistance, PR, and to the trigger force (i.e., that causes the movement of the last bar), Pin, respectively.
Generally, long reinforcement specimens show extensibility that induces a progressive mobilization of the interaction pullout mechanisms (Figure 9a).
Short reinforcement specimens show lower longitudinal strain (Figure 9b) and then an almost immediate development of the interaction mechanisms.
According to the research results, the extensibility effect is more evident for high values of the applied vertical stresses and for long specimens.
Figure 10, in which the peak pullout resistance is plotted as function of the vertical effective confining stress (open symbols indicate the confined tensile failure), shows the influence of the reinforcement stiffness and geometrical characteristics (effective bearing areas) on the peak pullout resistance.
By comparing the experimental results of the tests carried out on the two different mono-oriented geogrids with the same anchorage length and applied normal stress, independent of reinforcement extensibility and of dilatancy effects, it is possible to observe that (in the absence of interference phenomena) the maximum percentage differences of PR are very close to the percentage difference of the effective bearing areas (Ab) against which the passive resistance is mobilized.
Conclusions
The experimental results presented in this article have shown clearly the influence of vertical effective stress and of geosynthetic length on soil-reinforcement interaction under static pullout loading conditions.
On the basis of experimental results, the following conclusions can be drawn:
- The tensile strength in pullout conditions for the soil-geosynthetic interfaces investigated in the research is very close to the tensile strength obtained by in-air tensile tests performed at the same rate of displacement as pullout tests. This means that the influence of the soil confinement on reinforcement tensile strength is negligible.
- The pullout behavior depends on reinforcement length and on the applied vertical stress. In particular, the tests performed with “long” specimens (LR= 1.5 m) and confining stresses larger than 25 kPa show a strain-hardening behavior. The pullout interaction mechanisms develop progressively along the reinforcement specimens, with a progressive increase of the pullout resistance with an increase in displacement. The tests performed on “short” specimens (LR= 0.40 m) and on long specimens under low confining stress show a strain-softening behavior, with a progressive decrease of pullout resistance after peak load. In this case, the interaction mechanisms develop almost at the same time along the whole length of the specimen.
- The experimental results have also shown that the reinforcement extensibility has an influence on peak pullout strength. In particular, extensibility effects are more evident for long reinforcements and at high confining stresses.
- With the increase of the surface on which the passive interaction mechanisms can be mobilized, it is possible to note corresponding increase of peak pullout resistance under static pullout condition.
Acknowledgments
All authors have contributed in equal manner to this article. The research was funded by PON01_01869 TEMADITUTELA.